SEARCH

SEARCH BY CITATION

Keywords:

  • cerebellar granule neurons;
  • ethanol;
  • fetal alcohol spectrum disorder;
  • L1 cell adhesion molecule;
  • lipid rafts;
  • neurite outgrowth

Abstract

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

J. Neurochem. (2011) 119, 859–867.

Abstract

Fetal alcohol spectrum disorder is estimated to affect 1% of live births. The similarities between children with fetal alcohol syndrome and those with mutations in the gene encoding L1 cell adhesion molecule (L1) implicates L1 as a target of ethanol developmental neurotoxicity. Ethanol specifically inhibits the neurite outgrowth promoting function of L1 at pharmacologic concentrations. Emerging evidence shows that localized disruption of the lipid rafts reduces L1-mediated neurite outgrowth. We hypothesize that ethanol impairment of the association of L1 with lipid rafts is a mechanism underlying ethanol’s inhibition of L1-mediated neurite outgrowth. In this study, we examine the effects of ethanol on the association of L1 and lipid rafts. We show that, in vitro, L1 but not N-cadherin shifts into lipid rafts following treatment with 25 mM ethanol. The ethanol concentrations causing this effect are similar to those inhibiting L1-mediated neurite outgrowth. Increasing chain length of the alcohol demonstrates the same cutoff as that previously shown for inhibition of L1–L1 binding. In addition, in cerebellar granule neurons in which lipid rafts are disrupted with methyl-beta-cyclodextrin, the rate of L1-mediated neurite outgrowth on L1-Fc is reduced to background rate and that this background rate is not ethanol sensitive. These data indicate that ethanol may inhibit L1-mediated neurite outgrowth by retarding L1 trafficking through a lipid raft compartment.

Abbreviations used
CGN

cerebellar granule neurons

FAS

fetal alcohol syndrome

L1

L1 cell adhesion molecule

MBCD

methyl-beta-cyclodextrin

PBS

phosphate-buffered saline

PLL

poly-l-lysine

The use of ethanol during pregnancy can be detrimental to the offspring, the most severe result being the constellation of developmental defects and abnormalities commonly referred to as fetal alcohol syndrome (FAS). In the CNS, FAS is hallmarked by a wide range of defects in brain morphology, including hypoplasia of the corpus callosum and optic nerves, the formation of neuronal heterotopias, microcephaly, disorders in cortical lamination, and hypoplasia of the cerebellar vermis. These deficits can result in disorders such as motor dysfunction, hyperactivity, increased susceptibility to seizures, and deficits in learning and memory (Stratton et al. 1996). The resemblance of FAS to the spectrum of disorders caused by mutations in L1 cell adhesion molecule (L1) has led investigators to explore the possible role of L1 in FAS (Charness et al. 1994). L1 is a member of the immunoglobulin (Ig) superfamily implicated in a variety of processes in neurohistogenesis, including neurite elongation, axon fasciculation, and migration of neuronal precursors. It is a member of a subgroup of cell adhesion molecules which are related by structure and sequence. Each consists of six Ig-like domains, five fibronectin type III (FNIII)-like domains, and a highly conserved cytoplasmic tail (Kamiguchi 2003; Winckler and Yap 2011). L1 mediates cell–cell adhesion via a homophilic binding mechanism. In vitro L1 presented as a culture substrate stimulates neurite outgrowth by binding homophilically to L1 expressed on neurons. The neurite outgrowth of neurons plated on L1 is inhibited by ethanol (Bearer et al. 1999; Watanabe et al. 2004). We have previously shown that L1 does not signal through the fibroblast growth factor receptor but activates both c-Src (pp60src) and ERK1/2 (Tang et al. 2006). In addition, ethanol inhibits downstream signaling of L1 including activation of c-Src and ERK1/2, phosphorylation of L1 tyrosines, and the dephosphorylation of tyrosine 1176 in the cytoplasmic domain of L1 (Tang et al. 2006; Yeaney et al. 2009). Inhibition of c-Src alone inhibits all of these downstream events, suggesting that ethanol acts by preventing L1 activation of c-Src (Yeaney et al. 2009).

