Focal adhesion kinase (FAK) can play unique and opposing roles in regulating the morphology of differentiating oligodendrocytes

Abstract

During development cells of the oligodendrocyte lineage undergo significant changes in morphology by maturing from migratory oligodendrocyte progenitors, which are mostly bipolar, to post-migratory premyelinating oligodendrocytes, which extend complex and expanded process networks, to mature oligodendrocytes, which generate myelin sheaths required for efficient signal propagation within the nervous system. This extensive morphological remodeling occurs in the context of a complex extracellular environment and requires significant rearrangement of the cell’s cytoskeleton. The molecular mechanisms underlying this intricate integration of signals, however, remain poorly understood. A key regulator of extracellular matrix (ECM) to cytoskeleton signaling is the non-receptor tyrosine kinase FAK. Here, we report that FAK can regulate the morphology of differentiating post-migratory premyelinating oligodendrocytes in a unique and opposing fashion that is dependent on the nature of the ECM and mediated largely by FAK’s catalytic activity. More specifically, FAK was found to restrict process network expansion in the presence of fibronectin but to promote morphological maturation in the presence of laminin-2. In addition, FAK’s restraining role predominated for postnatal day (P)3-derived cells, while its maturation promoting role prevailed for P5-derived cells. Taken together, our findings reveal a complex role of FAK in regulating the morphology of post-migratory premyelinating oligodendrocytes.

Introduction

Focal adhesion kinase (FAK), also known as protein tyrosine kinase 2 (PTK2), is a ubiquitously expressed non-receptor tyrosine kinase that functions as an important regulator of cell shape and adhesion in response to environmental signals (Hanks & Polte 1997, Mitra et al. 2005, Parsons 2003, Schaller 2010, Schlaepfer et al. 1999). In particular, components of the extracellular matrix (ECM) are known to interact with transmembrane receptors of the integrin family and to subsequently recruit FAK to intracellular multi-molecular complexes, termed focal adhesions (Berrier & Yamada 2007, Geiger et al. 2009, Giancotti & Ruoslahti 1999, Schaller et al. 1992). FAK-containing focal adhesions function as key sensory machineries that integrate extracellular signals, interconnect them with the cell’s cytoskeleton and thus ultimately mediate complex cellular responses.

In the central nervous system (CNS) FAK expression has long been recognized to occur in neurons (Burgaya & Girault 1996, Burgaya et al. 1995, Burgaya et al. 1997, Contestabile et al. 2003, Grant et al. 1995, Stevens et al. 1996). Only more recently, however, its expression was characterized in cells of the oligodendrocyte lineage (Bacon et al. 2007, Kilpatrick et al. 2000). Oligodendrocytes, the myelin forming cells of the CNS, undergo extensive morphological remodeling when they differentiate from migratory bipolar oligodendrocyte precursor cells to post-migratory premyelinating oligodendrocytes, which extend a complex and expanded process network, and finally to mature oligodendrocytes, which generate the myelin sheaths (Baumann & Pham-Dinh 2001, Jackman et al. 2009, Pfeiffer et al. 1993). Completing these distinct steps of morphological maturation requires extensive remodeling of the cytoskeleton (Bacon et al. 2007, Bauer et al. 2009, Kim et al. 2006, Liang et al. 2004, Miyamoto et al. 2007, Richter-Landsberg 2008, Sloane & Vartanian 2007, Song et al. 2001, Southwood et al. 2007, Wang et al. 2008). Thus, FAK, as a key player in regulating cytoskeletal organization, is likely involved in the regulation of oligodendrocyte maturation and myelination. Indeed, FAK has been found to mediate process outgrowth from cells of the rat-derived oligodendrocyte cell line CG4 (Hoshina et al. 2007). In addition, phosphorylation of FAK at its autophosphorylation site, which represents a critical event for FAK’s activation and catalytic function, has been described to occur primarily in post-migratory differentiating oligodendrocytes compared to migratory oligodendrocyte progenitor cells (Liang et al. 2004). These findings suggest that FAK’s functional role in cells of the oligodendrocyte lineage is largely restricted to maturing cells. Further support for a crucial role of FAK in regulating oligodendrocyte maturation comes from studies in which FAK has been implicated in a number of signaling pathways found to regulate oligodendrocyte differentiation (Fox et al. 2004, Miyamoto et al. 2007, Rajasekharan et al. 2009, Wang et al. 2009). Most importantly, conditional knock-out of FAK in maturing oligodendrocytes was recently documented to result in an inhibition and/or delay of normal developmental myelination (Camara et al. 2009, Forrest et al. 2009). Taken together, these studies highlight the importance of FAK in regulating myelination. However, the molecular mechanisms that are mediated by FAK in maturing oligodendrocytes and are the basis of the observed phenotype in the conditional FAK knock-out mice are not fully understood.

Morphological maturation of cells of the oligodendrocyte lineage occurs in the context of a diverse extracellular environment. In particular two ECM proteins have been characterized with regard to their importance for oligodendrocyte differentiation and myelination, namely fibronectin and laminin-2. Both are present in the CNS during the time of normal developmental myelination (Colognato et al. 2002, Tom et al. 2003, Zhao et al. 2009). Fibronectin has been detected in the developing CNS in a likely diffuse manner, while non-basal lamina laminin-2 was found to be present on the axonal surface. Functionally, fibronectin was found to attenuate process outgrowth in oligodendrocytes, while laminin-2 has been implicated in stimulating myelin sheath formation (Buttery & ffrench-Constant 1999, Buttery & ffrench-Constant 2001, Chun et al. 2003, Colognato et al. 2005, Laursen & Ffrench-Constant 2007, Maier et al. 2005, Olsen & ffrench-Constant 2005, Siskova et al. 2009). The opposing effects seen in the presence of fibronectin versus laminin-2 raise the question of whether either one or both of the effects require FAK, and how these effects may relate to the in vivo phenotype seen in the conditional FAK knock-out mice.

