|Home | About | Journals | Submit | Contact Us | Français|
Chondroitin sulfate proteoglycans (CSPGs) are major components of the extracellular matrix in the central nervous system (CNS) that inhibit axonal regeneration after CNS injury. Signaling pathways in neurons triggered by CSPGs are still largely unknown. In this study, using well-characterized in vitro assays for neurite outgrowth and neurite guidance, we demonstrate a major role for myosin II in the response of neurons to CSPGs. We found that the phosphorylation of myosin II regulatory light chains is increased by CSPGs. Specific inhibition of myosin II activity with blebbistatin allows growing neurites to cross onto CSPG-rich areas and increases the length of neurites of neurons growing on CSPGs. Using specific gene knockdown, we demonstrate selective roles for myosin IIA and IIB in these processes. Time lapse microscopy and cytochemistry reveal a novel synergistic effect between the inhibition of myosin II activity and inhibition of cell spreading by CSPGs, which results in the rapid initiation of neurites on CSPGs, leading to increased neurite outgrowth.
Failure of neurons in the central nervous system (CNS) to regenerate their axons after injury is attributed at least partly to the formation of a glial scar soon after injury. Within the glial scar there is a rapid and large upregulation of chondroitin sulfate proteoglycans (CSPGs) that serve to limit axonal regeneration and recovery of function. As growth cones enter CSPG-rich regions, they stop, turn or even retract to avoid the repulsive CSPGs (Silver & Miller 2004). These inhibitory actions of CSPGs can be modeled in vitro: neurites in culture growing on permissive substrates turn at a CSPG boundary, and neurite outgrowth is reduced when dissociated neurons are plated onto uniform substrates of CSPGs (Snow & Letourneau 1992; Wang et al. 2008).
The response of neurons to guidance cues is orchestrated by signal transduction mechanisms that ultimately result in changes in cytoskeletal dynamics, of which the major components are actin and tubulin. After contact with CSPGs, microtubules are rearranged within growth cones to accomplish turning. Application of either cytochalasin B to block actin filament formation or low doses of taxol to stabilize microtubules results in a reduction in growth cone turning to avoid CSPGs, indicating that both actin filaments and dynamic microtubules are required for growth cone turning (Challacombe et al. 1996; Challacombe et al. 1997). We have previously demonstrated that the rate and direction of tubulin polymerization are locally altered only within growth cones that encounter CSPGs (Kelly et al. 2010). These clearly demonstrate the presence of local signaling pathways triggered by CSPGs. However, the underlying mechanisms for cytoskeletal reorganization remain to be elucidated.
Non-muscle myosin II (hereafter, myosin II) is an actin binding protein that has been implicated in many aspects of cell adhesion and neuronal motility. It contributes to retrograde actin flow and actin filament reorganization and therefore induces neurite extension/retraction and also controls the direction of growth (Brown & Bridgman 2003; Lin et al. 1996; Medeiros et al. 2006; Wylie et al. 1998). Mammalian neurons express three isoforms of myosin II heavy chains (IIA, IIB and IIC), among which myosin IIB is the predominant form (Golomb et al. 2004). The activity of each of these isoforms is regulated by the phosphorylation of a common pair of regulatory myosin light chains (RMLC) which are bound to the myosin II heavy chains (Vicente-Manzanares et al. 2009). Although myosin II isoforms are similar in structure and are capable of partially substituting for each other (Bao et al. 2007), each isoform has been implicated in the performance of distinct tasks essential for normal neuronal function: myosin IIA has been reported to be responsible for neurite retraction and myosin IIB is required for neurite outgrowth (Wylie & Chantler 2003; Wylie et al. 1998). Myosin IIA mediates growth cone collapse and neurite retraction in response to repulsive guidance molecule (RGMa) (Kubo et al. 2008) and myosin IIB mediates growth cone turning at the border between laminin (an attractive guidance cue) and poly-L-ornithine (PLO) (Turney & Bridgman 2005). Intriguing is the involvement of both myosin IIA and IIB in response to semaphorin 3A: myosin IIA mediates the first phase response which results in growth cone collapse and myosin IIB is required for the second phase response which drives neurite retraction (Brown et al. 2009).
