Retracted: Ethanol disrupts axon outgrowth stimulated by netrin-1, GDNF, and L1 by blocking their convergent activation of Src family kinase signaling

Authors

Errata

This article is corrected by:

  1. Errata: Retraction Volume 132, Issue 6, 756, Article first published online: 6 February 2015

Abstract

Pre-natal alcohol exposure causes fetal alcohol spectrum disorders (FASD), the most common, preventable cause of developmental disability. The developing cerebellum is particularly vulnerable to the effects of ethanol. We reported that ethanol inhibits the stimulation of axon outgrowth in cerebellar granule neurons (CGN) by NAP, an active motif of activity-dependent neuroprotective protein (ADNP), by blocking NAP activation of Fyn kinase and its downstream signaling molecule, the scaffolding protein Cas. Here, we asked whether ethanol inhibits the stimulation of axon outgrowth by diverse axon guidance molecules through a common action on the Src family kinases (SFK). We first demonstrated that netrin-1, glial cell line-derived neurotrophic factor (GDNF), and neural cell adhesion molecule L1 stimulate axon outgrowth in CGNs by activating SFK, Cas, and extracellular signal-regulated kinase 1 and 2 (ERK1/2). The specific SFK inhibitor, PP2, blocked the stimulation of axon outgrowth and the activation of the SFK-Cas-ERK1/2 signaling pathway by each of these axon-guidance molecules. In contrast, brain-derived neurotrophic factor (BDNF) stimulated axon outgrowth and activated ERK1/2 without first activating SFK or Cas. Clinically relevant concentrations of ethanol inhibited axon outgrowth and the activation of the SFK-Cas-ERK1/2 pathway by netrin-1, GDNF, and L1, but did not disrupt BDNF-induced axon outgrowth or ERK1/2 activation. These results indicate that SFK, but not ERK1/2, is a primary target for ethanol inhibition of axon outgrowth. The ability of ethanol to block the convergent activation of the SFK-Cas-ERK1/2 pathway by netrin-1, GDNF, L1, and ADNP could contribute significantly to the pathogenesis of FASD.

Abbreviations used
ADNP

activity-dependent neuroprotective protein

BDNF

brain-derived neurotrophic factor

BSA

bovine serum albumin

Cas

Crk-associated substrate

CGN

cerebellar granule neuron

ERK1/2

extracellular signal-regulated kinase 1 and 2

EtOH

ethanol

FCS

fetal calf serum

FASD

fetal alcohol spectrum disorders

GDNF

glial cell line-derived neurotrophic factor

HBSS

Hank's balanced salt solution

L1

neural cell adhesion molecule L1

L1-Fc

extracellular domain of human L1 fused to Fc from human IgG

MAP kinase

mitogen-activated protein kinase

NAP

NAPVSIPQ

PD7

post-natal day seven

PBS

phosphate-buffered saline

pERK1/2

dually phosphorylated (Thr183 and Tyr185) ERK1/2

pY410Cas

phosphorylated Tyr410 Cas

pY416SFK

phosphorylated Tyr416 Src family kinase

PP2

4-Amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo[3,4-d]pyrimidine

PLL

poly-l-lysine

SDS

sodium dodecyl sulfate

SFK

Src family kinase

TBST

Tris-buffered saline/0.01% Tween 20

Pre-natal alcohol exposure causes fetal alcohol spectrum disorders (FASD), a common condition characterized by cognitive and motor impairment, behavioral abnormalities, and brain and facial dysmorphology (Hoyme et al. 2005). The cognitive and neurological abnormalities of FASD arise in part from abnormal cerebellar neural circuit formation (Riley et al. 2004; Jaatinen and Rintala 2008). Cerebellar development is critically dependent on the genesis, migration, and axon outgrowth of cerebellar granule neurons (CGN), the most abundant cell type in the cerebellum (Wang and Zoghbi 2001); therefore, ethanol disruption of CGN development could contribute to the pathogenesis of FASD. Indeed, ethanol inhibits axon outgrowth in CGNs (Bearer et al. 1999; Chen and Charness 2008); however, the underlying molecular mechanisms remain unclear.

Three signaling molecules play an important role in the stimulation of axon outgrowth by diverse axon-guidance molecules: Src family kinases (SFK) (Sariola and Saarma 2003; Liu et al. 2004; Maness and Schachner 2007; Chen and Charness 2008), the scaffolding protein Crk-associated substrate (Cas) (Huang et al. 2006; Liu et al. 2007; Chen and Charness 2008), and mitogen-activated protein (MAP) kinases, especially extracellular signal-regulated kinases 1 and 2 (ERK1/2) (Schmid et al. 2000; Forcet et al. 2002; Chen et al. 2003; Loers et al. 2005; Tucker et al. 2008; Guimond et al. 2010). We recently demonstrated that ethanol inhibits axon outgrowth in CGNs induced by the peptide NAP (NAPVSIPQ), an active fragment of activity-dependent neuroprotective peptide (ADNP), by blocking NAP's sequential activation of Fyn kinase and Cas (Chen and Charness 2008). Subsequently, Yeaney and colleagues (Yeaney et al. 2009) showed that ethanol inhibits the sequential activation of Src and ERK1/2 by the L1 neural cell adhesion molecule. These results suggest that ethanol disrupts ligand induction of CGN axon outgrowth through actions on SFK, Cas, and ERK1/2. However, it remains uncertain whether one or more of these molecules are primary targets of ethanol and whether a specific pathway is required for ethanol disruption of axon outgrowth by diverse guidance molecules.

