Vitamin e does not protect against neona

Conclusions:In contrast to the previously reported neuroprotective potential of antioxidants on EtOH-mediated cerebellar damage, vitamin E supplementation did not diminish EtOH-induced s

Vitamin E Does Not Protect Against Neonatal Ethanol-Induced Cerebellar Damage or Deficits in Eyeblink

Classical Conditioning in Rats

Tuan D Tran, Holly D Jackson, Kristin H Horn, and Charles R Goodlett

Background:Rodent studies have shown that heavy binge-like ethanol (EtOH) exposure during the brain growth spurt [postnatal days (PD) 4 –9] causes cerebellar neuronal loss and deficits in cerebellar-mediated eyeblink classical conditioning (ECC) Oxidative stress has been implicated in EtOH-cerebellar-mediated brain damage, and studies using vitamin E have reported amelioration of EtOH-induced tissue damage, including protection in rats against EtOH-induced cerebellar Purkinje cell (PC) loss on PD 4 to 5 The purpose of this study was to determine whether dietary supplementation with vitamin E concurrent with binge EtOH exposure on PD 4 to 9 in rats would attenuate the cerebellar cell death and ECC deficits

Methods:Rat pups were given one of five different neonatal treatments: (1) intubation with EtOH in milk formula (twice daily, total dose 5.25 g/kg/day), (2) intubation with EtOH in milk formula supple-mented with vitamin E (12.26 mg/kg/feeding), (3) intubation with milk formula that contained vitamin E only, (4) sham intubations, or (5) normally reared unintubated controls Between PD 26 and 33, subjects received short-delay ECC for 3 consecutive days Unbiased stereological cell counts were performed on cerebellar PCs of left cerebellar lobules I to VI and neurons of the interpositus nucleus In a separate study with PD 4 pups, the effects of vitamin E on EtOH-induced expression of caspase-3 active subunits were assessed using Western blot analysis

Results:EtOH-treated groups showed significant deficits in acquisition of conditioned eyeblink re-sponses and reductions in cerebellar PCs and interpositus nucleus neurons compared with controls Vita-min E supplementation failed to protect against these deficits VitaVita-min E also failed to protect against increases in caspase-3 active subunit expression induced by acute binge EtOH exposure on PD 4

Conclusions:In contrast to the previously reported neuroprotective potential of antioxidants on EtOH-mediated cerebellar damage, vitamin E supplementation did not diminish EtOH-induced structural and functional damage to the cerebellum in this model of binge EtOH exposure during the brain growth spurt

in rats

Key Words:Fetal Alcohol Syndrome, Antioxidants, Eyeblink Conditioning, Cerebellum, Caspase-3

result in a spectrum of adverse outcomes to the fetus.

In humans, fetal alcohol syndrome (FAS) can be diagnosed

when the constellation of characteristic craniofacial

abnor-malities, growth retardation, and central nervous system

(CNS) damage is present (Hanson et al., 1978; Jones and

Smith, 1973; Streissguth et al., 1980; Wisniewski et al.,

1983) The spectrum of effects on CNS function can include

impairments of cognitive function, intelligence, social

be-havior, and sensory and motor function (Aronson et al.,

1997; Brown et al., 1991; Mattson and Riley, 1999; Olson et al., 1998, 1997; Roebuck et al., 1999; Streissguth et al., 1980; Uecker and Nadel, 1998), and these effects persist or are even exacerbated as the child grows into adulthood (Bookstein et al., 2002; Streissguth et al., 1991) Public educational efforts have not yet reduced the incidence of fetal alcohol spectrum disorders To the contrary, risk drinking during pregnancy, including binge drinking, actu-ally increased in the United States between 1991 and 1995 (Ebrahim et al., 1999, 1998).

An important goal of preclinical animal model studies of developmental exposure to alcohol is to identify interven-tions that can potentially prevent or attenuate ethanol’s (EtOH’s) effects on the CNS (Warren and Foudin, 2001; West et al., 1986) One approach has been to deliver mo-lecular interventions at the time of EtOH exposure to limit

or prevent the pathophysiological consequence of EtOH exposure during development For example, a promising recent discovery found that treatment with small peptides derived from larger endogenous glial proteins

(activity-From the Department of Psychology, Indiana University-Purdue University

at Indianapolis, Indianapolis, Indiana.

Received for publication March 10, 2004; accepted September 22, 2004.

Supported by grants R01 AA11945, U01 AA14829, and T32 AA07462 and

a Proctor & Gamble Company Graduate Fellowship (Society of Toxicology).

Reprint requests: Tuan D Tran, PhD, Department of Psychology, IUPUI,

402 N Blackford Street, LD-124, Indianapolis, IN 46202; Fax:

317-274-6756; E-mail: ttran@iupui.edu.

Copyright © 2005 by the Research Society on Alcoholism.

DOI: 10.1097/01.ALC.0000150004.53870.E1

dependent neuroprotective protein) (Brenneman and

Fos-ter, 1987; Brenneman et al., 2000) protected against fetal

death and growth restriction induced by binge EtOH

expo-sure on gestational day 7 in a C57BL/6 mouse model

(Spong et al., 2001) Although the mechanisms of this

protection are unknown, these peptides have potent

neu-roprotective and antioxidant capabilities in several

differ-ent model systems (Brenneman et al., 1998, 2000; Glazner

et al., 1999; Steingart et al., 2000).

Because multiple pathophysiological mechanisms are

im-plicated in the teratogenic effects of EtOH (Goodlett and

Horn, 2001; West et al., 1994), it is unlikely that a

molec-ular intervention targeting a single mechanism or

patho-logic process will be useful in preventing all prenatal

alcohol-induced damage Nevertheless, there is substantial

experimental evidence that at least some of the

develop-mental damage incurred by EtOH exposure is associated

with increased formation of reactive oxygen species (ROS)

and resulting toxicity related to oxidative damage in fetal

tissue induced by these oxygen radicals (Henderson et al.,

1999) Oxidative stress has the potential to disrupt cell

function in the developing organism and lead to cell death

(Grisham, 1992; Ramachandran et al., 2003; Valencia and

Moran, 2001) As demonstrated in many different model

systems, EtOH can enhance ROS accumulation beyond the

capacity of cellular antioxidant defenses, producing

peroxi-dation of biomacromolecules and disruption of cellular

function potentially leading to cell death (Montoliu et al.,

1995; Navasumrit et al., 2000; Sun et al., 1997) EtOH

induces the formation of ROS in cell lines (Davis et al.,

1990; Devi et al., 1993; Sun et al., 1997), whole-brain

homogenates (Uysal et al., 1989), and in vivo (Heaton et

al., 2002; Henderson et al., 1995; Reinke et al., 1987).

A substantial number of studies have reported successful

protection against EtOH-induced toxicity or damage with

exogenous application of antioxidants, frequently using

␣-tocopherol (vitamin E) Antioxidants (including vitamin

E) were protective in vitro against EtOH toxicity in

cul-tured whole mouse embryos, hepatocytes, PC12 cells,

neu-ral crest cells (Davis et al., 1990; Devi et al., 1993; Kotch et

al., 1995; Sun et al., 1997), and primary cultures of

hip-pocampal neurons, even at high EtOH concentrations

(Mitchell et al., 1999a,b).