In an attempt to better understand the role of FAK, as an integrator of ECM signaling, the current study investigated the role of FAK on the morphology of post-migratory premyelinating oligodendrocytes in the presence of fibronectin versus laminin-2 in a well defined in vitro system. The data presented here demonstrate unique and opposing roles of FAK that are dependent on the ECM substrate present and on the subtype of the maturing oligodendrocyte (P3- or P5-derived). Thus, these data provide novel insight into the role of FAK, and they highlight the multi-functionality of FAK in the context of oligodendrocyte differentiation and myelination.

Materials and methods

Antibodies

Hybridoma clone A2B5 (ATCC, Manassas, VA) was used for immunopanning of oligodendrocyte progenitor cells. Hybridoma clone O4 (gift from S. Pfeiffer) was used to identify post-migratory premyelinating oligodendrocytes (Bansal et al. 1989, Sommer & Schachner 1981). O4 supernatants, anti-MBP antibodies (SMI99; Covance, Princeton, NJ) and secondary Alexa 594 or Alexa 488-conjugated antibodies (Invitrogen/Molecular Probes, Carlsbad, CA) were used for immunostaining.

Animals

Sprague–Dawley female rats with early postnatal litters were obtained from Harlan Laboratories (Indianapolis, IN). All animal studies were approved by the Institutional Animal Care and Use Committee at Virginia Commonwealth University.

Primary oligodendrocyte cultures

Primary rat oligodendrocytes were isolated from postnatal day 3 or 5 rat brains in principal as described previously (Barres et al. 1992, Fox et al. 2003). Briefly, cerebral hemispheres were minced and incubated in Hank’s balanced salt solution (HBSS) supplemented with 0.25% trypsin (Invitrogen, Carlsbad, CA) and 1 μg/ml DNase (Sigma, St. Louis, MO). After gentle trituration single cells were collected by centrifugation, resuspended in Dulbecco’s Modified Eagle Medium (DMEM)/10% FCS (Invitrogen, Carlsbad, CA) and subjected to A2B5 immunopanning. Immunopanned cells were plated onto fibronectin (10 μg/ml)-coated 6-well tissue culture dishes and cultured in serum-free proliferation medium (DMEM containing 10ng/ml PDGF (R&D Systems; Minneapolis, MN) and 5ng/ml bFGF (Sigma, St. Louis, MO); DMEM/PDGF/bFGF) for 15-20 hrs followed by 20-24 hrs of culture in differentiation medium (DMEM containing 40 ng/ml tri-iodo-thyronine (T3; Sigma, St. Louis, MO) and 1× N2 supplement (Invitrogen, Carlsbad, CA); DMEM/T3/N2). Cells were then trypsinized and re-plated in differentiation medium onto ECM-coated coverslips. Cells were cultured for an additional 15-20 hrs and then analyzed.

siRNA-mediated knock-down of FAK expression

Oligodendrocytes were isolated by A2B5-immunopanning from postnatal day 3 or 5 rat brains and plated onto fibronectin (10 μg/ml)-coated 6-well tissue culture dishes. Cells were cultured in serum-free proliferation medium (DMEM/PDGF/bFGF) for 15-20 hrs. Subsequently, cells were switched into differentiation medium (DMEM/T3/N2) and transfected with siRNA using lipofectamine 2000 (Invitrogen, Carlsbad, CA). For siRNA transfection a siGLO green transfection indicator along with either a siRNA SMARTpool directed against rat FAK or a control non-targeting siRNA SMARTpool was used (all from Thermo Fisher Scientifc/Dharmacon Inc., Lafayette, CO). Transfection medium containing siRNA-lipofectamine complexes was replaced with serum-free differentiation medium (DMEM/T3/N2) after 3 hrs and cells were cultured for an additional 15-20 hrs. Under these conditions siRNA transfection efficiencies, as determined via the use of the siGLO green transfection indicator, were approximately 70% (data not shown). siRNA transfected cells were then re-plated onto fibronectin (10 μg/ml), laminin-2 (10 μg/ml) or mixed substrate (5 μg/ml fibronectin and 5 μg/ml laminin-2)-coated glass coverslips and cultured for an additional 15-20 hrs. No significant reduction in total FAK protein levels was detected at the time of re-plating (data not shown).

PF573228-mediated FAK inhibition

Oligodendrocytes were isolated by A2B5-immunopanning from postnatal day 3 or 5 rat brains and plated onto fibronectin (10 μg/ml)-coated 6-well tissue culture dishes. Cells were cultured in serum-free proliferation medium (DMEM/PDGF/bFGF) for 15-20 hrs followed by an additional 20-24 hrs in differentiation medium (DMEM/T3/N2). Oligodendrocytes were then re-plated onto fibronectin (10 μg/ml), laminin-2 (10 μg/ml), or mixed substrate (5 μg/ml fibronectin and 5 μg/ml laminin-2)-coated glass coverslips, allowed to settle for 2-4 hrs and cultured in the presence of 100nM PF573228, a specific inhibitor of FAK’s catalytic activity (Tocris Bioscience, Bristol, England), or vehicle (0.1% DMSO) as control. Cells were analyzed after an additional 15-20 hrs in culture.