In this study, we used a neurite guidance assay and a neurite outgrowth assay to demonstrate that treatment of dissociated cerebellar granule neurons (CGNs) with blebbistatin, a specific inhibitor of myosin II, or knockdown of myosin IIA or IIB reduces the inhibitory activity of CSPGs. In addition to the role of myosin II in growth cone guidance and neurite outgrowth, we demonstrate a novel synergistic effect of CSPGs and blebbistatin on the actin cytoskeleton that appears to promote neurite initiation on CSPGs.
Cultures of dissociated mouse CGNs were prepared from postnatal 5–8 day old C57BL/6 mice as described previously (Wang et al. 2008). Cells were directly seeded on coverslips or subjected to small interfering RNA (siRNA) transfection before plating. Dissociated cells were cultured in Neurobasal-A medium containing B27 (1: 50, v/v) supplement and 25 mM KCl.
For immunostaining analysis of neuronal RMLC phosphorylation induced by CSPGs, dissociated CGNs were plated on coverslips coated with poly-L-lysine (PLL) or 1 μg/mL chicken CSPGs (Millipore, Temecula, CA) on PLL. After 24 h, cells were fixed with 4% PFA and stained with an antibody against Ser19 phosphorylated RMLC (1:200, Cell Signaling, Danvers, MA). For western blot analysis of neuronal RMLC phosphorylation, dissociated CGNs were first plated on 3.5 cm dishes at a density of 1.5×106 cells/dish. After 24 h, cells were treated with indicated concentrations of soluble CSPGs or the same volume of PBS for 6 h and 24 h. Cell lysates were collected and subjected to SDS-PAGE and blotted with an antibody against Ser19-phosphorylated RMLC (1:1000) or an antibody against total RMLC (1:1000, Cell Signaling).
Freshly isolated CGNs were transfected with Smart Pool siRNA oligonucleotides targeting either the mouse myosin IIA gene (Myh9), the mouse myosin IIB gene (Myh10) or with a control non-targeting siRNA pool with an unrelated sequence (Dharmacon RNA Technologies, Lafayette, CO). CGNs (3.5 ×105 cells/well) were electroporated with 10 pmol of siRNA using the Nucleofector® 96-well shuttle system (Lonza, Gaithersburg, MD) with the Basic 96-well Nucleofector® Kit for Primary Mammalian Neurons and the program DC102 following the manufacturer's instructions. Some of the transfected cells were plated in PLL-coated 12-well culture plates at a density of 1 × 106 cells/well for quantitative PCR (qPCR) or western blot analysis of gene knockdown; the remainder were seeded on coverslips in 24-well plates at a density of 8 × 104 cells/well for the neurite guidance assay or 4 × 104 cells/well for the neurite outgrowth assay.
For qPCR analysis of siRNA knockdown, total RNA was extracted 4 days after siRNA transfection using the Absolutely RNA miniprep kit following the manufacturer's protocol (Stratagene, La Jolla, CA). Genomic DNA was removed by DNase I treatment. RNA was reverse transcribed using SuperScript III reverse transcriptase (Invitrogen, Carlsbad, CA) and subjected to qPCR in a Chromo4 (Bio-Rad, Hercules, CA) with the DyNAmo HS SYBR Green kit (New England Biolabs, Ipswich, MA). PCR conditions consisted of a 15 min hot start at 95 °C followed by 45 cycles of the following sequence: 94 °C for 15 s, 59 °C for 15 s, and 72 °C for 25 s. The primer sequences are as follows: myosin IIA forward, 5'-ATTGTTCGGAAAGGCACTGGCGAC-3'; myosin IIA reverse, 5'-CAGATTCTCAAAGGGTATGGAGGAC-3'; myosin IIB forward, 5'-GCTTAGGGGTGGGTTTGGATTG-3'; myosin IIB reverse, 5'-AAGCCAGACACTCAGGAGACCAC-3'; glyceraldehyde-3-phosphate dehydrogenase (GAPDH) forward, 5'-AAGGTGGTGAAGCAGGCATCTG-3'; and GAPDH reverse, 5'-TGGGTGGTCCAGGGTTTCTTAC-3'. All samples were run in triplicate, and all mRNA levels were normalized to the level of GAPDH. cDNAs for myosin IIA, myosin IIB and GAPDH were used as templates for PCR to obtain standard curves. To compare the endogenous expression level of myosin IIA and myosin IIB in normal CGNs, the absolute amounts of myosin IIA, myosin IIB or GAPDH mRNA were determined by comparing each qPCR result to its corresponding standard curve produced by qPCR from serial dilutions of a known amount of the DNA template. The amounts of myosin IIA and myosin IIB were normalized to the amount of GAPDH.