Most relevant to the current study, Heaton et al (2000)

reported that vitamin E protected against loss of Purkinje

cells (PCs) in lobule I of the cerebellar vermis in a neonatal

rat model involving binge EtOH exposure on postnatal

days (PD) 4 to 5 Many previous studies have shown that

binge EtOH exposure over PD 4 to 9 of neonatal rats, a

period of rapid brain (and cerebellar) growth comparable

to what occurs over fetal weeks 24 to 32 in humans (Zecevic

and Rakic, 1976), reliably causes significant,

dose-dependent loss of cerebellar neurons, typically

demon-strated by loss of postmitotic cerebellar PCs (Bonthius and

West, 1991; Goodlett et al., 1998, 1997; Hamre and West,

1993; Light et al., 2002) If vitamin E can protect against

neonatal EtOH-induced cerebellar cell loss in a relatively general manner, i.e., with multiple exposure episodes eval-uated after longer survival times, then vitamin E supple-mentation could provide a potentially effective approach to prevention of at least part of the brain damage and behav-ioral dysfunction associated with FAS.

Other than the Heaton et al (2000) report, we are unaware of any other in vivo EtOH studies evaluating the neuroprotective effects of vitamin E on cerebellar structure

or function The goal of this study was to assess whether vitamin E supplementation could protect against cerebellar damage in the full neonatal rat model of binge EtOH exposure during the “third trimester equivalent,” i.e., with daily episodes of binge exposure given over PD 4 to 9 Neuroanatomical outcomes (stereological counts of neuron number in two cerebellar populations) as well as functional outcomes (Pavlovian conditioning of eyeblink responses, a form of learning for which cerebellar neural circuits are required) were evaluated after the rats reached periadoles-cence (~PD 30) The EtOH treatments and vitamin E protocols matched that reported by Heaton et al (2000), except that the treatments were given on PD 4 to 9 instead

of just PD 4 to 5 and the rats survived to adolescence rather than being evaluated on PD 5.

Eyeblink classical conditioning (ECC) was used because significant and enduring deficits in conditioned response (CR) acquisition have been reliably demonstrated using this neonatal binge EtOH model (Green et al., 2002a,b, 2000; Stanton and Goodlett, 1998) Neuron counts were performed in cerebellar populations that are known to mediate ECC, i.e., the neurons of the interpositus nucleus (IP) and the PCs of lobules HI to HVI of the cerebellar hemisphere (Anderson and Steinmetz, 1994; Kim and Thompson, 1997; Lavond et al., 1993; Steinmetz, 2000) In addition, binge EtOH exposure triggers a pathogenesis cascade in several brain regions that leads to apoptotic cell death via activation of caspase-3, a critical protein involved

in the “execution” phase of the apoptotic pathway (Mooney and Miller, 2001; Olney et al., 2002a,b) Caspase-3 is gen-erated and stored as a proenzyme, and precursor cleavage must occur for enzyme activation The acute death of cer-ebellar PCs with binge EtOH exposure on PD 4 has been shown to involve activation of caspase-3, with a peak in active subunit expression 8 hr after the EtOH exposure (Light et al., 2002) Consequently, this study also assessed the potential of vitamin E to attenuate or prevent this EtOH-induced expression of the caspase-3 active subunit after an acute binge EtOH exposure on PD 4.

MATERIALS AND METHODS

Subjects

Litters of Long-Evans rats were produced from breeders obtained from Simonsen Laboratories (Gilroy, CA) The male breeders were mated with two to three female rats overnight in the vivarium of the Department of Psychology at Indiana University-Purdue University at Indianapolis Sperm smears were examined early the next morning, and female rats with

positive smears were designated as being on gestational day (GD) 0 of

pregnancy At birth, litters were assigned either to “intubation” conditions

or to unintubated (suckle control) conditions Litters were culled to eight

pups (four male and four female when possible), and handling and

treat-ment of litters began on PD 4 For the intubated litters, one male and one

female pup were randomly assigned within litter to each of four treatment

conditions administered on PD 4 to 9: (1) EtOH/milk (E), (2) vitamin

E/EtOH (VE-E), (3) vitamin E/Milk (VE-M), or (4) sham intubated (SI)

For the separate suckle-control litters [unintubated control (UC)], only

one male and one female pup per litter were selected at random for use

in this study

All rats remained with their mothers until PD 21, at which time they

were weaned and housed with same-sex littermates (two rats per cage)

until the day of surgery All animal care and experimental procedures were

reviewed and approved by the Institutional Animal Care and Use

Com-mittee before experimentation

Neonatal Treatment

All milk treatment formulas were prepared from a base milk formula

that was made according to the method of West, Hamre, and Pierce (West

et al., 1984) and were delivered via intragastric intubation as described

previously (Goodlett et al., 1997) For infusions with EtOH (AAPER,

Shelbyville, KY), the milk formula contained 11.9% (v/v) EtOH For

infusions with vitamin E (T1157, 1360 IU/g; Sigma, St Louis, MO), the

milk formula contained 60 IU of vitamin E/100 ml of formula, i.e., 0.441

mg of vitamin E/ml of solution With combined treatments, the milk

infusions contained 11.9% (v/v) EtOH and 0.441 mg of vitamin E/ml of

solution For each intubation, a total volume of 0.02778 ml was infused per

gram of body weight Each infusion with EtOH delivered 2.625 g/kg of

EtOH, and each infusion with vitamin E delivered 12.26 mg/kg of vitamin

E

The pups in the intubated litters (E, VE-E, and VE-M) received four

intubations each day (PD 4 –9), following the sequence reported by

Heaton et al (2000) Each intubation session was separated by 2 hr Group

E received an initial feeding of base milk (without EtOH or vitamin E),

followed by two feedings of EtOH in milk, and a final feeding of base milk

(without EtOH or vitamin E) Group VE-E received an initial feeding of

vitamin E in milk (no EtOH), followed by two feedings of vitamin E and

EtOH in milk, and a final feeding of vitamin E in milk (no EtOH) Group

VE-M received four feedings of vitamin E in milk (no EtOH) Pups from

the UC group (from separate litters) were weighed only daily The

addi-tional milk intubations were given because the intoxication produced by

the EtOH treatments impairs suckling behavior, limiting the intake of

calories for 12 to 18 hr Without the supplements of milk formula, the

EtOH pups would not receive sufficient calories and would show large

growth delays The SI group received only sham intubations (no infusion

of milk formula) because previous works using this model have shown that

control treatments involving intubation of isocaloric milk formula

pro-duces accelerated growth of the pups, compared both with EtOH-treated

littermates and with contemporaneous suckle control litters, as a result of

the added calories for the nonintoxicated controls (Goodlett et al., 1998,

1997) Because the intubation model cannot control the caloric intake of

the pups (as a result of a function of suckling and mother–pup

interac-tions), we and others have adopted the SI procedure (without isocaloric

milk) to limit the growth acceleration effects in control pups (Goodlett et

al., 1998, 1997; Lugo et al., 2002; Marino et al., 2002)

