Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Victims of fetal alcohol syndrome (FAS) have severe learning, emotional, motor, and other impairments that adversely influence behavior (Conry, 1990). The cerebellum is emerging as being more crucial for cognition than previously thought. With the use of new technology, including the trans-neuronal tracing of herpes simplex virus type 1, information is now available indicating that there exists an intricate cerebro-cerebellar system with both feed-forward and feedback connections (Middleton and Strick, 1997; Schmahmann and Pandya, 1997). These findings point to the importance of studying the cerebellum as a possible link to the understanding of FAS.

The hippocampus and cerebellum are vulnerable to ethanol intoxication-induced redox changes (Renis et al., 1996). Changes in the cerebellum brought about by free radicals are particularly interesting since many neurodegenerative conditions can be induced by this mechanism. DNA strand breaks in the cerebellum and hippocampus after chronic (but not acute) ethanol administration correlate with significant increases in lipid peroxidation (Renis et al., 1996). Among the various possible sources of free radicals, cytochrome P450 IIEI (CYP2E1), NADPH oxidase, and NADPH cytochrome P450 reductase are particularly important due to their inducibility by ethanol. CYP2E1 is widely expressed in the rat brain including the cerebellum, where P450 IIE immunoreactivity is present in glial cells and their processes (Hansson et al., 1990). It is thus conceivable that ethanol induces free radical formation via its induction of CYP2E1. One possible effect of free radicals generated by ethanol is at the level of transcriptional regulation. The DNA binding of two transcription factors, AP-1 and NF-B, which are regulated by cellular oxidation/reduction events (Dalton et al., 1999) in certain cell lines, were thus examined.

AP-1 consists either of a dimer of members of the Jun and Fos family proteins, or of two Jun units. These protein units function cooperatively as inducible transcription factors. The binding of the Fos-Jun heterodimer to the DNA element known to be the AP-1-binding site is redox-regulated (Abate et al., 1990; Dalton et al., 1999). The integrity of key thiol (-SH) groups within certain amino acids is essential for this binding. Specifically, oxidation or substitution of critical cysteine residues in the leucine zipper regions of these proteins decreases their binding to DNA, whereas reduction increases binding (Abate et al., 1990). This means that enhanced free radical activity could have a direct bearing on the function of this transcription factor.

NF-B is a heterotrimer (Baeuerle, 1991) made up of an inhibitory unit called IB, a p50 protein, and a p65 (Rel A) protein. It is the latter unit that is responsible for the induction of transcription (Wong et al., 1997). For NF-B activation, the inhibitory IB is released after it has been serine-phosphorylated on residues 32 and 36, then ubiquitinated and degraded by proteasome. This exposes nuclear translocation sequences. Activation can occur following oxidation events (Schreck et al., 1992) in the cytosol, leading to the degradation of IB. However, like AP-1, the DNA binding of NF-B is regulated by the redox state of a cysteine residue (cys-62 in the p50 subunit), requiring a reducing environment in the nucleus for binding.

Because ethanol is capable of generating free radicals, and both AP-1 and NF-B are sensitive to redox regulation, these studies sought to establish the effect of gestational ethanol on their activation (DNA binding) in developing offspring. They also sought to characterize related molecular events arising out of the chronic in utero ethanol exposure, as well as explore the means by which the observed changes occur.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. Consensus oligonucleotides for NF-B (5'-AGT TGA GGG GAC TTT CCC AGG-3') and AP-1 (5'-CGC TTG ATG AGT CAG CCG GAA-3') were obtained from Promega (Madison, WI). Anti-phospho-GSK-3 antibodies were obtained from Upstate Biotechnology (Lake Placid, NY). Phosphorylated c-jun antibody was from New England Biolabs (Beverly, MA). Other antibodies for Western blot analyses were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA).

Prenatal Ethanol Exposure Protocol. The Animal Resources Center at The University of Texas at Austin maintained, in a 12:12-h light/dark cycle, breeding colonies of 200- to 300-g male and female Sprague-Dawley rats. To minimize the pain to the animals, euthanasia was performed in accordance with recommendations of the Panel on Euthanasia of the American Veterinary Medical Association. On the days of successful mating, pregnant rats were randomly assigned to one of three experimental groups: an ethanol-exposed (E), a pair-fed (PF) control, or an ad libitum (AL) control group.