To answer these questions, we studied the effects of four axon guidance molecules on axon outgrowth in CGNs. Netrin-1, glial cell line-derived neurotrophic factor (GDNF), L1, and brain-derived neurotrophic factor (BDNF) each play independent and critical roles in the development of the nervous system (Airaksinen and Saarma 2002; Maness and Schachner 2007; Paratcha and Ledda 2008; Numakawa et al. 2010; Lai Wing Sun et al. 2011). All four molecules and their receptors are highly expressed in the developing cerebellum (Rathjen and Rutishauser 1984; Segal et al. 1995; Trupp et al. 1997; Burazin and Gundlach 1999; Livesey 1999). Each molecule promotes neurite outgrowth by binding to unique cognate receptors: netrin-1 to Deleted in Colon Cancer (DCC); GDNF to GFRα1 and Ret; L1 to L1 molecules in cis or trans; and BDNF to TrkB (Forcet et al. 2002; Maness and Schachner 2007; Paratcha and Ledda 2008; Numakawa et al. 2010). There are also important differences in the signaling of these molecules. Netrin-1, GDNF, and L1, each activate both SFK and MAP kinases, whereas BDNF activates MAP kinase through its interactions with TrkB. Here, we show that clinically relevant concentrations of ethanol inhibit axon outgrowth and the activation of the SFK-Cas-ERK1/2 signaling pathway by netrin-1, GDNF, and L1, but do not disrupt BDNF-induced axon outgrowth or ERK1/2 activation. These results indicate that SFK is a primary target for ethanol inhibition of ligand-stimulated axon outgrowth.

Materials and methods

Reagents and animals

Recombinant chicken netrin-1, human glial cell line-derived neurotrophic factor (GDNF), and human brain-derived neurotrophic factor (BDNF) were from R&D Systems (Minneapolis, MN, USA). NAP (NAPVSIPQ) was from New England Peptide (Gardner, MA, USA). Human L1-Fc protein, which contains the extracellular domain of human L1 fused to the Fc portion of human IgG (Haspel et al. 2001), was produced in HEK 293 cells and purified as described (Chen et al. 1999). Neurobasal medium, fetal calf serum (FCS), Hank's balance salt solution (HBSS), and L-glutamine were from Invitrogen (Carlsbad, CA, USA). Trypsin and DNase I were from Worthington Biochemical Corporation (Lakewood, NJ, USA). Polyclonal antibodies against phospho-Tyr416 Src family (pY416SFK) and phospho-Tyr410 Cas (pY410Cas) were from Cell Signaling (Danvers, MA, USA). Monoclonal antibodies against Fyn kinase and Cas were from BD Bioscience (San Diego, CA, USA). Polyclonal antibodies against Fyn kinase and ERK1/2 and protein A/G agarose were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Monoclonal antibodies against Tau-1 and ReBlot plus antibody stripping solution were from Chemicon (Temecula, CA, USA). Src family kinase inhibitor PP2 (4-amino-5-(4-chlorophenyl)-7-(dimethylethyl)pyrazolo[3,4-d]pyrimidine) and proteinase inhibitors cocktail were from Roche (Palo Alto, CA, USA). Mini-Protean Precast Gels and nitrocellulose membrane were from Bio-Rad (Hercules, CA, USA). Tris-buffered saline/0.01% Tween 20 (TBST) (50 mM Tris-HCl, pH 7.4; 150 mM NaCl; 0.1% Tween-20) and Laemmli's sodium dodecyl sulfate (SDS)-sample buffer were from Boston BioProducts (Ashland, MA, USA). All secondary antibodies were from Jackson ImmunoResearch Laboratories (West Grove, PA, USA). Halt phosphatase inhibitors cocktail and ECL Plus Western Blotting Substrate were from Thermo Scientific (Rockford, IL, USA). Ethanol, poly-l-Lysine (PLL), bovine serum albumin (BSA), insulin, transferrin, thyroxine, sodium selenite, aprotinin, human Fc, monoclonal antibodies against dually phosphorylated (Thr183 and Tyr185) ERK1/2 (pERK1/2), and all other reagents were obtained from Sigma-Aldrich (St. Louis, MO, USA), unless otherwise stated. Tissue culture plates were from Nunc (Roskilde, Denmark). Glass-bottom 24-well plates were from MatTek (Ashland, MA, USA). Newborn Sprague–Dawley (CD) rats were purchased from Charles River Laboratories (Wilmington, MA, USA) and were housed and treated according to the National Institutes of Health Guidelines for the Use and Care of Laboratory Animals and an approved Animal Care and Use Committee protocol from the Veterans Affairs Boston Healthcare System.

