Ethanol impairs insulin-stimulated neuronal survival in the developing brain: role of PTEN phosphatase.

Gestational exposure to ethanol causes fetal alcohol syndrome, which is associated with cerebellar hypoplasia. Previous in vitro studies demonstrated ethanol-impaired neuronal survival with reduced signaling through the insulin receptor (IRbeta). We examined insulin signaling in an experimental rat model of chronic gestational exposure to ethanol in which the pups exhibited striking cerebellar hypoplasia with increased apoptosis. Immunoprecipitation and Western blot analyses detected reduced levels of tyrosyl-phosphorylated IRbeta, tyrosyl-phosphorylated insulin receptor substrate-1 (IRS-1), and p85-associated IRS-1 but no alterations in IRbeta, IRS-1, or p85 protein expression in cerebellar tissue from ethanol-exposed pups. In addition, ethanol exposure significantly reduced the levels of total phosphoinositol 3-kinase, Akt kinase, phospho-BAD (inactive), and glyceraldehyde-3-phosphate dehydrogenase and increased the levels of glycogen synthase kinase-3 activity, activated BAD, phosphatase and tensin homolog deleted in chromosome 10 (PTEN) protein, and PTEN phosphatase activity in cerebellar tissue. Cerebellar neurons isolated from ethanol-exposed pups had reduced levels of insulin-stimulated phosphoinositol 3-kinase and Akt kinase activities and reduced insulin inhibition of PTEN and glycogen synthase kinase-3 activity. The results demonstrate that cerebellar hypoplasia produced by chronic gestational exposure to ethanol is associated with impaired survival signaling through insulin-regulated pathways, including failure to suppress PTEN function.

Previous experiments demonstrated that neuronal loss following ethanol exposure was mediated by apoptosis (8 -10) or mitochondrial dysfunction (10 -12), and recent studies correlated these adverse effects of ethanol to inhibition of growth factor-stimulated survival signaling (9,(11)(12)(13). In the developing CNS, insulin and insulin-like growth factor type 1 (IGF-1) 1 receptors are abundantly expressed (14 -16), and the corresponding growth factor-stimulated responses are critical mediators of neuronal growth, viability, energy metabolism, and synapse formation. Because insulin and IGF-1 signaling pathways are among the important targets of ethanol neurotoxicity in immature nervous system (9,13,17,18), neuronal loss associated with microencephaly in ethanol-exposed fetuses may be caused, in part, by ethanol inhibition of insulin/IGF-1stimulated survival mechanisms.
Our previous studies (11,13) using in vitro or in vivo models demonstrated that ethanol profoundly inhibits insulin-stimulated survival and mitochondrial function in cultured neuronal cells. The present study was designed to examine the effects of chronic gestational exposure to ethanol on insulin-, IRS-1-, and PI3 kinase-mediated signaling in the intact brain to determine the extent to which abnormalities in these pathways correlate with ethanol-induced developmental defects in the CNS. PTEN expression and activity were also examined, because PTEN dephosphorylates and negatively regulates PI3 kinase function (38), and the effects of ethanol on phosphatase and tensin homolog deleted in chromosome 10 (PTEN) were unknown. Cerebellar tissue was studied, because it represents a major in vivo target of ethanol neurotoxicity (2,39).

EXPERIMENTAL PROCEDURES
In Vivo Model of Chronic Ethanol Exposure-Long-Evans female rats were adapted to an ethanol-containing or isocaloric control liquid diet (BioServ, Frenchtown, NJ) over a 3-week interval, after which they were mated with normal males. Ethanol comprised 11.8, 23.6, and 35.4% of the caloric content of the feedings during the first, second, and third weeks of adaptation. The 35.4% ethanol-containing or control diets were maintained throughout pregnancy. Using this protocol, the serum ethanol concentrations in the rats ranged from 25 to 43 mM, which is within the range observed in human disease states (40). Rats were monitored daily to ensure equivalent caloric consumption and maintenance of body weight. Typically, in the ethanol-fed group, the litter sizes were reduced by 20%, and pup mean body weight was reduced by 10 to 15%. Studies were conducted with cerebella harvested from control and ethanol-exposed postnatal day 2 (P2) pups to evaluate the effects of ethanol in the early postnatal period and prior to the occurrence of any major compensatory developmental responses. Cerebellar tissue was fixed in Histochoice fixative (Amresco, Solon, OH) and embedded in paraffin. Histological sections were stained with hematoxylin and eosin to detect morphological abnormalities. Fresh cerebellar tissue was snap-frozen in liquid nitrogen and stored at Ϫ80°C for use in protein studies, assays of kinase or phosphatase activity, and measurement of PTEN mRNA levels.
