Neuronal susceptibility in rat models of developmental ethanol exposure

MISSING IMAGE

Material Information

Title:
Neuronal susceptibility in rat models of developmental ethanol exposure descriptions of cellular and molecular alterations
Portion of title:
Descriptions of cellular and molecular alterations
Physical Description:
x, 209 leaves : ill. ; 29 cm.
Language:
English
Creator:
Moore, David Blaine, 1972-
Publication Date:

Subjects

Subjects / Keywords:
Research   ( mesh )
Ethanol -- toxicity   ( mesh )
Cerebellum -- drug effects   ( mesh )
Choline Acetyltransferase -- drug effects   ( mesh )
Parvalbumins -- drug effects   ( mesh )
Prenatal Exposure Delayed Effects   ( mesh )
Gene Expression   ( mesh )
Gene Expression Regulation   ( mesh )
Genes, bcl-2   ( mesh )
Rats   ( mesh )
Department of Neuroscience thesis Ph.D   ( mesh )
Dissertations, Academic -- College of Medicine -- Department of Neuroscience -- UF   ( mesh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1998.
Bibliography:
Bibliography: leaves 181-208.
Statement of Responsibility:
by D. Blaine Moore.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002406584
oclc - 52003135
notis - AMB1536
System ID:
AA00009002:00001

Full Text










NEURONAL SUSCEPTIBILITY IN RAT MODELS OF
DEVELOPMENTAL ETHANOL EXPOSURE: DESCRIPTIONS OF
CELLULAR AND MOLECULAR ALTERATIONS














By '

D. BLAINE MOORE


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1998






























For Terri and my Mentors, who nurtured my interest in Biology














ACKNOWLEDGMENTS


There are many people to whom I am indebted, especially my wife, Terri, who is

my best friend and best critic. I am also grateful to my mentors, for their continued

support. I am indebted to my thesis advisory committee members, Drs. Marieta Heaton

(chair), Don Walker, Tony Yachnis, Kevin Anderson and John MacLennan, for their

efforts, and to Drs. Doug Anderson and Gerry Shaw for helpful advice. I benefited from

the assistance of Dr. Jim West and Jo Mahoney (Texas A&M University) in learning the

artificial rearing procedure. Likewise, Dr. Bill Farmerie and Regina Shaw (of the

Interdisciplinary Center for Biotechnology Research, UF), as well as Terri Edwards and

Dr. John MacLennan, provided excellent advice on molecular biology. Thanks to Heaton

and Walker lab members over the years, including Mike Paiva, Jeffrey Thinschmidt, Pat

Burnett, Steve Famworth, Leon Williams and Drs. Doug Bradley, Jean Mitchell, Meg

Davis, Mike King, and Doug Swanson. I am thankful for the support of the NIH during

my graduate training, including NIAAA F31 AA0550201 and NIAAA T32 AA07561.














TABLE OF CONTENTS
page


ACKNOW LEDGM ENTS .................................................................................................... iii

LIST OF TABLES ............................................................................................................. vi

LIST OF FIGURES........................................................................................................ vii

ABSTRACT....................................................................................................................... ix

CHAPTERS

1 INTRODUCTION ..........................................................................................................

Ethanol Exposure During Development and Resulting Nervous System
Alterations......................................................................................................................
Differential Temporal and Regional Vulnerability ...........................................................4...
Prenatal Ethanol Exposure and Neuroanatom ical Alterations..........................................5...
Postnatal Ethanol Exposure and Neuroanatom ical Alterations ........................................8...
Program ed Cell Death (PCD) and the bcl-2 Fam ily .......................................................11
Hypotheses Tested .......................................................................................................... 21

2 EFFECTS OF PRENATAL ETHANOL EXPOSURE ON PARVALBUMIN-
IMMUNOREACTIVE GABAERGIC NEURONS IN THE ADULT RAT
MEDIAL SEPTUM AND ANTERIOR CINGULATE CORTEX..............................25

Summ ary ............................................................................................................................25
Introduction........................................................................................................................26
M materials and M ethods.................................................................................................... 30
Results................................................................................................................................37
Discussion ..........................................................................................................................43

3 EFFECTS OF NEONATAL ETHANOL EXPOSURE ON CHOLINERGIC
NEURON S OF THE RAT M EDIAL SEPTUM ....................................................... 70

Summ ary ............................................................................................................................70
Introduction........................................................................................................................71
M ethods..............................................................................................................................73










Results................................................................................................................................80
D iscussion..........................................................................................................................83

4 EFFECTS OF NEONATAL ETHANOL EXPOSURE ON PURKINJE AND
GRANULE CELLS AND BCL-2 FAMILY MRNA LEVELS IN THE RAT
CEREBELLA R V ERM IS......................................................................................... 97

Sum m ary ............................................................................................................................97
Introduction........................................................................................................................98
M materials and M ethods................................................................................................... 102
Results.............................................................................................................................. 110
D iscussion........................................................................................................................ 116

5 CONCLUSIONS AND FUTURE DIRECTIONS................................................... 153

Recapitulation of Results and H ypotheses Tested ........................................................1... 53
Choice of A nim al M odels.............................................................................................1... 59
Choice of Cell Counting M ethods ................................................................................1... 63
O their M ethodological Considerations ..........................................................................1... 64
Future Directions for Developmental Ethanol Research ..............................................1... 65
Conclusion ....................................................................................................................... 179

REFEREN CES ........................................................................................................... 181

BIO G RA PH ICA L SK ETCH ......................................................................................... 209














LIST OF TABLES


Table page

2-1. Postnatal day 60 body and brain weight and brain to body weight ratio ................56

2-2. Number of sections, medial septum area per section, and parvalbumin-
immunoreactive neuronal density in the adult rat MS ............................................56

2-3. Number of sections, mean area per section, and parvalbumin-immunoreactive
neuronal density in the adult rat anterior cingulate cortex......................................57

3-1. Mean body weight, brain weight, and brain weight to body weight ratio of postnatal-
day 60 anim als ........................................................................................................ 92

3-2. Number of sections, medial septum area per section, and choline acetyltransferase-
immunoreactive neuronal density in the adult rat MS ............................................92

4-1. Lobule I cerebellar vermis volume, area, and cell nuclei diameter data at P21
following exposure on P4 only .............................................................................. 133

4-2. Lobule I cerebellar vermis volume, area, and cell nuclei diameter data at P21
follow ing exposure on P4-5 ................................................................................... 133