Oligodendrocyte morphology analysis

Oligodendrocyte morphology was analyzed as previously described (Dennis et al. 2008). siRNA or FAK inhibitor-treated cells along with their respective control cells were immunostained with O4 antibodies. For siRNA-treated cells only those containing the siGLO green transfection indicator were analyzed. Images of at least 25 cells were taken randomly and in a double-blinded fashion for each treatment group in each experiment (n ≥ 3) using an inverted fluorescent microscope (Olympus BX51; Olympus America Inc., Center Valley, PA). IP Lab imaging software (BD Biosciences Bioimaging, Rockville, MD) was used to determine process index (total area found to be O4-positive minus the cell body) and network area (total area within the radius of the process network surrounding the cell body minus the cell body). In addition, the number of primary processes (any process directly extending from the cell body) was counted for each cell. For the bar graphs representing network area, process index and primary process number, the mean value for cells cultured under control conditions was calculated. This mean value was set to 100% and adjusted, i.e. normalized, values for all cells were averaged for each experimental condition.

For the generation of representative images, confocal laser scanning microscopy was used (TCS SP2 AOBS, Leica Microsystems, Exton, PA). Images represent 2D maximum projections of stacks of 0.4 μm optical sections.

Live/Dead Viability Assay

siRNA or FAK inhibitor-treated cells were assayed for cell viability using 2μM calcein AM/4μM ethidium homodimer-1 as described by the manufacturer (Live/Dead Viability Assay kit, Invitrogen Corp., Carlsbad, CA) and used previously by others (p.e. Hahn et al. 2010, Silva et al. 2004). Images of 8 fields at 10x magnification were taken randomly for each treatment group in each experiment (n = 3) using an inverted fluorescent microscope (Olympus BX51; Olympus America Inc., Center Valley, PA), and the percentages of live (labeled with calcein) and dead (labeled with ethidium homodimer-1) cells were determined.

Results

FAK plays unique and opposing roles in regulating the morphology of P3-derived post-migratory premyelinating oligodendrocytes in the presence of fibronectin versus laminin-2

The ECM substrates laminin-2 and fibronectin have been previously described to differentially affect the maturation of post-migratory premyelinating oligodendrocytes (Buttery & ffrench-Constant 1999, Buttery & ffrench-Constant 2001, Olsen & ffrench-Constant 2005, Siskova et al. 2006, Siskova et al. 2009). To assess the effect of these ECM molecules on oligodendrocyte morphology in our tissue culture paradigm, we compared network area and process index as described by us previously (Dennis et al. 2008). In addition, we analyzed the effect of laminin-2 and fibronectin for two subtypes of differentiating oligodendrocytes that were both derived from oligodendrocyte progenitors isolated by A2B5 immunopanning. The first subtype was derived from postnatal day (P)3 rat brains (P3-derived post-migratory premyelinating oligodendrocytes) while the second one was from P5 rat brains (P5-derived post-migratory premyelinating oligodendrocytes). As previously established by us, the majority of oligodendrocyte progenitors derived from brains of rats older than P3 express both the progenitor cell marker recognized by the A2B5 antibody as well as the later stage surface antigen(s) recognized by the O4 antibody (Fox et al. 2004). In contrast, oligodendrocyte progenitors derived from P3 rat brains are to a large extent O4-negative. In the developing CNS the ECM protein fibronectin is likely distributed diffusely, while non-basal lamina laminin-2 seems restricted to axonal surfaces. Thus, it appears that oligodendrocyte progenitors start to differentiate in a fibronectin-containing environment. Only subsequently may differentiating oligodendrocytes encounter non-basal lamina laminin-2. To mimic the above proposed in vivo situation, both subtypes of progenitors were allowed to differentiate in the presence of fibronectin for 20-24 hours before re-plating onto ECM-coated glass coverslips. Cells were analyzed after an additional 15-20 hours of culture in differentiation medium and in the presence of the different ECM proteins. No apparent difference in the number of O4-positive cells was noted under any of the conditions.

As shown in Fig.1, P3-derived post-migratory premyelinating oligodendrocytes developed a much larger network area and process index in the presence of laminin-2 when compared to cells cultured in the presence of fibronectin. In addition, cells cultured in the presence of laminin-2 revealed an increased occurrence of membrane sheet-like structures as denoted by the arrows in Fig. 1a (right panel). The P5-derived post-migratory premyelinating oligodendrocytes showed a similar effect (network area fibronectin: 100% ± 7%, network area laminin-2: 155% ± 10%, p<0.05; process index fibronectin: 100% ± 7%, process index laminin-2: 153% ± 10%, p<0.05) with an even more obvious occurrence of membrane sheet-like structures (compare Fig. 2c (left panel) with 3c (left panel)). Taken together, these findings demonstrate that in agreement with previous studies the morphology of post-migratory premyelinating oligodendrocytes appears developmentally less advanced in the presence of fibronectin when compared to the morphology seen in the presence of laminin-2. Furthermore, this difference in oligodendrocyte morphology was found for both P3- and P5-derived oligodendrocytes.

Fig. 1.

The morphology of post-migratory premyelinating oligodendrocytes is uniquely regulated in the presence of fibronectin compared to laminin-2. Oligodendrocyte progenitors were isolated by A2B5 immunopanning from postnatal day 3 rat brains. After 2 days in culture, differentiating oligodendrocytes were re-plated onto fibronectin (Fn) or laminin-2 (Ln)-coated glass coverslips and allowed to differentiate for 20-24 hours. (a) Representative images of cells stained with the O4 antibody. (b) Bar graphs depicting quantitative analyses of the total area occupied by the cell’s process network (network area; left graph) and the total amount of O4-positive process surfaces per cell (process index; right graph). Mean network area and process index of cells cultured in the presence of fibronectin were set to 100% and values for the cells cultured in the presence of laminin-2 were adjusted accordingly. Means and standard errors are shown. Four independent experiments were performed and 25 cells per experiment and condition were analyzed. For both parameters, Student’s t-test analysis revealed an overall two-tailed significance level of p<0.05 (indicated by stars). Scale Bars: 20 μm.