On PD 4, 4 hr after the first EtOH feeding, a 20-␮l blood sample was

collected in a heparinized capillary tube from a tail-clip of each intubated

pup The tubes were centrifuged, and plasma was separated and frozen at

⫺70°C Blood EtOH concentrations (BECs) for E and VE-E rats were

determined using an oximetric procedure with an Analox GL5 Alcohol

Analyzer (Analox Instruments, Lunenburg, MA)

Surgery

Between PD 24 and 30, weanling rats were separated from their

littermates into individual cages with food and water ad libitum They

remained in individual cages for the duration of recovery and subsequent eyeblink training During surgery, an electromyographic (EMG) head-stage was implanted according to the method of Stanton et al (1992) under Isoflurane gas (Abbott Laboratories, Abbott Park, IL) The head-stage contained two stainless steel wires (size 3T; Medwire, Mt Vernon, NY) for measuring differential EMG activity of the eyelid and one stain-less steel wire that served as ground (size 10T; Medwire) The differential EMG recording wires were implanted on the upper eyelid muscle of the left eye, and the ground lead was placed subcutaneously, posterior to Lambda A bipolar stimulating electrode (MS303/2; Plastics One, Roanoke, VA), used to deliver the shock unconditioned stimulus (US), was placed subcutaneously with its tips in a V shape on the periorbital muscle immediately caudal to the left eye Cranioplast was added to secure all components to the cranium, and to anchor it, two 15-mm strips

of 24-G galvanized steel wire were implanted subcranially parallel to each other and at positions immediately posterior to Bregma and anterior to Lambda landmarks Surgeries lasted ~30 min for each rat, and after surgery, they were returned to their home cages and monitored for recovery from anesthesia

Apparatus

Animals were allowed to move freely in a modified test box (30.5⫻ 24.1⫻ 29.2 cm; Med-Associates, St Albans, VT) constructed with alumi-num and clear polycarbonate walls The floor was made of stainless steel rods (4.8 mm) in a polypropylene frame The test box was contained within

a sound-attenuated chamber (BRS-LVE, Laurel, MD) Each chamber was equipped with a fan (55- to 65-dB background noise level), house light (15 W), and two piezoelectric speakers (2- to 12-kHz range), one of which was used for presentation of the tone conditioned stimulus (CS) Both the fan and the house light were left running during training Each chamber was fitted with a commutator (AC267-20; Litton Systems, Blacksburg, VA) that was connected to peripheral equipment The commutator contained five lines of redundancy per channel and allowed a rat maximum mobility The US was produced by a constant-current, 60-Hz stimulus isolator (A365R; World Precision Instruments, Sarasota, FL) Two IBM-compatible computers (four chambers per computer) with custom-developed software controlled the delivery of stimuli and recording of EMG activity (JSA Designs, Raleigh, NC)

Design and Procedures

Between PD 26 and 33, rats were subjected to short-delay ECC using paired CS-US presentations Care was taken to ensure that age of surgery was balanced across all treatment groups for each surgical cohort The ECC procedure consisted of a tone CS (2.8 kHz, 80 dB) that preceded, overlapped, and co-terminated with a shock US (90 Hz) that lasted 100 msec The time between CS onset and US onset produced an interstimulus interval (ISI) of 280 msec Conditioned eyeblink training occurred over 3 consecutive days, and each day consisted of two sessions with 100 trials each (90 paired CS-US trials and 10 CS-alone trials per session) Each session was separated by a 5-hr interval, and the intertrial interval aver-aged 30 sec (range: 15– 45 sec) Eyelid EMG activity was amplified (⫻5000) and bandpass-filtered at 500 to 5000 Hz by a differential AC amplifier and then rectified and integrated by a DC integrator (⫻10) before being passed to a computer for storage

The US intensity used for conditioning was determined on an individ-ual rat basis such that during the first 20 trials of each session, the shock intensity was first set at 0.4 mA and the experimenter monitored emitted eyeblink responses to determine whether they consistently equaled or surpassed 1 arbitrary volt unit When the subject failed to exhibit reliable unconditioned responses (URs) during the first few trials, the shock intensity was increased in 0.2-mA increments until satisfactory URs were met After the first 20 trials (of each session), no further adjustments in shock intensity were performed

Histology and Cell Number Quantification

Fixation After completion of ECC (between PD 30 and 37), rats were

deeply anesthetized with Nembutal (100 mg/kg intraperitoneally), and

0.9% saline, followed by 1.0% (w/v) paraformaldehyde and 1.25% (v/v)

glutaraldehyde in 0.1 M of phosphate buffer, was perfused via the left

cardiac ventricle The brains were postfixed in the same perfusate at 4°C

Infiltration and Embedding The cerebella were divided into halves with

a midsagittal cut at the midline vermis (performed under a dissecting

microscope), and tissue blocks that contained the left cerebellar

hemi-sphere (ipsilateral to the eye subjected to eyeblink conditioning) were

processed The tissue blocks were dehydrated through a graded series of

EtOHs (70, 95, 95, and 100% for 2 hr each) before infiltration with a

graded series of glycolmethacrylate (GMA) resins without hardener (25,

50, 70, and 100%⫻4 days for 24 hr each day) Tissue blocks then were

embedded in GMA resin (Technovit 7100; Kulzer, Wehrheim, Germany)

Sectioning, Staining, and Mounting Embedded tissue blocks were cut

using a motorized rotary microtome (model 2155; Leica, Wetzlar,

Ger-many) using glass knives that were prepared using a Ralph glass

knife-maker (Energy Beam Sciences, Agawam, MA) Cerebella blocks

(exclud-ing flocculi) were cut exhaustively in the sagittal plane at a nominal

thickness of 30␮m, and a known fraction of one of every two sections was

saved (see “Optical Fractionator Method” below) A systematic random

sampling procedure was applied in collecting tissue sections, whereby the

first or second section was randomly saved, with every other section saved

thereafter Sections were transferred to an ice-cold solution that contained

20% EtOH in distilled water and were mounted onto plain glass slides

The sections were heat-fixed using a slide warmer at 60°C for at least 1 hr,

after which they were stained using a cresyl violet solution modified for

use with GMA-embedded tissue The staining solution contained 100 ml

of cresyl violet stock, 120 ml of 0.1 M of glacial acetic acid, and 80 ml of

0.1 M of sodium acetate buffer solution Sections were dehydrated using

a descending series of EtOHs, stained for 1 to 2 min at room temperature,

differentiated in dilute acetic acid in 100% EtOH for 3 to 4 min, and

briefly rinsed in an ascending series of EtOHs

Optical Fractionator Method PCs of lobules I to VI of the left

cerebel-lum, including left vermal lobules I to VI and left hemispheric lobules HI

