Alcohol Exposure Alters the Expression Pattern of Neural Cell Adhesion Molecules During Brain Development

Authors

  • R. Miñana,

  • E. Climent,

  • D. Barettino,

  • J. M. Segui,

  • J. Renau-Piqueras,

  • C. Guerri


  • Lippincott Williams & Wilkins, Inc., Philadelphia

  • Abbreviations used: BSA, bovine serum albumin; GFAP, glial fibrillary acidic protein; NCAM, neural cell adhesion molecule; NeuAc, N-acetylneuraminic acid; PAGE, polyacrylamide gel electrophoresis; PSA, polysialic acid; SDS, sodium dodecyl sulfate; ST, sialyltransferase.

Address correspondence and reprint requests to Dr. C. Guerri at Instituto de Investigaciones Citológicas, Amadeo de Saboya, 4, 46010-Valencia, Spain. E-mail:guerri@ochoa.fib.es

Abstract

Abstract: Neural cell adhesion molecules (NCAMs) play critical roles during development of the nervous system. The aim of this study is to investigate the possible effect of ethanol exposure on the pattern of expression and sialylation of NCAM isoforms during postnatal rat brain development because alterations in NCAM content and distribution have been associated with defects in cell migration, synapse formation, and memory consolidation, and deficits in these processes have been observed after in utero alcohol exposure. The expression of NCAM isoforms in the developing cerebral cortex of pups from control and alcohol-fed mothers was assessed by western blotting, ribonuclease protection assay, and immunocytochemistry. The highly sialylated form of NCAM [polysialic acid (PSA)-NCAM] is mainly expressed during the neonatal period and then is down-regulated in parallel with the appearance of NCAM 180 and NCAM 140. Ethanol exposure increases PSA-NCAM levels during the neonatal period, delays the loss of PSA-NCAM, decreases the amount of NCAM 180 and NCAM 140 isoforms, and reduces sialyltransferase activity during postnatal brain development. Neuraminidase treatment of ethanol-exposed neonatal brains leads to more intense band degradation products, suggesting a higher content of NCAM polypeptides carrying PSA in these samples. However, NCAM mRNA levels are not changed by ethanol. Immunocytochemical analysis demonstrates that ethanol triggers an increase in PSA-NCAM immunolabeling in the cytoplasm of astroglial cells, accompanied by a decrease in immunogold particles over the plasma membrane. These findings indicate that ethanol exposure during brain development alters the pattern of NCAM expression and suggest that modification of NCAM could affect neuronal-glial interactions that might contribute to the brain defects observed after in utero alcohol exposure.

One of the most severe consequences of maternal ethanol consumption is damage to the developing CNS, which is manifested by long-term cognitive and behavioral deficits in the offspring (Streissguth et al., 1994). Indeed, clinical and experimental evidence has clearly shown that ethanol produces various disruptions in many important developmental processes, including depression of neurogenesis, delayed and aberrant migration, and anomalous morphological development (Miller, 1992). Recent studies also indicate that glial cells are also profoundly affected by ethanol, suggesting that perturbation of neuron-glia interaction may participate in the developmental defects of the brain induced by ethanol (Vallés et al., 1996, 1997; Guerri and Renau-Piqueras, 1997).

Cell recognition molecules play crucial roles in cell-cell interactions, which form the basis for organization of the brain during development. Among the most prominent cell recognition molecules is neural cell adhesion molecule (NCAM), which plays an important role in the morphogenesis, plasticity, and regeneration of the nervous system.

NCAM is expressed in the brain as three main isoforms of 180, 140, and 120 kDa, generated by alternative splicing of a single gene (Gennarini et al., 1986). During embryonic development of the brain, NCAM undergoes posttranslational modifications that involve glycosylation leading to addition of several α-2,8 polysialic acid (PSA) residues in its extracellular domains (Edelman and Crossin, 1991). NCAM-mediated cell adhesion is developmentally regulated and occurs by a homophilic binding mechanism, the strength of which is inversely proportional to the amount of PSA on opposing molecules (Sadoul et al., 1983). During the period of cell migration and neurite extension, the embryonic highly polysialylated NCAM (PSA-NCAM) is expressed abundantly throughout the brain until the early postnatal period. In some regions of the nervous system, once migration is accomplished, PSA-NCAM is down-regulated, and the most abundant NCAM is the sialic acid-poor isoforms (Chuong and Edelman, 1984). This leads to an increase in homophilic and heterophilic binding, which seems of primary importance in determination of stable synapse formation when the structure is finally elaborated (Edelman and Chuong, 1982).