Fig. 2.

siRNA-mediated knock-down of FAK expression affects the morphology of P3-derived post-migratory premyelinating oligodendrocyte distinctively and in an opposing fashion when comparing cells cultured in the presence of fibronectin with cells cultured in the presence of laminin-2. Cells were isolated, differentiated and analyzed as described in Fig. 1. In addition, cells were treated with a siRNA pool against FAK (siFAK) or a control siRNA pool (siControl) 15-20 hours after initial plating. (a) Representative images of cells stained with the O4 antibody and cultured in the presence of fibronectin (Fn). (b) Bar graphs representing quantitative analyses of network area, process index and primary process number of cells cultured in the presence of fibronectin. (c) Representative images of cells stained with the O4 antibody and cultured in the presence of laminin-2 (Ln). (d) Bar graphs representing quantitative analyses of network area, process index and primary process number of cells cultured in the presence of laminin-2. (e) Bar graph representing the percentage of Hoechst-positive cells that are also MBP-positive. The inset in (e) shows a representative image of cells cultured in the presence of laminin-2 and stained with an antibody specific for myelin basic protein (MBP) (left panel) as well as with Hoechst to visualize nuclei (right panel). For the bar graphs in (b) and (d), mean control values were set to 100% and experimental (siFAK) values were adjusted accordingly; in (e) percentage values are depicted without normalization to control values. In all bar graphs, means and standard errors of at least three independent experiments are shown (25 cells each for all conditions). Stars indicate an overall two-tailed significance level of p<0.05 as determined by Student’s t-test analysis. Scale Bars: 20 μm (a,b); 50 μm (inset in e).

Fig. 3.

siRNA-mediated knock-down of FAK expression has no significant effect on the morphology of P5-derived post-migratory premyelinating oligodendrocyte cultured in the presence of fibronectin. In contrast, knock-down of FAK expression inhibits or delays morphological maturation of P5-derived post-migratory premyelinating oligodendrocyte cultured in the presence of laminin-2. Cells were isolated, differentiated, treated and analyzed as described in Fig. 2. (a) Representative images of cells stained with the O4 antibody and cultured in the presence of fibronectin (Fn). (b) Bar graphs representing quantitative analyses of network area, process index and primary process number of cells cultures in the presence of fibronectin. (c) Representative images of cells stained with the O4 antibody and cultured in the presence of laminin-2 (Ln). (d) Bar graphs representing quantitative analyses of network area, process index and primary process number of cells cultured in the presence of laminin-2. (e) Bar graph representing the percentage of Hoechst-positive cells that are also MBP-positive. The inset in (e) shows a representative image of cells cultured in the presence of laminin-2 and stained with an antibody specific for myelin basic protein (MBP) (left panel) as well as with Hoechst to visualize nuclei (right panel). For the bar graphs in (b) and (d), mean control values were set to 100% and experimental (siFAK) values were adjusted accordingly; in (e) percentage values are depicted without normalization to control values. In all bar graphs, means and standard errors of at least three independent experiments are shown (25 cells each for all conditions). Stars indicate an overall two-tailed significance level of p<0.05 as determined by Student’s t-test analysis. Scale Bars: 20 μm (a,b); 50 μm (inset in e).

In an attempt to identify a molecular player mediating the above described morphological difference, the extent to which one of the major known regulators of cellular morphology, namely FAK, may be involved was evaluated. P3-derived differentiating oligodendrocytes were treated with a siRNA pool against FAK or a non-targeting siRNA pool as control. Cells were cultured and analyzed as described above. As shown in Fig. 2a and b, siRNA-mediated FAK knock-down resulted in a significant increase in both network area and process index for cells cultured in the presence of fibronectin. In addition, the number of primary processes was found to be increased under these conditions. In contrast, when cells were cultured in the presence of laminin-2, all three parameters were found to be decreased upon siRNA-mediated FAK knock-down (Fig. 2c and d). These effects on cell morphology were unlikely due to an interference with initial cell spreading, since at the time of re-plating no significant differences in FAK protein levels were noted (data not shown). At the time of analysis, however, FAK protein levels were found decreased by approximately 50-60% in cells treated with the FAK-specific siRNA pool when compared to cells treated with the control siRNA pool (Fig. S1a).

It has been previously demonstrated that FAK not only acts as a regulator of cytoskeletal dynamics and cell shape but can also be involved in signaling events controlling cell survival (Cox et al. 2006, Lim et al. 2008a, Mitra et al. 2005, Westhoff et al. 2004). In addition, elevated expression of Pyk2, a cytoplasmic tyrosine kinase closely related to FAK, was found to functionally compensate for a loss of FAK in some cell types and to cause alterations beyond effects due to a loss of FAK under certain circumstances (Lim et al. 2008b, Sieg et al. 1998, Weis et al. 2008). Thus, we assessed both cellular viability and Pyk2 protein levels. No significant differences between control and siRNA-treated cells were observed at the time of analysis (Figs. S1c and S2a). In addition, no considerable difference in Pyk2 protein levels was noted between cells plated in the presence of fibronectin and those plated in the presence of laminin-2 (data not shown). These findings demonstrate that FAK regulates morphological maturation of post-migratory premyelinating oligodendrocytes without a significant change in cell viability or via a considerable increase in Pyk2 expression, and they thus suggest that FAK’s regulation occurs primarily by directly regulating the organization of the oligodendrocyte’s cytoskeleton.

