Morphological Differentiation of Oligodendrocytes Requires Activation of Fyn Tyrosine Kinase

Primary cultures of oligodendrocyte progenitor cells were isolated from P2 neonatal rat brain (Sprague Dawley; Taconic Farms) according to the procedure of Osterhout et al. (1997) with slight modifications. In brief, mixed dissociated cultures of neonatal brain were allowed to reach confluence, after which the oligodendrocyte progenitor population was removed by shaking. The progenitor cells were further purified by immunopanning with A2B5 (Osterhout et al., 1997), and plated onto polylysine (PLL)- coated dishes. Progenitors were lysed 24 h after plating and analyzed for Fyn protein expression and kinase activity (see below). Differentiation of progenitor cells was initiated by switching the media to a defined media lacking growth factors (Osterhout et al., 1997). To maintain the cells as progenitors, and prevent differentiation, PDGF and FGF were included in the differentiation media at concentrations that stimulate the maximal proliferation of the cells (10 ng/ml and 20 ng/ml, respectively; Osterhout et al., 1997). The mature oligodendrocytes were analyzed for Fyn expression and activity after 4–6 d in culture. The differentiation state of the cells, either progenitor or mature oligodendrocyte, was determined by immunocytochemical analysis of stage-specific markers such as A2B5 for progenitor cells, and O1, MAG, and MBP for differentiated oligodendrocytes (Pfeiffer et al., 1993). For the inhibitor studies, PP1, PP2, PP3, and PP4 (Calbiochem Novabiochem) were resuspended in DMSO, and included in the culture media at specified concentrations.

Cells were grown on PLL-coated 10-mm glass coverslips (Dynalab) for all immunochemistry. For oligodendrocyte lineage markers such as A2B5 and O1, the coverslips were rinsed in PBS (10 mM sodium phosphate, pH 7.2, 150 mM NaCl) and incubated in antibody diluted in PBS for 30 min at room temperature (RT). The coverslips were then fixed in 4% paraformaldehyde for 30 min. For Fyn localization, the cells were permeabilized with 0.2% Triton X-100, then incubated with a rabbit antibody to Fyn (Santa Cruz Biotechnology) overnight at 4°C. Fluorescein- and rhodamine-conjugated secondary antibodies (Vector Labs) were diluted 1:50 in PBS and left on the coverslips for 30 min. For each immunostain, one coverslip was incubated in the secondary antibody alone as a negative control for background immunofluorescence. The coverslips were rinsed in PBS and mounted on glass slides using Vectashield (Vector Labs). In all staining procedures, labeled cells were visualized using a Nikon Optiphot microscope or a Leica confocal microscope.

For analysis of process outgrowth, the distance from the center of the cell body to the tip of the longest process was measured, which in mature oligodendrocytes corresponds to the radius of the extensive process network that surrounds the cell body. For PP1-inhibited cells, this corresponds to the length of the longest process extending from the cell. If a process was not longer than the cell body diameter, it was not measured. For each condition, process outgrowth was measured for 200 individual cells from control and inhibitor-treated cells.

Cells were washed twice with cold STE (100 mM NaCl, 10 mM Tris, pH 7.4, 1 mM EDTA) and lysed in 1 ml of NP-40 lysis buffer (20 mM Tris, pH 8.0, 150 mM NaCl, 1% Nonidet P-40, 2.5 mM EDTA, 10 mM NaF, 1 mM Na₃VO₄, 10 μg/ml aprotonin, 10 μg/ml leupeptin, 1 mM PMSF). Cell lysates were then processed for immunoprecipitation as described (Wolven et al., 1997). In vitro kinase assays, enolase phosphorylation assays, and Western blotting have been described previously (Wan et al., 1997; Wolven et al., 1997). Quantitation of band intensity in both the Western analysis and kinase assays was accomplished using a Molecular Dynamics PhosphorImager.