Several studies have implicated NCAM in a wide variety of developmental effects in the nervous system, including growth, guidance, and bundling of axons, formation of cell layers in the retina and cerebellum, and neural crest cell migration (Fields and Itoh, 1996; Rutishauser, 1996). In addition, in vitro and in vivo experiments have shown that interference with NCAM binding and/or expression causes marked changes in morphogenesis. For example, recent studies have demonstrated that genetic deletion of NCAM 180 in mice produces defects in neuronal cell migration and causes subtle changes in hippocampal cytoarchitecture (Tomasiewicz et al., 1993), and knockout NCAM mice display an overall brain size reduction and persistent deficits in spatial learning (Cremer et al., 1994). Alterations of both NCAM levels and its sialylation state have also been observed in degenerative disease (Breen et al., 1998), developmental defects (Ibsen et al., 1983), and after exposure to some neurotoxins (Cookman et al., 1987). In fact, conversion from PSA-NCAM to adult isoforms of NCAM itself has been found to be delayed in staggerer mice in conjunction with the connectivity defects associated with the mutation (Edelman and Chuong, 1982) and in animals exposed to lead chloride during their development (Cookman et al., 1987).

Recently, we have analyzed the expression of NCAM isoforms in cortical rat astrocytes in primary culture (Miñana et al., 1998), and we have shown an accumulation of PSA-NCAM in cytoplasmic vesicles located around the Golgi apparatus and scarce expression on the cell surface of proliferating astrocytes. On maturation, both the sialyltransferase (ST) activity and PSA-NCAM decrease, and the distribution of other NCAM isoforms shifts to the cell surface. These results suggest that PSA-NCAM in astrocytes could be involved in brain morphogenesis during development.

Altered NCAM content and distribution could be a particularly attractive mechanism for explaining alterations in neuronal cell migration, neural—glial heterotopias, reduced brain growth, and other neurodevelopmental abnormalities observed after prenatal ethanol exposure. Therefore, the aim of the present study is to determine the possible relationship between ethanol-induced alterations in brain development and the pattern of expression and sialylation of NCAM isoforms. Moreover, because astrocytes are markedly affected by ethanol exposure, we have also determined whether ethanol could impair the content and distribution pattern of NCAM isoforms, participating in ethanol-induced disruption in NCAM-mediated neuronal—glial interactions during brain development.

MATERIALS AND METHODS

Animal treatment

Female Wistar rats weighing 200-250 g were used. All animals were maintained under controlled conditions of light and dark (12/12 h), temperature (23°C), and humidity (60%). Rats were either fed Lieber-DeCarli diet (Lieber and DeCarli, 1994) containing 5% (wt/vol) ethanol or were isocalorically balanced with dextrin-maltose for pair-fed controls. Female rats received the liquid diet (ethanol or control) for a minimum of 40 days before exposure to male rats. After mating, the dams were placed in separate cages and kept on ethanol or control liquid diets during gestation and lactation. During this treatment, the daily average ± SD alcohol intake of alcohol-fed dams was 14.3 ± 0.8 g of ethanol/kg of body weight. Average ± SD blood alcohol levels reached by pregnant rats and their fetuses were 22.5 ± 12 and 23.3 ± 4 mmol/L, respectively (Guerri and Sanchis, 1985; Sanchis et al., 1986). At different postnatal days animals were decapitated, brains were removed, and cerebral cortex was dissected, quickly frozen in liquid nitrogen, and stored at -70°C until used.

All experiments using rats were performed in strict compliance with the European Community Guide for the Care and Use of Laboratory Animals.

Sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) and western blot

Cerebral cortex samples were homogenized in 3 volumes of RIPA buffer (phosphate-buffered saline containing 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 1.25 mM phenylmethylsulfonyl fluoride, 40 μM leupeptin, 10 μg/ml aprotinin, and 1 mM sodium orthovanadate) and centrifuged at 100,000 g, and the supernatant was mixed with an equal volume of SDS buffer [0.125 M Tris-HCl (pH 6.8), 2% SDS, and 5% (vol/vol) 2-mercaptoethanol] and boiled for 3 min. Proteins were separated in SDS-polyacrylamide slab gels using the discontinuous gel and buffer system of Laemmli (1970), with 7% (wt/vol) polyacrylamide in the separation gel. After electrophoresis, proteins were transferred to nitrocellulose paper using a semidry electroblotter apparatus. The nitrocellulose membrane was incubated overnight with NCAM (monoclonal antibody, clone OB11; Sigma Chemical Co., St. Louis, MO, U.S.A.) and then incubated for 1 h with anti-mouse IgG-alkaline phosphatase conjugate (Promega). After 5-15 min of color development, the nitrocellulose sheets were washed and photographically reproduced. The intensity of the bands was quantified with SigmaGel version 1.0 image analysis software.

Neuraminidase treatment

To desialylate NCAM, cerebral cortex from control and alcohol-exposed pups were homogenized in 3 volumes of RIPA buffer, as described above. Then, samples (200 μg) were treated with 0.05 IU of Vibrio cholerae neuraminidase (Boehringer-Mannheim) at 37°C for 16 or 24 h in 50 mM sodium acetate (pH 5.0) containing 2 mM CaCl2. Control incubations were carried out under identical conditions but without enzyme. Enzymatic digestion was stopped by addition of SDS sample buffer (as described above) and boiled for 3 min. Western blotting was performed using an NCAM polyclonal antibody. Similar results were obtained after 16 and 24 h of treatment.