Recently, L1 and a similar family member, neurofascin, are identified in lipid rafts in the peripheral (P) domain of the growth cone (Ren and Bennett 1998; Nakai and Kamiguchi 2002). Lipid rafts are microdomains of the plasma membrane, composed of sphingolipids and cholesterol in the outer exoplasmic leaflet, connected to phospholipids and cholesterol in the inner cytoplasmic leaflet of the lipid bilayer. These microdomains are fluidic but more ordered and tightly packed than the surrounding bilayer (Simons and Ehehalt 2002; Simons and Gerl 2010). Current studies suggest that formation of lipid rafts is dynamic (Owen et al. 2010). L1 can be differentially associated with lipid rafts in different stages of development: L1 association with lipid rafts in cerebellum is only detectable between P8 and P28, but not detectable at P56 (Nakai and Kamiguchi 2002). The molecular basis for the regulation of L1 association with lipid rafts is currently unknown. Targeted disruption of the lipid raft reduces neurite outgrowth on both L1 and N-cadherin, but not laminin (Nakai and Kamiguchi 2002). Many proteins known to have substantial ethanol toxicity, such as the GABAA receptor (Mihic 1999), the NMDA receptor (Mihic 1999), the α-amino-3-hydroxy-5-methyl-4 isoxazole proprionic acid receptor (Mameli et al. 2005), the receptor for transforming growth factor beta (Luo and Miller 1999) and heterotrimeric G proteins (Zhao et al. 2003) have all recently been shown to be associated with lipid rafts. We hypothesize that a major mechanism of ethanol toxicity is through disruption of protein–lipid raft interactions.

In the present study, we investigate the redistribution of L1, N-cadherin, beta-tubulin, caveolin-1, c-Src as well as GABAA receptors in lipid rafts of cerebellar granule neurons (CGN) following ethanol exposure in vitro, and ethanol effects on L1-mediated neurite outgrowth in CGN in which lipid rafts have been disrupted. We present evidence that ethanol at pharmacological concentrations may retard L1 trafficking through lipid rafts, identifying L1–lipid raft interactions as a novel target for ethanol inhibition of neurite outgrowth.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Antibodies

Horseradish peroxidase-conjugated cholera toxin B subunits, mouse monoclonal anti-beta III tubulin are obtained from Sigma (St Louis, MO, USA). Mouse monoclonal anti-transferrin receptor antibody is from Invitrogen (Carlsbad, CA, USA). Goat polyclonal anti-neural cell adhesion molecule L1, rabbit polyclonal anti-caveolin-1 and mouse monoclonal anti-c-Src are from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Mouse monoclonal anti-N-cadherin antibody is from BD Biosciences (San Diego, CA, USA). Mouse monoclonal anti-beta III tubulin is from Invitrogen, and mouse monoclonal anti-GABAA receptor beta 2 subunit antibody is a gift from Dr Ruth Siegel, Department of Pharmacology, Case Western Reserve University. Alexa Fluor 488-conjugated cholera toxin B subunits and Alexa Fluor 633 conjugated donkey anti-goat IgG are from Invitrogen. All secondary antibodies for western blot analysis are from Jackson Immuno Research Laboratories (West Grove, PA, USA).

Preparation of CGN cells

All of the procedures involving animals that are used in the current studies are approved by the University of Maryland School of Medicine Institutional Animal Care and Use Committee and conform to NIH guidelines. Rat CGN are prepared from Sprague–Dawley rat pups (Zivic-Miller) on postnatal day 6 as previously described (Bearer et al. 1999). Viability of these dissociated cells is assessed with trypan blue and is > 90%. CGN are plated on prepared dishes, or coverslips in Dulbecco’s modified Eagle’s medium with 10% fetal bovine serum. For neurite outgrowth experiments, the media are changed to a serum free defined media at the time of addition of ethanol. Serum-free defined media consists of Neurobasal Media (Gibco, Rockville, MD, USA) with the following additions: 2% B27 supplement (Gibco), 20 mM l-glutamine, 6 g/l glucose, 20 mM HEPES, pH 7.2, penicillin/streptomycin. All the cultures are maintained in a humid atmosphere of 90% air, 10% CO2 at 37°C.