Cerebella. On specified postnatal days, cerebella of rat pups were removed for examination. Prenatally, on the 20th gestational day, pregnant dams were exposed to carbon dioxide for 90 s and euthanized by cervical dislocation. Their pups were subsequently surgically removed and decapitated for cerebellum removal.

Dietary Lithium. Separate groups of pregnant rats had dietary lithium chloride (30 mg/100 ml of liquid diet for an average daily dosage of 2 mEq/kg) during gestation. Other treatment groups had a combination of dietary lithium and 30% ethanol-derived calories. Cerebella were removed and examined as described above.

Nuclear Protein Extraction and Electrophoretic Mobility Shift Assay (EMSA). Assays were performed according to the procedure of Denison et al. (1988). The isolated cerebellar tissue was homogenized in HEGD buffer (12 µl of 25 mM HEPES, pH 7.5, 1 mM EDTA, 1 mM dithiothreitol, 10% glycerol, 0.75 µl of freshly made 1 M spermidine and 0.3 µl of freshly made 0.5 M spermine per ml, 0.1 mg/ml phenylmethylsulfonyl fluoride solubilized in dimethyl sulfoxide) in a tissue grinder (held on ice) with about 30 pestle strokes. The homogenate was centrifuged at 12,000g for 10 min at 4°C. The pellet was isolated, resuspended in HEGDK (0.5 M KCl in HEGD buffer), and held on ice for 1 h. The sample(s) were then spun at 16,000g for 10 min at 4°C. The supernatant was snap-frozen in liquid nitrogen until ready for use. A portion of this was assayed for protein content [Bio-Rad (Hercules, CA) DC protein assay].

Western Blot Analyses. Whole cerebellar tissue was homogenized in ice-cold buffer (50 mM Tris HCl, pH 7.4), 1% (v/v) Igepal CA-630, 1 µl/ml 50 mM sodium molybdate, 2.5 mM sodium pyrophosphate, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 µg/µl aprotinin, 1 µg/µl leupeptin, 1 µg/µl pepstatin-A, 2 mM NaF, 1 mM phenylmethylsulfonyl fluoride, 10 nM okadaic acid, p-bromotetramisole, 50 µM canthardin, 1 µM microcystin, and 2 mM sodium orthovanadate). Proteins were separated on a denaturing (sodium dodecyl sulfate) polyacrylamide gel (12% SDS-polyacrylamide gel electrophoresis) and transferred to a nitrocellulose membrane. The membrane was exposed to the primary antibody (1:1000) in TTBS for 1 h with shaking, and again subjected to the wash process. Each membrane was exposed to the horseradish peroxidase-labeled secondary antibody (1:3000) for 20 min to 1 h in TTBS. The membranes were subsequently washed and then exposed to enhanced chemiluminescence (Amersham Biosciences, Piscataway, NJ) detection reagents for 1 min, drained, and then exposed to film for up to 5 min.

Glutathione Assay. Total GSH and glutathione disulfide (GSSG) were determined using a modified method of Neuschwander-Tetri and Roll (1989). Isolated cerebella were homogenized in NaH₂PO₄, pH 6.0. Homogenate (250 µl) was added to 83 µl of 25 mM NaH₂PO₄ (pH 7.0) for total GSH determination, or 83 µl of N-ethylmaleimide for GSSG determination. In each case, 200 µl of the sample was then mixed with 200 µl of 25 mM dithiothreitol (in 25 mM NaH₂PO₄, pH 7.0) and 100 µl of Tris buffer (pH 8.5), and incubated on ice for 30 min. After addition of 0.5 ml of 2.5% (w/v) sulfosalicylic acid, the samples were centrifuged for 10 min at 5200g at 4°C. An aliquot (200 µl) of the supernatant was mixed with 200 µl of o-phthalaldehyde (5 mg/ml in 0.4 M potassium borate solution, pH 9.9) and incubated for 2 min at 25°C. Each sample was neutralized with 200 µl of 250 mM NaH₂PO₄, pH 7.0. The samples were then either kept on ice in the dark and analyzed immediately, or stored at 80°C overnight before analysis. Analysis involved separation of the glutathione-o-phthalaldehyde adduct using high performance liquid chromatography, followed by fluorometric detection and quantitation by integration of areas under peaks.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