RNA extraction and ribonuclease protection assay

Total RNA was extracted by using the procedure described by Chomczynski and Sacchi (1987). Frozen cerebral cortices were homogenized in guanidium thiocyanate solution [4.0 M guanidium thiocyanate, 100 mM Tris-C1 (pH 7.5), and 1% 2-mercaptoethanol] using an Ultraturrax homogenizer. The final RNA pellet was resuspended in 0.4 ml of 2 M ammonium, 0.9 ml of cold 100% ethanol was added, and the RNA was allowed to precipitate at -20°C overnight. The total amount of RNA extracted was calculated by measuring the absorbance at 260 nm.

The RNase protection analysis was performed with an RPAII ribonuclease protection assay kit (Ambion). The NCAM 180 probe was generated by RT-PCR. The upper primer (NC1) encompasses nucleotides 2,412-2,432 in exon 17 of the rat NCAM 140 cDNA (Small et al., 1987), and the lower primer (NC2) consists of nucleotides 94-114 of the sequence of exon 18 of mouse NCAM 180 cDNA (Barthels et al., 1988). Firststrand cDNA was prepared from rat embryonic day 21 brain total RNA using Moloney murine leukemia virus reverse transcriptase and (dT)12-18 as primer. This cDNA was used as a template for PCR with NC1 and NC2 oligonucleotide primers to generate a 350-bp rat NCAM 180 cDNA fragment covering the junction between exons 17 and 18. The cDNA probe was cloned in vector pGEM-T (Promega). To generate a cyclophilin probe to be used as loading control, a 126-bp NcoI fragment from rat cyclophilin cDNA (Danielson et al., 1988) was cloned into vector pBluescript-KS (Stratagene) (L. Ortiz and D. Barettino, unpublished data). Labeled probes were synthesized from linearized plasmid templates by in vitro transcription with T3 (cyclophilin) or T7 (NCAM) RNA polymerase in the presence of [α-32P]CTP (Amersham). Following digestion of the DNA template with 2 ml of RQ1 DNase (1 U/ml; Promega), the probes were purified through Sephadex G-50 spun columns equilibrated with 10 mM Tris-HCl, pH 7.5. The sizes of the resultant fragment following ribonuclease protection assay were 350, 249, and 126 nucleotides for NCAM 180, NCAM 140, and cyclophilin, respectively. The protected products were resolved on a 6% sequencing gel and visualized by autoradiography.

Microsomal preparation

Brain cortices from litters of control or alcohol-fed rats were homogenized at 0°C in 10 mM sodium cacodylate (pH 7.0) using a Potter-Elvehjem system with a Teflon pestle. The homogenates were centrifuged for 15 min at 9,000 g, and the resulting supernatant was then centrifuged for 60 min at 120,000 g. The pellet was rehomogenized in buffer to yield a concentration of 15-20 mg of brain protein/ml. These preparations were stored frozen at -20°C until use.

Enzymatic activities

Poly(α-2,8-ST) activity was determined by the procedure described by Easton et al. (1995), using colominic acid as an exogenous acceptor for ST. In brief, a microsomal suspension from brain cortical membranes (0.2-0.5 mg of protein) was incubated in a total volume of 50 μl containing 10 mg of colominic acid (Calbiochem), 25 nmol of CMP-N-acetyl[3H]neuraminic acid ([3H]NeuAc; 3.8 Ci/mol; DuPont-New England Nuclear), 5 μmol of sodium cacodylate buffer (pH 6.5), 0.25 μmol of MnCl2, and 0.15 μl of Triton X-100. Incubations were carried out for 16 h at 37°C. Reactions were stopped by addition of 200 μl of 0.05 M ammonium acetate (pH 5.4) and cooling on ice. Samples were then centrifuged in an Eppendorf centrifuge for 4 min at maximal speed. Pellets were washed once with the same amount of buffer and centrifuged again. The combined supernatants were passed over a column (0.7 × 50 cm) of Bio-Gel P-4 (200-400 mesh) equilibrated with 0.05 M ammonium acetate (pH 5.4) at a flow of 10 ml/h at room temperature. Fractions of 0.5 ml were collected and assayed for radioactivity by liquid scintillation counting. Incorporation was calculated from the sum of counts in the colominic acid-containing fractions. The values were corrected for the incorporation into endogenous acceptors by assaying incubations lacking colominic acid and subtracting the counts found in these samples.