In Situ Assays for Apoptosis-The terminal transferase dUTP endlabeling (TUNEL) assay was used to detect nicked or fragmented DNA in cryostat sections of cerebella. TUNEL assays were performed using fluorescein-labeled dUTP ([Fl]dUTP; Invitrogen) and terminal deoxynucleotide transferase (17). The labeled DNA was detected with biotinylated secondary antibody, alkaline phosphatase-conjugated Streptavidin, and 5-bromo-4-chloro-3-indolyl phosphate/nitroblue tetrazolium (BCIP/NBT) as the substrate. To detect nuclear pyknosis and fragmentation characteristic of apoptosis, adjacent sections were stained with Hoechst H33258 (1 g/ml in phosphate-buffered saline) for 2 min at room temperature. The slides were rinsed thoroughly in phosphate-buffered saline, cover-slipped with Vectashield mounting medium (Vector Laboratories, Burlingame, CA), and examined by fluorescence microscopy. Adjacent sections were immunostained with polyclonal antibodies to activated (cleaved) caspase-3 (Cell Signaling, Beverly, MA). Histological sections were prepared for immunostaining according to the manufacturer's protocol. Immunoreactivity was detected with biotinylated secondary antibody, alkaline phosphatase-conjugated Streptavidin, and the BCIP/NBT substrate.
Kinase Assays-PI3 kinase activity was measured in p85 immunoprecipitates (42) obtained from individual 500-g protein samples using rabbit polyclonal anti-p85 (1 g/ml) and protein A-Sepharose (Amersham Biosciences). Immunoprecipitates complexed with protein A-Sepharose were suspended in 50 l of TNE buffer (10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA), and reactions were initiated by sequentially adding 20 g of sonicated phosphatidylinositol (10 l), 10 l of 100 mM MgCl 2 , and 5 l of [␥-32 P]ATP working solution composed of 0.88 mM [␥-32 P]ATP (30 Ci of [␥-32 P]ATP/3000 Ci/mmol), 20 mM MgCl 2 , and 150 mM cold ATP. Reactions were incubated for 10 min at 37°C and 300 rpm and stopped by adding 15 l of 6 N HCl. Phospholipids extracted with chloroform/methanol were analyzed by thin layer chromatography using plates pre-coated with 1% potassium oxalate (Merck). PI3 kinase activity was detected by film autoradiography and quantified using the Kodak Digital Science Imaging Station.
PTEN Studies-PTEN expression was measured by Western blot analysis and real-time quantitative reverse-transcribed (RT)-PCR assays. In addition, PTEN phosphatase activity was measured in PTEN immunoprecipitates using the Biomol Green Reagent according to the manufacturer's protocol. Phosphatase inhibitors were omitted from the lysis buffer. For the real-time RT-PCR studies, total RNA was isolated from cerebellar tissue homogenates using TRIzol reagent (Invitrogen) according to the manufacturer's protocol. Samples containing 2 g of RNA were reverse-transcribed using the AMV First Strand cDNA synthesis kit (Roche Molecular Biochemicals) and random oligodeoxynucleotide primers. Highly conserved regions of PTEN and 18 S cDNAs isolated from rat cerebellar tissue by RT-PCR were cloned into the PCRII vector (Invitrogen) and used to generate standard curves for determining transcript abundance. PCR amplifications were performed using 25-l reaction volumes containing 20 ng of RT product, 0.4 M each of forward and reverse primers (Table I), and SYBR Green I PCR reagent (Applied Biosystems, Foster, CA). The amplified signals were detected continuously with the Bio-Rad iCycler and iCycler iQ Multi-Color Real Time PCR Detection System (Hercules, CA). The following real-time PCR amplification protocol was used: 1) initial denaturation, 95°C for 10 min; 2) a three-segment amplification and quantification In Vitro Experiments-In vitro experiments were used to examine the effects of ethanol on PTEN expression, phosphorylation, and phosphatase activity in CNS neurons and to determine whether the responses observed in vivo were mediated by impaired insulin or IGF-1 signaling. Primary neuronal cultures were generated with cerebellar tissue harvested from P2 pups (44). Fluorescence-activated cell sorting demonstrated that greater than 95% of cells isolated from control or ethanol-exposed pup cerebella were neuronal as evidenced by the immunoreactivity with antibodies to HuC/HuD neuron-specific RNA binding protein (45) (Molecular Probes, Eugene, OR). Cultures were maintained in Dulbecco's modified Eagle's medium supplemented with 5% fetal calf serum, 2 mM glutamine, 10 mM non-essential amino acid mixture (Invitrogen), 25 mM KCl, and 9 g/liter glucose. After overnight seeding, cells were treated with 6 M cytosine arabinoside to inhibit DNA synthesis. Cultures were exposed to 50 mM ethanol or nothing for 2 days using sealed humidified chambers (17,42), after which they were serum-starved for 12 h and then stimulated with 50 nM insulin (Humulin; Eli Lilly & Co., Indianapolis, IN) or 25 nM IGF-1 for 0, 15, 30, 60, or 120 min in the presence or absence of 50 mM ethanol. PTEN protein was detected by Western blot analysis. PTEN phosphorylation was evaluated by PTEN Western blot analysis of anti-phospho-serine/phospho-threonine immunoprecipitates. PTEN phosphatase activity was measured in PTEN immunoprecipitates using the Biomol Green Reagent (Cell Signaling, Beverly, MA).