Culture of CGNs

CGNs were prepared from cerebella of post-natal day 7 (PD7) rats, as described (Keilhauer et al. 1985), with modifications. Briefly, the cerebella were cut into small pieces, incubated in 1% trypsin/0.05% DNase for 16 min at 22°C, washed with HBSS, and resuspended in a 0.05% DNase solution. Cells were dissociated in DNase solution by mechanical trituration. The CGNs were isolated using a cushion made with 15% FCS in Neurobasal medium on the bottom and HBSS on the top and centrifuged at 800 g. The pellets of CGNs were washed once with HBSS and then culture medium before plating. Cells were plated on PLL-coated culture plates and maintained in Neurobasal medium supplemented with 2 mM l-glutamine, 1 mg/mL BSA, 12.5 μg/mL insulin, 4 nM thyroxine, 100 μg/mL transferrin, 30 nM sodium selenite, and 0.6 units of aprotinin.

Quantification of axon outgrowth and axon initiation

Dissociated CGNs were plated onto glass-bottom 24-well plates at a density of 105 cells/well for axon outgrowth assays. Netrin-1 (200 ng/mL), GDNF (50 ng/mL), or BDNF (50 ng/mL) and ethanol were added to culture medium 2 h after cell plating to avoid affecting initial adhesion of neurons to the substrate. L1-Fc (10 μg/mL) was coated on the PLL-coated plate as a substrate for L1-Fc-treated CGN axon outgrowth. Human Fc (10 μg/mL) was used as a control for L1-Fc experiments. The CGNs were maintained in culture for an additional 20 h and processed 22 h after plating for immunofluorescence with anti-Tau antibody. The corresponding images of the bright field, the Tau-immunofluorescence, and the stained nuclei were captured from each field. Axon outgrowth and axon initiation were determined by axon length, and by the percentage of neurons that bore axons, respectively. Axon length was measured in 10–20 image fields using Openlab software (Improvision, Waltham, MA, USA), as described (Chen and Charness 2008). Axon initiation was calculated based on the percentage of CGNs with axons of at least one cell diameter among all CGNs in 10–20 image fields. The great majority of CGNs expressed a single Tau-positive neurite. At least 50 CGNs were analyzed in each experiment, and each experiment was performed at least three times.

Immunofluorescence

CGNs were washed with cold phosphate-buffered saline (PBS) and fixed in 4% paraformaldehyde for 30 min, washed with PBS, and blocked for 1 h at 22°C in PBS containing 0.1% Triton X-100 and 3% BSA. The CGNs were incubated overnight at 4°C with anti-Tau-1 antibody in PBS containing 1% BSA, washed, and incubated for 1 h at 22°C with Cy3-donkey anti-mouse IgG secondary antibody. After washing, the cells were incubated for 2 min with 1 μg/mL of Hoechst 33342 to visualize nuclei. Images were acquired using a Nikon fluorescence microscope using a digital camera (Hamamatsu, Model C4742-95) and imaging software (OpenLab). The representative images from each experiment were exported and assembled using Adobe Photoshop.

Cell lysate preparation

Dissociated CGNs were plated onto each well of a six-well tissue culture plate at a concentration of four million cells per 3 mL. Twenty-four hours after plating, CGNs were treated for 30 min with netrin-1 (200 ng/mL), GDNF (50 ng/mL), BDNF (50 ng/mL), L1-Fc (10 μg/mL), or NAP (10-12M) in the presence and absence of ethanol (25 mM or 100 mM) or PP2 (5 μM). Cells were washed with ice-cold PBS and lysed on ice for 30 min in 20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mm EDTA, 1% Triton X-100 with cocktails of proteinase inhibitors and phosphatase inhibitors, sonicated for 1 min on ice, and centrifuged at 16 000 g for 20 min. The supernatants were collected, and the protein concentration in the supernatant was determined by the bicinchinonic acid method (Thermo Scientific). Cell lysates were processed for western blot analysis or kept frozen at −20°C until use.