During normal development morphological maturation of post-migratory premyelinating oligodendrocytes is associated with an expansion of the cell’s process network and occurs concurrent with the establishment of a protein expression profile characteristic for mature oligodendrocytes (Dugas et al. 2006, Emery et al. 2009, Hardy & Friedrich 1996a). Under experimental conditions, however, an independent regulation of these two processes has been observed indicating that morphological and gene expression profile maturation may be regulated by distinct molecular mechanisms (Buckinx et al. 2009, Buttery & ffrench-Constant 1999, Osterhout et al. 1999, Sloane & Vartanian 2007, Younes-Rapozo et al. 2009). To assess the extent to which the effect of siRNA-mediated FAK knock-down on cell morphology was associated with a change in the oligodendrocyte’s protein expression profile, the number of cells expressing one of the most extensively studied proteins associated with oligodendrocyte maturation, namely myelin basic protein (MBP), was determined (Dubois-Dalcq et al. 1986, Hardy & Friedrich 1996b, Zeller et al. 1985). As shown in Fig. 2e, no significant changes in the number of the MBP-expressing cells were noted under either of the conditions analyzed. At these relatively early stages of oligodendrocyte differentiation investigated here, changes in MBP isoform expression are unlikely to complicate the above analysis (Campagnoni 1988).

Taken together, the above data demonstrate that FAK can regulate the morphology of a P3-derived post-migratory premyelinating oligodendrocyte in a unique and opposing manner depending on the presence of the predominant ECM protein. In particular, FAK limits the expansion of the oligodendrocyte’s process network in the presence of fibronectin, while it promotes morphological maturation in the presence of laminin-2. In both cases, this regulation appears to be independent of survival and/or the expression of the myelin protein MBP, and it thus appears mediated primarily by direct effects on the organization of the cytoskeleton.

For P5-derived post-migratory premyelinating oligodendrocytes, FAK’s role in regulating cellular morphology in the presence of fibronectin appears largely absent while it is persistent in the presence of laminin-2

Our previous in vivo data demonstrated a significant decrease in the number of primary processes upon induction of FAK knock-out (Forrest et al. 2009), an effect resembling the one seen for the P3-derived post-migratory premyelinating oligodendrocytes in the presence of laminin-2 but not fibronectin. The in vivo data were obtained from P14 optic nerves and thus at more advanced developmental stages than those analyzed in Fig. 2. These observations suggest that the role of FAK in regulating cellular morphology may be altered during development. To explore this idea, oligodendrocyte progenitors were isolated from a later postnatal age, namely postnatal day 5 rat brain, and treated/analyzed as described above. Compared to the P3-derived post-migratory premyelinating oligodendrocytes, the P5-derived cells were found to express similar levels of FAK mRNA and protein (our unpublished data and Dugas et al. 2006). In addition, there was no significant difference in the efficiency of siRNA-mediated FAK knock-down between the two subtypes of oligodendrocytes (Fig. S1a and b). Furthermore, as observed for the P3-derived post-migratory premyelinating oligodendrocytes, viability, Pyk2 expression and the number of MBP expressing cells remained unchanged for P5-derived post-migratory premyelinating oligodendrocytes upon siRNA-mediated FAK knock-down (Figs. S1d, S2b and 3e). In addition, siRNA-treated P5-derived post-migratory premyelinating oligodendrocytes displayed a similar difference in morphology upon FAK knock-down in the presence of laminin-2 as seen for the P3-derived cells (compare Fig. 2c and d with Fig. 3c and d). However and different from the P3-derived post-migratory premyelinating oligodendrocytes, there was no discernable change in network area, process index and primary process number when analyzing P5-derived cells upon siRNA-mediated FAK knock-down in the presence of fibronectin (Fig. 3a and b).

The above data demonstrate that in the presence of fibronectin FAK’s capacity and/or effectiveness in restricting the expansion of the process network is absent when differentiating post-migratory pre-myelinating oligodendrocytes are derived from P5 rat brains. FAK’s role in promoting the morphological maturation of post-migratory pre-myelinating oligodendrocytes in the presence of laminin-2, however, appears to be similar for both subtypes of differentiating oligodendrocytes.

For P3-derived post-migratory premyelinating oligodendrocytes, FAK’s constraining role on process network expansion predominates while for the P5-derived cells FAK’s stimulatory role on morphological maturation prevails

Both, fibronectin and laminin-2 have been found present in the developing CNS during the time when oligodendrocyte maturation occurs (Fig. S2g and Colognato et al. 2002, Fox et al. 2004, Tom et al. 2003). Thus, some maturing oligodendrocytes likely encounter both ECM proteins at the same time. For this reason, the role of FAK in regulating the morphology of post-migratory premyelinating oligodendrocytes was determined under conditions, in which both fibronectin and laminin-2 were present in the environment. P3- and P5-derived post-migratory premyelinating oligodendrocytes were cultured, treated and analyzed as described above, with the exception that cells were re-plated on a mixed substrate of equal concentrations of fibronectin and laminin-2 (instead of the individual substrates). Interestingly, even though FAK was found to regulate morphological maturation of P3-derived post-migratory premyelinating oligodendrocytes in an opposing manner on fibronectin versus laminin-2, its role on the mixed substrate appeared to be primarily a constraining one, i.e. similar to the one observed on fibronectin alone (compare Fig. 2a and b with Fig. 4a and b). For the P5-derived post-migratory premyelinating oligodendrocytes, in which FAK’s capacity and/or effectiveness to regulate oligodendrocyte morphology in the presence of fibronectin was absent, the role of FAK was expectedly a stimulatory one, i.e. similar to the one seen on laminin-2 alone (compare Fig. 3c and d with Fig. 4c and d).