To examine the effects of ethanol on insulin-stimulated viability and kinase activity, primary neuronal cultures were generated with cerebella harvested from control or ethanol-exposed P1 pups. However, to mimic the in vivo model in which the ethanol exposure was discontinued after birth, the cultured cells were not further exposed to ethanol. Micro-cultures (96-well plates) in which 5 ϫ 10 4 viable cells (determined by Trypan Blue exclusion) were seeded per well were used for viability assays, and 60-mm Petri dish cultures were used for kinase or PTEN phosphatase assays. To measure insulin-stimulated responses, after 1 day in culture, the cells were serum-starved for 12 h and then stimulated with 50 nM insulin (Humulin; Eli Lilly & Co., Indianapolis, IN). Viability was measured after 24 h of insulin stimulation using the crystal violet assay (13,17,42). IRS-1-associated PI3 kinase, total PI3 kinase, Akt kinase, GSK-3 kinase, and PTEN phosphatase activities were measured in corresponding immunoprecipitates from cells stimulated with insulin for 0, 5, 15, or 30 min.
Source of Reagents-Monoclonal antibodies to phospho-tyrosine (PY20) and phospho-serine were purchased from Transduction Laboratories (Lexington, KY). Polyclonal antibodies to p85, IR␤, IRS-1, and PTEN phosphatase were purchased from Upstate Biotechnology, Inc. (Lake Placid, NY). Phospho-specific antibodies to Akt, GSK-3␤, and BAD were obtained from Cell Signaling (Beverly, MA). Protein A-Sepharose was purchased from Amersham Biosciences. Monoclonal antibodies to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and GSK-3␤ were purchased from Chemicon Corp. All other reagents were purchased from CalBiochem or Sigma-Aldrich.
Statistical Analysis-Data depicted in the graphs represent the mean Ϯ S.D. of results. Inter-group comparisons were made with the Student's t test or analysis of variance and the Fisher least significance post hoc significance test using Number Cruncher Statistical Systems (Dr. Jerry L. Hintze, Kaysville, UT).

Ethanol-induced Cerebellar
Hypoplasia-Histological studies of the cerebellar tissue revealed well developed folia and distinct laminar architecture in control pups (Fig. 1A, C, and E) and hypoplasia with marked simplification of the folia and poor lamination of the cortex in ethanol-exposed pups (Fig. 1, B, D, and F). Adjacent sections stained with H33258 revealed uniform nuclear morphology and only occasional apoptotic bodies in control cerebella (1G) and reduced cell densities with increased nuclear condensation, pyknosis, and fragmentation in ethanol-exposed cerebella (Fig. 1H). Correspondingly, TUNEL assays revealed conspicuously increased nuclear labeling consistent with genomic DNA damage (Fig. 2, A and B), and immunohistochemical staining detected increased levels of activated caspase-3 (Fig. 2, C and D) in cerebella from ethanolexposed pups. In the ethanol-exposed brains, the activated caspase-3 was mainly localized within nuclei. Previous studies showed that nuclear translocation is required for caspase induction of apoptosis (46).