Sialidase activity was determined as previously described (Ghosh et al., 1998) in the microsomal fraction. The assay was assessed in a total volume of 100 μl containing 1 mmol/L substrate as bound sialic acid (ganglioside mixture from bovine brain; Calbiochem), 0.1 mg/ml sodium deoxycholate, 20 μg bovine serum albumin (BSA), 45 mmol/L acetate, 135 mmol/L NaCl, 8 mmol/L CaCl2, and 0.2-0.4 mg of microsomal protein. Incubation was carried out for 4 h at 37°C. The reaction was stopped by adding 0.2 ml of 0.25 mol/L periodate. Free NeuAc content was determined by the thiobarbituric acid method of Warren (1959). Enzyme and substrate controls were incubated simultaneously, and their values were subtracted from the experimental values.

Immunocytochemistry

Rat pups at postnatal day 7 were perfused with 0.5% glutaraldehyde and 4% formaldehyde in 0.1 M PIPES buffer, pH 7.2. Brains were then removed by dorsal craniotomy, immersed in the same fixative solution for 60 min at 4°C, washed in buffer, incubated for 60 min in 50 mM NH4Cl, dehydrated in methanol, and embedded in Lowicryl K4M as previously described (Renau-Piqueras et al., 1989a, 1997).

PSA-NCAM and NCAM isoform location was carried out using the immunogold procedure as previously described (Renau-Piqueras et al., 1989a,b; Iborra et al., 1992; Miñana et al., 1998). Ultrathin sections (80 nm) mounted on Formvarcoated nickel grids were floated for 30 min on 0.1% BSA-Tris buffer [20 mM Tris-HCl and 0.9% NaCl (pH 7.4) containing 0.1% BSA, Type V] supplemented with inactivated fetal calf serum and then transferred to droplets of 0.1% BSA-Tris buffer containing 1% fetal calf serum, an anti-glial fibrillary acidic protein (GFAP) antibody, polyclonal or monoclonal, used to select astroglial cells (Sigma), and anti-PSA-NCAM (1:300) or anti-NCAM (1:200) antibodies. The sections were incubated in a moist chamber for 180 min at 37°C. Although under these conditions the number of gold particles was reduced, this procedure results in a very specific, repetitive labeling for both antibodies. After three rinses with 0.1% BSA-Tris buffer for 10 min each, the grids were placed on droplets of 0.1% BSA-Tris buffer containing 0.5% Tween 20, 5% fetal calf serum, and anti-mouse and rabbit IgG-gold complexes (1:10; Sigma). In all the experiments 10-nm particles were used to localize NCAM, and 5-nm particles were used to localize GFAP. The incubation time was 60 min at room temperature, as above. After two 30-min rinses with 0.1% BSA-Tris buffer and a rinse in double distilled water, the sections were air-dried and finally counterstained with lead citrate (5 s) and uranyl acetate (20 min). Controls were incubated without the first antibody. Clumping of gold particles in the anti-IgG-gold complexes was checked routinely. Astroglial cells were recognized on the basis of morphological and/or immunocytochemical criteria.

Micrographs were taken of germinal matrix and subventricular zone cells showing both anti-NCAM (10-nm gold particles) labeling and intermediate filament bundles decorated with anti-GFAP (5-nm gold particles).

Statistical analysis

Results are reported as mean ± SD values. Data were analyzed using two-way ANOVA.

RESULTS

Effect of ethanol on developmental profiles of NCAM isoforms in cerebral hemispheres during postnatal brain development

We first analyzed the profiles of NCAM isoforms in cerebral cortex from control litters as a function of postnatal age. Western blot analysis of cerebral hemispheres from control litters revealed that during the neonatal period (days 5-7) only a broad band of high-molecular-mass isoform (200-250 kDa) or PSA-NCAM appeared (Fig. 1). However as the brain matures (days 7-30 of age), PSA-NCAM content gradually diminished in parallel with the increase in intensity of other bands of lower molecular masses corresponding to NCAM 180 and NCAM 140 (Fig. 1). It is noteworthy that under our experimental conditions and the antibodies used, we could not detect the NCAM 120 isoform. Densitometric analysis of the NCAM 180 and NCAM 140 immunoreactive bands obtained from four different control litters indicates that although NCAM 180 increased twofold between day 7 to day 21 and then decreased, NCAM 140 increased up to day 30 of the postnatal period, although it decreased in the adult stage (data not shown). This gradual conversion of the embryonic form of PSA-NCAM to less sialylated or adult isoforms occurring during vertebrate development confirmed previous reports suggesting it is a primary determinant of stable cell-cell contact in the adult nervous system (Edelman and Chuong, 1982).

Figure 1.

Western blot and densitometric analysis of NCAM [200 (A), 180 (B), and 140 kDa (C)] from control and ethanol-exposed rats during postnatal cerebral cortex development. Each lane contains 100 μg of total protein from homogenates. A representative western blot is shown. A-C: Data are mean ± SD values of two separate determinations from four different litters. *P < 0.05 in A, *P < 0.01 in B, *P < 0.001 in C for treatment effect by two-way ANOVA.