4-3. Lobule I cerebellar vermis volume, area, and cell nuclei diameter data at P21
follow ing exposure on P7-8 ................................................................................... 133

4-4. Lobule I cerebellar vermis volume, area, and cell nuclei diameter data at P5
follow ing exposure on P4-5 ................................................................................... 134














LIST OF FIGURES


Figure page

2-1. The mean total number of parvalbumin-expressing neurons detected on alternate
sections through the medial septum of postnatal-day 60........................................58

2-2. Photomicrographs of 40 gpm coronal sections through the medial septum of ethanol-
treated (A) and sucrose-treated (B) postnatal-day 60 female rats...........................59

2-3. The mean number of parvalbumin-expressing neurons per section in alternate
sections through the medial septum of postnatal-day 60........................................61

2-4. The mean diameter of parvalbumin-expressing neurons on alternate sections through
the medial septum of postnatal-day 60 rats............................................................. 62

2-5. The mean medial septum (MS) volume of postnatal-day 60 rats exposed throughout
gestation to an ethanol-containing liquid diet, sucrose-containing liquid diet, or lab
chow and w ater. ...................................................................................................... 63

2-6. The mean total number of parvalbumin-expressing (PA+) neurons detected on
alternate sections through the anterior cingulate cortex of postnatal-day 60.............64

2-7. Photomicrographs of 40 pum coronal sections through the anterior cingulate cortex of
ethanol-treated (A) and sucrose-treated (B) postnatal-day 60 rats..........................65

2-8. The mean number of parvalbumin-immunoreactive neurons per section counted in
the anterior cingulate cortex.................................................................................... 67

2-9. The mean diameter of parvalbumin-immunoreactive neurons in the anterior
cingulate cortex of postnatal-day 60 rats. ................................................................ 68

2-10. The mean cingulate gyrus volume of postnatal-day 60 rats exposed throughout
gestation ....................................................................................................................69

3-1. Mean total number of choline acetyltransferase-immunoreactive neurons in the
medial septum of postnatal day 60 rats................................................................... 93









3-2. Mean number of choline acetyltransferase-immunoreactive neurons per section
detected on alternate 40 p.m sections through the medial septum of postnatal day 60
rats..............................................................................................................................9 4

3-3. Mean somatic cross sectional area (pm2) of choline acetyltransferase-
immunoreactive neurons in the medial septum of postnatal day 60 rats................... 95

3-4. The mean medial septum volume of postnatal-day 60 rats.................................... 96

4-1. Effects of ethanol delivered on P4 only on P21 body weight (panel A), P21 brain
weight (panel B) and P21 brain to body weight ratio (panel C). ............................135

4-2. Effects of ethanol delivered on P4-5 on P4, P5, and P21 body weight (panel A),
brain weight (panel B) and brain to body weight ratio (panel C). ..........................137

4-3. Effects of ethanol delivered on P7-8 on P7, P8, and P21 body weight (panel A),
brain weight (panel B) and brain to body weight ratio (panel C). ..........................139

4-4. Ethanol delivered during the first postnatal week reduces mean Purkinje and granule
cell number per section in lobule I of the cerebellar vermis ................................. 141

4-5. Ethanol delivered during the second postnatal week does not reduce mean Purkinje
or granule cell number per section in lobule I of the cerebellar vermis.................143

4-6. Ethanol delivered during the first postnatal week on days 4-5 reduces mean Purkinje
but not granule cell number per section in lobule I of the cerebellar vermis as
determ ined on postnatal day 5 (P5)........................................................................ 144

4-7. mRNAs encoding pro-apoptotic molecules of the bcl-2 family are upregulated
following acute ethanol delivered on postnatal day 4 ........................................... 145

4-8. A further ethanol exposure on P5 does not significantly alter the expression of bcl-2
fam ily m R N A s.. ................................................... .................................................. 147

4-9. Effects of acute ethanol delivered on postnatal day 7 on bcl-2 family gene
expression. ..........................................................................................................149

4-10. An additional ethanol exposure on postnatal day 8 increases mRNAs encoding the
pro-apoptotic m olecule bax.................................................................................... 151














Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

NEURONAL SUSCEPTIBILITY IN RAT MODELS OF
DEVELOPMENTAL ETHANOL EXPOSURE: DESCRIPTIONS OF
CELLULAR AND MOLECULAR ALTERATIONS

By

D. Blaine Moore

December 1998

Chairman: Marieta B. Heaton, Ph.D.
Major Department: Neuroscience

Developmental disorders arising from maternal consumption of ethanol during

pregnancy are collectively termed the fetal alcohol syndrome. Ethanol exposure during

development induces abnormalities in particular brain regions, and is known to alter the

expression of particular genes and their protein products. The present body of work

sought to further document neuronal populations in the brain which display vulnerability

to developmental ethanol exposure. A further goal of this work was to investigate cell

death gene expression shortly after ethanol insult in the cerebellum in order to test a

specific hypothesis about the cellular mechanism of ethanol neurotoxicity.

A prenatal exposure model was used to examine long term changes in protein

expression patterns of parvalbumin (a marker for gamma-aminobutyric acid- (GABA)

expressing neurons) in the rat brain. Deficiencies in the mean number of immuno-

positive cells per section were noted in the medial septum (in a sexually dimorphic








manner) and anterior cingulate cortex of ethanol-treated rats. This represents the first

documentation of GABAergic neuronal susceptibility to ethanol in either brain region. A

neonatal exposure model was used to examine long-term changes in the expression

pattern of choline acetyl-transferase (ChAT, a marker for cholinergic neurons) in the

medial septum; no significant ethanol-induced changes in the mean number of immuno-

positive cells per section were noted.

A similar neonatal exposure paradigm was used to document Purkinje and granule

cell numbers in the cerebellar vermis during known periods of ethanol sensitivity and

insensitivity, and to investigate mRNA levels of the bcl-2 family of cell death molecules.

First postnatal week ethanol treatment significantly reduced Purkinje and granule cell

number, while second week exposure did not. bcl-2 family gene expression was

measured in the vermis shortly after ethanol treatment to determine whether alterations in

these genes might correlate with the noted cell death. Transcripts encoding the pro-

apoptotic molecules bax and bcl-xs were up-regulated following both first and second

week exposure. Thus, a positive correlation between altered bcl-2 expression and

cerebellar cell death was not found. Suppression of these pro-apoptotic processes may be

the critical determinant of cerebellar susceptibility. These findings suggest new avenues

of research on the intracellular consequences of such expression changes.