Reduced PY of IR␤ in Cerebella of Ethanol-exposed Rat Pups-Previous in vitro studies linked ethanol-induced neuronal loss to inhibition of insulin signaling through its receptor (17,47). To characterize potential mechanisms by which chronic gestational exposure to ethanol inhibits insulin signaling in vivo, the levels of IR␤ and PY-IR␤ were examined in cerebellar tissue homogenates by direct Western blot analysis and Western blot analysis of immunoprecipitates, respectively. Digital image quantification of results obtained with 12 brains in each group demonstrated significantly reduced levels of PY-IR␤ in ethanol-exposed relative to control samples (p Ͻ 0.01; see Fig. 3, A and C) but similar levels of total IR␤ protein in control and ethanol-exposed cerebella (Fig. 3, B and D). . Cerebella from P2 ethanol-exposed pups were hypoplastic (B, arrow) with simplified folia (D and F, arrows), and incomplete cortical lamination (F). Adjacent sections stained with Hoechst H33258 revealed uniform nuclear morphology and rare condensed nuclei (G) in control cerebella and conspicuous nuclear condensation and pyknosis, consistent with apoptosis (H, arrows) in cerebella from ethanol-exposed pups.
FIG. 2. Ethanol-induced neuronal apoptosis. Histological sections were either labeled using the TUNEL assays to detect nicked or fragmented DNA (A and B) or immunostained to detect activated caspase-3 (C and D). Reaction products were detected with biotinylated secondary antibody, alkaline phosphatase-conjugated Streptavidin, and BCIP/NBT substrate (blue precipitate). Cerebella from control pups (A and C) had only scattered TUNELϩ or activated caspase-3 immunoreactive cells in the cortex. In contrast, cerebella from ethanol-exposed pups had conspicuous TUNELϩ (B) and activated caspase-3 immunoreactive (D) nuclei. Insets in Panels A and B show higher magnification images of nuclei in sections labeled using the TUNEL assay.

Ethanol Inhibition of Insulin-responsive Gene Expression-
To examine the downstream effects of ethanol on insulin signaling, we measured cerebellar tissue expression of GAPDH, which is an important insulin-responsive gene product (48). Western blot analysis using denaturing and reducing conditions detected a single ϳ37-kDa protein corresponding to monomeric GAPDH (Fig. 4A). Digital image quantification of results obtained with 12 brains in each group demonstrated significantly reduced GAPDH expression in ethanol-exposed relative to control cereballa (p Ͻ 0.01; see Fig. 4B).
Reduced Levels of PY-IRS-1 in Cerebella of Ethanol-exposed Pups-Major effects of insulin are mediated through IRS-1, which transmits signals downstream to regulate growth, survival, and energy metabolism. To determine whether gestational exposure to ethanol inhibits tyrosyl phosphorylation of IRS-1 in cerebellar tissue, PY-IRS-1 and IRS-1 protein levels were examined by immunoprecipitation/Western blot analysis (n ϭ 12 per group). The studies demonstrated significantly reduced mean levels of PY-IRS-1 following gestational exposure to ethanol (p Ͻ 0.001; see Fig. 5, A and C) but similar mean levels of IRS-1 protein in ethanol-exposed and control cerebella (Fig. 5, B and D).
Reduced PI3 Kinase Signaling in Cerebella from Ethanolexposed Pups-Neuronal survival is mediated through PI3 kinase, which can be activated through both IRS-1-dependent and IRS-1-independent mechanisms. IRS-1-associated PI3 kinase activity was assessed by measuring p85 interactions with PY-IRS-1 by Western blot analysis of IRS-1 immunoprecipitates, and total PI3 kinase activity was measured in p85 immunoprecipitates. Chronic gestational exposure to ethanol resulted in significantly reduced levels of both IRS-1-associated p85 (p Ͻ 0.001; see Fig. 6, A and C) and total PI3 kinase (p Ͻ 0.001; see Fig. 6, E and F) activities. In contrast, p85 protein levels were similar in control and ethanol-exposed cerebella (Fig. 6, B and D).
PTEN Phosphatase Expression-PTEN phosphatase is an important negative regulator of PI3 kinase signaling downstream through Akt (27,38). Therefore, high levels of PTEN activity would be expected to inhibit survival signaling downstream of PI3 kinase. However, the effect of ethanol on PTEN expression in the brain has not been investigated. Western blot analysis detected the expected ϳ45-kDa PTEN protein in all samples (Fig. 7A). Digital image analysis of the Western blot signals demonstrated significantly higher mean levels of PTEN in ethanol-exposed relative to control cerebella (p Ͻ 0.001, n ϭ 12 per group; see Fig. 7B). Correspondingly, the mean level of PTEN phosphatase activity was significantly higher in cerebellar tissue from ethanol exposed relative to control pups (p Ͻ 0.001) (Fig. 7C).