Fig. 4.

siRNA-mediated knock-down of FAK expression in the presence of a mixed fibronectin/laminin-2 substrate affects the morphology of P3- and P5-derived post-migratory premyelinating oligodendrocyte in a distinct and opposing fashion. Cells were isolated, differentiated, treated and analyzed as described in Fig. 2. (a,c) Representative images of cells stained with the O4 antibody. P3-derived post-migratory premyelinating oligodendrocytes are shown in (a), while (c) depicts P5-derived cells. Scale Bars: 20 μm. (b,d) Bar graphs representing quantitative analyses of network area, process index and primary process number. Means and standard errors of at least three independent experiments are shown (25 cells each for all conditions). Stars indicate an overall two-tailed significance level of p<0.05 as determined by Student’s t-test analysis.

These data demonstrate that FAK’s role in regulating cellular morphology follows different priorities in at least the two different subtypes of oligodendrocytes analyzed here, namely a predominantly process network restraining one for P3-derived and a mostly morphology promoting one for P5-derived post-migratory premyelinating oligodendrocytes.

FAK’s unique and opposing roles in regulating morphological maturation of post-migratory premyelinating oligodendrocytes are mediated by its catalytic activity

FAK’s role in regulating the morphology of various cell types has at least in vitro been well documented to depend on its catalytic, i.e. autophosphorylation, activity (p.e. Cary et al. 1996, Hanks et al. 2003, Parsons 2003, Playford & Schaller 2004, Schlaepfer et al. 1999). However, recent in vitro and in vivo studies highlight the importance of both autophosphorylation-dependent and autophosphorylation-independent mechanisms in mediating FAK’s physiological functions (Cance & Golubovskaya 2008, Corsi et al. 2009, Lim et al. 2008a). To assess the extent to which the above described effects of FAK on oligodendrocyte maturation were dependent on its catalytic activity, the FAK kinase inhibitor PF573228 was used and differentiating oligodendrocytes were treated with this inhibitor following initial attachment to prevent potential effects on initial cell spreading. PF573228 has been shown in a variety of cell types to effectively block FAK autophosphorylation and downstream signaling without significantly affecting cell survival (Chen et al. 2009, Slack-Davis et al. 2007). In agreement with these findings, treatment of differentiating oligodendrocytes with PF573228 reduced autophosphorylation at FAK’s Y397 site considerably without significantly affecting cell viability (Fig. S3).

As shown in Fig. 5, for both P3- and P5-derived post-migratory premyelinating oligodendrocytes treatment with PF573228 resulted in changes in cell morphology similar to the ones seen upon FAK knock-down. The concentration of 100 nM used here was previously reported to yield approximately half-maximal inhibition of FAK autophosphorylation in various cell types, and in agreement with these data it was found here to reduce FAK autophosphorylation in post-migratory premyelinating oligodendrocytes by approximately 50% (Fig. S3a-b and Slack-Davis et al. 2007). Thus, in both paradigms, siRNA-mediated FAK knock-down and PF573228 pharmacological inhibition, residual FAK activity remains. Therefore, the data presented here suggest that a significant downregulation of FAK expression and/or FAK catalytic activity is sufficient to block the majority of FAK’s roles in regulating the morphology of post-migratory premyelinating oligodendrocytes. Interestingly, for P5-derived post-migratory pre-myelinating oligodendrocytes treatment with PF573228 in the presence of fibronectin resulted in a slight increase in the number of primary processes, while network area and process index remained unchanged. This residual effect suggests that FAK’s role in the presence of fibronectin is not completely absent but significantly diminished in the P5-derived cells. In addition, it indicates that FAK’s role in limiting initial process outgrowth in the presence of fibronectin may be particularly sensitive to the inhibition of FAK’s catalytic activity.

Fig. 5.

Inhibition of FAK’s catalytic activity affects the morphology of P3- and P5-derived post-migratory premyelinating oligodendrocyte in a similar fashion as siRNA-mediated knock-down of FAK expression. Cells were isolated, differentiated and analyzed as described in Figs. 2-4. Cells were, however, treated with the FAK kinase inhibitor PF573228 (PF228; 100 nM) or vehicle (Control; 0.1% DMSO). (a,b) Bar graphs representing quantitative analyses of network area, process index and primary process number of P3- (a) and P5- (b) derived post-migratory premyelinating oligodendrocytes in the presence of fibronectin (Fn) or laminin-2 (Ln). Means and standard errors of at least three independent experiments are shown (25 cells each for all conditions). Stars indicate an overall two-tailed significance level of p<0.05 as determined by Student’s t-test analysis.

Discussion

The current study demonstrates that FAK can regulate the morphology of post-migratory premyelinating oligodendrocytes in a unique and opposing fashion that is dependent on the nature of the ECM substrate present, i.e. fibronectin or laminin-2, and that appears mediated primarily by FAK’s catalytic activity. In addition, the regulatory role of FAK on the morphology of post-migratory premyelinating oligodendrocytes was found to be distinct for different subtypes of maturing oligodendrocytes. More specifically, for P3-derived post-migratory premyelinating oligodendrocytes FAK’s constraining role on process network expansion predominated, while its stimulatory role on morphological maturation prevailed for P5-derived cells of the oligodendrocyte lineage (Fig. 6). Taken together, the data presented here provide novel insight into the complexity of the role that FAK plays in regulating the morphology of differentiating post-migratory premyelinating oligodendrocytes in the context of developmental changes and spatial differences in the molecular composition of the extracellular environment.