to HVI, and neurons of the IP were quantified in an unbiased manner

using the optical fractionator method (Green et al., 2002b; Gundersen,

1986; Gundersen et al., 1988; Oorschot, 1994; Tran and Kelly, 2003; West

et al., 1991) This method involves counting nuclei (or nucleoli) with

optical dissectors in a uniform and systematic sample that constitutes a

known fraction of the neural region that is of concern All sampling,

counting, and computational rules according to the method of West et al

(1991) were applied to obtain unbiased estimates of total neuron number

We used this approach to quantify neurons in recent reports (Green et al.,

2002b; Tran and Kelly, 2003) and briefly describe it below Investigators

were blind to the identity of rats in each treatment group

For achieving an efficient sampling scheme that had minimal precision

error in individual estimates [i.e., the coefficient of error (CE)] and

produced an average of 1 to 2 neurons sampled per frame and 100 to 200

counts per cerebellar or IP region (Gundersen and Jensen, 1987; West et

al., 1991), initial pilot work with juvenile rats of similar age was used to

determine the counting parameters for the experimental animals From

this pilot work, it was determined that the optimal section sampling

fraction (sections saved/sections cut) was one half for the IP and one sixth

for left cerebellar lobules I to VI This yielded ~14 to 20 sagittal sections

through the entire extent of the IP and ~12 to 22 sagittal sections through

the entire extent of left cerebellar lobules I to VI The distance between

dissector samples on each section (X,Y stepping distance) was 200␮m2for

the IP and 800␮m2for left cerebellar lobules I to VI The counting frame

size (a[frame]) varied according to the neuron size in each region; for the

IP, it was 3.03⫻ 103␮m2, and for left cerebellar lobules I to VI, it was 6.06

⫻ 103␮m2 For both the IP and cerebellar lobules, counting was always

conducted using a guard height of 3␮m and confined to a dissector height

of 15␮m A minimum of three section thickness (t) measurements (for

three or more dissector samples) were obtained for each section by

determining the top and bottom surfaces with a Z-axis encoder while

viewing with a⫻100 oil immersion lens (Plan Apo, 1.4 N.A.) An average section thickness computed across all sections was used as a factor in the estimate of total neuron number

Caspase-3 Active Subunit Expression Subjects Forty-eight PD 4 Long-Evans rat pups were used for Western

blot analysis of caspase-3 active subunit expression In addition, 16 pups were used for immunocytochemical localization in cerebellar sections (using the same antibody as for the Western blots) Each litter was culled

to eight, with one male and one female pup assigned to each of four treatment groups: (1) milk only (M), (2) vitamin E/milk (VE-M), (3) EtOH/milk (E), or (4) vitamin E/EtOH (VE-E) There was an exception for one of the six litters, in which unequal sex distribution resulted in both animals in the E and VE-E treatments being the same sex (female) On

PD 4, pups were treated in the same manner as described for the eyeblink and anatomy portions of this study (i.e., with 2.625 g/kg of EtOH and 12.26 mg/kg of vitamin E each feeding) with the following exception: control pups for this experiment received milk without vitamin E (M), replacing the SI group of pups

Protein Isolation Eight hours after the initial EtOH intubation on PD

4, animals were decapitated and cerebella were dissected, weighed, frozen

in liquid nitrogen, and stored at⫺70°C until the protein isolation was performed Cytosolic protein was isolated following a modified procedure described previously (Oberdoerster and Rabin, 1999) In brief, cerebellar tissue was sonicated in 200␮l of extraction buffer (20 mM of HEPES, 0.4

M of NaCl, 20% glycerol, 5 mM of MgCl2, 0.5 mM of EDTA, 0.1 mM of EGTA, 1% NP-40, 1␮g/␮l of Pepstatin A, 5 mM of dithiothreitol, and 0.5

mM of phenylmethylsulfonyl fluoride) and incubated on ice for 30 min Sonicates then were centrifuged at 14,000 RPM (Helmer, Noblesville, IN) for 20 min, and the supernatant was collected Protein concentrations were determined using a Pierce BCA protein assay (Rockford, IL) and quantified at an absorbance of 562 nm on a Smart Spec 3000 (BioRad, Hercules, CA) Protein was stored at⫺70°C until Western blotting was performed

Western Blot Analysis Sixty micrograms of protein was mixed with

laemmli buffer, boiled for 5 min, loaded onto a 15% SDS acrylamide/ bisacrylamide gel, and run overnight at 10 mA (Laemmli, 1970) After transfer to polyvinylidene difluoride membrane (BioRad), blots were stained with Coomassie blue to verify lane loading and blocked for 1 hr in 5% nonfat dry milk solution in TTBS (25 mM of Tris, 0.15 M of NaCl, 0.1% Tween 20, concentrated HCl, and 0.001% Thimerosal) Blots then were incubated with antibody that recognizes caspase-3 active subunit (1:500 dilution in TTBS with 1%␥-globulin; Cell Signaling, Beverly, MA) for 20 min at room temperature and then overnight at 4°C The following day, blots were washed in TTBS and incubated in 5% milk that contained horseradish peroxidase–conjugated anti-rabbit secondary antibody (1:2000; Santa Cruz Biotechnologies, Santa Cruz, CA) for 1.5 hr at room temperature Blots were washed, and protein detection was performed using ECL techniques (Amersham, Piscataway, NJ) Actin expression was used for blot standardization Blots were blocked overnight, incubated with Actin antibody (C-2, 1:100 dilution; Santa Cruz) for 1.5 hr, washed, and incubated with goat anti-mouse horseradish peroxidase–conjugated secondary antibody (1:2000; Santa Cruz) for 1.5 hr, and protein was detected using ECL techniques Densitometric data were obtained for each gel using a GS-710 Densitometer (BioRad)

Immunocytochemistry Frozen sections (40␮m) were cut and washed in PBS, and endogenous peroxidases were quenched with hydrogen perox-ide Sections then were washed in PBS, blocked for 1 hr in 5% NGS with 1% Triton X-100, and incubated overnight at 4°C in primary antibody (1:1000; Cell Signaling) The following day, sections were washed and incubated with secondary antibody (Santa Cruz; 1:200) for 1 hr at room temperature Sections then were washed, incubated with ABC reagent (Vector Laboratories, Burlingame, CA) for 1 hr at room temperature, and washed again, and caspase-3 was visualized with diaminobenzidine tetra-hydrochloride (DAB)