The results of the EMSAs indicate that at birth there is a decrease in the DNA binding of both NF-B and AP-1 in the in utero ethanol-exposed (E) group relative to both pair-fed and ad libitum controls (Fig. 1). Counts per minute, determined by a microarray detector (Packard Instrument Co., Inc., Downers Grove, IL), showed that in the E neonates, AP-1 DNA binding was decreased to 21% of ad libitum control levels (n = 3, p < 0.01); NF-B DNA binding was decreased to 58% of ad libitum control values (n = 4, p < 0.01). To confirm the identity of the bands, before incubation with labeled oligonucleotide at room temperature, 1 µg of nonspecific IgG antibody was incubated with nuclear extract for 30 min on ice for each lane. In addition, 1 µg of various antibodies to the transcription factor subunits was added during this time. Antibodies for p50 and p65 each reduced NF-B DNA binding, as did antibodies for c-Jun and c-Fos for AP-1 DNA binding (Fig. 1). By binding to their respective proteins, these antibodies reduced the amount of the transcription factor available for binding to the DNA in a way that nonspecific IgG could not.

View larger version (70K): [in this window] [in a new window]   Fig. 1.   Rat cerebellar NF-B (left panel) and AP-1 (right panel) DNA binding. Representative EMSAs for NF-B (four experiments) and AP-1 (three experiments) showing decreased DNA binding in the in utero ethanol-exposed rat cerebella on the first postnatal day (PD1). Each sample consisted of pooled cerebella of PD1 rats. The assays were performed as described within the text. However, before incubation with labeled oligonucleotide, 1 µg of nonspecific IgG antibody was incubated with nuclear extract for 30 min on ice for each lane. In the specified lanes, 1 µg of the specified antibodies was also added during this time, to help determine specific DNA-bound bands. The E, PF, and AL control groups have been described under Experimental Procedures. In the ethanol-exposed groups, NF-B DNA binding was decreased to 58% of ad libitum control values (n = 4, p < 0.01); AP-1 DNA binding was decreased to 21% of ad libitum control levels (n = 3, p < 0.01). The PF control results were intermediate between those of the E and AL groups.

Because both AP-1 and NF-B are redox-regulated, the effect of ethanol on cerebellar GSH levels and catalase activity was investigated. Figure 2 illustrates that GSH and GSSG were not significantly altered at birth relative to controls in this FAS model. Catalase activity also did not differ significantly between the groups (data not shown).

View larger version (41K): [in this window] [in a new window]   Fig. 2.   Glutathione levels on the first postnatal day under the FAS model. As described under Experimental Procedures, alcohol treatment was withdrawn 2 days before birth, and the levels of cerebellar glutathione were measured on PD1 in the E, PF, and AL control groups. Neither GSH nor GSSG levels in the in utero ethanol-exposed rats were significantly different between the E and control groups (n = 3).

To determine whether the diminished transcription factor activation observed on PD1 was due to rebound effects, prenatal cerebella were examined before the withdrawal of the treatments described above. EMSAs performed on embryonic day 20 (EM20) cerebella for AP-1 and NF-B DNA binding in the in utero ethanol-exposed rat cerebella showed average decreases of 68% and 50%, respectively, relative to AL controls (Fig. 3). The trends were similar to those observed on PD1.

View larger version (36K): [in this window] [in a new window]   Fig. 3.   Rat cerebellar AP-1 and NF-B DNA binding on EM20. EMSAs for AP-1 and NF-B show decreased DNA binding in the in utero ethanol-exposed rat cerebella. The E, PF, and AL control groups have been described under Experimental Procedures. Each sample consisted of pooled cerebella of EM20 rats. Ethanol and ethanol II, and their corresponding pair-fed controls represent pools of cerebella obtained from different individual experiments (adult pregnant rats). The trends were similar to those observed on PD1. The PF control results were intermediate between those of the E and AL groups.

To assess whether the decreased DNA binding was the result of reductions in the quantity of transcription factors present in the cerebella at birth, Western blot analyses of the various subunits of AP-1 were conducted. As Fig. 4A shows, cerebellar c-Fos and c-Jun protein levels were not different between the groups. Similarly, the quantities of the PD1 NF-B subunits (cytosolic IB, nuclear p50, and nuclear p65) were not significantly changed (Fig. 4B). Phosphorylated c-JUN levels were, however, diminished in the E group relative to controls (Fig. 4A).

View larger version (12K): [in this window] [in a new window]   Fig. 4.   A, AP-1 protein amounts on the first postnatal day. In utero ethanol exposure does not decrease the levels of c-Fos and c-Jun proteins in neonatal cerebella. Phosphorylated c-jun (Ser 73) levels were, however, decreased in accordance with the decrease in AP-1 DNA binding (Fig. 1). B, NF-B protein amounts on the first postnatal day. In utero ethanol exposure does not decrease the levels of NF-B p50, p65, and IB- subunit proteins in neonatal cerebella.