FIG. 1.

Exposure to alcohol during brain development increased PSA-NCAM levels and delayed the loss of PSA-NCAM expression and its conversion into less sialylated NCAM isoforms. At the same time, the levels of NCAM 180 and NCAM 140 were reduced at the end of postnatal development (Fig. 1).

To confirm the higher levels of PSA-NCAM in cerebral cortex from alcohol-exposed litters, we treated cortical homogenates with neuraminidase, an enzyme that cleaves sialic acid from PSA-NCAM (Rothbard et al., 1982; Toikka et al., 1998). In agreement with previous findings in chicken, frog, and mouse brains (Gennarini et al., 1986; Sunshine et al., 1987), after enzymatic treatment of samples containing PSA-NCAM (days 5 and 7), in both control and alcohol-exposed brains the PSA smear disappeared, and three bands of apparent molecular masses of ∼180, 140, and 100 kDa were observed on SDS-PAGE (Fig. 2A). These bands were more intense in samples from alcohol-exposed animals (Fig. 2B), suggesting a higher content of NCAM polypeptide chains carrying PSA in ethanol-exposed samples compared with samples from untreated control animals. Similar results were obtained after 16 or 24 h of enzymatic treatment.

Figure 2.

Western blot of neuraminidase-treated PSA-NCAM from control and alcohol-exposed rat cerebral cortex at days 5 and 7 of postnatal development. Samples were treated with neuraminidase as described in Materials and Methods. After enzymatic treatment three polypeptides of apparent molecular masses of ∼180, 140, and 100 kDa appeared in SDS-PAGE. Molecular sizes of standard proteins assayed in parallel are indicated. A: Representative western blotting before and after enzymatic treatment. B: Densitometric analysis of the degradation products of NCAM polypeptides. Data are mean ± SD (bars) values of three different experiments. *P < 0.05, **P < 0.001 versus control samples.

FIG. 2.

Changes in NCAM splicing patterns during brain development: effect of ethanol exposure

Levels of mRNAs encoding NCAM were measured to determine whether the ethanol-induced changes in the developmental profile in NCAM isoforms are the consequence of posttranscriptional or posttranslational modifications. We used a sensitive ribonuclease protection assay to quantify the levels of mRNAs encoding the NCAM 180 and NCAM 140. Figure 3 shows that the levels of mRNAs encoding both isoforms decreased during postnatal brain development. Alcohol exposure did not cause significant variations in mRNA levels, suggesting that translational modifications might be involved in the changes in NCAM levels observed in brains exposed to alcohol.

Figure 3.

NCAM mRNA levels in cerebral cortex (cortical hemispheres) from control and alcohol-exposed rats during postnatal development. A: Representative autoradiogram of RNase protection assay illustrates the levels of NCAM 180 and NCAM 140. B: Histogram illustrates the relative levels of mRNAs encoding NCAM 180. C: NCAM 140 levels obtained by densitometric analysis of the bands. The values obtained were normalized to cyclophilin levels. Data are average ± SD (bars) values of two separate determinations from four different litters.

FIG. 3.

ST and sialidase activities in developing brain

Previous results have demonstrated that NCAM polysialylation is primarily regulated at the biosynthesis level and that the Golgi enzyme ST participates in this process. Therefore, to analyze further the changes in PSA-NCAM following ethanol exposure, we measured the levels of ST in the microsomal fraction from developing brains. In agreement with previous results (Breen et al., 1987), ST activity decreased during the course of brain development (Fig. 4). This decline in activity has been correlated with a reduction in the polysialylation on NCAM. It is surprising that exposure to alcohol during brain development significantly decreases the levels of ST activity.

Figure 4.

Changes in CMP-NeuAc(NeuAcα2-8)n (colominic acid) ST activity in microsomes from brain cortices of control and ethanol-exposed rats during postnatal development. Data are average ± SD (bars) values of five microsomal preparations from different litters. *P < 0.001 versus control for treatment effects by two-way ANOVA.

FIG. 4.

However, because the processes of sialylation-desialylation seem to be relevant to the expression of adhesion molecules, we then analyzed the levels of membranebound sialidase to determine whether the developmental changes in NCAM pattern following ethanol exposure are due to an increase in sialidase activity. Figure 5 shows that the sialidase activity from microsomal membranes increased from postnatal day 5 to 21 and that ethanol exposure causes minor changes in the levels of this enzymatic activity.

Figure 5.

Changes in sialidase activity in microsomes from brain cortices of control and ethanol-exposed rats during postnatal development. Data are average ± SD (bars) values, of five microsomal preparations from different litters.

FIG. 5.