CHAPTER 1
INTRODUCTION


Ethanol Exposure During Development and Resulting Nervous System Alterations


Ethanol's teratogenic actions have been recognized throughout recorded history

and considerable research has defined its deleterious actions (West et al., 1994). With the

recognition of the fetal alcohol syndrome (FAS), and the linkage of ethanol to

malformations in children of alcoholic mothers, Jones and Smith (1973) spurred a

plethora of studies on alcohol-induced fetal abnormalities (West et al., 1994).

Consequently, prenatal ethanol exposure has been shown to result in serious

developmental alterations, including intrauterine growth deficiencies, facial dysmorphias,

mental retardation (Abel, 1984), attention deficiencies and autistic-like syndromes

(Aronson et al., 1997), and lowered IQ (Mattson et al., 1997). Ethanol readily crosses the

placental and blood-brain barriers, diffuses into all aqueous components of the

developing fetus where it can interact with membrane proteins and lipids (Zajac and

Abel, 1992), and has been shown to affect protein synthesis, placental nutrient transport,

fetal glucose availability, fetal oxygen levels, generation of reactive oxygen radicals, and

neurotrophic factor activity (Abel and Hannigan, 1995; Bonthius and West, 1990; Heaton

and Bradley, 1995; Henderson et al., 1995; Mukherjee and Hodgen, 1982; West et al.,

1994; Zajac and Abel, 1992).








Zajac and Abel (1992) have characterized fetal alcohol exposure as "the leading

known cause of mental retardation in the Western world." Indeed, despite this

recognition, the incidence of FAS in the United States has increased six-fold between

1979 and 1993 (Prevention, 1995) to 1.95/1000 live births, and 43.1/1000 live births

among heavy drinkers (Abel and Hannigan, 1995). Sampson et al. (1997) recently

estimated the incidence of FAS and alcohol-related neurodevelopmental disorders

between 1975-1981 in Seattle as 9.1/1000 live births. As such, FAS remains a significant

health problem in the United States. Further research describing the neurodevelopmental

changes induced by ethanol and the cellular and molecular changes induced by ethanol is

needed, especially since the mechanism of ethanol teratogenicity remains unknown.

The toll that ethanol exacts on the development of the central nervous system

(CNS) represents its most potent danger. Children exposed prenatally to ethanol show,

among other abnormalities, reduction in corpus callosum area (Riley et al., 1995) and

malformations in cerebellar structure (Clarren et al., 1978; Sowell et al., 1996;

Wisniewski et al., 1983). Animal models have been developed which recapitulate many

of the neurodevelopmental alterations and behavioral outcomes seen in humans with

FAS, and these models have allowed for the characterization of various nervous system

changes resulting from developmental exposure to ethanol (Hannigan, 1996). While

various model systems have been employed, including the chick (Bradley et al., 1997)

and mouse (Schambra et al., 1990) the most common choice for investigators is the rat.

Rats exposed during embryonic development show long-lived learning

impairment (Clausing et al., 1995), behavioral alterations (Riley, 1990), microencephaly

(West and Pierce, 1986), changes in neuronal proliferation and migration (Miller, 1986;








Miller, 1995b; Miller, 1996), reductions of neuronal number (Barnes and Walker, 1981;

Miller, 1995a; Napper and West, 1995b), alterations in neuronal circuitry (West et al.,

1981), delays in synaptogenesis (Hoff, 1988), permutations in neuromorphological

development (Burrows et al., 1995; Davies and Smith, 1981; Kotkoskie and Norton,

1989), changes in neurochemistry (Black et al., 1995; Swanson et al., 1995), and

alterations in the levels of specific mRNA species (Lee et al., 1997) and specific receptor

molecules such as the high affinity trkA nerve growth factor receptor and the low affinity

neurotrophin receptor p75 (Dohrman et al., 1997).

The present body of work sought to further document neuronal populations in the

brain which display vulnerability to developmental ethanol exposure. Another goal of

this work was to investigate cell death gene expression shortly after ethanol insult in the

cerebellum in order to test a specific hypothesis about the cellular mechanism of ethanol

neurotoxicity. To accomplish these goals the following studies were performed: first, a

prenatal rat exposure model was used to examine long term changes in protein expression

patterns of parvalbumin (a marker for GABAergic neurons) in the rat medial septum and

cingulate cortex by counting parvalbumin-immunoreactive neurons in these structures;

second, a neonatal rat exposure model was used to examine long-term changes in the

expression pattern of choline acetyltransferase (ChAT, a marker for cholinergic neurons)

in the medial septum by counting ChAT-immunoreactive neurons in this region; and

third, a similar neonatal rat exposure paradigm was used to document Purkinje and

granule cell numbers in the cerebellar vermis during known periods of ethanol sensitivity

and insensitivity, and to investigate mRNA levels of the bcl-2 family of cell death

molecules.








Differential Temporal and Regional Vulnerability


An interesting property of ethanol and its effect on the developing CNS is the fact

that particular brain regions are differentially affected in both human FAS and rodent

models of FAS (West and Pierce, 1986). Fortunately, animal models of FAS provide for

rigorously controlled studies of both regional and temporal susceptibilities, and allow for

the identification of the neuroanatomical substrates that underlie FAS. Although the

brain develops throughout the entire prenatal and early postnatal period in humans and

rats, there are both regional and temporal susceptibilities to ethanol. In terms of regional

vulnerability, the cerebral cortex (Miller, 1986), and hippocampus (Barnes and Walker,

1981) have been demonstrated to be severely affected with chronic prenatal ethanol

exposure. In terms of temporal vulnerability, CNS development during the human third-

trimester equivalent is especially tenuous. During this period rapid, global brain

development is occurring and this dynamic phase, which occurs during the human third

trimester, is often termed the brain growth spurt. Although all mammals develop in a

similar manner, the timing of this intensified growth is different across species. The rat,

for example, from which most animal data regarding FAS are derived, undergoes its brain

growth spurt during the first two postnatal weeks, with a peak at postnatal day 4-10 (P4-

10) (Dobbing and Sands, 1979). Ethanol exposure during this vigorous period of brain

development affects both mature and proliferating neurons (West and Pierce, 1986). The

principal neurons of the cerebellar cortex provide a good example, as neonatal exposure

reduces the number of differentiating Purkinje cells as well as proliferating granule cells

(Bonthius and West, 1990), albeit in a temporally restricted manner (see below).