Real-time quantitative RT-PCR was used to determine whether PTEN mRNA levels were increased in the ethanolexposed relative to control brains. PTEN and 18 S RNA transcripts were quantified in parallel reactions using RT products generated from 20 ng of total RNA. Serial dilutions of plasmid DNA (0.002 to 20 ng) containing PTEN or 18 S cDNA target sequences were used as standards in all experiments. Graphs relating the threshold cycle values and input target DNA amounts (ng) resulted in standard curves that were linear over FIG. 5. Reduced levels of PY-IRS-1 in cerebella from ethanolexposed relative to control rat pups. A, IRS-1 protein was immunoprecipitated from cerebellar homogenates using polyclonal anti-IRS-1. Immunoprecipitates were subjected to Western blot analysis using monoclonal anti-PY. B, direct Western blot analysis of cerebellar tissue was used to detect IRS-1 protein. Each corresponding lane in Panels A and B corresponds to an individual brain. Immunoreactivity corresponding to PY-IRS-1 (C) or IRS-1 (D) was quantified with digital imaging, and results obtained from 12 pups/group are depicted graphically (mean number of pixels Ϯ S.D.; *, p Ͻ 0.001).
the 5 orders of magnitude tested and had correlation coefficients (r 2 ) of 0.99 for 18 S and 0.97 for PTEN. Equations generated from the standard curves were used to calculate 18 S and PTEN mRNA abundance in the samples. No template and total RNA controls were included in all experiments. In addition, studies showed that rat genomic DNA could not be amplified using the conditions employed for real-time RT-PCR. All reactions were performed in triplicate, and all assays were repeated at least three times.
The mean threshold cycle value for PTEN was 25.4 Ϯ 1.43 for the control group and 26.2 Ϯ 1.2 in the ethanol-exposed group. The mean threshold cycle value for 18 S was 12.8 Ϯ 3.2 in the control group and 12.1 Ϯ 1.5 in the ethanol-exposed group. The PTEN mRNA/18 S RNA ratios were calculated for each sample to normalize for small difference in RNA content in the initial reactions. Inter-group comparisons made using Student's t test analysis demonstrated that the mean percentages of PTEN/18 S for control (1.35 Ϯ 0.06) and ethanol-exposed (1.55 Ϯ 0.1) brains were not statistically significant (p ϭ 0.183; see Fig. 7D).
To more directly evaluate the roles of ethanol and impaired insulin/IGF-1 signaling in relation to PTEN expression and function, experiments were conducted using primary cerebellar neuron cultures. Insulin was found to be more effective than IGF-1 in stimulating Ser phosphorylation of PTEN (Fig. 8, A  and B) and suppressing PTEN phosphatase activity (data not shown). In contrast, PTEN protein expression was not modulated by short term insulin or IGF-1 stimulation (Fig. 8A). In the in vitro ethanol-treated cultures, insulin-stimulated levels of phospho-PTEN were reduced by 70 -80% (p Ͻ 0.001), and IGF-1-stimulated levels of phospho-PTEN were reduced by 15-20% (p Ͻ 0.01) relative to control (Fig. 8, C and D). Correspondingly, in vitro ethanol treatment was associated with reduced insulin and IGF-1 suppression of PTEN phosphatase activity, with greater impairment of insulin compared with IGF-1-mediated responses (Fig. 8, E and F).
Ethanol Inhibits Survival Signaling Downstream of PI3 Kinase-PI3 kinase mediates survival by phosphorylating Akt and GSK-3␤, which results in activation of Akt kinase (28,49) and inhibition of GSK-3␤ activity (50). Akt kinase also phosphorylates and inactivates GSK-3␤, as well as BAD (29,31,32), which are both pro-apoptotic. To determine whether the hypoplasia and increased apoptosis observed in cerebella from ethanol-exposed pups were associated with impaired signaling downstream of PI3 kinase, we examined Akt, phospho (p)-Akt, GSK, pGSK, BAD, and pBAD protein levels by Western blot analysis and measured Akt and GSK-3␤ kinase activities in corresponding immunoprecipitates. Chronic gestational exposure to ethanol resulted in significantly reduced levels of pAkt, pGSK-3␤, pBAD, and Akt kinase activity and increased levels of total BAD protein and GSK-3␤ kinase activity (p Ͻ 0.05 or p Ͻ 0.001; see Fig. 9). In contrast, the levels of Akt and GSK-3␤ proteins were similar in control and ethanol-exposed cerebella.