Immunocytochemistry of NCAM in brain astroglial cells

We have recently demonstrated that during the proliferation of cortical astrocytes (which occurs during the neonatal period in the rat) (Guerri et al., 1990), these cells expressed PSA-NCAM in small cytoplasmic vesicles located in or near the Golgi apparatus. As the cells differentiated, PSA-NCAM gradually disappeared, and the mature form, NCAM 140, is expressed at the cell surface (Miñana et al., 1998). In addition, previous findings from our laboratory indicate that ethanol induces retention of several secretory and surface glycoproteins in astrocytes (Renau-Piqueras et al., 1992; Valles et al., 1994). Therefore, to investigate whether the increase in PSA-NCAM content observed in ethanol-exposed brains is due to an accumulation of this isoform in the cytoplasm of astroglial cells, we performed an in situ study using control and alcohol-exposed brains.

Immunogold analysis of anti-total NCAM and anti-PSA-NCAM binding sites in those cells showing bundles of intermediate filaments immunolabeled with anti-GFAP (Fig. 6A-E, 5-nm particles) from the germinal matrix or ventricular and subventricular zone on postnatal day 7 brains revealed that gold particles corresponding to anti-PSA-NCAM appeared scattered over the plasma membrane, Golgi apparatus, cytoplasm, and vesicles (Fig. 6A and F). In some sections, PSA immuno-reactivity on the plasma membrane where glial cells and neurons are in contact was also observed (Fig. 6A). However, as previously shown (Miragall et al., 1990), PSA-NCAM labeling was absent in the bundles of intermediate filaments (Fig. 6A and E). On the other hand, we also found that alcohol exposure induced an accumulation of anti-PSA-NCAM binding sites in the cytoplasm and a decrease of labeling over the plasma membrane (Fig. 6B). Quantitative analysis of the labeling distribution at postnatal day 7 revealed that whereas 62% of the gold particles appeared in the cytoplasm and/or near the Golgi apparatus, 38% of PSA-NCAM immunolabeling was detected at the plasma membrane. Alcohol exposure induces a significant increase (two- to fivefold) of PSA-NCAM immunolabeling (Table 1) and changes its distribution, inducing an accumulation of gold particles in the cytoplasm and decreasing the labeling over the plasma membrane (Table 1).

Figure 6.

Double immunocytochemical analysis of the intracellular distribution of (A, B, and F) total PSA-NCAM and (C and D) total NCAM in control (A and C) and prenatally exposed to alcohol (B and D) astroglial cells in the germinal, ventricular, and subventricular layers of postnatal day 7 rat brain. Astroglial cells were characterized by the presence of intermediate filaments (IF) and anti-GFAP labeling (5-nm particles, small arrows). The 10-nm particles locate total NCAM or PSA-NCAM. Neurons were identified by morphological criteria and by the absence of IF and GFAP labeling. In control astroglial cells, PSA-NCAM and total NCAM appeared frequently located over the plasma membrane (A and C). In contrast, few gold particles corresponding to anti-PSA-NCAM (B) or total NCAM (D) binding sites were found over the plasma membrane of alcohol-exposed cells, where particles accumulate in the cytoplasm (see also Table 1). In a few cell sections we have observed PSA-NCAM and total NCAM immunoreactivity (A and C) on the plasma membrane where the glial cells and neurons (asterisk) are in contact (arrowheads). As shown (A-C and E in detail), astroglial areas containing bundles of IF lacked PSA-NCAM or total NCAM labeling. F: PSA-NCAM label over the Golgi complex in control cells. A similar labeling was observed after incubation with the anti-total NCAM antibody. Similar results were also found in cells exposed to alcohol. Bars = 0.2 μm.

Table 1. Gold particle density corresponding to anti-PSA-NCAM and anti-total-NCAM binding sites in astroglial cells of postnatal day 7 rat brains
  Postnatal day 7
Group Postnatal day 7 (no. of particles/μm2) % cytoplasm% membrane
  1. Data are mean ± SD values of a minimum of 20 electron micrographs per case. The distribution, as a percentage, of label over cytoplasm and plasma membrane of postnatal day 7 cells is also given.

  2. ap < 0.05, control versus ethanol treatment.

Control
PSA-NCAM0.065 ± 0.02962 ± 738 ± 4
Total NCAM0.063 ± 0.03764 ± 636 ± 3
Ethanol
PSA-NCAM0.188 ± 0.046a82 ± 9a18 ± 3a
Total NCAM0.146 ± 0.079a73 ± 8a27 ± 2a

FIG. 6.

TABLE 1.

Polyclonal anti-NCAM antibody, which recognizes the different NCAM isoforms (including PSA-NCAM), was reactive with all the astroglial cells examined. The labeling pattern was very similar to that described for PSA-NCAM, and gold particles were localized in both control and ethanol-treated cells over the cytoplasm, plasma membrane (Fig. 6C and D), Golgi apparatus, and processes (data not shown). As occurred in sections incubated with anti-PSA-NCAM, in some sections gold particles appeared over the plasma membrane where two cells are in contact (Fig. 6C). Using this antibody we also observed that alcohol exposure increased the density of anti-NCAM binding sites and induced an accumulation of gold particles in the cytoplasm, decreasing the immunolabeling over the plasma membrane (Fig. 6D and Table 1). Similar results were observed in astrocytes in primary culture obtained from control and animals prenatally exposed to alcohol (data not shown).