Prenatal Ethanol Exposure and Neuroanatomical Alterations


Experimental studies, utilizing rodent models of FAS, have been successful in

reproducing many of the behavioral and morphological changes found in FAS, and have

described numerous CNS alterations resulting from in utero ethanol exposure. Some of

these include alterations in a variety of developmental processes in the CNS, including

perturbed neuronal generation and migration (Miller, 1986; Miller, 1995b; Miller, 1996),

permutations in neuromorphological development (Burrows et al., 1995; Davies and

Smith, 1981; Kotkoskie and Norton, 1989), changes in neurochemistry (Black et al.,

1995; Swanson et al., 1995), delayed synapse turnover in the hippocampus (Hoff, 1988),

alterations in neuronal number in hippocampus (Barnes and Walker, 1981), and long-

lasting detrimental effects on learning and behavior (Clausing et al., 1995; Riley, 1990).

Septohippocampal system

The septohippocampal (SH) system of basal forebrain afferents and hippocampal

targets has been shown to be sensitive to prenatal ethanol exposure. Reductions in CAl

pyramidal neuronal number (Barnes and Walker, 1981; Wigal and Amsel, 1990),

alterations in hippocampal mossy fiber organization (West et al., 1981) and dendritic

arborization (Davies and Smith, 1981; Smith and Davies, 1990), delays in synaptogenesis

(Hoff, 1988), changes in neurochemistry (Black et al., 1995; Swanson et al., 1995), and

deficits in hippocampal synaptic plasticity (Sutherland et al., 1997) have all been

demonstrated in the rodent SH system following exposure to ethanol in utero. Given the

importance of the hippocampus in learning and memory and the role of the SH system in

generating and maintaining electrical activity in the hippocampus (Dutar et al., 1995), the








possibility exists that ethanol-induced changes in this system can have detrimental effects

on offspring exposed to ethanol prenatally.

The SH system is a pathway of cholinergic and GABAergic fibers originating

from the medial septal (MS) nucleus and the horizontal and vertical limbs of the nucleus

of the diagonal band of Broca (DBB) which synapse on pyramidal neurons, granule cells

and interneurons of the hippocampus (Dutar et al., 1995; Freund and Antal, 1988).

Evidence exists for an ethanol-induced alteration of the cholinergic component of the SH

pathway. Severe abnormalities were noted in the basal forebrain of fetal mice following

an acute ethanol dose at gestational day 7 (G7) (Sulik et al., 1984), and Schambra et al.

(1990) found a reduction in the number of ChAT+ neurons in fetal mice following an

acute ethanol administration on G7. Arendt et al. (1988) reported similar findings in

adult rats following chronic ethanol treatment, although their experimental design suffers

from lack of pair-fed controls.

Studies from our laboratory utilizing a rat model of chronic prenatal ethanol

treatment have shown an ethanol-induced delay in the normal ontogeny of ChAT enzyme

activity in the SH pathway, but have not revealed an effect on ChAT-immunoreactive

neuronal number (Swanson et al., 1995; Swanson et al., 1996). It was a goal of the

present work to determine if the other major population of neurons in the medial septum,

the GABAergic neuronal population, is affected by developmental ethanol exposure.

Thus, studies were performed to test the hypothesis that GABAergic neurons in the SH

pathway are susceptible to chronic prenatal ethanol treatment, in order to determine

whether this subset of neurons in the medial septum is affected by ethanol. To achieve

this aim, an antibody which recognizes parvalbumin (PA), a calcium binding protein








commonly found in GABAergic neurons, was used and PA-immunoreactive neurons

were counted.

PA and the SH system

Although the GABAergic component of the SH projection has not been as

extensively characterized as the cholinergic component, it is an important part of this

system. Approximately 33% of the neurons in the MS/DBB region display

immunoreactivity to PA, an 11.8 kDa member of a Ca2+ binding superfamily of proteins

(Kiss et al., 1990a; Kiss et al., 1990b; McPhalen et al., 1994) commonly found in rapidly-

firing GABAergic neurons where it influences the activity of Ca2+-dependent K+ channels

(Plogmann and Celio, 1993). In rats, PA ontogeny begins in the MS at G21 and coincides

with the beginning of physiological activity such as spontaneous firing and excitatory

synaptic input (Lauder et al., 1986; Solbach and Celio, 1991). The PA-expressing MS

neurons innervate inhibitory interneurons in the hippocampus (Freund and Antal, 1988).

Although CNS regions differ in the extent to which PA and GABA co-localize (Alonso et

al., 1990; Brauer et al., 1991; Kiss et al., 1990a), within the MS nucleus most, if not all,

of the hippocampal-projecting GABAergic neurons are parvalbumin-immunoreactive

(Freund, 1989; Krzywkowski et al., 1995). Thus, PA-immunoreactivity serves as a

reliable marker for hippocampal-projecting GABAergic neurons in the MS nucleus and

identifies a subpopulation of the total GABAergic neuronal pool within the basal

forebrain.

PA and cingulate cortex

Quantitative analyses of neuroanatomical changes in the cerebral cortex following

prenatal ethanol exposure have been conducted and have suggested ethanol-induced








neuroanatomical alterations. Specifically, alterations in the generation and proliferation

of neurons have been noted (Miller, 1986; Miller, 1996). Another goal of the present

work was to determine whether prenatal ethanol exposure alters the neuroanatomy of

limbic cortex by testing the hypothesis that the number of GABAergic intemeurons

expressing PA are altered in the cingulate cortex following prenatal ethanol exposure.

Although little is known about the cingulate cortex, it is clear that the cingulate is a relay

center of the limbic lobe and is important for emotion and memory (Kupfermann, 1991).

PA expression in the cingulate cortex begins during the first postnatal week in rats and

coincides with the functional maturation of cerebral intemeurons (de-Lecea et al., 1995).

The cingulate cortex was chosen to extend the earlier observations in the cerebral

cortex following prenatal ethanol exposure because of the observed behavioral problems

in children with FAS, including poor judgment, distractibility, and difficulty perceiving

social cues (Streissguth et al., 1991). Additionally, many alcoholics who develop

Korsakoff's syndrome have deficiencies in glucose utilization within the cingulate cortex,

potentially contributing to learning and memory defects due to interruption of Papez'

circuitry (Joyce et al., 1994). It is conceivable that alterations in PA expression patterns

might contribute to behavioral anomalies and/or learning and memory deficiencies.