Fig. 6.

Proposed model for the role of FAK in regulating the morphology of post-migratory premyelinating oligodendrocytes. The data presented here demonstrate that FAK can have unique and opposing roles depending on the subtype of differentiating oligodendrocyte and the nature of the prevalent ECM molecule that the cell encounters. For P3-derived post-migratory premyelinating oligodendrocytes FAK restrains morphological process network expansion when cells encounter fibronectin. In contrast, it promotes morphological maturation when cells encounter laminin-2. Based on the mixed substrate data, it is the restraining role of FAK that predominates for the P3-derived post-migratory premyelinating oligodendrocytes. For the P5-derived cells, FAK’s role in restraining process network expansion in the presence of fibronectin is significantly diminished. In contrast, FAK’s promoting role on morphological maturation still remains active. Taken together, these data uncover a yet underappreciated complexity of FAK’s role in regulating the morphology of post-migratory premyelinating oligodendrocytes, and they suggest that proper oligodendrocyte maturation requires a well coordinated balance between mechanisms that restrain and those that promote the expansion of the oligodendrocyte’s process network.

FAK’s role in limiting the establishment of an expanded process network in the presence of fibronectin was found to be significantly diminished or even absent for P5-derived post-migratory premyelinating oligodendrocytes (Fig. 6). These findings are suggestive of a functional difference between different developmental stages of maturing oligodendrocytes. However, in initial studies no discernable differences were noted in the expression of a variety of proteins known to change with differentiation of cells of the oligodendrocyte lineage (data not shown). Thus, a comprehensive proteomic analysis seems necessary to determine the extent to which the two subtypes of oligodendrocytes may represent different maturation stages of the same lineage. Alternatively, heterogeneity within cells of the oligodendrocyte lineage has been postulated (Anderson et al. 1999, Butt et al. 1995, Fanarraga & Milward 1997, Kitada & Rowitch 2006, Parras et al. 2007, Spassky et al. 2000). At this point, the possibility that isolating oligodendrocyte progenitors from different postnatal ages of rat brains leads to an enrichment of inherently heterogeneous subtypes of oligodendrocytes can, therefore, not be excluded. From a merely functional point of view, the differences between the two subtypes of oligodendrocytes are likely a result of alterations in signaling up- and/or downstream of FAK. Oligodendrocytes of known types and at all developmental stages express and display on their surface fibronectin receptors, that is the αv integrins αvβ1, αvβ3, αvβ5 and/or αvβ8 (Cahoy et al. 2008, Dugas et al. 2006, Milner & Ffrench-Constant 1994, Milner et al. 1997). Thus, while the functional role of FAK in the presence of fibronectin likely involves the activation of integrin type receptors, it seems unlikely that a reduced expression of fibronectin receptors in P5-derived cells is the primary cause of the observed absence of FAK’s process network restricting role in the presence of fibronectin. Interestingly, fibronectin-attenuated oligodendrocyte process outgrowth has been associated with an inhibition of matrix metalloproteinase (MMP)-9 activity along processes, an effect that may be mediated by modulation of MMP-9 association with ganglioside- and integrin-dependent membrane microdomains (Siskova et al. 2009). FAK could be a critical component of this mechanism and thus be dependent on a specific membrane lipid composition. Ganglioside composition undergoes significant changes during oligodendrocyte development and may, therefore, be the feature most critical for distinguishing the two oligodendrocyte subtypes used here (Mack et al. 1981; Satoh et al. 1996). However, multifaceted studies will be necessary to better understand the precise differences between the two subtypes of differentiating oligodendrocytes and the molecular mechanism(s) responsible for their distinct behavior.

It has been previously reported that FAK can also limit the size of the axonal arbor for a variety of neuronal cell types (Rico et al. 2004). In addition, FAK was found to limit the length and number of dendritic protrusions by directly regulating the organization of the cytoskeleton (Moeller et al. 2006, Shi et al. 2009). Both functions of FAK were found to be, at least in part, dependent on FAK’s catalytic activity. Most importantly, they were implicated in controlling the establishment of the final pattern of connections between neurons and their targets by enabling efficient pruning of overproduced and/or ‘weak’ connections. With regard to the findings described here, it is, therefore, tempting to speculate that the process network limiting function of FAK observed in the presence of fibronectin, an ECM molecule that is expressed broadly during development, allows for efficient pruning of ‘non-functional’ oligodendroglial processes and represents part of a regulatory mechanism that determines the final number of myelinated segments generated by an individual mature oligodendrocyte.