DISCUSSION

Development of the nervous system is a complex process involving both the formation of new structures and their precise interconnections. A large body of evidence indicates that NCAMs participate in several aspects of neural development, establishing specific patterns of cell-cell recognition and complex connective patterning suggesting that NCAM may be involved in morphogenesis (Edelman and Crossin, 1991). In accordance with previous findings, our results show that during brain development and maturation NCAM isoforms are differentially temporally expressed. Thus, at early postnatal rat brain development (days 5-10), PSA-NCAM content is high and then is progressively down-regulated as the adult NCAM isoforms (NCAM 180, NCAM 140, and NCAM 120) began to appear throughout the postnatal period. Unfortunately, under our conditions we could not detect NCAM 120, although it has been shown that NCAM 120 is the latest isoform to appear during development (Chuong and Edelman, 1984; Nagata and Schachner, 1986). Posttranslational modifications seem to be involved in the developmental pattern of NCAM expression because a decrease in the total NCAM message was observed during brain development. Differences in the turnover of NCAM isoforms and/or mRNA stabilization might account for this developmentally regulated pattern (Gennarini et al., 1986). The addition of PSA to NCAM polypeptides appears to be the main posttranslational process controlling the sialylation state of NCAM, and Golgi enzyme STs are involved in this process (Breen et al., 1987). In fact, we show here that ST is developmentally regulated, with its activity being high during the early neonatal period and decreasing progressively during brain development. This decrease in activity has been correlated with the polysialylation state of NCAM.

Polysialylation of NCAM seems to be a critical functional feature for NCAM-mediated cell interaction and function. High expression of PSA-NCAM, during early development, appears to play a permissive role allowing structural remodeling by decreasing cell adhesion mediated by NCAM and thereby facilitating the guidance and targeting of axons and migration of neuronal and glial precursors (Rutishauser, 1996). However, down-regulation of PSA and NCAM appearance on the cell surface in both astroglial (Miñana et al., 1998) and developing neurons might create a suitable environment to reduce the rate of axonal growth at or near the target, making NCAM available in its highly adhesive state to mediate or modulate glial-neuronal interaction-dependent processes. Indeed, alterations in the pattern of NCAM polysialylation or failure to down-regulate PSA-NCAM during the postnatal synaptic elaboration results in gross structural brain deficits (Edelman and Chuong, 1982), as occurs in staggerer mice or after exposure to some toxicants. Chronic lead exposure impairs the expected decrease in NCAM polysialylation state in the period when final synapse formation is completed (Cookman et al., 1987), leading to persisting neuroplastic deficits associated with memory consolidation (Murphy and Regan, 1999).

In the present work we demonstrate that ethanol exposure during brain development increases PSA-NCAM levels in the neonatal period and delays its down-regulation when neuronal outgrowth and synaptic elaboration are accomplished. Our data on neuraminidase treatment demonstrate that neonatal brain has a higher content in NCAM polypeptides carrying PSA. An increase in PSA-NCAM has also been observed in the developing chick brain exposed to alcohol (Kentroti et al., 1995). It is surprising that in our results the increase in PSA-NCAM is accompanied by a decrease in ST activity. Although at present we do not know the reasons for the latter results, some evidence (Kiss and Rougon, 1997) indicates that, in addition to ST activity, other biosynthesis-independent processes might participate in the regulation of PSA-NCAM expression (see below).

Although the increase of PSA-NCAM induced by ethanol could be derived from neurons or other types of cells, we investigate here the involvement of astroglial cells in the PSA-NCAM up-regulation because these cells are important targets of alcohol toxicity (Guerri and Renau-Piqueras, 1997), they accumulate surface glyco-proteins when exposed to alcohol (Renau-Piqueras et al., 1992; Vallés et al., 1994), and they express PSA-NCAM during their proliferation (Miñana et al., 1998), which in the rat takes place during the neonatal period. Using immunocytochemical analysis the present results demonstrate that ethanol exposure during brain development induces an accumulation of PSA-NCAM in astroglial cells. Alterations in NCAM polysialylation and its distribution at the cell surface may affect neuronal-glial interactions that may lead to deficits in cell migration, neuronal growth patterns, and synaptic connections observed after in utero alcohol exposure (Miller, 1992).

In addition to the alteration of PSA-NCAM temporal pattern, ethanol exposure also causes a significant reduction in the levels of NCAM 180 and NCAM 140 at the end of the postnatal period, when synapse formation is completed (Guerri, 1987). Because NCAM 180 and NCAM 140 are primarily located at the synapses and they are involved in synaptic stabilization (Persohn and Schachner, 1987), a reduction in the levels of these proteins might affect cell-cell interactions, leading to a reduction in axonal growth and synaptic remodeling and stabilization. In fact, ethanol has been shown to alter neuronal growth patterns and neuritic branching in culture (Kentroty et al., 1995), and in vivo alcohol exposure affects synaptogenesis (Guerri, 1987; Inomata et al., 1987).