Western blot analysis

CGN cell lysates were boiled in Laemmli's SDS-sample buffer, separated on a 4–15% gradient pre-cast gel, and transferred to nitrocellulose membranes. The membranes were blocked for 1 h at 22°C with 3% low-fat milk powder and 2% BSA in TBST and incubated overnight at 4°C in blocking solution containing primary antibodies against pY416SFK, pY410Cas, or pERK1/2. Membranes were washed with TBST and incubated for 1 h at 22°C with horseradish peroxidase-conjugated secondary antibodies. Immunolabeling was detected with Pierce ECL Plus Western Blotting Substrate. The same membrane was stripped with ReBlot plus antibody stripping solution and reprobed with antibodies against total Fyn, Cas, or ERK1/2. Western blots were scanned, assembled using Adobe Photoshop, and exported as TIFF images. For selected gels, lanes that were not relevant to the reported findings were digitally removed to allow clearer presentation of pertinent data. The densitometry analysis of western blots captured as TIFF images was performed using TINA Image software (Version 2.0, Raytest, Straubenhardt, Germany).

Statistical analysis

Data are presented as the mean ± SEM from at least three independent experiments performed in triplicate in axon outgrowth studies and at least three independent experiments in quantification of western blotting. Statistical analysis of the data was performed using the Student's t-test.

Results

Ethanol inhibits axon outgrowth mediated by netrin-1, GDNF, and L1, but not BDNF

Although all four molecules increase axon outgrowth in different regions of the central nervous system, neither netrin-1 nor GDNF has been shown to do so in CGNs. Therefore, we first characterized the effects of these four molecules on axon outgrowth in CGNs maintained under identical culture conditions. CGNs from PD7 rats were cultured on PLL in serum-free medium and treated with netrin-1, GDNF, L1-Fc, and BDNF. Cells were fixed 20 h after treatment and stained with Tau-1 antibody to label axons.

The majority of neurons bore a single neurite labeled with Tau-1 antibody; under our culture conditions, the Tau-1 antibody also stained the CGN cell body (Fig. 1a). We quantified the effects of each molecule on axon elongation, as determined by axon length, and axon initiation, as determined by the percentage of neurons that bore axons. The average length of axons in control CGNs was 23.8 ± 2.8 μm (n = 8), and 37.2 ± 1.4% (n = 9) of control CGNs' extended axons. All four molecules significantly increased axon initiation and elongation (Fig. 1; Table S1).

Figure 1.

Effect of ligands and ethanol on axon outgrowth. Cerebellar granule neurons (CGNs) from PD7 rats were cultured under serum-free conditions for 20 h in the absence and presence of 200 ng/mL netrin-1, 50 ng/mL glial cell line-derived neurotrophic factor (GDNF), 10 μg/mL L1-Fc, 50 ng/mL brain-derived neurotrophic factor (BDNF), and 25 mM ethanol (EtOH)(same concentrations in all figures). (a) Representative immunofluorescence micrographs of CGNs stained with Tau-1 antibody following different ligand treatments. Scale bar, 10 μm. (b, c) Bars depict the mean ± SEM for (b) axon length of individual neurons (four to eight independent experiments); and (c) the percentage of neurons with neurites (four to nine independent experiments), following the indicated treatments. For both b and c: open bar, control; black bar, ligand; gray bar, ligand plus ethanol. *p < 0.05; **p < 0.01; compared with untreated control; #p < 0.05 differences in the absence or presence of ethanol.

The effect of ethanol on CGN axons was tested using 25 mM, a concentration attained in the blood of women after ingesting an average of three alcoholic beverages within 1 h (Fisher et al. 1987). Treatment of CGNs for 20 h with either 25 mM or 100 mM ethanol did not decrease cell survival (data not shown). Ethanol significantly decreased the induction of axon initiation and elongation by netrin-1, GDNF, and L1, but had no effect in the absence of these ligands (Fig. 1; Table S1). In contrast, neither 25 mM ethanol (Fig. 1) nor 100 mM ethanol (data not shown) altered BDNF-mediated axon initiation and elongation. These results indicate that ethanol inhibition of axon outgrowth is specific for selected ligands. To investigate whether the effects of ethanol on axon outgrowth are mediated through specific ligand-activated signaling pathways, we first characterized the effects of netrin-1, GDNF, L1, and BDNF on the activation of the SFK-Cas-ERK1/2 signaling pathway in CGNs.

SFK is activated by netrin-1, GDNF, and L1, but not BDNF

Activation of SFK is critical for the stimulation of neurite outgrowth by NAP and L1 in CGNs (Ignelzi et al. 1994; Loers et al. 2005; Maness and Schachner 2007; Chen and Charness 2008), netrin-1 in cortical neurons and commissural neurons (Liu et al. 2004), and GDNF in dopaminergic neurons (Airaksinen and Saarma 2002). However, it is unknown whether netrin-1 and GDNF activate SFK in CGNs. Therefore, we measured the activation of SFK in CGNs from PD7 rats following 30 min of exposure to netrin-1, GDNF, L1-Fc, or BDNF. Activation of SFK was analyzed by immunoblotting with an antibody against pY416SFK and reprobing with antibodies against Fyn kinase, the principal SFK in CGNs (Umemori et al. 1992). The ratio of pY416SFK to Fyn was used as an index of SFK activity.