Eyeblink Conditioning Data Collection and Statistical Analyses

The EMG signals were sampled in 2.5-msec bins during each 800-msec

trial epoch Each trial epoch was divided into four time periods: (1)

pre-CS period, a 280-msec baseline period before the onset of the tone

CS; (2) startle response (SR) period, first 80 msec after CS onset (EMG

activity relating to a nonassociative short-latency startle reaction elicited

by the CS); (3) CR period, EMG activity that occurred during the 200

msec of CS presentation that preceded onset of the US; and (4) UR

period, EMG activity that occurred from the onset of the US to the end of

the trial (240 msec) Using criteria described by Skelton (1988) and

Stanton et al (1992), any EMG responses during the SR, CR, or UR

periods that exceeded 0.4 arbitrary units above the pre-CS baseline mean

were registered

Relative frequency and amplitudes of SRs, CRs, and URs were

ana-lyzed using mixed ANOVAs (sex⫻ treatment ⫻ session) and, when

appropriate, with two-way (treatment⫻ session) ANOVAs Mean CR

amplitude was regarded as a measure of response magnitude that included

activity not registered during the CR period as a result of failure to reach

threshold (CR amplitude⫽ 0) Total cerebellar PC and IP numbers were

analyzed using one-way ANOVAs All main treatment group differences

were subjected to Tukey’s HSD post hoc tests, and means were reported

as mean⫾ SEM Physical and age data were analyzed using t tests (BECs),

a mixed factorial ANOVA (PD 4 –9 body weights), a between-subjects

ANOVA (PD 21), a one-way ANOVA (cerebellar tissue weight), or a

Kruskal-Wallis one-way ANOVA (age on first training day)

The effects of vitamin E administration on EtOH-induced increases in

caspase-3 active subunit expression were evaluated using a three-way

ANOVA with treatment (EtOH versus milk) and vitamin E

supplemen-tation (no vitamin E versus vitamin E) and sex (male, female) as fixed

factors Post hoc comparisons among neonatal treatment groups were

conducted using Tukey’s HSD, and means were reported as mean⫾ SEM

RESULTS

BECs and Growth

Data from a total of 44 rats (E: m ⫽ 6, f ⫽ 5; VE-E:

m ⫽ 6, f ⫽ 3; VE-M: m ⫽ 5, f ⫽ 3; SI: m ⫽ 3, f ⫽ 5;

UC: m ⫽ 4, f ⫽ 4) were considered for all analyses, with

the exception of cell counts (see below) Fifty-four rats

were initially obtained for the ECC study, but 10 rats (E

⫽ 1, VE-E ⫽ 4, VE-M ⫽ 3, SI ⫽ 2, and UC ⫽ 0) were

excluded from the analyses because their UR EMG

sig-nals across six sessions did not meet criteria for

reliabil-ity, typically as a result of suboptimal placement of the

bipolar electrodes or loss of signals from the EMG

electrodes.

The mean BEC of all EtOH pups on PD 4 was 391 ⫾ 18

mg/dl A t test comparing E (393 ⫾ 29 mg/dl) and VE-E rat

pups (388 ⫾ 22 mg/dl) revealed no significant difference

between treatments For pups in the study of active

caspase-3 subunit expression on PD 4, the BECs averaged

367 ⫾ 17 mg/dl and did not differ between E and VE-E groups (369 vs 363 mg/dl, respectively).

A sex ⫻ treatment ⫻ day mixed ANOVA with day as the within-subjects variable on body weights over the neonatal treatment period (PD 4 –9) yielded the expected significant

main effect of treatment [F(4,34) ⫽ 2.79, p ⬍ 0.05], along

with the expected significant increase in body weight as a

function of PD [F(5,170) ⫽ 692.07, p ⬍ 0.0001] and a significant interaction of these two factors [F(2,0,170) ⫽

3.34, p ⬍ 0.0001] There were no other significant main or interactive effects The main effect of treatment was re-stricted to lower body weights in EtOH-treated groups (E and VE-E) compared with the SI group The treatment ⫻ day interaction was due mainly to the slower increases in body weights for either EtOH group (E or VE-E) com-pared with the SI group during PD 6 to 9 By PD 21, body weights among treatment groups no longer differed signif-icantly, as indicated by a two-way ANOVA with sex and treatment as between-subjects factors Cerebellar tissue weight (taken at the end of ECC training) was expressed as

a ratio of whole cerebellum (in mg) to final body weight (in g) to minimize bias as a result of differences in the age at which rats were perfused The one-way ANOVA indicated

a significant effect of treatment on these ratios [F(4,44) ⫽

3.58, p ⬍ 0.05] Post hoc analysis with Tukey’s HSD showed that rats of group E had significantly lower ratios than those of groups VE-M and SI, which did not differ from each other; VE-E and UC rats did not differ significantly compared with any group or each other Mean body weight and ratio data are shown in Table 1.

US Intensity

The shock US intensity used for eyeblink training was adjusted accordingly for each subject during the first 20 trials of each training session (see “Materials and Meth-ods”) Because of the variability in working US intensities, the mean US intensity during the first 20 trials of each session was subjected to a mixed ANOVA (treatment ⫻ session) with session as the repeated factor This analysis indicated that the mean shock intensity values (first 20 trials/session) were not significantly different among treat-ment groups (E: 1.40 ⫾ 0.08; VE-E: 1.51 ⫾ 0.18; VE-M: 1.20 ⫾ 0.08; SI: 1.40 ⫾ 0.12; and UC: 1.6 ⫾ 0.14; mean of

Table 1 Mean (⫾ SEM) Body Weights Recorded During PD 4 to 9 and on PD 21, and Cerebellar to Final Body Weight Ratios for Rats that Underwent ECC Group

Postnatal body weights (g)

CB to body weight ratio (mg/g)

PD 4 PD 5 PD 6 PD 7 PD 8 PD 9 PD 21

E (n⫽ 11) 9.7 ⫾ 0.3 10.5 ⫾ 0.3 11.8 ⫾ 0.4 13.5 ⫾ 0.4 a

15.2 ⫾ 0.5 a

16.7 ⫾ 0.6 a

44.4 ⫾ 2.3 1.12 ⫾ 0.07 b

VE-E (n⫽ 9) 9.5 ⫾ 0.3 10.2 ⫾ 0.5 11.7 ⫾ 0.6 a 13.5 ⫾ 0.8 15.4 ⫾ 1.0 17.5 ⫾ 1.0 45.1 ⫾ 2.1 1.24 ⫾ 0.07

VE-M (n⫽ 8) 9.3 ⫾ 0.4 10.9 ⫾ 0.4 13.0 ⫾ 0.5 15.1 ⫾ 0.5 17.0 ⫾ 0.6 19.3 ⫾ 0.9 45.5 ⫾ 2.6 1.45 ⫾ 0.11

SI (n⫽ 8) 10.2 ⫾ 0.2 11.6 ⫾ 0.4 13.8 ⫾ 0.5 16.0 ⫾ 0.6 18.0 ⫾ 0.6 20.5 ⫾ 0.7 48.4 ⫾ 1.8 1.49 ⫾ 0.07

UC (n⫽ 8) 9.8 ⫾ 0.4 11.2 ⫾ 0.5 13.1 ⫾ 0.5 15.2 ⫾ 0.5 16.8 ⫾ 0.6 18.5 ⫾ 0.7 48.1 ⫾ 1.6 1.37 ⫾ 0.09

aSignificantly different from SI controls, p⬍ 0.05.

b

Significantly different from VE-M and SI controls, p⬍ 0.05.

six sessions ⫾ SEM), and there were no significant effects

of session or interactive effects of treatment and session.