Lipid raft isolation

CGN are cultured in a 150 mm dish grown for 40 h. Three hours prior to harvest, the cells are serum starved. For CGN treated with ethanol, the ethanol is added 1 h before harvest. Prior to harvest, all equipment is cooled to 4°C. Cultures are taken to the cold room where media is removed, and cells are washed with cold Hank’s balanced salt solution. The cells are scraped, and lysed for 30 min on ice in 4 mL Tris-buffered saline containing 0.5% Triton X-100, 10 mM Na vanadate, 2 μM aprotinin, 100 μMphenylmethylsulfonyl fluoride, 1 μM leupeptin, 10 μg/mL turkey trypsin inhibitor, 1 μg/mL pepstatin A, 100 pM cypermethein, phosphatase inhibitor cocktail I and phosphatase inhibitor cocktail II. The lysates are centrifuged at 13 000 g for 30 min at 4°C. After centrifugation, 4 mL of the supernatant is overlaid on a sucrose density gradient with 40%, 32% and 5% layers. The gradient is centrifuged at 180 000 g, for 24 h at 4°C. Sequential 1 mL fractions are drawn off the top of the gradient. An aliquot from each fraction is analyzed by western blot analysis for both transferrin receptor and GM1 ganglioside. All GM1 ganglioside containing fractions are combined (lipid raft fractions). All remaining fractions are combined (non-lipid raft fractions). Proteins from each pool are precipitated using chloroform/methanol and resuspended in sample buffer for western blot.

Sodium dodecyl sulfate–polyacrylamide gel electrophoresis and immunoblotting analysis

Cell lysates from each sample are resolved by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (4–15% Tris gradient gel) then transferred to a polyvinylidene difluoride membrane. The membrane is incubated with washing buffer (20 mM Tris-buffered saline and 0.1% Tween 20) containing 5% non-fat dry milk (Bio-Rad Laboratories, Hercules, CA, USA) for 1 h to block non-specific protein binding. Primary antibodies are diluted in washing buffer containing 5% non-fat dry milk and are applied to the polyvinylidene difluoride membrane at 4°C overnight. After being washed, the blots are incubated with the appropriate horseradish peroxidase-conjugated secondary antibodies (diluted 1 : 1000 or 1 : 2000 in washing buffer with 5% milk) for 1 h at 23°C. Immunoreactive bands are visualized with the enhanced chemiluminescence system (GE Healthcare, Piscataway, NJ, USA). Where indicated, blots are stripped to be reprobed with additional primary antibodies. Densitometric quantitation is carried out with the software Kodak ID.

Immunofluorescence and confocal microscopy

CGN plated on poly L-lysine (PLL) coated coverslips are rinsed in phosphate-buffered saline (PBS) three times. The cells are fixed in 4% paraformaldehyde for 30 min at 23°C, followed by three more washes of PBS. Permeabilization/blocking solutions (3% bovine serum albumin/0.2% TX-100/PBS) are added to each coverslip. Coverslips are incubated for 1 h at 37°C or overnight at 4°C. Primary antibody is added, and the cells are incubated for 2 h at 37°C and washed. Secondary antibody is added and incubated for 1 h at 37°C. Cells are washed extensively. Confocal microscopy is performed with a Leica TCS SP2 confocal microscope with 63× objective lens using laser excitation at 488 and 633 nm. The colocalizations between two emissions channels are measured using the software Image J. Pixel intensity values greater than threshold level are defined as intense signals. A pixel with intense signals in both wavelength channels is regarded as indicating colocalization.

Preparation of L1-Fc

L1-Fc containing plasmid is prepared as previously described (Tang et al. 2006). Briefly, the plasmid is stably expressed in NIH 3T3 cells using lipofectamine. Cells are co-transfected with the Neomycin-resistant plasmid, pMAMneo (ratio 1 : 9) (a gift from Susanne Brady-Kalnay, Case Western Reserve University). Transfected cells are selected with G4 18 [Gibco Advanced Bioreactors (Kensington, MD, USA)] at 0.9 mg/mL. Clones are selected and subcloned by limited dilution. Supernatants are assayed for L1-Fc. An L1-Fc expressing stably transfected 3T3 clone is propagated using the CELLMAX system [Cellco Advanced Bioreactors (Kensington, MD, USA)] using manufacturer’s instructions. Serum in the media in the CELLMAX system is gradually removed. L1-Fc is purified from the harvested serum free media using affinity chromatography using protein A-AffiGel (Bio-Rad).