Netrin-1, GDNF, and L1-Fc significantly increased the phosphorylation of SFK, whereas BDNF had no significant effect (Fig. 2; Table S2). PP2, a specific inhibitor of SFK, did not significantly reduce basal levels of SFK, but prevented the phosphorylation of SFK induced by netrin-1, GDNF, and L1. PP2 did not significantly reduce SFK phosphorylation in CGNs treated with BDNF. These findings indicate that SFK is involved in the signaling of netrin-1, GDNF, and L1, but not BDNF, in CGNs.

Figure 2.

Ligand activation of Src family kinase (SFK). Cerebellar granule neurons (CGNs) were treated for 30 min with (a) vehicle; (b) netrin-1; (c) glial cell line-derived neurotrophic factor (GDNF); (d) L1-Fc; or (e) brain-derived neurotrophic factor (BDNF) in the absence and presence of 5 μM PP2. The ratio of pY416SFK to Fyn kinase was normalized to control values in each experiment. Shown is the mean ± SEM normalized ratio following the indicated treatments. Representative immunoblots show phosphorylated Y416SFK, detected with pY416SFK antibody (top panels), and total Fyn kinase, detected with Fyn kinase antibody (bottom panels). **p < 0.01; compared with control; #p < 0.05 between indicated groups (n = three to nine independent experiments).

Activation of SFK is required for netrin-1, GDNF, and L1 activation of Cas

The scaffolding protein Cas regulates a variety of biological processes, including axon guidance (Defilippi et al. 2006; Huang et al. 2006, 2007; Liu et al. 2007). Activation of SFK induces tyrosine phosphorylation of Cas, thereby promoting axon outgrowth in CGNs and netrin-1-mediated axon guidance in cortical and commissural neurons (Huang et al. 2006; Liu et al. 2007). To determine whether netrin-1, GDNF, L1, and BDNF activate Cas in CGNs, we incubated CGNs with each ligand for 30 min and then measured the ratio of pY410Cas to total Cas as an index of Cas activation (Chen and Charness 2008). Netrin-1, GDNF, and L1-Fc each significantly increased Cas activation in CGNs, whereas BDNF had no effect. We further asked whether SFK activation is necessary for netrin-1, GDNF, and L1-Fc activation of Cas. PP2 did not reduce basal Cas activity; however, PP2 abolished the activation of Cas by netrin-1, GDNF, and L1-Fc (Fig. 3, Table S2). These findings indicate that activation of SFK is required for netrin-1, GDNF, and L1 activation of Cas. In contrast, BDNF, which does not activate SFK, also does not activate Cas.

Figure 3.

Ligand activation of Cas. Cerebellar granule neurons (CGNs) were treated for 30 min with (a) vehicle; (b) netrin-1; (c) glial cell line-derived neurotrophic factor (GDNF); (d) L1-Fc; or (e) brain-derived neurotrophic factor (BDNF) in the absence and presence of 5 μM PP2. The ratio of pY410Cas to Cas was normalized to control values in each experiment. Shown is the mean ± SEM normalized ratio following the indicated treatments. Representative immunoblots show phosphorylated pY410Cas, detected with pY410Cas antibody (top panels), and total Cas, detected with Cas antibody (bottom panels). *p < 0.05, **p < 0.01 compared with control; #p < 0.05 between indicated groups (n = 4–10 independent experiments).

Activation of SFK is required for netrin-1, GDNF, and L1 activation of ERK1/2

MAP kinases, including ERK1/2, have been implicated in the downstream signaling of SFK in response to a number of molecules that stimulate neurite outgrowth, including netrin-1, GDNF, and L1 (Schmid et al. 2000; Tucker et al. 2008; Pantera et al. 2009; Guimond et al. 2010). BDNF activation of MAP kinases has been widely accepted as a pivotal signaling pathway in BDNF-mediated neuronal protection, but its contribution to BDNF-mediated axon outgrowth is less well defined [for review see (Numakawa et al. 2010)]. To investigate whether netrin-1, GDNF, L1, NAP, and BDNF activate MAP kinase in CGNs under our conditions, we incubated cells with each ligand for 30 min and then measured the ratio of phosphorylated ERK1/2 to total ERK1/2. All five ligands significantly increased ERK1/2 activity (Fig. 4; Table S2). To determine whether activation of ERK1/2 was downstream of SFK activation, we exposed CGNs to each ligand in the absence and presence of PP2. PP2 alone did not significantly reduce basal levels of ERK1/2 activity, but abolished the activation of ERK1/2 by netrin-1, GDNF, L1, and NAP (Fig. 4). In contrast, PP2 had no effect on BDNF activation of ERK1/2 (Fig. 4d). These results indicate that netrin-1, GDNF, NAP, and L1 activate a common signaling pathway in which Cas and ERK1/2 are downstream of SFK. In contrast, BDNF activates ERK1/2 without first activating SFK.

Figure 4.