The level of the active form of GSK-3, tyrosine-phosphorylated GSK-3, which regulates the activities of key transcription factors during development, was studied on PD1. Total protein extracts from rat pups were examined by Western blot for GSK-3 phosphorylation on EM20 and PD1. There was decreased tyrosine-phosphorylated GSK-3 (Y279/Y216) and slightly elevated or unchanged phosphorylated stress-activated protein kinase-1 (SAPK-1) on EM20 (Fig. 5). In individual rats from EM20 with the most severely depressed NF-B and AP-1 DNA binding, there was more depression of the active form of GSK-3 and more SAPK-1 activity increases. On the contrary, PD1 Western blots (Fig. 6) showed increased amounts of phosphorylated GSK-3 (Y216 and Y279) in the E group relative to controls. There were no corresponding increases in the relative amounts of phosphorylated GSK-3 (Ser 21), the inactive form.

View larger version (38K): [in this window] [in a new window]   Fig. 5.   Active forms of GSK-3 and SAPK-1 in the cerebellum on EM20. Western blots showing decreased tyrosine-phosphorylated GSK-3 and slightly elevated phosphorylated SAPK-1 on EM20. The blots were stripped and reprobed with -actin to control for loading.

View larger version (53K): [in this window] [in a new window]   Fig. 6.   Representative Western blots (three experiments) showing phosphorylated GSK-3 (Y216 and Y279) in the E group relative to controls. There were no corresponding increases in the relative amount of phosphorylated GSK-3 (Ser 21).

To determine the importance of GSK-3 activity for DNA binding, separate groups of pregnant Sprague-Dawley rats had lithium chloride added to their daily gestational diet to inhibit the activity of GSK-3 (Salinas and Hall, 1999; Ryves and Harwood, 2001). As Fig. 7 shows, AP-1 DNA binding was negligible in lithium-exposed pups. Similarly, NF-B DNA binding in E-alone pups was 5-fold above the level when lithium was present in the maternal diet. The effect was reversed when both ethanol and lithium were present in the maternal diet.

View larger version (49K): [in this window] [in a new window]   Fig. 7.   Effect of inhibiting GSK-3 activity with lithium on AP-1 and NF-B DNA binding on PD1. Groups of pregnant Sprague-Dawley rats were fed as described under Experimental Procedures. EMSAs showed diminished AP-1 and NF-B DNA binding in pup cerebella when lithium was present in maternal diet. Ethanol, ethanol II, lithium, lithium II, lithium + ethanol, and lithium + ethanol II each represent pooled cerebella samples from the separate sample groups. In each case, lithium reduced DNA binding. The effect was intermediate when both ethanol and lithium were present in the maternal diet.

Adult female rats subjected to E, PF, AL, and lithium treatment, as described above, were also examined (Fig. 8). The adult E, PF, and AL results contrasted with those found in pups. AP-1 DNA binding was enhanced 3-fold in the E adult cerebellum relative to AL control. Lithium treatment alone brought about no change in DNA binding (Fig. 8). However, concomitant lithium and E treatment enhanced DNA binding relative to AL control, but less than the level in the E treatment group.

View larger version (31K): [in this window] [in a new window]   Fig. 8.   Effect of ethanol on AP-1 activation in adult female rats. Rats were subjected to E, PF, AL, and lithium treatment, as described under Experimental Procedures. A representative EMSA (of three replicates) shows AP-1 DNA binding enhanced 3-fold in the E adult cerebellum relative to AL control. Lithium treatment alone had no effect on DNA binding. However, concomitant lithium and E treatment resulted in DNA binding over 2-fold greater than that in AL control, but less than E group DNA binding.

The decreased activation of the transcription factors seen on PD1 disappeared as the neonates underwent further development and were no longer exposed to ethanol. By the eighth postnatal day (PD8), the DNA binding of NF-B was restored to normal (Fig. 9A). However, in the case of AP-1, by PD8 (Fig. 9B) there was only a partial restoration of the DNA binding (68% and 84% of AL values, respectively for E and PF). By PD15, AP-1 activation in the ethanol-exposed group had rebounded in excess of control values (Fig. 9C).