FIG. 6. Reduced levels of IRS-1-associated and total PI3 kinase activities in cerebellar tissue from ethanol-exposed rat pups. A, IRS-1 protein was immunoprecipitated from cerebellar homogenates and analyzed by Western blotting with polyclonal anti-p85. B, p85 protein was detected by direct Western blot analysis. Each lane in Panels A and B corresponds to an individual brain. Western blot signals corresponding to p85-associated IRS-1 (C) and p85 (D) protein were quantified using the Kodak Digital Image Station, and the results from 12 pups/group are depicted graphically (mean number of pixels Ϯ S.D.; *, p Ͻ 0.001). E, PI3 kinase activity was measured in p85 immunoprecipitates. Radiolabeled phospholipids were separated by thin-layer chromatography and detected by autoradiography. Each lane corresponds to an individual brain. F, the autoradiographic signals corresponding to PI3 kinase activity were quantified by densitometry using the Kodak Digital Image Station, and the results from 12 pups/group are depicted graphically (mean number of pixels Ϯ S.D.; *, p Ͻ 0.001) .   FIG. 7. Ethanol modulation of PTEN. A, PTEN protein (ϳ45 kDa) expression was examined in cerebellar tissue by Western blot analysis. Each lane corresponds to an individual sample. B, immunoreactivity was quantified using the Kodak Digital Science Image Station, and the results from 12 pups/group are depicted graphically (mean number of pixels Ϯ S.D.; *, p Ͻ 0.001). C, phosphatase activity was measured in PTEN immunoprecipitates from cerebellar tissue. Phosphate release was measured using the BioMol malachite green-based assay. A standard curve generated with known amounts of phosphatase was used to calculate phosphate release (reflecting phosphatase activity) in the unknown samples. The graphed values depict the mean Ϯ S.D. of results for each group (*, p Ͻ 0.001). D, real-time quantitative RT-PCR was used to measure PTEN mRNA and 18 S RNA expression in cerebellar tissue. Standard curves generated with known copy numbers of recombinant plasmid DNA were used to quantify transcripts. The PTEN mRNA/18 S RNA percentages (based on ng of cDNA) were calculated for each sample, and results obtained from 10 samples per group are depicted graphically (mean Ϯ S.D.; p ϭ 0.183).
Chronic Gestational Exposure to Ethanol Impairs Insulinstimulated Survival Signaling in the Brain-In vitro studies were used to help validate the in vivo observations and demonstrate that chronic gestational exposure to ethanol inhibits insulin-stimulated survival mechanisms in CNS neurons. Cultures generated from ethanol-exposed pups had significantly reduced levels of insulin-stimulated neuronal viability (p Ͻ 0.005), IRS-1-associated PI3 kinase, total PI3 kinase, and Akt kinase (all p Ͻ 0.001) activities and increased levels of GSK-3 activity (p Ͻ 0.001) relative to neuronal cultures generated from control pups (Fig. 10, A-C, E, and F). In addition, insulin significantly suppressed PTEN phosphatase activity in control cultures (p Ͻ 0.01), whereas in cultures generated from ethanol-exposed pups, the basal levels of PTEN were significantly higher than control (p Ͻ 0.001), and insulin failed to suppress the phosphatase activity (Fig. 10D). DISCUSSION Ethanol impairs neuronal growth and viability in the developing CNS by inhibiting growth factor-stimulated signaling mechanisms (9,11,13,18,51). Previous in vitro studies showed that ethanol inhibition of insulin-stimulated cell growth and viability were mediated by reduced activation of mitogen-activated protein kinase (42,43) and PI3 kinase (11,47). Although similar inhibitory effects of ethanol have been demonstrated with respect to IGF-1 (9,18), our studies suggested that ethanol may have a more potent inhibitory effect on insulin compared with IGF-1-stimulated signaling in neuronal cells (17). Therefore, in the present studies, we utilized an in vivo model to evaluate the effects of chronic gestational exposure to ethanol on the integrity of insulin signaling through pathways that mediate neuronal survival in the CNS. The gestational exposure to ethanol produced major structural abnormalities in the CNS including cerebellar hypoplasia with associated incomplete foliation and poor lamination of the cortex and increased apoptosis with activation of caspase-3. These effects resemble the abnormalities described in human cases and other experimental models of fetal alcohol syndrome (4,8).