Acquisition of CRs: Paired CS-US Training

Acquisition of CRs across the six training sessions was

measured by CR percentage and amplitude (arbitrary units,

in V) during the CR period (200 msec before the onset of

the shock US) As shown in Figure 1A and 1B, both groups

that were given EtOH on PD 4 to 9 (E and VE-E) were

impaired in their acquisition of conditioned responding A

mixed ANOVA of the percentage of CRs (Fig 1A)

re-vealed a significant main effect of treatment [F(4,34) ⫽

9.05, p ⬍ 0.0001] and session [F(5,170) ⫽ 91.21, p ⬍ 0.0001]

and a significant interaction between treatment and session

[F(2,0,170) ⫽ 1.66, p ⬍ 0.05] There were no main or

interactive effects of sex Post hoc analysis of main effects

using Tukey’s HSD test of group mean percentage of CRs

(with sexes combined) revealed that groups E (33.6 ⫾

5.1%) and VE-E (27.8 ⫾ 6.0%) expressed significantly

lower mean percentage of CRs (collapsed over session)

than the VE-M (63.1 ⫾ 6.0%), the SI (56.2 ⫾ 6.2%), or the

UC (65.4 ⫾ 6.0%) groups, which did not differ from each

other.

The significant treatment ⫻ session interaction seemed

to be due to the slower acquisition and lower asymptotic

performance of the EtOH-treated groups compared with

the UC and VE-M controls We first evaluated whether the

three control groups differed using a 3 (treatment) ⫻ 6

(session) mixed ANOVA on the CR percent data from UC,

SI, and VE-M groups, and this analysis confirmed that they

did not differ significantly from each other A second

follow-up 2 (treatment) ⫻ 6 (session) mixed ANOVA

in-dicated that the two EtOH-treated groups (E, VE-E) did

not differ significantly from each other, confirming that

neonatal vitamin E supplements did not improve eyeblink

follow-up two-way ANOVAs were also performed to com-pare the VE-M control group with the E group and with the VE-E group These ANOVAs confirmed that the VE-M controls showed significantly more CRs over training than

either of the EtOH-treated groups [F(1,17) ⫽ 15.1, p ⬍ 0.001 for the E group analysis; [F(1,15) ⫽ 18.6, p ⬍ 0.001

for the VE-E group analysis] Similar ANOVAs comparing the SI control group with the EtOH-treated groups also had confirmed that the SI group had significantly higher

CR percentages over training than either the E group

[F(1,18) ⫽ 7.71, p ⬍ 0.05] or the VE-E group [main effect,

F(1,16) ⫽ 9.52, p ⬍ 0.01; treatment ⫻ session interaction,

F(5,80) ⫽ 2.34, p ⬍ 0.05].

Analysis of CR amplitude across the six training sessions

also revealed main effects of treatment [F(4,34) ⫽ 8.81, p ⬍ 0.0001] and session [F(5,170) ⫽ 59.44, p ⬍ 0.0001]

Fur-thermore, differences among treatment groups were de-pendent on the training session (treatment ⫻ session

inter-action, [F(2,0,170) ⫽ 4.18, p ⬍ 0.0001]) There were no

main or interactive effects of sex Post hoc analysis of treatment main effects using Tukey’s HSD test (with sexes combined) revealed that groups E (0.8 ⫾ 0.4 V) and VE-E (0.8 ⫾ 0.5 V) expressed significantly lower average CR amplitudes (collapsed over sessions) than either VE-M (3.2 ⫾ 0.5 V) or UC groups (3.8 ⫾ 0.5 V), which did not differ from each other (Fig 1C) Unlike the percentage of CRs, rats in group SI (2.0 ⫾ 0.5 V) did not differ signifi-cantly in mean CR amplitude from any treatment group The significant treatment ⫻ session interaction for CR amplitude was examined with a series of follow-up mixed ANOVAs on sessions 1 to 6 to identify the source of the interaction A 3 (treatment) ⫻ 6 (session) mixed ANOVA

on the three control groups (UC, SI, and VE-M) confirmed that they did not differ significantly from each other A 2

Fig 1 CR acquisition (⫾ SEM) during paired (CS-US) and probe (CS alone) trials as

a function of training session (S1–S6) (A and C) The frequencies of CRs expressed across sessions were significantly lower in EtOH-exposed rats (E and VE-E) compared with rats

in the three control groups (VE-M, SI, and UC) (B and D) CR amplitude also was significantly lower in EtOH-exposed rats (E and VE-E) com-pared with rats in the three control groups.

(treatment) ⫻ 6 (session) mixed ANOVA showed that the

two EtOH-treated groups (E and VE-E) also did not differ

significantly from each other, indicating again that the

neonatal vitamin E supplements did not improve eyeblink

performance of EtOH-treated pups Additional follow-up

two-way ANOVAs indicated the VE-M group acquired

significantly higher CR amplitudes than either the E group

[F(1,17) ⫽ 20.95, p ⬍ 0.0001; treatment ⫻ session

interac-tion, F(5,85) ⫽ 10.29, p ⬍ 0.0001] or the VE-E group

[F(1,15) ⫽ 11.41, p ⬍ 0.01; treatment ⫻ session interaction,

F(5,75) ⫽ 5.33, p ⬍ 0.0001] The SI control group also

acquired significantly higher CR amplitudes than the E

group [F(1,18) ⫽ 7.99, p ⬍ 0.05; treatment ⫻ session

interaction, F(5,90) ⫽ 3.30, p ⬍ 0.01], but the differences

between the SI and VE-E groups on CR amplitudes did not

reach significance.

Mixed ANOVAs (sex ⫻ treatment ⫻ session) were

con-ducted on SRs to the tone CS, a measure of nonassociative

responding to the CS The percentage of SRs significantly

differed across sessions [F(5,170) ⫽ 6.44, p ⬍ 0.001], but

there were no significant differences among treatment

groups or between sexes or significant interactions

Like-wise, SR amplitudes increased significantly as training

pro-gressed [F(5,170) ⫽ 4.99, p ⬍ 0.0001], but there were no

significant treatment or sex effects.

Measures of nonassociative elicitation of eyeblinks in

re-sponse to the US shock stimulation (URs) were analyzed for

the first session using a 2 ⫻ 5 ANOVA with sex and treatment

as grouping factors Only the data from the first session were

analyzed because the UR amplitude measure can change over

training as a result of summation effects of the emerging CR.

No significant main effects or interactions on URs were

evi-dent The mean UR amplitudes during session 1 of training

(final 80 trials) for each treatment group are as follows: E: 4.6

⫾ 0.6 V; VE-E: 3.1 ⫾ 0.7 V; VE-M: 5.2 ⫾ 0.7 V; SI: 3.7 ⫾ 0.7

V; and UC: 5.2 ⫾ 0.7 V.

Acquisition of CRs: Probe Trials

On every 10th trial within a session, a probe trial that

consisted of the presentation of the tone CS alone, without

subsequent delivery of the shock US, was implemented to

assess further the acquisition of CRs among treatment groups.