In contrast to FAK’s process network limiting role discussed above, its role in stimulating the establishment of a more mature morphology in the presence of laminin-2 was found operational in both subtypes of post-migratory premyelinating oligodendrocytes, even though it predominated for the P5-derived cells. Due to the more restricted presence of laminin-2 on the axonal surface (Colognato et al. 2002), this function of FAK may play a more spatially confined role than the one seen in the presence of fibronectin. Notably, the interaction between laminin-2 and the integrin receptor α6β1 has been well demonstrated to be involved in regulating the morphological maturation of post-migratory premyelinating oligodendrocytes in vitro and in vivo (Barros et al. 2009, Buttery et al. 1999, Camara et al. 2009, Chun et al. 2003, Colognato et al. 2002, Laursen & Ffrench-Constant 2007, Lee et al. 2006, Olsen & ffrench-Constant 2005, Relvas et al. 2001). Most importantly, hypomyelinating phenotypes observed upon reducing α6β1 integrin signaling in maturing oligodendrocytes resemble those seen upon inducing FAK knock-out in maturing oligodendrocytes (Camara et al. 2009, Forrest et al. 2009). Thus, in the presence of laminin-2 FAK likely functions as a downstream target of laminin-2-α6β1 integrin interactions. In addition to regulating morphological maturation of differentiating oligodendrocytes laminin-2-α6β1 integrin interactions have been described to promote the transition from oligodendrocyte progenitor cells to mature MBP-expressing oligodendrocytes and to enhance survival of oligodendrocytes in response to limiting concentrations of growth factors (Baron et al. 2005, Colognato et al. 2002, Colognato et al. 2007, Frost et al. 1999). However, no significant changes in the number of MBP-expressing or surviving cells were noted in the studies presented here. These finding are in agreement with the data obtained from conditional FAK knock-out mice, in which no obvious changes in the number of oligodendrocytes were noted (Camara et al. 2009, Forrest et al. 2009). Thus, FAK, as a downstream target of laminin-2-α6β1 integrin interactions, is unlikely to play a significant role in regulating myelin protein gene expression and/or survival for the more advanced developmental stages of the oligodendrocyte lineage investigated here. This interpretation may be supported by the discovery that the src family kinase (SFK) regulatory proteins Csk and Cbp, protein components of a likely alternate signaling pathway to FAK, are critically involved in mediating the transition from oligodendrocyte progenitor cells to MBP-expressing oligodendrocytes downstream of laminin-2-α6β1 integrin interactions (Colognato et al. 2004, Relucio et al. 2009). Taken together, these findings, therefore, strongly suggest a crucial role of FAK in regulating the morphological maturation of post-migratory premyelinating oligodendrocytes by functioning as a downstream target of spatially restricted laminin-2-α6β1 integrin interactions that affect the organization of the cell’s cytoskeleton. Interestingly, dystroglycan has been recently identified as a second laminin receptor functionally involved in promoting oligodendrocyte maturation and myelination (Colognato et al. 2007). However, knock-down of dystroglycan was found to significantly affect both myelin gene expression as well as morphological maturation of oligodendrocytes. Thus, it is unclear to what extent laminin-2-dystroglycan interactions may contribute to the FAK-mediated mechanism described here.

The role of FAK in regulating oligodendrocyte morphology in the presence of both fibronectin and laminin-2 was seen to be largely dependent on FAK’s catalytic activity. Phosphorylation of FAK at its autophosphorylation site creates a high affinity binding site for Src-homology 2 (SH2) domain containing proteins, in particular SFKs (Cobb et al. 1994, Mitra et al. 2005, Mitra & Schlaepfer 2006, Parsons 2003). The binding of SFKs to FAK can lead to the formation of a FAK - activated SFK complex, which has the capability to act as a regulator of cell shape. In addition, it has been demonstrated that FAK can directly phosphorylate SFKs at their Y418 activation site (Wu et al. 2008). Two SFKs have been described to be expressed at significant levels in maturing oligodendrocytes, namely Fyn and Lyn (Colognato et al. 2004, Kramer et al. 1999, Osterhout et al. 1999). In agreement with these studies, both Fyn and Lyn were found expressed in post-migratory premyelinating oligodendrocytes at the maturation stages analyzed here (Fig. S2c-f). The SFK Fyn has been well documented to be involved in regulating morphological maturation of differentiating oligodendrocytes via the laminin-2-α6β1 integrin pathway (Colognato et al. 2004, Relucio et al. 2009). Taken together with the data presented here, morphological maturation of post-migratory premyelinating oligodendrocytes appears thus to be mediated by the concerted action of FAK and Fyn as downstream targets of laminin-2- α6β1 integrin interactions (see also Hoshina et al. 2007). In contrast to the laminin receptor α6β1 integrin, which has been shown to associate with Fyn, the fibronectin receptor αvβ3 integrin was found to associate with the SFK Lyn and not Fyn (Colognato et al. 2004). This association has been functionally implicated in the regulation of PDGF-mediated oligodendrocyte progenitor proliferation. The data presented here, could indicate an additional role for this association, namely a role in limiting the expansion of the process network extended by post-migratory premyelinating cells via a fibronectin-αvβ3 integrin-FAK-Lyn pathway.

The current study provides novel insight into the complex and diverse roles that FAK can play in regulating the morphology of post-migratory premyelinating oligodendrocytes. In vivo FAK knock-out was found to lead to a delay in myelination, a phenotype seen to be associated with a reduction in the number of primary oligodendroglial processes (Camara et al. 2009, Forrest et al. 2009). Based on the data presented here, this phenotype appears at least in part to be due to a lack of FAK’s maturation promoting role in the presence of laminin-2. In the context of demyelinating diseases, such as Multiple Sclerosis, FAK, therefore, presents itself as a good therapeutic target for promoting remyelination. However, demyelinated lesions are characterized by an extracellular environment that is different from the one found in the developing CNS. In particular, high levels of fibronectin were noted in lesions of Multiple Sclerosis patients (Satoh et al. 2009, Sobel & Mitchell 1989, van Horssen et al. 2007). Thus, FAK signaling may at least in part contribute to the limited repair seen in such lesions. Taken together, the findings presented here highlight the complexity of FAK’s role in regulating the maturation of post-migratory premyelinating oligodendrocytes, and they emphasize the importance of a better understanding of the signaling pathways involving FAK in order to be able to design effective therapeutic strategies for promoting remyelination under pathological demyelinating conditions.