The finding that chronic gestational exposure to ethanol resulted in reduced levels of PY-IR␤ yet had no significant effect on IR␤ protein expression corresponds with previous in vitro results (47,52) and indicates that the inhibitory effects of ethanol on insulin signaling in the developing brain begin at the level of IR␤ function rather than IR␤ expression. These results are consistent with findings in a previous report (53) demonstrating that chronic gestational exposure to ethanol impaired IGF-1 receptor function but not the receptor expression. Potential consequences of impaired insulin signaling include inhibition of insulin-responsive gene expression and cellular functions such as proliferation, neurite outgrowth, glucose utilization, phospholipid metabolism, and amino acid transport. In this regard, we demonstrated that ethanol-exposed cerebella had significantly reduced levels of an important insulin-responsive gene, GAPDH (48), suggesting that CNS glucose utilization and energy metabolism may have been impaired. In addition to regulating glucose utilization, insulin stimulates glucose uptake. Glucose uptake is mediated by glucose transporter molecules termed GLUTs (54), and previous studies (55)(56)(57) demonstrated that chronic gestational exposure to ethanol inhibits expression and function of glucose transporter molecules GLUT1 and GLUT3 in the brain. Together, these observations potentially link ethanol-impaired insulin signaling and neuronal loss to deficiencies in uptake and utilization of glucose in the developing brain.
Our finding of reduced PY-IRS-1 levels in cerebellar tissue from ethanol-exposed pups indicates that chronic gestational exposure to ethanol impairs IRS-1 signaling, which could lead to reduced activation of growth and survival mechanisms. In this regard, we also detected reduced binding of the p85 subunit of PI3 kinase to IRS-1 in cerebellar tissue and reduced levels of insulin-stimulated IRS-1-associated PI3 kinase activity in cultured cerebellar neurons from ethanol-exposed pups. These results indicate that IRS-1-mediated survival mechanisms, including those stimulated by insulin, were significantly impaired by ethanol, consistent with previous results obtained using in vitro ethanol exposure models (9,17,42,47,58,59). However, further studies demonstrated that ethanol caused even greater reductions in total PI3 kinase activity, both in vivo and following insulin stimulation in vitro. Therefore, PI3 kinase was likely inhibited through both IRS-dependent and IRS-independent, e.g. PTEN, pathways. Indeed, impaired signaling through IRS-2 is another candidate for mediating the IRS-dependent effects, because insulin-stimulated tyrosyl phosphorylation of IRS-2 can activate PI3 kinase (60), IRS-2 is expressed in the brain and functionally active in CNS neurons (61), and ethanol inhibits growth factor signaling through FIG. 8. PTEN modulation by insulin and IGF-1 in cultured primary cerebellar granule neurons. Cultures were serum-starved for 12 h and then stimulated with only 50 nM insulin or 25 ng/ml IGF-1 for 0 -60 min. A, p-PTEN detected by PTEN Western blot analysis of anti-phospho-Serine immunoprecipitates and total PTEN protein was detected by direct Western blot analysis. B, the autoradiographic signals from studies shown in Panel A were quantified using the Kodak Digital Science Image Station, and the calculated mean ratios of p-PTEN/PTEN observed in insulin-or IGF-1-stimulated cultures were depicted graphically. Inter-group comparisons were made using analysis of variance (*, p Ͻ 0.001). C, p-PTEN and total PTEN levels were examined in control and ethanol-treated cells that were stimulated with insulin or IGF-1 for 5 or 15 min as described. Prior to the stimulation studies, the cultures were exposed to 50 mM ethanol or nothing for 2 days (see "Experimental Procedures"). Ethanol exposure was maintained throughout the study. D, mean levels (number of pixels Ϯ S.D.) of p-PTEN measured in four replicate cultures by densitometric analysis of the Western blot signals (*, p Ͻ 0.001; †, p Ͻ 0.01). E and F, phosphatase activity (nmol of phosphate released) was measured in PTEN immunoprecipitates from control and ethanol-treated, insulin-(E) or IGF-1-stimulated (F) cultures using the Biomol Green reagent (see legend for Fig. 7C and "Experimental Procedures") (*, p Ͻ 0.001; †, p Ͻ 0.01; ࡗ, p Ͻ 0.05).