There were 10 probe trials per session CRs in a probe trial

may occur during the typical CR EMG collection period (200

msec), during the period that allowed collection of UR EMGs

on paired CS-US trials (140 msec), or during both EMG

collection periods The analysis of these trials was treated in a

similar manner to the analysis of paired CS-US trials

de-scribed previously Results on the means (across sessions) for

both the percentage of CR and CR amplitude measures

in-dicated patterns of significant treatment group differences

that paralleled the pattern of significant differences observed

in the paired CS-US trial means (Fig 1B and 1D) For

per-centage of CRs (with session as the repeated factor), there

were significant group differences between EtOH-treated (E

and VE-E) and control groups (VE-M, SI, and UC); the E and VE-E groups were equally impaired in acquiring CRs compared with the three control groups, which did not differ

among each other [main effect of treatment, F(4,34) ⫽ 8.94, p

⬍ 0.0001; main effect of session, F(5,170) ⫽ 59.30, p ⬍ 0.0001;

no sex effect or interactions] In terms of CR amplitude, the E and VE-E groups were not significantly different from each other but had significantly lower amplitudes than the VE-M and UC groups (which did not differ) Like the paired CS-US trials, SI rats were intermediate, not differing significantly from the E and VE-E groups or from the VE-M and UC

groups [main effect of treatment, F(4,34) ⫽ 7.12, p ⬍ 0.0001; main effect of session, F(5,170) ⫽ 47.23, p ⬍ 0.0001; session ⫻ treatment, F(2,0,170) ⫽ 1.84, p ⬍ 0.05; no sex effect].

Age at First Day of Training

Because of the variation in age of subjects on the first day

of eyeblink training, i.e., PD 26 to 33, there may be concern about the role of age-related differences in the observed group differences A Kruskal-Wallis one-way ANOVA con-ducted with age at the first test day as the dependent variable indicated that there were no significant differences among

treatment groups (p ⬎ 0.50), confirming that age did not influence the group differences in acquisition of ECC or cerebellar cell number among groups (see below) The me-dian and range of ages (in days) were as follows: E: mean ⫽

30, range ⫽ 26 to 33; VE-E: mean ⫽ 30, range ⫽ 26 to 31; VE-M: mean ⫽ 31, range ⫽ 26 to 33; SI: mean ⫽ 29, range ⫽

26 to 33; and UC: mean ⫽ 30, range ⫽ 27 to 32.

Neuron Number

Cerebella from 28 rats were randomly selected from among those that underwent eyeblink training and were processed for histology and neuron number estimation (E:

m ⫽ 4, f ⫽ 4; VE-E: m ⫽ 6, f ⫽ 3; VE-M: m ⫽ 2, f ⫽ 3; UC:

m ⫽ 2, f ⫽ 4) Sham controls (SI) were omitted from this analysis because of the time- and labor-intensive require-ments of the cell-counting process Estimates of the total number of neurons were counted within the IP and of Purkinje neurons in left cerebellar lobules I to VI, and analyzed with one-way ANOVAs As shown in Figures 2 and 3, significant treatment group differences in total cell

number were observed in both the IP [F(3,24) ⫽ 10.95, p ⬍ 0.0001] and left cerebellar lobules I to VI [F(3,24) ⫽ 13.68,

p ⬍ 0.0001] Post hoc analyses using Tukey’s HSD con-firmed that in both neural regions, the two EtOH-exposed groups (E and VE-E) did not differ significantly from each other, and both had significantly fewer IP and Purkinje

neurons than VE-M and UC rats (p ⬍0.005), which did not differ from each other.

Cell counts were obtained with a high level of precision,

as the CE (measure of stereological precision) for each treatment group was well within the recommended upper limit of 0.10 (West et al., 1991) The mean CE for counts made within the IP for group E was 0.052, whereas the CEs

for the other groups were ⬍0.05 The mean CE for PC

counts made within left cerebellar lobules I to VI was 0.052

for group E and 0.051 for group VE-E and was ⬍0.05 for

the two control groups.

Mean section thickness ( ␮m) measured optically during

the counting procedure for the IP and left cerebellar

lob-ules I to VI neurons was also subjected to a one-way

ANOVA No significant treatment group differences in

section thickness were found The mean ( ⫾ SEM) section

thickness measurements were as follows: E: 26.5 ⫾ 0.4;

VE-E: 26.4 ⫾ 0.5; VE-M: 26.0 ⫾ 0.3; and UC: 25.5 ⫾ 0.2.

Correlations: Neuron Number and Eyeblink Conditioning

Pearson correlational analyses of cell number and CR

amplitude were conducted to determine the degree of

re-lationship between these variables CR amplitude was

cho-sen as the dependent measure for learning as opposed to percentage of CRs because it is a more continuous variable that may provide a more sensitive index of learning For example, it may better capture the interaction between cerebellar PCs and IP neurons, because cerebellar PCs have been hypothesized to modulate the gain of the learned eyeblink response produced by the IP (Berthier and Moore, 1986; Gould and Steinmetz, 1996) To capture the rate of acquisition as a single index of learning, we calcu-lated the slope of the regression line based on CR ampli-tude for each rat across sessions 1 to 3 Sessions 1 to 3 were chosen because CR amplitude did not significantly increase after session 3 for any of the treatment groups Bivariate correlations between the slope and neuron number (in IP and left cerebellar lobules I–VI) were computed Learning rate based on the slopes was significantly correlated with

both IP neuron number (r adjusted ⫽ 0.58, p ⬍ 0.001) and

PC number in left cerebellar lobules I to VI (r adjusted ⫽

0.62, p ⬍ 0.0001) In addition, total IP neuron number and

PC number were significantly correlated (r adjusted ⫽ 0.85,

p ⬍ 0.001).

Caspase-3 Active Subunit Expression

As shown in Figure 4, the EtOH-treated pups showed the expected increases in caspase-3 active subunit expression 8

hr after the first intubation, and vitamin E supplementation did not significantly alter this effect of EtOH treatment A

confirmed that EtOH significantly increased expression of the active subunit of caspase-3 [main effect of EtOH,

F(1,40) ⫽ 130.2, p ⬍ 0.001] and that there were no

signif-icant main or interactive effects of vitamin E

supplemen-tation or sex (p ⬎0.57) There was no evidence that vitamin

E supplementation protected against EtOH-induced in-creases in caspase-3 active subunit expression on PD 4 This observation was further supported by extensive caspase-3 labeling of cerebellar PCs, using DAB, in the two EtOH-treated groups but not in the control groups (see Fig 5).