Neurite outgrowth

Control coverslips are coverslips coated with PLL. To prepare an L1 substratum, control coverslips are coated with L1-Fc. At 1 h post-plating of CGN, 4 mM methyl-beta-cyclodextrin (MBCD) is added to some cultures. At 2 h post-plating, ethanol is added to some cultures. Cells are fixed in 4% paraformaldehyde for 30 min at 12 h following plating and then labeled fluorescently for beta III tubulin. A blinded investigator identifies eligible neurites for measurement using an a priori design. Eligible neurites are the longest neurite of a CGN that are longer than the length of the cell soma, from single isolated neurons and that do not touch other neurites. A Nikon Optiphot-2 fluorescence microscope and the SPOT camera software are used for measurements of neurite outgrowth.

Statistics

Means, standard deviations and p-values are calculated using Microsoft Excel software.

Results

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Ethanol alters L1 distribution in lipid rafts, but not N-cadherin distribution

To investigate the possibility that L1–lipid raft interactions are disrupted by ethanol, CGN from rat pups on postnatal day 6 are cultured for 40 h and serum starved for 1 h before treated with 25 mM ethanol for 1 h. No L1 is detectable in the pellet from the initial low speed centrifugation (data not shown). Lipid rafts are separated from non-lipid rafts in CGN lysate by Triton X-100 extraction and sucrose density gradient centrifugation. As shown in Fig. 1a, ethanol pre-treatment does not cause a significant change in the separation of the marker for the lipid raft compartment, GM1 ganglioside, and the marker for the non-lipid raft compartment, transferrin receptor. The sucrose density gradient fractions containing GM1 ganglioside are pooled as the lipid raft fraction, and the remaining sucrose gradient fractions are pooled into a non-lipid raft fraction. Thus, all L1 is either in the lipid raft or non-lipid raft pools. The proteins are precipitated from these fractions, reconstituted in equal volumes of sample buffer, and immunoblotted with antibody to the cytoplasmic domain of L1. As can be seen in Fig. 1b and c, the amount of L1 in lipid rafts from cells pre-treated with 25 mM ethanol for 1 h increases to 43% of total L1, compared with 10% from control cells.

image

Figure 1.  (a) Ethanol does not alter the separation of lipid rafts from cell lysates. Immunoblots of two sucrose density gradients, control and EtOH treated. Top panel (Control) is of neurons not exposed to ethanol. Bottom panel (EtOH) is of neurons exposed to 25 mM ethanol for 1 h prior to lysis. Fractions of the sucrose density gradients were blotted for both cholera toxin subunit B (CTxB) as a marker of lipid raft fractions, and for transferrin receptor (TfR) as a marker of non-lipid raft fractions. (b) Ethanol shifts L1 into lipid rafts while not affecting N-cadherin distribution. CGN incubated with or without 25 mM ethanol (EtOH) for 1 h are separated into lipid raft (LR) and non-lipid raft (N) pools. Proteins from each pool are precipitated, and reconstituted into equal volumes of sample buffer. Equal volumes from each pool are loaded onto two lanes, LR and N, and immunoblotted for L1, then stripped and blotted for N-cadherin and GM1 ganglioside using cholera toxin subunit B (CTxB) as a marker for lipid rafts. (c) Densitometric analysis of blots (= 4). The total pixels in the L1 or N-cadherin bands in both the LR and N lanes is determined for both control and ethanol-exposed cultures. The percentage of cell adhesion molecules (CAM) in the LR is calculated by taking the pixels in the LR band and dividing by the sum of the pixels in the N and LR bands. The mean ± SD is shown. Statistically significant difference is indicated (*paired t-test, < 0.04).

Download figure to PowerPoint

To confirm these results, CGN cells are fluorescently labeled for L1 (red) and GM1 ganglioside (green) and viewed by confocal microscopy. The colocalization of L1 and GM1 ganglioside are determined. The percentage of overlap of the red and green channel is measured in at least 20 cells in three separate experiments. As shown in Fig. 2, 40.2 ± 1.4% of labeled L1 colocalizes with GM 1 ganglioside in ethanol treated cells, significantly more than in control cells (27.6 ± 1.7%, < 0.02).

image

Figure 2.  Confocal images confirm ethanol’s effect on colocalization of L1 and GM1 ganglioside. CGN grown on PLL are pre-treated with or without 25 mM ethanol for 1 h, then fixed and labeled for L1 (red) and GM1 ganglioside (green). Cells are visualized using a Leica TCS SP2 confocal microscope. Images are merged to visualize overlapping of L1 and lipid rafts (yellow). (a) CGN not exposed to ethanol. No yellow signal is seen. (b) CGN treated with ethanol. Extensive overlap is seen, as shown by the yellow signal. (c) The percentage of overlap of the red and green channels is measured in at least 20 cells in three separate experiments. The mean overlap is calculated for each condition in each experiment. The mean ± SD of the individual means is shown. Statistically significant difference is indicated (*paired t-test, < 0.02).