For western blot analysis of siRNA knockdown, cell lysates were collected 4 days after siRNA transfection and then subjected to SDS-PAGE and blotted with rabbit anti-myosin IIA polyclonal antibody (1:500, Sigma), or rabbit anti-myosin IIB polyclonal antibody (1:500, Sigma), or a monoclonal anti-β-actin antibody (1:5000, sigma) as an internal control.
Neurite guidance assays were performed as described previously (Wang et al. 2008; Meiners et al. 1999). An interface between PLL and CSPGs was created by placing a 5 μL drop of chicken CSPG together with Texas Red in the center of a PLL-coated glass coverslip in a 24-well plate. Texas Red was used to visualize the interface and was used alone for negative control experiments. For blebbistatin treatment, dissociated CGNs were seeded onto the coverslips at a density of 6 ×104 cells/well, and cells were treated with 50 μM blebbistatin or an equal volume of dimethyl sulfoxide (DMSO) 4 h after plating. Cells were cultured for 2 days before fixed with 4% PFA. In case of siRNA knockdown experiments, siRNA-transfected cells were seeded on spotted coverslips at a density of 8 × 104 cells/well and cultured for 4 days before fixation. Cells were then stained with anti-β-tubulin III monoclonal antibody (1:1000, Sigma), followed by incubation with FITC-conjugated Goat anti-mouse IgG antibody (1:200, Jackson ImmunoResearch Laboratories, West Grove, PA). Fluorescence images were acquired on a Nikon E1000 upright microscope equipped with a Zeiss Axiocam HRc camera (Carl Zeiss, Thornwood, NY) controlled by Zeiss Axiovision software. The number of neurites crossing onto the entire CSPG spot was divided by the total number of neurites within 20 μm of the CSPG-PLL interface to get a ratio of neurite crossing. Only single, non-fasciculate neurite was considered for the analysis. In addition, only neurites growing toward the boundary were counted and no neurite whose soma was sitting on the interface was scored. Each experiment was performed in triplicate and repeated at least three times. Relative neurite crossing was obtained by normalizing to 0 μg/mL of CSPG.
For the neurite initiation assay, dissociated CGNs were treated with 50 μM blebbistatin or the same volume of DMSO and seeded on the PLL- or CSPG- (1 μg/mL) coated coverslips at a density of 2 × 104 cells/well and cultured for 20 h before fixation with 4% PFA. For the neurite elongation assay, dissociated CGNs were seeded on the PLL or CSPG (1 μg/mL) coated coverslips at a density of 2 × 104 cells/well. Twenty-four hours later, cells were treated with 50 μM blebbistatin or the same volume of DMSO and cultured for additional 24 h before fixation with 4% PFA. In case of siRNA knockdown experiments, transfected cells were seeded on the coverslips at a density of 4 × 104 cells/well and cultured for 4 days. Cells were stained with anti-β-tubulin III antibody and relative neurite length was measured by a stereological method (Rønn et al. 2000) as described previously (Wang et al. 2008). At least 30 random images from three independent experiments for each group were measured using the ImageJ program (available at: http://rsb.info.nih.gov/nih-image/). Data were normalized to either the DMSO-treated group growing on PLL or the control siRNA-transfected group growing on PLL.