Ligand activation of extracellular signal-regulated kinase 1 and 2 (ERK1/2). Cerebellar granule neurons (CGNs) were treated for 30 min with (a) vehicle; (b) netrin-1; (c) glial cell line-derived neurotrophic factor (GDNF); (d) L1-Fc; (e) brain-derived neurotrophic factor (BDNF), or (f) 10-12M NAP in the absence and presence of 5 μM PP2 or 25 mM ethanol. The ratio of pERK1/2 to ERK1/2 was normalized to control values in each experiment. Shown is the mean ± SEM normalized ratio following the indicated treatments. Representative immunoblots show phosphorylated ERK1/2, detected with phospho-ERK1/2 antibody (top panels), and total ERK1/2, detected with ERK1/2 antibody (bottom panels). *p < 0.05, **p < 0.01, compared with control; #p < 0.05 between indicated groups (n = 4–10 independent experiments).

PP2 differentially blocks stimulation of axon outgrowth and axon initiation by netrin-1, GDNF, and BDNF

Our previous work implicated SFK in the induction of CGN axon outgrowth by L1 and NAP (Loers et al. 2005; Chen and Charness 2008); therefore, we asked whether SFK also mediates the induction of CGN axon outgrowth by netrin-1 and GDNF. PP2 abolished the induction of axon initiation and elongation by netrin-1 and GDNF, but did not significantly reduce axon outgrowth in otherwise untreated cultures (Fig. 5, Table S1). In contrast, PP2 had no effect on BDNF-induced axon elongation and initiation. These results and reports indicate that activation of SFK is necessary for the stimulation of axon outgrowth by netrin-1, GDNF, L1, and NAP, but not BDNF.

Figure 5.

Effect of PP2 on ligand-stimulated axon outgrowth. Cerebellar granule neurons (CGNs) were cultured under serum-free conditions and treated for 20 h with netrin-1, glial cell line-derived neurotrophic factor (GDNF), and brain-derived neurotrophic factor (BDNF) in the absence and presence of 5 μM PP2, as indicated. (a) Mean ± SEM axon length of individual neurons; (b). Mean ± SEM percentage of neurons with neurites. For both a and b: open bar, control; black bar, ligand; gray bar, ligand plus PP2. *p < 0.05, **p < 0.01 compared with control; #p < 0.05, differences in the absence or presence of PP2 (n = three to six independent experiments).

Ethanol inhibits activation of the SFK-Cas-ERK1/2 pathway by netrin-1, GDNF, and L1

We next investigated the hypothesis that ethanol inhibits the actions of netrin-1, GDNF, and L1 by blocking their activation of the SFK-Cas-ERK1/2 pathway. We first determined that 25 mM ethanol did not significantly alter basal levels of SFK in CGNs cultured in serum-free medium (Fig. 6a). In contrast, 25 mM ethanol completely blocked the activation of SFK by netrin-1, GDNF, and L1 (Fig. 6; Table S2). Likewise, ethanol did not significantly alter basal levels of Cas or ERK1/2 activation, but completely abolished activation of Cas and ERK1/2 by netrin-1, GDNF, and L1 (Figs 7, 8; Table S2). Ethanol also blocked NAP activation of ERK1/2 (Fig. 4f). As noted, BDNF did not activate SFK or Cas, and ethanol did not reduce SFK or Cas activity in the presence of BDNF. Although BDNF robustly activated ERK1/2, ethanol had no effect on BDNF activation of ERK1/2 (Fig. 8e). These findings indicate that ethanol inhibits ligand activation of ERK1/2 by blocking the upstream activation of SFK.

Figure 6.

Effect of ethanol on ligand activation of Src family kinase (SFK). Cerebellar granule neurons (CGNs) were treated for 30 min with (a) vehicle; (b) netrin-1; (c) glial cell line-derived neurotrophic factor (GDNF); (d) L1-Fc; or (e) brain-derived neurotrophic factor (BDNF) in the absence and presence of 25 mM ethanol. The ratio of pY416SFK to Fyn kinase was normalized to control values in each experiment. Shown is the mean ± SEM normalized ratio following the indicated treatments. Representative immunoblots show phosphorylated Y416SFK (top panels) and total Fyn kinase (bottom panels). **p < 0.01 compared with control; #p < 0.05 between indicated groups (n = five to nine independent experiments).

Figure 7.

Effect of ethanol on ligand activation of Cas. Cerebellar granule neurons (CGNs) were treated for 30 min with (a) vehicle; (b) netrin-1; (c) glial cell line-derived neurotrophic factor (GDNF); (d) L1-Fc; or (e) brain-derived neurotrophic factor (BDNF) in the absence and presence of 25 mM ethanol. The ratio of p410Cas to Cas was normalized to control values in each experiment. Shown is the mean ± SEM normalized ratio following the indicated treatments. Representative immunoblots show phosphorylated pY410Cas (top panels) and total Cas (bottom panels). *p < 0.05, **p < 0.01 compared with control; #p < 0.05 between indicated groups (n = 5–10 independent experiments).