Download figure to PowerPoint

Our previous study shows that N-cadherin-mediated neurite outgrowth is not inhibited by pharmacologic concentrations of ethanol (Tang et al. 2006). Here, we test whether ethanol also changes the distribution of N-cadherin in lipid rafts. As shown in Fig. 1b and c, 38.1 ± 10.2% of N-cadherin is in lipid rafts in control cells and no shift occurs after ethanol treatment.

Ethanol decreases the association of beta-tubulin with lipid rafts and has no effect on the lipid raft association of caveolin-1

L1 endocytosed in the C domain of the growth cone is recycled to the leading edge in vesicles carried by microtubules (Kamiguchi and Lemmon 2000; Dequidt et al. 2007). Microtubules regulate the lipid raft localization of some proteins (Head et al. 2006). If ethanol traps L1 in the lipid rafts, it may increase the association of L1 with beta III tubulin, thus increasing the beta III tubulin associated with lipid rafts. As shown in Fig. 3, the amount of beta III tubulin decreased in lipid rafts exposed to ethanol (51% ± 5% vs. 40% ± 8%), opposite to the effect on L1. Chronic alcohol abuse has been shown to increase levels of caveolin-1 in liver, and increased its binding to eNOS (Wang and Abdel-Rahman 2005). We examined whether our brief, 1 h exposure to ethanol changed caveolin-1 redistribution in lipid rafts potentially trapping L1, and found no effect (Fig. 4).

image

Figure 3.  Ethanol has no effect on caveolin-1 association with lipid rafts. (a) CGN are treated as described, either no addition (Control) or 25 mM ethanol for 1 h (EtOH), prior to preparation of lipid rafts. Lipid rafts (LR) were isolated from non-lipid rafts (N) by sucrose density gradient, pooled, protein precipitated and immunoblotted for caveolin-1. Blots were stripped and reblotted for L1 and CTxB. (b) Blots (= 3) were quantified by densitometry. Mean values ± SD are shown.

Download figure to PowerPoint

image

Figure 4.  Ethanol shifts beta III tubulin out of lipid rafts, opposite to the effect on L1. (a) CGN are treated as described in Fig. 3. Lipid rafts (LR) were isolated from non-lipid rafts (N) by sucrose density gradient, pooled and western blotted for beta III tubulin. (b) Blots were stripped and reblotted for L1. (c) Blots from four separate experiments quantified by densitometry. Shown are mean values ± SD. *< 0.05, paired t-test.

Download figure to PowerPoint

Butanol redistributes L1 in lipid rafts but pentanol does not, in parallel with effects seen on L1–L1 homophilic binding

Ethanol inhibits L1 adhesive properties in some cells (Charness et al. 1994; Ramanathan et al. 1996) and not in others (Vallejo et al. 1997; Bearer et al. 1999). In three separate L1-expressing cell lines, ethanol inhibits 55% of cell aggregation with a half-maximal inhibition at 5 to 10 mM ethanol (Wilkemeyer and Charness 1998). Short chain alcohols (C < 5) mimic the inhibitory effect, but longer chain alcohols (C > 5) do not (Charness et al. 1994). To determine if the effect on L1–lipid raft interactions also parallels the observed alcohol chain length dependence of L1–L1 homophilic binding, we tested the ability of butanol and pentanol at equipotent concentrations to alter L1 distribution in lipid rafts. If the L1–lipid raft disruption causes the reduction of L1 adhesivity, then the same cut off between butanol and pentanol should exist for the effect on lipid rafts (Charness et al. 1994). As shown in Fig. 5, the redistribution of L1 also occurs with an equipotent concentration of butanol, but not pentanol.

image

Figure 5.  Butanol redistributes L1 in lipid rafts, but pentanol does not. (a) CGN pre-treated with equipotent concentrations of ethanol (EtOH), butanol (BtOH) and pentanol (PtOH) are separated into lipid rafts and non-lipid rafts as described, and immunoblotted for L1. N, non-lipid raft pool; LR, lipid raft pool. (b) Blots (= 3) are quantified as described in Fig. 1. Mean ± SD are shown. Statistically significant differences are indicated (*< 0.02, **< 0.003, paired t-test).