Glass coverslips in 24-well plates or 35 mm dishes (MatTek, Ashland, MA) were coated with PLL or 1 μg/mL CSPGs in PLL. Dissociated CGNs pre-treated with 50 μM blebbistatin or the same volume of DMSO were seeded on the coated coverslips at a density of 6 × 104 cells/well. For cell attachment assay, cells were allowed to adhere to coverslips for 30 min and rinsed 3 times with PBS followed by fixation with 4% PFA. Cells were stained with DAPI. The number of attached cells in each random field was counted using ImageJ. For visualization of cell morphology, cells were directly fixed with 4% PFA without a pre-wash with PBS at 30 min, 2 h and 2.5 h after plating. Cells were stained with TexasRed-phalloidin and DAPI. For time-lapse recording, cells were plated into the 35 mm glass bottom dishes, and immediately placed onto a temperature-controlled stage on a Nikon Eclipse microscope and kept at 37°C and 5% CO2 for the duration of recording. Imaging with a 40 × 0.95 NA planapochromat phase contrast objective was begun immediately and images were taken at 1 min intervals using an Orca-ER cooled CCD camera (Hamamatsu Photonics, Bridgewater, NJ) under the control of Metamorph (Molecular Devices, Sunnyvale, CA) software.
Statistical analyses were performed either by one-way ANOVA with a Tukey post hoc test or by two-way ANOVA with Bonferroni post-tests to compare differences between individual groups using GraphPad Prism (GraphPad, San Diego, CA). Results were considered statistically significant if p < 0.05.
Myosin II activity is mainly regulated through phosphorylation of RMLC primarily on Ser19 (Vicente-Manzanares et al. 2009). To understand whether myosin II is involved in the inhibitory actions of CSPGs, we first evaluated changes in the phosphorylation of RMLC in response to CSPGs. Neurons growing on PLL showed a very low level of immunoreactivity to phospho-RMLC. In contrast, increased RMLC phosphorylation was detected both in the cell body and the neurites, especially in the growth cone area, of neurons plated on the 1μg/mL CSPG substrate for 24 h (Fig.1A, B).
To further confirm increased phosphorylation of RMLC by CSPGs, we performed western blot analysis of RMLC phosphorylation in response to soluble CSPGs. Because high concentrations of CSPGs inhibit cell adhesion, CGNs were allowed to adhere to PLL for 24 h before the addition of soluble CSPG. Under these conditions, there was no obvious cell loss. Phosphorylation of RMLC was increased 6 h after addition of 10 μg/mL CSPG and stayed high for at least 24 h (Fig. 1C). At lower concentrations (1μg/mL), the increase in RMLC phosphorylation was delayed, showing an increase after 24 h (Supplementary Fig.1).
Since CSPGs induced the activation of myosin II, we next sought to determine whether inhibiting myosin II activity could block the actions of CSPGs on neurons. We first introduced blebbistatin, a selective inhibitor of myosin II ATPase activity (Kovacs et al. 2004), to pharmacologically block myosin II activity. Application of blebbistatin itself showed no effect on phosphorylation of RMLC (Supplementary Fig.1A). We then tested if blebbistatin could alter neurite behavior at a boundary between PLL and CSPGs. Dissociated CGNs were plated onto PLL-coated coverslips that were spotted with CSPGs mixed with Texas Red to allow visualization. Control spots contained Texas Red alone. After 48 h, cultures were fixed and stained for β-tubulin III to visualize neurites. The behavior of neurites that were within 20 μm of the boundary was scored as either crossed or not crossed, and the percentages of neurites that crossed were calculated (Wang et al. 2008). As the CSPG concentration in the spot increased, the number of neurites that crossed was decreased dramatically, indicating increased repellant activity of CSPGs. Application of blebbistatin significantly promoted neurite crossing onto CSPG spots at all concentrations of CSPGs except the highest used (5 μg/mL), which showed few crossings both in control and treated groups (Fig. 2). Blebbistatin did not completely overcome the inhibitory activity of CSPGs: about 40% of the neurites in 1.25 μg/mL CSPGs and more than a half of the neurites in 2.5 μg/mL CSPGs still turned when encountering CSPGs at the border (Fig. 2B). This suggests that there may be other mechanisms in addition to myosin II controlling growth cone turning in response to CSPGs.
There are three isoforms of myosin II in neurons and the predominant isoforms are myosin IIA and myosin IIB (Rochlin et al. 1995). As in the whole mouse cerebellum (Ma et al. 2010), myosin IIB was more abundant than myosin IIA in cultured mouse CGNs. The mRNA level of myosin IIB was about 25 fold higher than that of myosin IIA by qPCR (Fig. 3A). To investigate whether myosin IIA or IIB (or both) isoforms are involved in growth cone turning induced by CSPGs, we selectively reduced the levels of myosin IIA or myosin IIB using siRNAs targeting each isoform into neurons. Western blot analysis 4 days after transfection into CGNs confirmed the specificity of each siRNA for its target (Fig. 3B).