Postnatal Ethanol Exposure and Neuroanatomical Alterations


SH system

The SH system of basal forebrain afferents and hippocampal targets also exhibits

susceptibility to ethanol during neonatal development. While numerous studies have

described the effects of neonatal ethanol exposure on neurons within the hippocampus








(Bonthius and West, 1990; Bonthius and West, 1991; Greene et al., 1992; Pierce and

West, 1987; West and Pierce, 1986) and have documented deficits in spatial learning

following neonatal ethanol exposure (Goodlett and Peterson, 1995; Kelly et al., 1988), the

effect of ethanol exposure during the brain growth spurt on the cholinergic neurons of the

medial septum is unknown. A previous study from our laboratory documented an

ethanol-induced delay in the normal ontogeny of ChAT enzyme activity in the SH

pathway, but did not reveal an effect on ChAT-immunoreactive neuronal number

following chronic prenatal ethanol treatment (Swanson et al., 1995; Swanson et al.,

1996). It was a goal of the present work to determine the long-term effects of ethanol

exposure on the number of ChAT+ neurons in the rat MS when ethanol is delivered

during the brain growth spurt to determine whether this neuronal population is sensitive

to neonatal ethanol exposure.

Cerebellum

Evidence of cerebellar vulnerability to developmental ethanol exposure comes

from human studies demonstrating size reduction in the cerebellar vermis of children

exposed prenatally to ethanol (Sowell et al., 1996). In rodent studies, the cerebellum

displays a pattern of differential temporal susceptibility to ethanol in the brain growth

spurt. This effect has been demonstrated in multiple laboratories utilizing a variety of

ethanol-exposure techniques. Purkinje cell number is reduced following exposure to

ethanol postnatally, during differentiation, but not following exposure to ethanol

prenatally during neurogenesis (Marcussen et al., 1994). Within the postnatal period,

Purkinje cells have been shown to be particularly vulnerable to ethanol in the first

postnatal week (Bauer-Moffett and Altman, 1977; Bonthius and West, 1991; Goodlett








and Eilers, 1997; Pauli et al., 1995; Pierce et al., 1993). Purkinje cells are generated in

the rat cerebellum between embryonic day 14 and 17; the period of Purkinje cell death in

the cerebellum begins late in gestation and peaks in the first postnatal week (Cragg and

Phillips, 1985). Purkinje cell susceptibility to ethanol within the cell death period has

been well documented, with ethanol-accelerated Purkinje cell loss found as early as 12

hours following a postnatal day 3 ethanol insult (Cragg and Phillips, 1985).

Exposure to comparable levels of ethanol in the second postnatal week, however,

has been shown to have no effect on Purkinje cells (Goodlett et al., 1997; Hamre and

West, 1993; Pauli et al., 1995) or very little effect (Thomas et al., 1998). In contrast to

Purkinje cells, granule cells are generated during the rat brain growth spurt (Altman,

1969). However, like Purkinje cells, granule cells display a differential pattern of

susceptibility to ethanol, with loss occurring following ethanol exposure in the first

postnatal week but not in the second postnatal week (Hamre and West, 1993). One aim

of the present work was to determine the pattern of cerebellar neuronal susceptibility in

our laboratory following first and second postnatal ethanol treatment, in order to validate

the model system for use in our laboratory. The main objective, however, was to

determine whether neonatal ethanol treatment altered mRNA levels of members of the

bcl-2 family of cell death regulators in the cerebellar vermis (and thereby influenced the

survival or death of neurons in this region). Experiments were performed in order to test

the specific hypothesis that altered bcl-2 family gene expression ensues following ethanol

exposure.

A role in maintaining cerebellar neurons has previously been demonstrated for the

bcl-2 family. Gillardon et al. (1995) investigated bcl-2 and bax gene expression in the








cerebella of Purkinje-cell-degeneration mice (mutants that lose nearly all of their Purkinje

cells between P22-28 following otherwise normal development). They found that bcl-2

mRNA levels decreased while bax mRNA levels remained unchanged beginning on P22.

In addition, thyroid hormone-induced upregulation of bcl-2 protects early-differentiating

cerebellar granule cells from apoptosis in vitro (Muller et al., 1995), and transgenic mice

overexpressing bcl-2 contain more cerebellar Purkinje and granule cells than controls

(Zanjani et al., 1996; Zanjani et al., 1997). The following section provides background

on the bcl-2 family and its involvement in cell death regulation.

Programmed Cell Death (PCD) and the bcl-2 Family


PCD, a developmental form of apoptotic cell death, is a common process in the

animal kingdom (Ellis et al., 1991). In the vertebrate nervous system the regulation of

neuronal survival is essential for the correct formation of synapses and for the survival of

the appropriate number of neurons (Oppenheim, 1991). Only the most appropriate

connections are maintained, making axon-target interactions maximally efficient and

simultaneously ensuring that cells which are generated in excess, develop poorly, are

functionally inadequate, or are harmful, do not endure in adult organisms (Ellis et al.,

1991). Competition for a limited supply of target-derived neurotrophic factors is thought

to determine which neurons survive the period of naturally occurring PCD (Davies,

1994). The mechanism by which non-essential neurons are eliminated is consistent with

the apoptotic form of cell death (Johnson and Deckwerth, 1993), with morphological

characteristics consisting of chromatin condensation, cell shrinkage, cleavage of DNA

into oligonucleosomal fragments (Edwards et al., 1991), and phagocytosis of dead cells








(Ellis et al., 1991) without induction of the inflammatory response (Columbano, 1995).

Most mammalian cells constituitivly express the proteins essential for the cell death

program (Davies, 1995; Raffet al., 1993).

Cytotoxicity due to ethanol in vitro has been shown to be apoptotic in fetal

hypothalamic neurons (De et al., 1994) and in thymocytes (Ewald and Shao, 1993).

Additionally, ethanol-induced cell death in the cerebellum appears to proceed through an

apoptotic mechanism. Cerebellar neurons undergo apoptosis in vitro and in vivo in

response to ethanol (Bhave and Hoffman, 1997; Liesi, 1997; Renis et al., 1996; Singh et

al., 1995) and ethanol induces nuclear DNA strand breaks in the cerebellum after chronic

adult exposure (Renis et al., 1996).