Although the precise mechanisms by which ethanol exposure during brain development induces abnormalities in the pattern of expression and sialylation of NCAM isoforms are as yet unknown, there are several mechanisms by which ethanol could participate in these alterations. First, ethanol might affect the transcription of the different NCAM isoforms. However, our results on NCAM mRNA splicing patterns would not support this possibility because we did not observe mRNA level changes between control and alcohol-exposed brains.

The second possibility is that alcohol may alter the glycosylation process modifying the content of sialic acid on NCAM isoforms. This hypothesis is supported by findings demonstrating that chronic ethanol intake and prenatal ethanol exposure cause alterations in the glycosylation process and reduce galactosyltransferase and ST activities in liver (Renau-Piqueras et al., 1989b, 1997; Guasch et al., 1992) and brain (Hale et al., 1998). NCAM polysialylation seems to be primarily regulated by the Golgi enzyme ST. It is interesting that two different polysialyltransferases have recently been cloned and characterized (Eckhardt et al., 1995; Kojima et al., 1995). Both enzymes are able to catalyze the synthesis of PSA, and they are differently expressed and regulated during development (Hildebrandt et al., 1998); however, the specific role of each of these enzymes remains unclear. We show in the present work that ST activity decreases throughout the course of brain development and that alcohol exposure reduces this activity. Therefore, a possible alteration in the glycosylation of these sialoglycoproteins by ethanol might account for some of the changes observed in NCAM expression, e.g., NCAM 180 and NCAM 140, during brain development.

The third possible mechanism is that alcohol may alter cell surface PSA-NCAM and/or NCAM expression. The mechanisms that control the levels of PSA-NCAM and the molecular and cellular sequence of events contributing to PSA-NCAM expression in the cell surface remain to be clarified. However, some results suggest that biosynthesis-independent mechanisms involving Ca2+-dependent processes, together with the exo/endocytosis pathway, might participate in this process (Kiss and Rougon, 1997; Bruses and Rutishauser, 1998). Alterations in exocytosis (Miñana et al., 1998) and/or in internalization and recycling of PSA-NCAM could explain the effects of prenatal alcohol exposure on the accumulation or on the up-regulation of PSA-NCAM during early postnatal development. Our results with neuraminidase treatment suggest that NCAM polypeptides carrying PSA are up-regulated during the neonatal period. Up-regulation of PSA-NCAM might occur by inhibiting its transport to the plasma membrane by monensin (Scheidegger et al., 1994), and preliminary data in our laboratory with astrocytes treated with monensin also support this conclusion. We show here that alcohol exposure increases the cytoplasmic levels of PSA-NCAM and reduces its expression at the cell surface in in situ astroglial cells, suggesting an alteration in PSA-NCAM transport to the plasma membrane. Similar results were observed in astrocytes in primary culture exposed to alcohol (authors' unpublished data). Calcium influx associated with electrical activity, transmembrane signaling (Kiss and Rougon, 1997), and intracellular and extracellular calcium levels have been suggested as regulating the mobilization of PSA-NCAM to the cell surface (Kiss et al., 1994), and alcohol seems to modify calcium homeostasis affecting voltage- and receptor-operated calcium channels (Catlin et al., 1998). In addition, alcohol seems to alter the endocytosis process in liver (Tuma et al., 1991, 1996) and in astrocytes (Megías et al., 2000), inducing an accumulation of glycoproteins in astroglial cells (Renau-Piqueras et al., 1992; Vallés et al., 1994). These results suggest that alterations in the processes of exocytosis and endocytosis may be a general mechanism of ethanol toxicity in different cell types.

Alternatively, the capacity of ethanol to interact with the membrane component might promote an antiadhesive environment, disrupting homophilic and heterophilic interactions with other cell adhesion molecules, such as L1, and thus affecting cell surface expression of NCAM. It is known that NCAM enhances the homophilic trans-binding activity of L1 (Horstkorte et al., 1993), and ethanol, at low concentrations, is known to alter the L1-dependent neurite outgrowth (Bearer et al., 1999) and inhibit cell adhesion or aggregation (Charness et al., 1994). Therefore, ethanol-induced disruption in NCAM-NCAM or NCAM-L1 interactions might also participate in the observed changes in NCAM levels during development.

Finally, considering that NCAMs are involved in morphogenesis and plasticity of the nervous system by regulating cell-cell interaction, ethanol-induced alteration in their pattern of expression during critical periods of brain development might underlie the defects in neural migration, glial development, synaptogenesis, and plasticity observed after in utero alcohol exposure and in fetal alcohol syndrome.

Ancillary