View larger version (43K): [in this window] [in a new window]   Fig. 9.   A, transcriptional activation of NF-B at eight postnatal days (upon withdrawal of ethanol treatment). The figure shows a representative EMSA (three experiments) for cerebellar NF-B on the 8th postnatal day (PD8). By PD8, there was no significant difference in NF-B activation in both E and PF groups relative to AL control levels. B, transcriptional activation of AP-1 on later postnatal days (upon withdrawal of ethanol treatment). EMSAs for AP-1 on PD8 and PD15 show gradual restoration of AP-1 activation in the in utero ethanol-exposed rat cerebella (E) on PD8. By PD15, E group transcriptional activation had rebounded in excess of control levels. The figure shows the in utero ethanol-exposed group, the pair-fed control, and the ad libitum control groups, respectively, as described in the text.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

AP-1 is a sequence-specific transcription activator (Karin, 1995). Importantly, the promotion regions of many cellular genes have AP-1-binding sites. Thibault et al. (2000), using DNA array studies on SH-SY5Y neuroblastoma cells, demonstrated that some 42 genes had mRNA levels altered after 3 days of ethanol exposure. Our analysis of the promoter regions of many of these genes showed AP-1-binding sites. The fact that, after chronic exposure to the in utero ethanol environment of FAS, there is less (Fig. 1) AP-1 and NF-B DNA binding may have a significant bearing on the ability of brain cells to develop properly and survive. Interestingly, similar PD1 DNA-binding trends were observed with the general transcription factor II D as well as cAMP response element-binding protein (unpublished data).

Neurotrophic factors can protect cells from programmed death (Cui et al., 1997; Luo et al., 1997; Zhang et al., 1998; Ikonomidou et al., 1999). There is evidence that some neurotrophic factors activate AP-1 and/or NF-B. Gaiddon et al. (1996) demonstrated that brain-derived neurotrophic factor stimulates AP-1 activation. Similarly, Maggirwar et al. (1998) have shown that nerve growth factor activates both AP-1 and NF-B. These transcription factors could, therefore, be involved in protecting neurons from apoptosis.

NF-B protects certain cell types from apoptosis possibly by inducing a number of antiapoptotic proteins including cIAP-1 and cIAP-2 (Wang et al., 1996). Beg and Baltimore (1996), using 3T3 cell lines derived from p65 +/+ and p65 / embryonic fibroblasts, found the former to be better protected from tumor necrosis factor--induced cytotoxicity. There was protection in the p65 / cells only after transfection with p65. Wang et al. (1996) used a human fibroblast cell line, HT10180I, containing a mutant "super-repressive" I-B, and a control version, HT10180V. They found tumor necrosis factor-mediated apoptosis to be blocked in the normal cell line but not in the mutant one.

By itself, malnutrition impairs fetal growth (Kennedy, 1984; Schenker et al., 1990). The under-nutrition associated with ethanol use was controlled for by way of the PF group. As Fig. 1 shows, the pair-fed control rats, which had been exposed to identical amounts of calories as the ethanol-fed rats, also had reductions in the DNA binding of AP-1 and NF-B. However, the reductions were to a lesser extent than those observed in the E group.

Reactive oxygen species readily react with biomacromolecules, either directly damaging them or starting chain reactions resulting in extensive damage to cellular structures. Thus, aerobic cells have developed an array of effective antioxidant defenses including superoxide dismutase, catalase, glutathione, glutathione peroxidase, and thioredoxin. Excessive production of highly reactive free radicals puts cells under "oxidative stress". Under such conditions, this antioxidant arsenal can be overwhelmed, and cell death results. Alternatively, there can be an induction of compensatory changes in antioxidant activity. Changes in glutathione levels or catalase activity would, therefore, be appropriate indices of oxidative stress (Bondy, 1992).

The pregnant rats used in these studies had their ethanol-containing diet withdrawn 2 days before parturition. The levels of glutathione were not statistically different between the groups on PD1. There were also no significant differences in the activity of catalase between the groups. These observations, although important, do not rule out oxidative stress on PD1. Neither do they preclude such stress occurring during the period of actual ethanol exposure in utero, or at points during the developmental process.

Adult AP-1 DNA-binding trends were opposite to observations in pups. The E treatment increased adult cerebellum AP-1 DNA binding to 3-fold AL control levels. This observation underscores the contrast between certain other important ethanol effects on the mature brain versus that in the developing brain (Snell et al., 1996; Hughes et al., 1998).