Antibodies raised against amino acids 1–148 of Fyn for immunoprecipitation have been previously described (Wolven et al., 1997). Rabbit polyclonal antibodies to Src, Fyn, Lck, and Lyn were purchased from Santa Cruz Biotechnology. Antiphosphotyrosine antibodies 4G10 and PY20 were purchased from UBI and Transduction Labs. MAG antibodies were a generous gift from Dr. Marie Filbin (Hunter College, New York). O1 antibodies were a generous gift from Drs. Steven Pfeiffer and Rashmi Bansal (University of Connecticut, Farmington, CT). Antibodies to neomycin phosphotransferase II (NPTII) were purchased from 5′→ 3′.

The K299M Fyn construct contains a point mutation in the ATP binding pocket (lysine 299 was changed to methionine), rendering the kinase inactive (Twamley-Stein et al., 1993). The K299M Fyn cDNA was generated in SP65 using wild-type Fyn as a substrate, and subsequently cloned into the retroviral pLJ vector (Wolven et al., 1997). The mutation was verified by DNA sequencing. ψ2 cells were transfected with K299M or pLJ plasmids, and cell supernatants containing the retrovirus were collected and used to infect oligodendrocyte progenitor cultures.

Primary cultures of oligodendrocyte progenitors typically contain bipolar cells that are characterized by the expression of the surface markers A2B5 (80–90%) and O4 (10– 20%). These progenitors can be induced to differentiate by switching to a serum-free differentiation media 24 h after plating (see Materials and Methods). After several days in culture, mature cells are characterized by marked process outgrowth, with an increase in the number of processes emanating from the cell body and extensive branching of these processes. The result is an intricate lacework of processes that surrounds the oligodendrocyte cell body. The morphological changes are accompanied by the appearance of the cell surface galactocerebroside O1, as well as myelin proteins such as MAG and MBP. At later stages, the extensive process outgrowth gives rise to the formation of myelin-like membrane sheets, which can be visualized by staining for O1 (Sommer and Schachner, 1981). Fyn is expressed in both the progenitors and differentiated cell populations (Figs. 1 and 2, immunocytochemistry and Western analysis, respectively). In progenitors, Fyn is localized primarily to the cell body and found along the length of the processes. (Fig. 1 A). Likewise in the mature oligodendroycte, Fyn is observed in the cell body and throughout the processes (Fig. 1 B). Staining for O1 and Fyn also confirmed that the majority of Fyn immunoreactivity can be visualized in the cell body and processes of the mature cell (Fig. 1 C).

To investigate whether Fyn protein levels changed during differentiation, Western analysis of oligodendrocyte cell populations was performed. The bipolar progenitors were harvested 24 h after plating. Mature oligodendrocytes were harvested after allowing progenitors to differentiate for 5–6 d. Fyn protein is expressed at levels two- to threefold higher in the mature cells compared with progenitors (Fig. 2 A). Moreover, Fyn tyrosine kinase activity in mature oligodendrocytes is 10–30-fold higher than that of the progenitors (Fig. 2, B, D, and E), as measured by autokinase activity and enolase phosphorylation. This suggests that the specific activity of Fyn is increased 3–15-fold per unit of Fyn protein in the mature cells. Since oligodendrocyte differentiation occurs over 5–6 d in vitro, Fyn activity was examined each day after the plating of oligodendrocyte progenitors to establish the time course of Fyn upregulation during differentiation. Fyn autokinase activity was used as an indication of Fyn activation, as it mimics the results obtained by measuring enolase phosphorylation (Fig. 2, D and E). The observed increase in Fyn activity occurs within 24 h after switching to differentiation media, which corresponds to 48 h after plating (Fig. 2 C). Remarkably, cells at this time point are indistinguishable from those progenitors harvested at 24 h after plating, both in morphology and in the expression of stage specific markers. Since Fyn activation precedes any morphological changes that accompany differentiation, these observations suggest that the upregulation of Fyn is a very early event in the maturation of oligodendrocyte progenitors.