Almost a decade ago, a gene, bcl-2, was discovered that appeared to modulate

apoptosis. The name bcl-2 is an acronym for B-cell lymphoma/leukemia-2 gene (Reed,

1994) and the identification of this gene family resulted from studies examining the

t(14;18) chromosomal translocation in human follicular non-Hodgkin's B-cell

lymphomas (Tsujimoto et al., 1985). The protein product encoded by the bcl-2 gene, Bcl-

2, is a 25 kDa protein found predominately in mitochondrial membranes (Hockenbery et

al., 1990) but it is also found in endoplasmic reticula and outer nuclear membranes (Akao

et al., 1994). In recent years, a number of new genes similar to bcl-2 have been

characterized and added to the diverse bcl-2 family of genes. These include bcl-xl (Boise

et al., 1993), bcl-xs (Boise et al., 1993), bax (Oltvai et al., 1993), bad (Yang et al., 1995)

al (Lin et al., 1993), mcl-1 (Kozopas et al., 1993), bak (Chittenden et al., 1995), bcl-w

(Gibson et al., 1996), brag-1 (Das et al., 1996), bok (Hsu et al., 1997), and bim (O'connor

et al., 1998). Some members of the bcl-2 gene family serve to inhibit cell death (e.g. bcl-








2, bcl-xl, mcl-1, al) and others have been found to promote cell death (e.g. bcl-xs, bax,

bad, bak, bok).

Bcl-2 (and the similar anti-apoptotic protein Bcl-xl) functionally blocks apoptotic

death in neurons (Allsopp et al., 1993; Garcia et al., 1992) by inhibiting caspase

activation (Shimizu et al., 1996), regulating mitochondrial membrane potential, proton

flux across mitochondrial membranes (Shimizu et al., 1998), and by preserving

mitochondrial outer membrane integrity (Vanderheiden et al., 1997). The pro-apoptotic

molecules of the bcl-2 family (such as the proteins Bad and Bak) function by inhibiting

the ability of the anti-apoptotic molecules of the bcl-2 family to function. They do so by

preventing the necessary homodimerization of the protective molecules through direct

competition for binding to the ligand binding regions of these proteins (Ottilie et al.,

1997). The intracellular mechanisms of bcl-2 family function are described in detail

below.

Expression patterns

Given the identification of these molecules and the fact that apoptosis is known to

occur in the developing nervous system, the possibility that bcl-2 family members and

their protein products might modulate PCD in the nervous system has been explored.

Castren et al. (1994) have shown that bcl-2 mRNA is expressed in high levels in the

prenatal rat neuroepithelium and cortical plate, with a late-prenatal peak and expression

reaching lower adult levels during postnatal development. In situ hybridization also

revealed that bcl-2 expression is retained in the olfactory bulb, hippocampus, pons,

cerebellum, and ependymal cells of the adult rat brain (Castren et al., 1994). Bcl-2

protein is also widely expressed in the developing nervous system of mice, rhesus








monkeys, and humans, especially during embryonic development, but also during the

period of PCD (Merry et al., 1994). Neuroepithelial cells of the ventricular zone,

postmitotic cells of the cortical plate, Purkinje and granule cells of the cerebellum,

hippocampus, and spinal cord all express Bcl-2 (Gleichmann et al., 1998; Merry et al.,

1994).

Expression of bcl-xl follows a similar pattern, with mRNA levels increasing at the

beginning of the PCD period. In the brain, in contrast to Bcl-2, Bcl-xl expression

increases after birth in neurons of the cortex and olfactory bulb, as well as in Purkinje

cells, to reach a high level in the adult brain, suggesting a role for Bcl-xl in the adult CNS

(Frankowski et al., 1995; Sohma et al., 1996). Although previous reports have suggested

that bcl-xs is not expressed in the rat CNS, newer studies using more sensitive techniques

have detected bcl-xs mRNA in the adult rat brain (Dixon et al., 1997; Rouayrenc et al.,

1995). Although the protein product of bcl-xs is able to weakly bind Bcl-2 and Bcl-xl, it

appears that Bcl-xs affects apoptosis by a distinct mechanism that, unlike other family

members, does not involve direct protein interactions with cell death repressor proteins

(Minn et al., 1996). Both long and short forms of bcl-x are also found in cerebellar

granule cells during development (Gleichmann et al., 1998).

bax is a gene in the bcl-2 family whose protein product, Bax, is a 21 kDa protein

with amino acid homology (21%) with Bcl-2 (Oltvai et al., 1993). Bax forms

heterodimers with Bcl-2, Bcl-xl, Mcl-1, and Al (Sedlak et al., 1995), and Bax can

counteract the protective actions of Bcl-2 when overexpressed (Oltvai et al., 1993). It has

been suggested that the ratio of Bcl-2 to Bax may determine survival or death from

apoptosis by serving as a sort of rheostat: when Bcl-2 is in excess, cells survive, but








when Bax is in excess, cells die (Oltvai et al., 1993). bax mRNA has been found in the

developing and adult rat brain. Similarly, in the adult mouse, the Purkinje and granule

cells of the cerebellum, cerebral cortical neurons, and sympathetic neurons all express

Bax protein (Gleichmann et al., 1998; Krajewski et al., 1994; Oltvai et al., 1993;

Vekrellis et al., 1997). While Bax protein expression is high in the neonatal cerebral

cortex and cerebellum, protein levels drop off dramatically after the PCD period,

suggesting that neurons regulate their sensitivity to apoptosis during development by

regulating expression of Bax (Vekrellis et al., 1997).

Consistent with the rheostat hypothesis, the ratio of Bcl-2:Bax correlates with cell

survival in the mature rat hippocampus following global ischemia. bax mRNA and

protein are both constitutively expressed in the ischemia-sensitive CAl hippocampal

neurons, whereas Bcl-2 is not expressed in these neurons (Chen et al., 1996). In

hippocampal region CA3, a population more resistant to ischemia than cells in region

CAl, Bcl-2 protein expression, but not that of Bax, is high (Chen et al., 1996).