Figure 8.

Effect of ethanol on ligand activation of extracellular signal-regulated kinase 1 and 2 (ERK1/2). Cerebellar granule neurons (CGNs) were treated for 30 min with (a) vehicle; (b) netrin-1; (c) glial cell line-derived neurotrophic factor (GDNF); (d) L1-Fc; or (e) brain-derived neurotrophic factor (BDNF), in the absence or presence of 25 mM ethanol. The ratio of pERK1/2 to ERK1/2 was normalized to control values in each experiment. Shown is the mean ± SEM normalized ratio following the indicated treatments. Representative immunoblots show phosphorylated ERK1/2 (top panels) and total ERK1/2 (bottom panels). *p < 0.05, **p < 0.01, compared with control; #p < 0.05 between indicated groups (n = 4–10 independent experiments).

Discussion

The principal finding of this study is that ethanol disrupts axon outgrowth stimulated by netrin-1, GDNF, and L1 by blocking their convergent activation of SFK. Inhibition of axon outgrowth and SFK activity by ethanol was associated with inhibition of SFK's downstream signaling elements Cas and ERK1/2. Importantly, ethanol did not inhibit axon outgrowth mediated by BDNF, a ligand that activates ERK1/2 without first activating SFK or Cas. These results suggest that SFK, but not ERK1/2, is a primary target for ethanol inhibition of axon outgrowth (Fig. 9).

Figure 9.

Src family kinase (SFK) is a primary target in ethanol disruption of cerebellar granule neurons (CGN) axon outgrowth. Axon outgrowth begins when diverse ligands bind to their unique receptors; a membrane-bound receptor for NAP has not yet been identified. The ligand-bound receptors activate SFK, thereby activating its downstream signaling molecules, Cas and extracellular signal-regulated kinase 1 and 2 (ERK1/2). Activation of the Cas and ERK1/2 kinases is coupled to remodeling of the cytoskeleton and axon outgrowth. Ethanol inhibits ligand stimulation of axon outgrowth by blocking ligand activation of SFK. Brain-derived neurotrophic factor (BDNF) stimulates axon outgrowth and activates ERK1/2 without activating SFK, and ethanol does not inhibit BDNF-mediated axon outgrowth or activation of ERK1/2. Thus, SFK appears to be a primary target in ethanol inhibition of axon outgrowth stimulated by diverse axon-guidance molecules and growth factors.

We first demonstrated that each molecule stimulates axon outgrowth in CGNs. This observation was novel for netrin-1 and GDNF and confirmed previous findings for BDNF and L1 (Segal et al. 1995; Dahme et al. 1997). Together with our earlier work on NAP (Chen and Charness 2008), our model system allowed us to characterize under identical conditions the effects of ethanol on five different ligands that regulate CGN axon outgrowth. There were important commonalities in the downstream signaling events that follow treatment of CGNs with netrin-1, GDNF, L1, and NAP. Their activation of SFK was blocked by PP2, a selective inhibitor of SFK. These results extend previous reports that NAP and L1 activate SFK in CGNs (Chen and Charness 2008; Yeaney et al. 2009). The observation that PP2 also blocks ligand activation of Cas and ERK1/2 indicates that Cas and ERK1/2 are activated downstream of SFK for netrin-1, GDNF, L1, and NAP, establishing SFK as a critical control point in the action of these ligands and demonstrating a consistent signaling cascade from SFK to Cas and ERK1/2.

Cas is a scaffolding protein that coordinates the interactions of multiple proteins (Defilippi et al. 2006); hence, it is possible that Cas serves as a platform for the assembly of SFK, ERK1/2, and additional molecules into an active signaling complex. Even in the resting state, Cas is physically associated with Fyn kinase in CGNs and in cerebellum (Nishio et al. 2001; Chen and Charness 2008), and SFK colocalizes with or is physically associated with ERK1/2 in retinal ganglion cells (Ramseger et al. 2009) and B cells (Pleiman et al. 1993). It is not clear whether SFK activation of Cas and ERK1/2 is sequential or concurrent, and further experiments are required to determine whether SFK activates ERK1/2 directly or through an interaction with a protein complex that includes ERK1/2, Cas, and other molecules.

Our work also demonstrates that the SFK-Cas-ERK1/2 signaling cascade plays a fundamental role in the physiological response of CGNs to netrin-1, GDNF, L1, and NAP. Activation of SFK-Cas-ERK1/2 was necessary for axon outgrowth induced by each of these ligands, because PP2 blocked their induction of axon initiation and elongation. We reported previously that PP2 or siRNA knockdown of Fyn kinase also reduces NAP stimulation of CGN neurite outgrowth, suggesting that at least for NAP, the effects of PP2 are not because of non-specific effects on other kinases (Chen and Charness 2008). Taken together, these findings suggest that an early signaling event in the stimulation of axon outgrowth is the activation of the SFK-Cas-ERK1/2 cascade by a variety of ligands.