Download figure to PowerPoint

Other known protein targets of ethanol change their lipid raft distribution in the presence of ethanol

In this study, the effects of ethanol on the lipid raft association of other proteins known to be targets of ethanol toxicity, such as the tyrosine kinase c-Src (Suvarna et al. 2005) and GABAA receptor (Mihic 1999) are investigated. As can be seen in Fig. 6, ethanol decreases the distribution of c-Src in lipid rafts in contrast to its effect on L1. Similarly, experiments with GABAA receptor beta 2 subunit show the same effect as c-Src, a decrease of GABAA receptor beta 2 subunit in the lipid raft in the presence of ethanol (Fig. 7). This subunit is particularly important for receptor function as it is required for sensitivity to GABA. The alteration in L1, c-Src and GABAA receptor beta 2 subunit distributions in lipid rafts, known targets for ethanol action, suggests that disruption of protein–lipid raft interactions may be an underlying mechanism of toxicity.

image

Figure 6.  Ethanol shifts c-Src out of lipid rafts, opposite to the effect on L1. (a) CGN are treated as described, either no addition (Control) or 25 mM ethanol for 1 h (EtOH), prior to preparation of lipid rafts. Lipid rafts (LR) were isolated from non-lipid rafts (N) by sucrose density gradient, pooled, protein precipitated and immunoblotted for Src. Blots were stripped and reblotted for L1. (b) Blots (= 3) were quantified by densitometry. Mean values ± SD. *< 0.05, paired t-test.

Download figure to PowerPoint

image

Figure 7.  Ethanol shifts GABAA receptor out of lipid rafts, opposite to the effect on L1. (a) CGN are treated as described in Fig. 5. Lipid rafts (LR) were isolated from non-lipid rafts (N) by sucrose density gradient and western blotted for GABAA receptor beta 2 subunit. (b) Blots from three separate experiments quantified by densitometry. Shown are mean values ± SD. *< 0.05, paired t-test.

Download figure to PowerPoint

Redistribution of L1 into lipid rafts occurs at low concentrations of ethanol

Ethanol inhibition of L1-mediated neurite outgrowth has an IC50 of 3–5 mM (Bearer et al. 1999). If the inhibition of L1-mediated neurite outgrowth is a result of altered trafficking through lipid rafts, then the concentrations of ethanol that induce redistribution of L1 into lipid rafts should be similar to that which inhibits neurite outgrowth. As shown in Fig. 8, the redistribution of L1 occurs even at the lowest concentration tested, 5 mM ethanol, indicating a similar sensitivity to ethanol of L1-mediated neurite outgrowth.

image

Figure 8.  L1 redistribution to lipid rafts is dependent on ethanol concentration. (a) CGN are prepared as described and cultured overnight. Ethanol is added at the indicated concentrations for 1 h, then cells are harvested and lipid rafts isolated. Both the lipid raft pool (LR) and non-lipid raft pool (N) are immunoblotted for L1 (L1) and cholera toxin B (CTxB). (b) The percent of L1 in lipid rafts (% L1 in LR), calculated as described in the text, are plotted for each concentration of ethanol. Shown are means ± SD (= 3).

Download figure to PowerPoint

Removal of L1 from the lipid raft compartment abolishes ethanol inhibition of neurite outgrowth

Neurite outgrowth on L1 is inhibited by pharmacological treatments that deplete cellular cholesterol or sphingolipids, essential components for lipid rafts (Nakai and Kamiguchi 2002). Remaining neurite outgrowth on L1 is independent of lipid rafts. If the effects of ethanol are on the lipid rafts, then this remaining neurite outgrowth will be insensitive to ethanol. To directly address this hypothesis, lipid rafts are disrupted by cholesterol depletion using MBCD. The survival of CGN grown in the presence of 4 mM MBCD is first determined. Results from four different experiments are shown in Table 1. Cells remain viable up to 12 h of exposure to MBCD with and without 25 mM ethanol with no significant difference to control cell survival at 12 h.