Several members of the Src family of tyrosine kinases are expressed in the neonatal brain, including Src, Fyn, Lyn, and Yes (Fig. 2 A, lanes denoted T). However, analysis of oligodendrocyte cell lysates (Fig. 2, D and E) revealed that Fyn is the only member of the Src family with significant kinase activity in either cultured oligodendrocytes or progenitor cells. Src is not expressed or catalytically active in these cells; Yes protein is barely detectable in differentiated cells, but is not active (Fig. 2, A, D, and E). Lyn protein is expressed in both the progenitors and mature oligodendrocytes (Fig. 2 A), but Lyn tyrosine kinase activity was not detected in either population of cells (Fig. 2, D and E). Analysis of kinase activity was performed using two different substrates, including autokinase as well as enolase phosphorylation for all Src family members (Fig. 2, D and E). These findings distinguish oligodendrocytes as one of the few cell systems in which Fyn is the major active Src family member. These cells therefore can provide a useful system in which to study Fyn function.

Growth factors such as PDGF and FGF stimulate progenitor cell proliferation, regulate cell migration and modulate differentiation in the oligodendrocyte lineage (Noble et al., 1988; Besnard et al., 1989; Osterhout et al., 1997). In vitro, progenitors maintained in the presence of these growth factors fail to differentiate; withdrawal from the cell cycle and terminal maturation is triggered by removal of the growth factors. In the protocol used for this study, progenitors are harvested soon after plating; under these conditions, the progenitors have a low basal level of Fyn kinase activity (Fig. 2, D and E). Since PDGF and FGF can activate Src and Fyn in other cell types (Twamley-Stein et al., 1993; Langren et al., 1995), we examined the possibility that PDGF and FGF are required to activate Fyn in the progenitor cell population. Fyn kinase activity was examined after acute treatment with FGF (20 ng/ml) and PDGF (10 ng/ml) for 5, 10, and 20 min. There was no increase in Fyn activity above basal levels (Fig. 3, A and B). Since differentiation can be prevented by maintaining progenitors in the presence of growth factors (McKinnon et al., 1990), we also examined the effects of chronic treatment with growth factors. After 3 d in culture with either PDGF (10 ng/ml) or FGF (20 ng/ml), Fyn activity remained at basal levels (Fig. 3, A and B). Fyn activity was not upregulated until the growth factors were withdrawn, which triggered differentiation.

It is also possible that the increase in Fyn activity could be influenced by specific media components. Fyn kinase activity was assayed in cells cultured under several different media formulations, each of which will allow oligodendrocyte differentiation, but vary in their composition (Barres et al., 1993; Osterhout et al., 1997; Yoon et al., 1998). There were no differences in the levels of Fyn kinase activity in the mature oligodendrocytes under these different culture conditions (data not shown). These data indicate that the activation of Fyn is closely associated with the start of progenitor cell differentiation and not stimulated by FGF or PDGF. The lack of growth factor regulation suggests that activation of Fyn in oligodendrocytes is achieved via a novel regulatory mechanism, which is an early event in the terminal differentiation of oligodendrocytes.

The upregulation of Fyn kinase activity during differentiation suggests that enhanced tyrosine phosphorylation of potential Fyn substrates should be evident in mature oligodendrocytes. This supposition was verified by the results depicted in Fig. 4 A, in which an increase in tyrosine phosphorylated proteins was observed in both total cell lysates and Fyn immunoprecipitates of differentiated cells. Several proteins were phosphorylated and associated with Fyn as oligodendrocytes differentiate, and these phosphorylated proteins were also observed in Fyn autokinase assays on oligodendrocyte lysates (Fig. 2, C and D). These proteins were not phosphorylated under the assay conditions required for enolase phosphorylation by Fyn (Fig. 2 E). Interestingly, these proteins were not phosphorylated in autokinase assays using progenitor cell lysates even though there is a low level of Fyn activity, which suggests that phosphorylation of these proteins is part of the maturation program. These phosphorylated bands appear to be novel and specific to oligodendrocytes. They do not correspond to any known Src or Fyn substrates, including Fak, Cbl, p130Cas, annexins, PI3 kinase, or cortactin, and the identification of these potentially novel proteins is currently under investigation.