Furthermore, high levels of Bax and concomitant low levels of Bcl-2 have been found in

other populations of neurons that are sensitive to cell death induced by ischemia, such as

Purkinje cells (Krajewski et al., 1995). Additional evidence supporting the bcl-2 rheostat

hypothesis comes from studies in mouse brain that have demonstrated up-regulation of

Bax and down-regulation of Bcl-2 associated with kainate-induced apoptosis (Gillardon

et al., 1995), as well as studies from rat that have shown a decrease in the Bcl-2:Bax ratio

in motoneurons following sciatic nerve transaction (Gillardon et al., 1996) and ischemia

(Isenmann et al., 1998). Additional evidence from Dixon et al. (1997) shows that the pro-

apoptotic bcl-xs mRNA is upregulated shortly after global ischemia in rats. Thus, it is








clear that the balance between pro- and anti-apoptotic molecules (at the protein and

mRNA levels) during development and following injury in adulthood determines the

survival of the cell.

Genetically engineered mice

Studies using genetically engineered mice have been fruitful in describing a role

for various members of the bcl-2 gene family in vivo. In transgenic mice overexpressing

Bcl-2 protein in the nervous system, a reduction in developmental cell death of facial

motoneurons and retinal ganglion cells, and a general hypertrophy of the nervous system

are observed (Martinou et al., 1994). Neurons from these animals are more resistant to

ischemia, neurotrophic factor withdrawal, and axotomy (Dubois-Dauphin et al., 1994;

Farlie et al., 1995). Similarly, facial motoneurons from mice overexpressing bcl-xl are

resistant to axotomy during the postnatal period, indicating a role for the bcl-xl gene in

the survival of postnatal CNS neurons (Parsadanian et al., 1998). Studies using

genetically engineered mice lacking the bcl-2 gene have been similarly instrumental in

defining the normal actions of Bcl-2 in vivo. bcl-2 -/- mice live through gestation,

display massive apoptosis in lymphoid organs, and produce gray hair follicles (Veis et al.,

1993). The brains of these animals, however, appear grossly normal at birth although fine

analyses of specific neuronal populations have not been performed. The lack of massive

cell death in neonatal bcl-2 knockouts has been attributed to redundancy, as the bcl-2

family is so large (Motoyama et al., 1995). However, analysis of neuronal populations

after the PCD period reveal loss relative to controls of motoneurons, sympathetic, and

sensory neurons, demonstrating a role for bcl-2 in maintaining these neuronal populations

(Michaelidis et al., 1996).








Mice lacking the bcl-x gene have also been generated, but die around E13. Upon

examination, extensive apoptotic cell death is evident in neurons of the brain, spinal cord,

and dorsal root ganglion (Motoyama et al., 1995), revealing the importance of bcl-x in

embryonic life. bax-deficient mice have been generated and have helped define the

actions of cell-death promoting molecules in particular neuronal populations (Deckwerth

et al., 1996). Early postnatal sympathetic and facial motoneurons from bax knockouts

survive growth factor deprivation and axotomy. Additionally, superior cervical ganglia

and facial nuclei of bax knockouts possess more surviving neurons in vivo. Thus, a role

for bax has been demonstrated in cell death associated with growth factor deprivation and

axotomy (Deckwerth et al., 1996). Shindler et al. (1997) have generated mice deficient in

both bcl-xl and bax. While bax deficiency does not prevent the embryonic lethality of

bcl-xl deficient mice, the double knockout did demonstrate the interplay between these

pro- and anti-apoptotic molecules in vivo. Specifically, bax and bcl-xl deletion produced

less apoptosis in the brainstem and spinal cord when compared with bcl-xl knockouts.

It is important to note that while studies using genetically engineered mice are

interesting and often quite suggestive, they suffer from several caveats. For instance,

overexpression of a particular gene or the creation of a mutant lacking a particular gene

may cause compensatory up- or down- regulation of other gene products which can lead

to developmental alterations in attempts to compensate for the altered gene (Gerlai,

1996). Moreover, mutant mice may differ in non-targeted gene loci and certain

behavioral alterations may be due to differences in genetic background (Gerlai, 1996).

Thus, while suggestive of the role of particular genes, studies utilizing genetically

engineered mice must be analyzed in light of these caveats.








Intracellular mechanisms

Recent evidence has shed light on the intracellular mechanism by which bcl-2

family members operate and has revealed a "double identity" for the family as both ion

channel and adapter protein (Reed, 1997). Bcl-xl, Bcl-2, and Bax have all been shown to

form functional ion channels in lipid membranes (Antonsson et al., 1997; Minn et al.,

1997; Schendel et al., 1997). Also, Bax channel-forming activity is inhibited by Bcl-2

(Antonsson et al., 1997). These ion channels appear to influence cell survival by

regulating the permeability of the intracellular membranes in which they are anchored. In

particular, the Bcl-2 and Bcl-xl membrane complexes inhibit the release of cytochrome C

from mitochondria (Yang et al., 1997), regulate mitochondrial membrane potential and

proton flux across mitochondrial membranes (Shimizu et al., 1998), and help preserve

mitochondrial outer membrane integrity (Vanderheiden et al., 1997). Bax channels, in

contrast, are known to promote the release of cytochrome C from mitochondria (Rosse et

al., 1998).

Cytochrome C is a known activator of the PCD effector molecules, the caspases

(Yang et al., 1997). Permeability transition across mitochondrial membranes appears to

be an early event in PCD, and is related to the activation of the release of cytochrome C

and the resultant activation of the effector phase of PCD (Petit et al., 1996). This

activation is brought about by cytochrome C-dependent cleavage and activation of

caspases 3 (Li et al., 1997) and 9 (Zou et al., 1997). In keeping with the adapter protein

role of the anti-apoptotic molecules, recent evidence shows that Bcl-xl binds to Apaf-1,

the newly discovered mammalian homolog of the C. elegans CED-4 (Zou et al., 1997),

and these two proteins, along with the uncleaved and inactive caspase 9, exist in a ternary








complex (Pan et al., 1998). Cytochrome C release is thought to promote activation of the

caspase cascade by promoting the dissociation of this ternary complex through an

undefined mechanism. Once caspases are activated, they act upon a variety of

intracellular substrates, including inhibitors of caspase-activated deoxyribonuclease

(ICAD). Caspase 3 is known to cleave ICAD and result in the elimination of its normally

inactivating effect on caspase-activated deoxyribonuclease (CAD). CAD functions

downstream in the proteolysis cascade and its reduced inhibition results in DNA

degradation characteristic of apoptotic cell death (Sakahira et al., 1998).