We then performed a neurite guidance assay by plating siRNA-transfected cells on PLL-coated coverslips with spots of different concentrations of CSPG (0–4 μg/mL) and allowed neurons to grow for 4 days. In each experiment, a specific knockdown of the level of myosin IIA or IIB was confirmed using qPCR in parallel cultures (Fig. 3C). No difference in neurite crossing onto the control spots without CSPGs was found among neurons transfected with control, myosin IIA, or IIB siRNAs. At 2 μg/mL of CSPGs, both siRNA targeting IIA and IIB resulted in an increased percentage of neurite crossing compared to control siRNA (Fig. 3D, E). The inhibitory activity of 4 μg/mL CSPGs was maximal, and, similarly to blebbistatin, we found no crossing in all the groups. Thus, both myosin IIA and IIB appear to be involved in the process of neuronal guidance in response to CSPGs.
Neurite guidance and neurite outgrowth are different phenomena and may be controlled through different processes (Powell & Geller 1999; Powell et al. 1997). We therefore evaluated whether myosin II activity is required for the inhibition of neurite outgrowth by CSPGs. It is noteworthy that CSPGs are highly negatively charged and hence reduce cell adhesion in vitro. In fact, we found a significant loss of CGNs when the concentration of CSPGs used for coating was higher than 4 μg/mL (data not shown). The anti-adhesive action also resulted in cells forming clusters in the CSPG-treated group, which prevented accurate measurement of neurite outgrowth.
However, at 1 μg/mL CSPG, there was no obvious difference in cell density between neurons plated onto PLL or CSPG substrates. We therefore used 1 μg/mL CSPGs for the subsequent neurite outgrowth assays. CGNs were treated with either blebbistatin or same volume of DMSO and plated on coverslips coated with PLL or PLL plus CSPGs. In the DMSO control group, 51.20 ± 9.57% of the neurons growing on PLL extended neurites that were longer than one cell-body diameter 20 h after plating, while the percentage of neurite bearing cells growing on CSPGs was significantly reduced to 33.90 ± 8.99% (Fig. 4). However, when treated with blebbistatin, the majority of the neurons either on PLL or CSPGs had neurites extending longer than one cell body. Blebbistatin treatment increased the percentage of neurite-bearing cells on PLL and CSPGs to 84.56 ± 8.36% and 81.12 ± 8.47%, respectively (Fig. 4B). Since the percentage of neurons bearing neurites 20 h after plating reflects the extent of neurite initiation, blebbistatin appeared to promote neurite initiation regardless of substrate.
A similar result was obtained in terms of neurite length (Fig. 4C): neurons plated on CSPGs extended shorter neurites with more than a 40% reduction in neurite length compared to PLL 20 h after plating. In contrast, application of blebbistatin significantly promoted neurite outgrowth both on PLL and on CSPG substrates (Fig. 4C). Notably, neurite length on inhibitory CSPGs was comparable to that on PLL after treatment of neurons with blebbistatin, suggesting that inhibition of myosin II activity by blebbistatin not only completely overcomes the effect of CSPGs on neurite outgrowth but also keeps promoting neurite growth irrespective of the presence of inhibitory CSPGs. It is also noteworthy that application of blebbistatin 1 day after cell plating increased neurite outgrowth both on PLL and CSPGs (Figure 5). The finding that neurite outgrowth is promoted no matter when blebbistatin is applied suggests that inhibition of myosin II promotes neurite elongation as well as initiation.
To further investigate which isoforms of myosin II are responsible for neurite outgrowth inhibition induced by CSPGs, we transfected neurons with siRNAs targeting myosin IIA or myosin IIB and plated them on either PLL or CSPG substrates. On PLL, depletion of myosin IIA significantly increased neurite outgrowth while myosin IIB depletion showed a decrease in neurite length. However, when neurons were plated on CSPGs, both IIA siRNA and IIB siRNA promoted neurite outgrowth (Fig. 6). This suggests that myosin IIB is serving to promote neurite extension on PLL, while both myosin IIA and IIB attenuate outgrowth on inhibitory CSPGs.