To determine whether activation of Fyn was required for differentiation, we examined the fate of progenitors cultured in the presence of the tyrosine kinase inhibitors PP1 and PP2. These pyrazolopyrimidine derivatives are potent and specific inhibitors of Src family kinases in T cells and compete with ATP to prevent phosphorylation (Hanke et al., 1996). They have different affinities for each member and do not inhibit other tyrosine kinases at the concentrations effective for the Src kinase family (Hanke et al., 1996). Since Fyn is the only Src family kinase active in this cell system, Fyn should be the only Src family member affected by the inhibitors. Fyn kinase activity in lysates from mature oligodendrocytes was significantly lower in the presence of PP1 and PP2 (Fig. 4 B), and a corresponding reduction in tyrosine phosphorylation of total cellular protein was observed (Fig. 4 C). Inhibition of Fyn kinase activity by PP1 was concentration dependent, with total inhibition occurring at concentrations of 5 μM and higher. The effects of PP1 are specific since the inactive compound PP3 could not inhibit Fyn kinase (Hanke et al., 1996). The activity levels of Fyn in oligodendrocytes in the presence of PP3 were comparable to control cells (Fig. 4 B). Tyrosine phosphorylation of proteins in the total cell lysates was reduced but not eliminated in the presence of PP1 (Fig. 4 C). At low concentrations, where PP1 has limited effects on Fyn activity (Fig. 4 B), there was a slight reduction in tyrosine phosphorylation, including Fyn autophosphorylation. As the concentration of PP1 increased, the levels of tyrosine phosphorylation measurably decreased. Fyn autophosphorylation was reduced by 70% at 5 μM PP1 (Fig. 4 C), a concentration at which Fyn kinase activity is totally inhibited (Fig. 4 B). Similar reductions in tyrosine phosphorylation were observed in immunoblots probed with anti-phosphotyrosine antibodies, including proteins at 100–120, 80–85, and 40–45 kD, which may represent Fyn substrates in oligodendrocytes. The observation that tyrosine phosphorylation was not totally eliminated in the presence of PP1 suggests that while Fyn is an important kinase in these cells, other tyrosine kinases may be active and not inhibited by the PP1 inhibitor family. One possibility is Csk, which phosphorylates Fyn at tyrosine 527 and acts as a negative regulator of Fyn (Chow et al., 1993).

If activation of Fyn tyrosine kinase is indeed responsible for differentiation, then treatment of cells with PP1 should block this process. Progenitors induced to differentiate in the presence of PP1 retracted their processes and clumped together in the dish (Fig. 5). This retraction was evident after 1–2 d in differentiation media, just as the control cells started to extend multiple processes. This effect was dose dependent; at low concentrations (1 μM), there was a 10% reduction in process outgrowth after 4 d, as assessed by measuring the radius of the extensive process network that surrounds the oligodendrocyte cell body. At concentrations <1 μM, there were no observable effects on cell morphology (data not shown). In the presence of 5 μM PP1, process outgrowth was significantly inhibited (>95%). All of the cells extended only one or two short, unbranched processes which were generally less that one cell body diameter in length. The treated cells lacked the extensive process network that is characteristic of the differentiated phenotype, and morphologically resembled oligodendrocyte progenitor cells. None of the cells formed the myelin-like membrane sheets observed in cultures of mature oligodendrocytes. The degree of process retraction observed with increasing concentrations of the inhibitors can be correlated with the levels of Fyn kinase inhibition observed in the in vitro kinase assays (Fig. 4 B), which suggests that Fyn kinase activity is required for process extension during oligodendrocyte differentiation.

While the inhibition of Fyn activity clearly retarded process outgrowth, it did not affect the expression of the cell surface marker galactocerebroside O1 (Fig. 5, B, D, F, and H) and myelin proteins MAG and myelin basic protein (Fig. 6). The process retraction was reversible, as removal of PP1 resulted in renewed process extension as early as 24 h after the inhibitor removal (Fig. 7, A and B). To control for any non-specific effects, cells were treated with the related inactive analogues PP3 and PP4, which had no effect on Fyn activity (Fig. 4 B) and no effect on differentiating cells in culture (Fig. 5 G). Finally, when progenitors were maintained in FGF and PDGF, mitogens that block differentiation (Noble et al., 1988; Bogler et al., 1990; McKinnon et al., 1990), PP1 had no effect on process outgrowth or cell division (Fig. 7, C and D). Progenitor cells maintained in 10 μM PP1, a concentration that completely inhibits Fyn kinase activity (Fig. 4 B) proliferated normally in the presence of FGF and PDGF, and maintained a bipolar morphology. These data suggest that the PP1 family of inhibitors are not toxic to the cells, but are acting solely to inhibit Fyn kinase activity. Thus, inhibition of Fyn activity appears to block morphological differentiation of progenitors into mature oligodendrocytes, without affecting the synthesis of stage-specific proteins that are characteristically expressed by mature oligodendrocytes.