Table 1.   CGN survival in MBCD and/or 25 mM ethanol
ConditionMean ± SE, % survivalp-value vs. Control, 0 hp-value vs. Control, 12 h
Control, 0 h90.75 ± 1.38n/an/a
Control, 12 h80.50 ± 2.630.007n/a
MBCD, 12 h76.75 ± 2.170.0070.22
MBCD + EtOH, 12 h74.75 ± 3.330.0150.19

To examine whether L1 is still associated with lipid rafts after cholesterol depletion, CGN are grown overnight, then exposed to MBCD for 1 h. Cells are harvested and lipid rafts isolated by sucrose density gradients. Fractions containing non-lipid raft proteins are identified by transferrin receptor immunoreactivity. Each fraction is immunoblotted for L1. As shown in Fig. 9a, L1 was strictly located in fractions containing transferrin receptor. This result means that all L1 in the MBCD treated cells are located in non-lipid raft compartments.

image

Figure 9.  L1-mediated neurite outgrowth is dependent on lipid rafts. (a) Immunoblot of L1 in sucrose density gradient fractions of CGN treated with or without MBCD. All detectable L1 colocalizes with the transferrin receptor (TfR), a marker for the non-lipid raft fraction. (b) Neurite length of CGN in the presence of various reagents. Cells are plated on either PLL alone (Control) or PLL and L1-Fc (L1-Fc). Ethanol (25 mM) and/or MBCD (4 mM) are added 2 h after plating. Cells are fixed 12 h after plating, and neurite length is measured. Shown is mean neurite length ± SD. Paired t-test: a< 0.0002 from control; b< 0.0002 from L1-Fc; c< 0.02 from Control; dp-value not significant from Control; ep-value not significant from L1-Fc+MBCD nor Control.

Download figure to PowerPoint

To determine the effect on neurite outgrowth, CGN are cultured on PLL (control), or PLL with L1-Fc. L1-Fc is a chimeric protein containing the extracellular domain of L1 (Bearer et al. 1999; Tang et al. 2006). One hour after plating, 4 mM MBCD and/or 25 mM ethanol are added to the plates. Cells are fixed 10 h after addition of MBCD and/or ethanol and neurite length is measured. Results are shown in Fig. 9b. As can be seen, cells grown on L1-Fc have longer neurites. Ethanol reduces neurite lengths of CGN grown on L1-Fc by 70% (L1-Fc vs. L1-Fc+EtOH). MBCD reduced neurite length of CGN grown on L1-Fc to that of PLL alone (L1-Fc+MBCD vs. Control). Addition of ethanol to CGN grown on L1-Fc and in the presence MBCD has no further effect on neurite length (L1-Fc+MBCD vs. L1-Fc+MBCD+EtOH).

Discussion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

There is growing interest in identifying lipid rafts as targets for ethanol action. Ethanol reportedly blocks the lipopolysaccharide redistribution of CD14 to a lower density fraction of the lipid raft (Dai et al. 2005). Ethanol inhibits the lipopolysaccharide-mediated redistribution of toll-like receptor 4 to the lipid rafts, but has no effect on the peptidoglycan-mediated redistribution of toll like receptor 2 (Dolganiuc et al. 2006). Another study has shown that insulin treatment of adipocytes from rats recruits cCbl, a multifunctional molecule with ubiquitin ligase activity and a protein adaptor function, and TC10, a Rho family small GTPase, to lipid rafts. However, in adipocytes from rats chronically fed ethanol, insulin treatment does not translocate cCbl and TC10 to the lipid raft (Sebastian and Nagy 2005). The underlying mechanism for the effects of ethanol on the redistribution of proteins is yet to be investigated. L1-mediated neurite outgrowth is dependent on lipid rafts as shown by targeted disruption of lipid rafts in the P but not C domain of growth cones (Nakai and Kamiguchi 2002) and our results reported here (Fig. 9a and b).