For BDNF, the molecular events leading to axon outgrowth and activation of ERK1/2 are clearly different from those induced by netrin-1, GDNF, L1, and NAP. BDNF did not activate SFK or Cas, and PP2 did not inhibit BDNF activation of ERK1/2 or its induction of axon outgrowth. These data indicate that BDNF stimulation of CGN axon outgrowth is not dependent on the activation of SFK and Cas.

Clinically relevant concentrations of ethanol completely blocked the effects of netrin-1, GDNF, L1, and NAP on SFK-Cas-ERK1/2 signaling and axon outgrowth. The ability of ethanol to block the physiological actions of at least four developmentally critical molecules at a convergent point of signaling might contribute significantly to the pathogenesis of FASD. We have noted previously that children with mutations in the gene for L1 have brain lesions similar to those of children with FASD, including agenesis or hypoplasia of the corpus callosum, hydrocephalus, and dysplasia of the first five cerebellar lobules (Fransen et al. 1995; Ramanathan et al. 1996). Netrin-1-knockout mice die shortly after birth and also display agenesis of the corpus callosum (Serafini et al. 1996; Lai Wing Sun et al. 2011). Mice lacking GDNF and its receptor GFRα1 die shortly after birth because of the absence of enteric and parasympathetic neurons and kidney agenesis (Airaksinen and Saarma 2002; Paratcha and Ledda 2008); GDNF heterozygous mice display impaired learning (Gerlai et al. 2001). Deletion of the ADNP gene results in neural tube defects and embryonic death after 8–9 days of gestation (Pinhasov et al. 2003). Of note, pre-natal ethanol exposure also causes neural tube defects (Chen et al. 2005). Pre-natal ethanol exposure typically causes less severe developmental abnormalities than the deletion of any of these genes, most likely because the effects of ethanol are limited to discrete periods of development.

Further research is required to determine the mechanism by which ethanol disrupts the SFK-Cas-ERK1/2 signaling cascade. Ethanol might directly disrupt the interaction of growth factor receptors with SH2 or SH3 domains on SFK. Ethanol could also activate inhibitors of SFK, as occurs with H-ras inhibition of Src in hippocampal neurons (Suvarna et al. 2005). Likewise, ethanol might alter protein trafficking to prevent the interaction of SFK and its activating proteins within specialized cellular microdomains. Tang and colleagues (Tang et al. 2011) recently reported that treatment of CGNs with ethanol shifted L1 into lipid rafts, while shifting Src kinase out. It remains to be investigated whether a similar mechanism underlies the actions of netrin-1, GDNF, and NAP.

The effects of ethanol on axon outgrowth differ as a function of brain region, stimulating ligands, developmental stage, and culture conditions. For example, ethanol inhibits L1-mediated neurite outgrowth in CGNs (Bearer et al. 1999), but not in cortical neurons (Hoffman et al. 2008). In contrast, L1 actually rescues ethanol inhibition of cortical axonal growth cone responsiveness to guidance cues of semaphorin-3 and netrin-1 (Sepulveda et al. 2011). Ethanol inhibits BDNF-induced axon initiation, but potentiates BDNF-induced axon elongation in hippocampal pyramidal neurons (Lindsley et al. 2003). Ethanol was reported to inhibit BDNF activation of ERK1/2 in CGNs (Ohrtman et al. 2006); however, these experiments used CGNs after 3 days in vitro, in contrast to 1 day in our experiments. These findings suggest that the actions of ethanol depend on neuronal cell type and cellular context, including intracellular signaling events and extracellular guidance cues. Conceivably, the differential expression and regulation of SFK-Cas-ERK1/2 signaling in diverse neuronal populations and at different developmental times account for some of these variable responses to ethanol.

Acknowledgements

We thank Carrie Menkari, Devon M. Fitzgerald, and Lazaros Yiannos for their excellent technical assistance. This study was supported in part by the ABMRF Award (S.C.), the National Institute on Alcohol Abuse and Alcoholism Grant 2R01AA012974 (M.E.C.), NIAAA U24-AA014811, as a component of the Collaborative Initiative on Fetal Alcohol Spectrum Disorders (M.E.C) and the Medical Research Service, Department of Veterans Affairs (M.E.C). Dr. Charness is a member of the scientific advisory board for Allon Therapeutics, which is developing clinical applications for NAPVSIPQ (NAP). Dr. Charness holds stock options in Allon Therapeutics, and has two patents related to the study to declare. This does not alter the authors' adherence to all the Journal of Neurochemistry policies on sharing data and materials. Conceived and designed the experiments: SC, MC; Performed the experiments: SC; Analyzed the data: SC, MC; Wrote the paper: SC, MC.

Ancillary