Once neurons are plated on dishes, they attach, spread, and initiate neurites, followed by elongation of neurites. The roles of myosin II in neurite outgrowth and guidance have been extensively studied in different types of neuronal cells as well as on different substrates, most of which focused on changes in growth cone behaviors (Brown et al. 2009; Hur et al, 2011; Turney & Bridgman 2005). Myosin II regulates cytoskeleton reorganization within the growth cone which controls turning and also the rate of neurite outgrowth. However, the role of myosin II in neurite initiation from the cell body has been less well documented. Our finding showed that blebbistatin promoted neurite initiation both on PLL and CSPGs (Figure 4). Given blebbistatin treatment directly promotes neurite initiation on PLL, this accelerated neurite initiation on CSPGs by blebbistatin could be at least partly due to an effect on neuronal attachment or spreading as negatively charged CSPGs might inhibit or delay these processes. To exclude this possibility, we first analyzed cell attachment with the same images used for neurite outgrowth assays. No difference was found in the numbers of attached cells among blebbistatin treated or untreated groups plating on either PLL or 1 μg/mL CSPGs after 20 h incubation. However, it is still possible that neurons attach and spread more slowly on CSPGs than on PLL and therefore start neurite initiation later as compared to PLL. We therefore performed cell attachment assays on both PLL and CSPG substrates at short time points. CGNs pretreated with blebbistatin were plated and incubated at 37°C for 30 min, and unattached cells were removed by rinsing with PBS. The number of cells attached to CSPGs was significantly reduced as compared to that on PLL 30 min after plating and blebbistatin treatment failed to increase neuronal attachment to CSPGs (Supplementary Fig. 2). This result together with our neurite outgrowth data indicates inhibition of myosin II is effective in initiating neurites both from firmly attached and loosely attached neurons.
To further study the effect of myosin II inhibition on neuronal attachment, spreading, and neurite initiation on the CSPG substrate, we monitored initial adhesion and outgrowth using time-lapse microscopy. CGNs rapidly spread on PLL (within minutes) by elaborating a lamellum (Fig. 7 and Supplementary Movie 1). Microspikes were observed to actively extend and retract from the cell body but fail to give rise to a neurite. In contrast, it took a much longer time for neurons plated on CSPGs to attach. Even after neurons attached, most of them remained round with limited spreading and no neurites were formed by the end of imaging (Fig. 7 and Supplementary Movie 2). Neurons treated with blebbistatin accelerated neurite initiation both on PLL and CSPGs. Some neurons on PLL first spread out to form a lamellum and then generate neurites quickly, and some even initiated neurites directly without the formation of a lamellum (Fig. 7 and Supplementary Movie 3). Blebbistatin treated neurons on CSPGs did not exhibit a lamellum but extended neurites rapidly after attaching (Fig. 7 and Supplementary Movie 4).
In order to more fully evaluate this process, cultures were fixed at 30 min, 2 or 2.5 h after plating, and stained with Texas Red-phalloidin and DAPI. Images of the stained cells are presented in Fig. 8A. Consistent with the time-lapse observations, most of the neurons plated onto PLL were highly spread 30 min after plating, forming an actin-rich lamellum with a pancake-like shape. In contrast, neurons plated onto CSPGs remained round with less spreading or no spreading. At 2 h and 2.5 h time points, many well spread neurons on PLL or low spread neurons on CSPGs had extended somatic microspikes or filopodia, but only a few neurons had generated neurites. However, a significant number of neurons exposed to blebbistatin had initiated neurites either on PLL or CSPGs at this time points (Fig. 7A).