To verify the requirement for Fyn in oligodendrocyte differentiation, a kinase inactive, dominant negative Fyn was introduced into cultures of differentiating oligodendrocyte progenitors. The construct contains a point mutation in the ATP-binding site in which lysine 299 was mutated to methionine (Fig. 8 A). This mutation results in a Fyn protein that cannot bind ATP and is catalytically inactive (Twamley et al., 1992; Twamley-Stein et al., 1993). The Fyn K299M mutant can still bind to substrate molecules and acts in a dominant negative fashion when expressed with wild-type Fyn in cells (Twamley et al., 1992). The Fyn K299M cDNA was inserted into the viral pLJ vector containing an IRES sequence linked to the Neo gene, which confers resistance to geneticin, and allows for both selection and immunocytochemical detection of cells infected with the recombinant retrovirus (Fig. 8).

Oligodendrocyte progenitors were infected with a vector control or Fyn kinase inactive vector using the protocol of Wolven et al. (1997) and maintained in differentiation media. Infected cells were detected by immunostaining for neomycin phosphotransferase II (NPTII). Attempts to select for stable cell lines expressing the construct failed, even in low concentrations of G418. However, in transient infections, many cells were positive for NPTII, although the levels of staining varied considerably. Several effects were evident in the cultures infected with the Fyn K299M vector (Fig. 8, B and D). Cells infected with the pLJ vector alone proceeded to differentiate normally, extending branched processes after culture in differentiation media (Fig. 8, A and C). Cultures infected with the inactive Fyn kinase construct showed considerable inhibition of process outgrowth after 3 d (Fig. 8, B and D). During the time course allowed for differentiation, many cells rounded up and detached from the plate, and it is not known whether these cells were expressing the mutant Fyn protein. Others extended processes less than one cell diameter in length and resembled those cells treated with the PP1 inhibitors. A few NPTII positive cells extended processes similar to the vector control. Overall, there was a 50% reduction (Fig. 8 E) in the number of cell extending processes longer than one cell body in diameter in the cells expressing the Fyn K299M virus compared with control. Since these data are derived from transient infections in which only 10– 30% of the cells were infected, it was not feasible to assay the level of Fyn kinase activity in the cells expressing the mutant construct. However, the reduction in process outgrowth observed in cultures expressing the dominant negative Fyn supports the hypothesis that Fyn kinase activation is required for the formation of processes that occurs during oligodendrocyte differentiation.

The initial observation that Fyn-deficient mice showed reduced myelination in the central nervous system suggested that this tyrosine kinase could be one of the signals involved in regulating myelination during development (Umemori et al., 1994). However, the underlying mechanism for the failure to myelinate in these mice was unclear. The present study directly examines the potential role of Fyn in myelination using primary cultures of differentiating oligodendrocytes. This is a well-characterized cell system; differentiation proceeds along a defined pathway, marked by discrete stages that can be identified by both morphological changes and the expression of cell surface markers and myelin proteins (Pfeiffer et al., 1993). Our results indicate that Fyn tyrosine kinase is activated very early after the oligodendrocyte progenitor cells are induced to differentiate (Fig. 2). This upregulation occurs at a point before the appearance of MAG in these cells, indicating that another signal triggers Fyn in these cells. Inhibition of the kinase activity during differentiation prevents the extensive process outgrowth and formation of membrane sheets normally observed in cultures of mature oligodendrocytes, suggesting that Fyn may regulate cytoskeletal rearrangements that control process extension.