IRS-2 (62). The results from both the in vivo and in vitro experiments showed that ethanol inhibited PI3 kinase activity and suggest that inhibition of insulin signaling through IRS-dependent and IRS-independent pathways contributed to the observed impairments in survival signaling and development of cerebellar hypoplasia. Primary cerebellar neuron cultures were generated from control and ethanol-exposed postnatal day 1 pups. To mimic the in vivo model, the cells were not further exposed to ethanol. After 12 h of serum deprivation, the cultures were stimulated with 50 nM insulin for either 24 h to measure viability or 0 -30 min to assess insulin-stimulated signaling. A, viability was measured using the crystal violet (CV) assay. B, IRS-1-associated and total PI3 kinase activities were measured in IRS-1 (IRS-1/PI3K) and p85 (p85/PI3K) immunoprecipitates, respectively. Radiolabeled phospholipids were separated by thinlayer chromatography. C, PI3 kinase autoradiographic signals were quantified by densitometry. The graph depicts the mean number of pixels (ϮS.D.) detected in each group. D, PTEN phosphatase activity was measured in PTEN immunoprecipitates as described in the legend to Fig. 7C. The † above the 15-min control bar indicates significant difference from control values at other time points. Akt kinase (E) and GSK-3␤ (F) activities were measured in immunoprecipitates using cAMP-response element-binding protein and Crosstide as substrates for phosphorylation. Incorporation of [␥-32 P]ATP was quantified in a TopCount machine, and results from four samples per group are depicted graphically (mean cpm Ϯ S.D.) (*, p Ͻ 0.001; **, p Ͻ 0.005; †, p Ͻ 0.01; ࡗ, p Ͻ 0.05). PI3 kinase promotes survival by phosphorylating Akt and activating its kinase (27,28), which then phosphorylates Ser and Thr residues on target pro-apoptosis molecules, including GSK-3 (50), BAD (31), and caspase-9 (63,64), and rendering them inactive. Our experiments demonstrated that ethanol impaired survival signaling downstream of PI3 kinase, as was manifested by the reduced levels of phospho-Akt, phospho-BAD, phospho-GSK-3␤, and Akt kinase activity, and increased levels of activated BAD and GSK-3␤ activity. Further in vitro studies showed that insulin activation of Akt kinase and inhibition of GSK-3␤ were impaired by ethanol. Therefore, these studies linked ethanol impairment of insulin-stimulated survival signaling downstream of PI3 kinase to the increased cell loss, apoptosis, and cerebellar hypoplasia observed in vivo.
Further studies revealed that PTEN phosphatase activation represents an additional non-IRS-1 mechanism by which survival signaling downstream of PI3 kinase could be inhibited by ethanol in CNS neurons. PTEN dephosphorylates and reverses the activation of PI3 kinase (38,65), whereas inactivation of PTEN promotes membrane recruitment of Akt, leading to increased Akt phosphorylation and kinase activity (38,65). PTEN expression and phosphatase activity are negatively regulated by phosphorylation of PTEN protein (66); however, the effects of ethanol exposure and growth factor stimulation on PTEN expression and function were previously unknown. Our in vivo studies demonstrated significantly higher levels of PTEN protein and phosphatase activity, without correspondingly increased PTEN mRNA expression in ethanol-exposed cerebella, indicating post-transcriptional modulation of PTEN by ethanol. In addition, the magnitude of increased PTEN expression was similar to that of GSK-3␤ activity, corresponding with the expected inhibitory effects of PTEN on survival signaling downstream of PI3 kinase.
In vitro studies were used to characterize the effects of ethanol on the levels of PTEN protein, phosphorylation, and phosphatase activity and to determine whether the adverse effects of ethanol on PTEN were linked to impaired insulin signaling. Insulin, and to a lesser extent IGF-1, stimulated PTEN phosphorylation and correspondingly inhibited PTEN phosphatase activity in control cerebellar neuron cultures. Chronic gestational exposure to ethanol impaired the insulin-stimulated phosphorylation of PTEN and suppression of PTEN phosphatase activity. Therefore, the adverse effect of ethanol on insulin signaling to inhibit the function of PTEN in CNS neurons probably represents an important mechanism of impaired survival and increased apoptosis.
The results from our in vivo and in vitro experiments linked chronic gestational exposure to ethanol to aberrantly increased expression and enzymatic activity of PTEN, which has a pivotal role in regulating PI3 kinase-activated survival signaling. Although PTEN has been shown to inhibit insulin-stimulated Akt phosphorylation without affecting IR␤ activation of IRS-1 (67), other more recent data (68) suggest that PTEN can also inhibit insulin-stimulated IRS-1 phosphorylation and the attendant downstream signaling. Therefore, in the ethanol-exposed cerebella, the impaired survival signaling may have been mediated by the combined effects of the following: 1) ethanol inhibition of IR␤ phosphorylation and function, 2) PTEN inhibition of IRS-1 phosphorylation, and 3) PTEN inhibition of Akt kinase. These studies suggest important mechanisms by which ethanol-impaired insulin signaling could promote the development of cerebellar hypoplasia in fetal alcohol syndrome.