The classic lipid raft hypothesis postulates that lipid rafts containing a given set of proteins can change their size and composition in response to intra- or extracellular stimuli. This ability favors specific protein–protein interactions, resulting in the activation of signaling cascades (Simons and Toomre 2000). Proteins with lipid anchors or transmembrane domains have been shown to dynamically partition into and out of lipid rafts such that only a fraction of proteins are clustered in the lipid rafts (Simons and Vaz 2004). In the axonal growth cone, L1-mediated adhesive interactions may be dynamically regulated to create a front-versus-rear asymmetry in growth cone-substrate adhesion which is believed to be necessary for growth cone migration (Kamiguchi 2003; Dequidt et al. 2007). Therefore, the L1–lipid raft interactions should be considered as dynamic, requiring both L1 trafficking to the rafts, and L1 trafficking out of the rafts to the C domain of the growth cone where it is endocytosed (Kamiguchi and Lemmon 2000; Schaefer et al. 2002). Disruption of this trafficking leading to accumulation of L1 in the lipid raft would reduce L1-mediated neurite outgrowth. Although ethanol-induced increase of L1 association with lipid rafts may increase L1-mediated cell adhesion, there are several reports of ethanol decreasing L1–L1 cell adhesion (Charness et al. 1994; Ramanathan et al. 1996; Wilkemeyer and Charness 1998). However, this effect seems both cell type and cell clone dependent (Vallejo et al. 1997; Wilkemeyer and Charness 1998; Bearer et al. 1999). The changes in adhesive strength of lipid raft associated L1 will need to be investigated.

An increase in the amount of L1 in the lipid raft could be caused either by an increase in L1 trafficking to the lipid raft, or a decrease in L1 trafficking out of the lipid raft. Proteins which are temporally associated with rafts do so through the dynamic regulation of the following processes: (i) S-palmitoylation, (ii) initial partial immobilization by binding to the cytoskeleton, (ii) phosphorylation and dephosphorylation, (iv) increased affinity of the protein for raft associated proteins or (v) length of the transmembrane domain (Lucero and Robbins 2004). A previous study suggested that targeting of L1 family members to lipid rafts is in part dependent on palmitoylation (Ren and Bennett 1998). Ethanol may increase L1 in lipid rafts by increasing palmitoylation of L1 (Carpenter-Hyland and Chandler 2006). In addition, we have previously shown that ethanol inhibits the tyrosine phosphorylation of L1 (Yeaney et al. 2009). Phosphorylation of a highly conserved tyrosine in the cytoplasmic domain of neurofascin abolishes ankyrin binding (Garver et al. 1997; Tuvia et al. 1997). Thus, ethanol may increase L1 binding to the cytoskeleton and therefore increase its lipid raft association. The changes in the phosphorylation state of L1 alone may increase its association with lipid rafts. L1 may have increased affinity for a lipid raft associated protein, such as c-Src (Lu et al. 2008). While we have shown that L1 activation of c-Src occurs prior to changes in L1 tyrosine phosphorylation (Yeaney et al. 2009), a direct interaction of L1 with c-Src has not been reported. It is unclear whether a direct interaction of L1 with c-Src within the lipid raft has been investigated.

Finally, it must be considered that ethanol has a direct effect on the lipid raft. Ethanol is a solvent and may have therefore altered the characteristics of the lipid raft bilayer membrane. Experiments to isolate effects of ethanol on proteins or on the lipid raft are needed. It is intriguing to speculate that ethanol effects on protein–lipid raft interactions may be a common mechanism of ethanol’s toxicity. Here, we show that both c-Src and GABAA receptor, known targets for ethanol toxicity, shift out of lipid rafts in the presence of ethanol. More work is needed to define the effect of ethanol on the lipid raft proteome.

Numerous proteins require trafficking through a lipid raft compartment to initiate signal transduction. Data is accumulating on the importance of lipid rafts in basic cellular biology, implying that impairment of protein–lipid raft interactions plays a role in toxicology/clinical disease (Simons and Gerl 2010). The involvement of lipid raft dysfunction in human disease is yet relatively unexplored but may represent a novel target for therapeutic action (Escriba et al. 2008). FAS may be one example of a clinical disease caused at least in part by a perturbation of protein–lipid raft interactions. Further investigations will have important diagnostic and therapeutic implications for FAS, as well as other clinical diseases caused by protein–lipid raft disruptions.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

This work was supported by National Institutes of Health Grant AA-0 16398 (to C.F. B.). C.F. Bearer holds the Mary Gray Cobey Professorship in Neonatology. The authors have no conflicts of interest to disclose.

References

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References