The observed increase in Fyn activity is triggered as progenitor cells are induced to differentiate by maintaining them in a defined differentiation media. Activation of Src family kinases in other cell types is commonly triggered by external signals, such as growth factor–receptor binding, or integrin-substrate interactions (reviewed by Thomas and Brugge, 1997). Since cells in the oligodendrocyte lineage are responsive to growth factors, it was surprising that Fyn activation was not affected by growth factor treatment. This finding suggests that another external signal or a novel regulatory mechanism is responsible for Fyn activation. Oligodendrocytes do express integrins on their cell surface that can modulate migration of these cells (Milner et al., 1996, 1997); however it is unlikely that integrins are responsible for this response, as there are no abrupt changes in cell adhesion or extracellular substrates when Fyn is activated. Thus, it is likely that a novel regulatory mechanism is responsible for this activation, potentially involving release of a negative regulatory mechanism when the cells are placed in media lacking mitogens. It is possible that Csk, a negative regulator of Fyn, may be highly active in oligodendrocyte progenitors treated with growth factors; thus Fyn activity would be low until the growth factors are removed and Csk is downregulated. Alternatively, the withdrawal of growth factors causes the progenitors to exit the cell cycle and begin to differentiate (Casaccia-Bonnefil et al., 1997; Durand et al., 1997). Signal transduction pathways that modulate differentiation may be activated once the cells leave the cell cycle, resulting in subsequent activation of Fyn. Elucidation of the upstream signal that triggers Fyn activation may ultimately reveal the specific cellular pathways involved in oligodendrocyte differentiation.

Since Fyn is the primary Src family kinase activated during oligodendrocyte differentiation, these cells provide a model system in which to study biological effects of Fyn activation. The inhibition of Fyn, by either a pharmacological inhibitor or by expression of a dominant negative Fyn protein, suggests that Fyn can regulate the cytoskeletal rearrangements that accompany oligodendrocyte differentiation. The localization of Fyn to processes further supports this hypothesis. The precise mechanism underlying this effect is unknown at present; while it is clear that several proteins are phosphorylated when Fyn is activated in these cells, the identity of these proteins is unknown. It is not clear that phosphorylation of any one specific protein is the critical event that modulates the observed process retraction. Examination of tyrosine phosphorylated proteins in the presence of the PP1 inhibitors reveals that there is a general reduction in all phosphorylated proteins, not one protein in particular. This may indicate that Fyn activates another signal transduction pathway that ultimately contributes to the cytoskeletal rearrangement. The dominant negative studies confirm the hypothesis that Fyn is essential for oligodendrocyte process formation. Identification of Fyn substrates in these cells will be necessary to further characterize the intracellular events responsible for the observed biology reported here.

The results obtained in this study may in part explain the partial loss of myelin proteins in the Fyn-deficient transgenic mice. If Fyn directs the formation of myelin sheet formation in vivo as it does in vitro, the myelin in these animals may show a reduction in lamellar structure, or oligodendrocyte cell volume. This would lead to a subsequent reduction in the total amount of myelin proteins. It will be interesting to determine if the loss of myelin protein leads to a reduction of the myelin sheath formation in Fyn-deficient mice. The initial phenotypic characterization of the Fyn-deficient mice did not suggest a defect in myelination, as the animals exhibited normal motor function (Stein et al., 1992, 1994). However, normal nerve conduction and motor function can be realized even if myelin sheath is reduced to 25% of normal (Waxman, 1997, 1998). Further examination of the myelin content and the state of oligodendrocyte differentiation from these mice will provide more details on the role of Fyn during myelination.

The authors would like to thank Dr. Wouter van't Hof for helpful discussions and Dr. Victor Friedrich for assistance with confocal microscopy.

This work was supported by grants from American Cancer Society, National Institutes of Health (NIH GM57966 to M.D. Resh), American Paralysis Association (OB1-9801-2 to D.J. Osterhout), and National Multiple Sclerosis Society (FA 1171A1/TT to D.J. Osterhout). M.D. Resh is an Established Scientist of the American Heart Association.

FAK

focal adhesion kinase

MAG

myelin-associated glycoprotein

MBP

myelin basic protein

NPTII

neomycin phosphotransferase II

PLL

polylysine

RT

room temperature