The PD1 transcriptional observations are not attributable to changes in amounts of transcription factor proteins since they were all unchanged. However, to be transcriptionally active, c-JUN has to be phosphorylated at sites such as serine 73. When c-JUN is phosphorylated on Thr-231 or Ser-243 (by casein kinase II or a DNA-dependent kinase), or on Ser-149 (by extracellular signal-regulated kinase), binding to DNA is inhibited. Removing such phosphate groups by a protein kinase C-activated phosphatase facilitates DNA binding. Phosphorylation of Ser-63 and Ser-73 by stress-activated protein kinases-1 (SAPK-1), on the other hand, facilitates DNA binding (Karin, 1995). As illustrated in Fig. 5A, trends in c-JUN phosphorylation (serine 73) mirror the observations made in the EMSAs. Western blots further showed that in utero ethanol exposure does not decrease the levels of c-FOS proteins in neonatal cerebella.

The effects of in utero ethanol exposure on NF-B and AP-1 DNA binding could be viewed in terms of their impact on cell development. GSK-3 is crucial in cell fate decisions both during development and in adults (Kim and Kimmel, 2000). Besides glycogen synthase, GSK-3 has several other substrates, including NF-B and a number of other transcription factors (Saskela et al., 1992; Fiol et al., 1994; Ross et al., 1999; Hoeflich et al., 2000; Xavier et al., 2000). It also suppresses AP-1 DNA binding (Boyle et al., 1991). GSK-3, when phosphorylated on serine/threonine residues, is inactive and unable to suppress normal cell development and proliferation. However, when it is phosphorylated on tyrosine 216 (GSK-3) or tyrosine 279 (GSK-3), it is active and able to regulate transcriptional activity (Wang et al., 1994; Murai et al., 1996; Markuns et al., 1999). Hoeflich et al. (2000) have shown that GSK-3 is required for NF-B-mediated activation of genes important for cell survival. As Fig. 5 shows, the active form of GSK-3 is diminished in EM20 E rat cerebella. By PD1, however, the active form of GSK-3 activity is increased in the E group (Fig. 6). Thus, increased PD1 GSK-3 activity could be a compensatory response to the suppression of NF-B DNA binding, especially in view of the fact that, in this model, the NF-B DNA binding itself is restored within 8 days postnatal (Fig. 9). However, increased GSK-3 activity leads to less AP-1 transcriptional activity. In this model AP-1 DNA binding is restored in the E group between PD8 and PD15 (Fig. 9).

Lithium, a noncompetitive specific inhibitor of GSK-3 (Salinas and Hall, 1999; Ryves and Harwood, 2001), when introduced into the gestational diet, greatly diminishes the cerebellar DNA binding of AP-1 and NF-B in the rat pup on PD1 (Fig. 7). This may be an indication of the importance of GSK-3 activity in the transcriptional effects observed in this fetal alcohol syndrome model. In this regard, there was a contrast between the developing and the mature cerebellum: lithium treatment did not alter AP-1 DNA binding in the adult cerebellum (Fig. 8). It must be pointed out, however, that lithium has other effects on the brain. Chronic lithium treatment leads to increased glutamate uptake and increased GABA_B receptor levels. Therapeutic doses of lithium inhibit enzymes that recycle inositol in cells. Lithium also affects the release of 5-hydroxytryptamine (serotonin) and dopamine (reviewed by Salinas and Hall, 1999). Thus, although gestational lithium blocks GSK-3, other targets may have been affected simultaneously.

In individual rat EM20 cerebella that have more severe depletion of AP-1 and NF-B DNA binding, SAPK-1 activity was elevated. Active SAPK-1 phosphorylates c-JUN and makes it transcriptionally active (Karin, 1995). Concomitant reduction in the active form of GSK-3 similarly should result in increased c-JUN transcriptional activation. This would be a bid to increase and restore AP-1 transcriptional activity in the face of continued suppression. Indeed, upon parturition, when the ethanol environment is withdrawn, cerebellar transcriptional activity is restored (Fig. 9). Thus, the effects of chronic in utero ethanol exposure on NF-B and AP-1 were reversible. Although the suppression of transcriptional activation is reversible upon ethanol withdrawal, the consequences could be longer-lasting because this is a period of neurodevelopment.

Our findings show that one of the ways by which chronic in utero ethanol exposure may undermine the development of the cerebellum in FAS is at the level of regulation of the activities of transcription factors. Studies probing further the consequences of these effects, and the mechanisms through which cerebellar transcriptional activity gets altered by ethanol during neurodevelopment, are currently under way.