DISCUSSION

Vitamin E supplementation in the present study, follow-ing a protocol previously reported to protect against PC loss in 5-day-old pups induced by EtOH exposure on PD 4

to 5 (Heaton et al., 2000), failed to protect against cerebel-lar structural and functional damage (measured as juve-niles) when the same neonatal binge EtOH treatments were administered daily over PD 4 to 9 Regardless of the vitamin E condition, the PD 4 to 9 EtOH treatments in-duced significant deficits in standard delay ECC along with permanent loss of neurons in the cerebellar populations known to be essential for this form of associative learning (Kim and Thompson, 1997; Steinmetz, 2000; Woodruff-Pak and Steinmetz, 2000) In this study, no sex differences in the two measures of CR acquisition (percentage and ampli-tude) were observed, but this does not preclude the

possi-Fig 2 Estimates of total neuron number (mean⫾ SEM) using the stereological

optical fractionator method in juvenile rats that were trained on ECC (counts from

the left cerebellar hemisphere, which is ipsilateral to the eye used for

condition-ing) (A) The two EtOH-treated groups [EtOH/milk (E) and vitamin E/EtOH (VE-E)]

showed comparable deficits in PC number ( ⫻10 5 ) compared with controls (B)

Total IP neuron number (⫻10 3

) was also not protected from EtOH with vitamin E supplementation (VE-E) during the neonatal treatment period.

bility that there may be sex differences as a result of low

statistical power within each of the treatment groups

Al-though we did not pursue measures of serum levels of

vitamin E or its bioavailability to brain and other tissues,

the supplementation protocol that we followed ensured

that vitamin E was on board in advance of the EtOH

treatment each day The daily doses used (~67 IU/kg/day)

were approximately three times higher (on a per-kilogram

basis) than the typical amounts of supplements for humans

(up to 20 IU/kg/day).

The current results did confirm and extend our previous

findings that binge-like neonatal EtOH exposure during

PD 4 to 9 significantly impairs short-delay ECC (Green et

al., 2002a,b, 2000; Stanton and Goodlett, 1998) and

de-pletes neurons in the specific cerebellar populations that

express learning-related neuronal plasticity necessary for

acquisition of ECC (Green et al., 2002b; Tran et al., 2000).

Vitamin E supplementation failed to affect EtOH-induced

deficits in either measure of learning (percentage and

am-plitude of CRs), and the terminal performance (sessions

5– 6) of both EtOH groups never reached 60% CRs, in

contrast to the ⬎85% CRs for control groups Consistent

with the lack of protection against cerebellar-dependent

learning, vitamin E also failed to protect against

EtOH-induced reductions in PCs in left cerebellar lobules I to VI

(56% of controls) or in IP neurons (69% of controls) The

EtOH-induced expression of active caspase-3 subunits also

replicated the findings of Light et al (2002), but vitamin E

also had no effect on acute increases in expression of the

active subunit of caspase-3 on PD 4 In the current neonatal

rat model, vitamin E did not provide any neuroprotection

against functional, structural, or pathophysiological indica-tors of EtOH-induced cerebellar damage.

Our findings with EtOH exposure on PD 4 to 9 stand in contrast to the previous report that vitamin E protected against the early neonatal EtOH-induced reductions in PC density in cerebellar vermal lobule I in 5-day-old rat pups after binge EtOH exposure on PD 4 to 5 (Heaton et al., 2000) Although the same total daily EtOH dose (5.25 g/kg) and the same dosing regimen of vitamin E were used in the two studies, several differences may be involved in the discrepant findings.

Vitamin E may protect against relatively limited episodic exposure (e.g., PD 4 –5) but not against multiple, repeated binges (PD 4 –9) Although the period of greatest vulnera-bility to EtOH for rat PCs is during PD 4 to 6 compared with PD 7 to 9 or later (Goodlett and Lundahl, 1996; Hamre and West, 1993), EtOH exposure that extends the entire period (PD 4 –9) does produce more severe loss of PCs than similar exposure limited to just PD 4 to 6 (Goodlett and Lundahl, 1996) It is possible that the cumu-lative effects of repeated daily binges may have over-whelmed the short-term protective advantage afforded by vitamin E supplementation.

In the Heaton et al (2000) study, the cerebella were collected for cell counts within 6 hr of the onset of the second daily binge treatment (on PD 5) In that case, vitamin E simply may have delayed the PC death rather than prevented it However, our caspase-3 active subunit data suggest that any potential delay in apoptotic cell death

by vitamin E is not likely to be associated with delayed EtOH-induced expression of the active form of this

“exe-Fig 3 Digital photomicrographs of a

mid-sagittal section from a juvenile rat cerebellum

(top left, ⫻40 total magnification) and of

rep-resentations ( ⫻400 total magnification) of the

PC layer in hemispheric lobule I (solid box) for

the VE-M group (top right), VE-E group

(bot-tom left), and E group (bot(bot-tom right) The

ar-rows indicate cerebellar PCs.

cutioner” protease in the PD 4 cerebellum Also, the

den-sity counts performed by Heaton et al (2000) were limited

to just a small portion of the cerebellar vermis (lobule I),

and vitamin E may be effective in protecting only certain

portions of the cerebellum (e.g., the anterior vermis), areas

that were not included in our more extensive

three-dimensional counts of total numbers of Purkinje neurons in

lobules I to VI of the left cerebellar hemisphere If so, then

any cerebellar neuroprotection imparted by supplements of

vitamin E is likely to be of only limited therapeutic

effec-tiveness against alcohol-induced cerebellar injury during

the third trimester.

Another possibility for the discrepancies between the two

studies is the difference in peak BECs reported Although

both studies used similar intubation procedures, EtOH

concentrations, and temporal sequences of vitamin E and

EtOH administration, Heaton et al (2000) reported mean

peak BECs that never surpassed 270 mg/dl, whereas our

mean peak BECs were ~100 mg/dl higher (~380 mg/dl).

We have consistently observed peak BECs in excess of 300

mg/dl with these and similar treatments (Goodlett et al.,

1997; Green et al., 2002a,b, 2000; Stanton and Goodlett,

1998) Peak BEC is correlated with the extent of

EtOH-induced teratogenesis (Goodlett et al., 1990; Pierce and

West, 1986a,b; West et al., 1990), and higher BECs achieved in this study may have resulted in the induction of

a higher level of free radicals that may not have been ameliorated with the dose of vitamin E administered It is not clear why the peak BECs in our hands differed from

Fig 4 Caspase-3 subunit expression in whole cerebellum 8 hr after initial

EtOH treatment Supplementation of EtOH with vitamin E did not protect against

EtOH-induced increases in caspase-3 active subunit expression (Top)

Repre-sentative Western blots showing expression of active subunit in the cerebella of

the four EtOH-treated pups (lanes E and VE-E) but not in the four control-treated

pups (lanes M and VE-M) The left lane ( ⫹C) contained caspase-3 active subunit

as a positive control for the Western blot analysis (Bottom) Quantitative

densi-tometric analysis of the active peptide subunit of caspase-3, determined as the

ratio of the density of the active subunit to the density of actin Data are expressed

as mean ratios⫾ SEM (n ⫽ 6 per group) *Significantly different from both control

groups (Tukey’s post hoc test, p⬍ 0.05).

Fig 5 Representative sections that were immunoprocessed for caspase-3 in

hemispheric lobule I of PD 4 rat pups Caspase-3 labeling of PCs using DAB (arrows) was observed in EtOH-treated pups (E and VE-E) but not in the control pups (only the M group is shown here).