To quantify this phenomenon, we classified these morphologies into four phases (Fig. 8B): phase 0, a round cell without spreading; phase 1, cell spreading with “pancake” type morphology; phase 2, somatic microspikes or filopodia but without neurites; and phase 3, neurite initiation. The classification data are presented in Fig. 8C. Thirty min after plating on PLL, the majority of the neurons are in phase 1. Treatment with blebbistatin speeded up neurite initiation on PLL, so that at 30 min after plating about 25% of the neurons are at phase 2 or phase 3. Thirty min after plating on CSPGs, about half the cells are in phase 0, while another half are in phase 3, with only a few cells at phase 2. This means cell spreading was greatly inhibited by CSPGs, as a result, these less spread cells tend to extend many filopodia (Phase 2) but they failed to form neurites probably due to the neurite outgrowth inhibitory activity of CSPGs. Interestingly, treating cells with blebbistatin did not overcome the inhibition of CSPGs on cell spreading but did overcome the inhibition of CSPGs on neurite outgrowth, which resulted in increased numbers of cells in Phase 2 and Phase 3. This effect of blebbistatin on neurite initiation was more apparent at 2 h, where nearly 40% of the cells plated on CSPG had generated neurites (Fig. 8C). By 2.5 h, compared to less than 10% in DMSO controls, about 40% blebbistatin treated neurons on PLL and about 35% blebbistatin treated neurons on CSPGs had extend neurites longer than one diameter of cell body (Fig. 8D). These results indicate that high cell spreading virtually delays neurite initiation. In contrast, inhibition of cell spreading by CSPGs synergizes with neurite growth promotion following myosin II inhibition to accelerate neurite initiation.
After injury to the mammalian CNS, CSPGs are one of the major components that create a barrier to axonal growth, preventing effective regeneration and recovery of function. Growing axons respond to CSPGs by altering their rate and direction of extension, both of which require coordinated cytoskeleton reorganization in neurons. We present several lines of evidence showing that myosin II plays pivotal roles in initiation, outgrowth and guidance of neurites in response to inhibitory CSPGs. First, CSPGs induced phosphorylation of RMLC, which activates myosin II and is one of the main factors involved in the regulation of cytoskeleton dynamics. Second, pharmacological inhibition of myosin II activity by blebbistatin caused neurites to ignore the negative guidance cues presented by CSPGs, resulting in increased crossing at the interface between CSPGs and PLL. Third, blebbistatin promoted neurite outgrowth irrespective of the presence of inhibitory CSPGs as substrates. Fourth, gene silencing of individual isoforms of myosin II revealed that both myosin IIA and IIB are involved in axonal guidance and neurite outgrowth. Finally, although blebbistatin did not alter neuronal cell attachment to PLL and CSPGs, the blockage of myosin II activity accelerated the initiation of neurites.
There are three isoforms of myosin II: IIA, IIB and IIC and myosin IIB is predominant in neurons. Which isoforms are involved in control of neurite growth and response to guidance cues are still a matter of intense investigation. Using gene silencing approaches, we found that the combination of individual myosin isoforms and substrates used for neuron culture is important: knockdown of myosin IIA promoted neurite outgrowth while knockdown of myosin IIB inhibited outgrowth on permissive cues such as PLL and laminin, consistent with previous publications (Bridgman et al. 2001; Wylie & Chantler 2003; Wylie et al. 1998; Turney & Bridgman 2005; Hur et al. 2011). In contrast, knockdown of either myosin IIA or IIB reduced the response to CSPGs, which is in agreement with a very recent report by Hur et al. (2011) on the role of myosin II in DRG neurons plated onto laminin. Other data suggest that the role of myosin II in the response to inhibitory molecules may not be consistent: Kubo et al. (2008) reported that knockdown of myosin IIA, but not IIB, reduced the inhibition of neurite outgrowth by RGMa, while Brown et al. (2009) reported neurons with a knockout of myosin IIB showed a significantly reduced retraction rate in response to semaphorin 3A. Overall, these findings indicate that myosin IIA is generally involved in slowing growth both on permissive or non-permissive substrates, while myosin IIB has opposite actions depending upon the substrate.
Time-lapse microscopy revealed a surprising synergism in neurite initiation between the inhibition of myosin II activity by blebbistatin and inhibition of cell spreading by CSPGs. Similarly, after blebbistatin treatment, axons of DRG neurons on CSPGs grow longer than axons growing on laminin (Hur et al. 2011). Myosin II activity accelerates retrograde actin flow and actin depolymerization (Medeiros et al. 2006), and is also necessary for the stabilization of focal contacts (Chrzanowska-Wodnicka & Burridge 1996). Both processes inhibit neurite growth: stabilization of focal adhesions reduces mobility, and retrograde actin flow inhibits the extension of microtubules into the periphery of the growth cone and is involved in growth cone retraction (Kalil & Dent 2005). Blebbistatin antagonizes both actions of myosin II, reducing adhesion and retrograde actin flow (Koch et al. 2011). CSPGs also inhibit cellular adhesion and cells on CSPGs do not form the actin-rich lamella that is the basis for retrograde actin flow. It appears that under the circumstances of poor adhesion, microtubule protrusions are not prevented from extending, and so neurites are generated much earlier than on more adhesive substrates. This action, coupled with a reduction in adhesion, then may explain the ability of blebbistatin and myosin II inhibitors to promote neurite outgrowth on CSPGs. Further studies are required to clarify the correlation between the extent of cell spreading and the speed of neurite initiation.
Unlike many other inhibitory molecules which induce growth cone collapse, such as myelin-associated inhibitors and semaphorins (Bandtlow et al. 1993; Jin & Strittmatter 1997), growth cones that encounter CSPG boundaries turn to avoid the inhibitory cue but do not collapse (Snow & Letourneau 1992; Wang et al. 2008). Growth cone turning requires the coordination of the dynamics of both actin localized at the periphery and microtubule at the central region of the growth cone. This implies the presence of a local signaling cascade triggered by CSPGs in the growth cone to guide the direction of growing neurite. Our previous data demonstrate such a local mechanism for tubulin polymerization at a boundary (Kelly et al. 2010). Attractive guidance cues do initiate a local promotion of actin polymerization on the side of the attractive stimulus (Marsick et al. 2010); a similar increase in F-actin can be achieved by altering retrograde actin flow (Schaefer et al. 2008). It might be expected that repulsive cues would have an opposite effect, either locally decreasing F-actin polymerization or increasing retrograde actin flow. Since myosin II increases retrograde actin flow in growth cones (Lin et al. 1996), a local activation of myosin II by CSPGs would cause growth cone turning away from the stimulus. In contrast, blebbistatin reduces F-actin levels (Hur et al. 2011), which is consistent with a requirement of polymerized actin for turning.
The mechanisms by which CSPGs or other repulsive molecules slow neurite growth are not well understood. Recently, the receptor tyrosine phosphatases σ (RPTPσ) and LAR have been identified as putative receptors for CSPGs (Fry et al. 2010; Shen et al. 2009; Fisher et al. 2011). However, the intracellular signal transduction pathway downstream of RPTPσ is not known. We found that exposure to soluble CSPGs produced a sustained increase in myosin light chain phosphorylation. Many guidance cues affect the contraction of myosin II by modulating the balance between Rho and Rac and/or Cdc42 activities (Amano et al. 1998; Driessens et al. 2001). However, it was recently shown that lysophosphatidic acid (LPA), which enhances Rho-dependent myosin contractility, induces growth cone retraction independent of peripheral retrograde actin flow but rather driven by the contraction of more central actin structures (Zhang et al. 2003). We found that treatment with the Rho kinase inhibitor Y-27632 or the myosin light chain kinase inhibitor ML-7 did not alter myosin light chain phosphorylation in response to CSPGs (P. Y, unpublished). A lack of involvement of Rho is also supported by the minimal change in axonal growth on CSPGs when neurons are treated with Y-27632 (Hur et al. 2011). These results suggest that CSPGs may affect myosin light chain kinase or phosphatase through pathways that do not involve Rho or MLCK.
Our results, together with others, indicate that myosin II activity is involved in growth cone turning and neurite outgrowth in response to both positive and negative guidance cues. Although blockage of myosin II activity overcomes inhibition caused by CSPGs and other inhibitors which prevent axonal regeneration after CNS injury, it might not be a good target for promoting axonal regeneration, since myosin II also mediates the positive response towards permissive cues like laminin. Further investigations into more selective inhibitors of myosin II isoforms or other members of the signaling cascade that involves myosin II are therefore warranted.
We acknowledge the expert assistance of the Light Microscopy Imaging Core of the National Heart, Lung, and Blood Institute Division of Intramural Research. We acknowledge Dr. Abigail Mabe and Elisa Gutierrez for the technical assistance.