Another piece to this intriguing puzzle has been provided by Yang et al. (1998),

who have shown that CED-4 (and likely the mammalian counterpart Apaf-1) promotes

CED-3 (the C. elegans caspase) processing and activation by promoting the aggregation

of unprocessed CED-3. This "induced proximity" is thought to sequester inactive caspase

proenzymes to increase their local concentration and promote conformational changes

which will increase the likelihood of their activation (Hengartner, 1998), and is brought

about by oligomerization of the CED-4:CED-3 complex (Yang et al., 1998). Thus, an

important role of Bcl-xl and Bcl-2 appears to be to prevent the association of these

proenzymes by preventing CED-4 (and by association, Apaf-1) oligomerization

(Hengartner, 1998). While a similar role for Apaf-1 has not been definitively

demonstrated, this will surely be one of the important future discoveries in this

blossoming field.

Newly uncovered evidence also indicates that the anti-apoptotic molecules of the

bcl-2 family can contribute to the cell's demise under certain intracellular conditions. For

example, if caspase activation proceeds to a critical point, caspases will act on the Bcl-2








and Bcl-xl proteins as substrates (Cheng et al., 1997; Clem et al., 1998). Indeed, caspase

cleavage of these protective molecules converts them into Bax-like death-promoting

molecules. This cleavage is thought to act as a feed-forward mechanism for further

caspase activation, and should ensure cell death.

Besides these important discoveries, several others have demonstrated that

apoptosis interacts with signal transduction intracellularly and that this interaction is

mediated, in part, by the bcl-2 family (Gajewski and Thompson, 1996). The application

of IL-3 to the FL5.12 cell line leads to phosphorylation and inactivation of the pro-

apoptotic molecule Bad. The phosphorylation of Bad leads to its association with the

cytosolic protein 14-3-3, prevents Bad:Bcl-xl heterodimerization, and promotes Bcl-xl

homodimerization, which results in cell survival (Zha et al., 1996). Additionally, Wang

et al. (1996) showed that Bcl-2 could target Raf-1 to mitochondrial membranes which

results in the phosphorylation of Bad.

Two groups independently determined that the protein kinase Akt is also

responsible for growth factor-mediated Bad phosphorylation (Datta et al., 1997; Delpeso

et al., 1997). Akt (also known as protein kinase B, or PKB) is an important intracellular

molecule and its phosphorylation (and activation) is brought about by growth factor-

mediated phosphorylation of PI3 kinase, followed by PI3 kinase phosphorylation of Akt

(Kahn, 1998; Zhou et al., 1997). Thus, it is apparent that a complex interaction between

extracellular growth and survival signals integrates with the complex intracellular biology

of individual neurons to determine whether a particular cell will survive a developmental

process or injury.








Hypotheses Tested


As mentioned previously, neuronal populations are not uniform in their

vulnerability to ethanol, nor are developmental stages uniform in susceptibility to ethanol

treatment. And while a variety of CNS populations have been demonstrated to be

adversely affected by prenatal ethanol exposure, the full extent of nervous system

vulnerability to ethanol is unknown. Likewise, the cellular effects of ethanol exposure,

such as changes in gene expression, are poorly understood. It is with this in mind that the

research reported in this document was undertaken.

Because of the observed learning and memory deficits in children exposed in

utero to ethanol (Streissguth et al., 1991), the SH system is a natural region in which to

extend studies in rodents of neuroanatomy following ethanol treatment. While cell loss in

the hippocampus has been documented following prenatal exposure (Barnes and Walker,

1981), cellular changes in the cholinergic basal forebrain component of the SH system

have not been found after a similar pattern of exposure (Swanson et al., 1996). Thus, it

was an aim of the present work to determine whether the GABAergic component of the

SH pathway is susceptible to prenatal ethanol treatment. Additionally, the cingulate

cortex was analyzed because of the observed behavioral problems in children with FAS

(Streissguth et al., 1991) and the decreased glucose utilization in the cingulate cortex of

alcoholics (Joyce et al., 1994). Thus, the following hypothesis was explored: chronic

prenatal ethanol exposure will lead to alterations in the number of neurons expressing

PA in the adult rat medial septum and cingulate cortex.








Rat studies have also identified the early postnatal period as a developmental time

during which particular CNS structures are sensitive to ethanol's toxic effects. For

example, the Purkinje cells of the cerebellar cortex are reduced in number when ethanol

is delivered during the early neonatal period, but not when ethanol is given in utero

(Marcussen et al., 1994). The SH system is a region which exhibits susceptibility to

ethanol during both the prenatal and the neonatal periods. Neonatal ethanol exposure

disrupts the normal development of the SH system in rodent models of FAS, and much

research has documented neuroanatomical changes in the hippocampus (Bonthius and

West, 1990; Bonthius and West, 1991; Greene et al., 1992; Pierce and West, 1987; West

and Pierce, 1986). However, the effect of neonatal ethanol exposure on the MS has not

been investigated. Because the number of cholinergic neurons of the medial septum are

not altered by prenatal ethanol treatment, it was an objective of the present work to

determine the long-term effects of neonatal ethanol exposure on the cholinergic neurons

in the rat. The following hypothesis was investigated: early postnatal ethanol exposure

will lead to alterations in the number of neurons expressing ChAT in the adult rat CNS.

While descriptions of susceptible neuronal populations following developmental

ethanol exposure has been, and continues to be, a fruitful avenue of research, further

investigation into the molecular consequences of ethanol treatment (e.g. gene expression

changes resulting from ethanol) are warranted. Given the previously described temporal

pattern of Purkinje and granule cell loss in the postnatal cerebellum, a survey of the bcl-2

literature raises the intriguing possibility that the differential temporal teratogenicity of

ethanol on cerebellar cells may be related to changes in the levels of expression of PCD

repressor and inducer genes. It is significant that proteins translated from bcl-2 and bax








mRNA dimerize, and as noted above, cell death or survival depends on the relative

amounts (ratio) of these proteins (Oltvai et al., 1993). Similar cellular survival outcomes

have been linked to the relative ratio of bcl-2 to bax gene expression (Basile et al., 1997;

Chen et al., 1996).

The observation that cerebellar neurons are susceptible to ethanol neurotoxicity as

a function of the timing of the ethanol insult suggests that ethanol may act in vivo to

modulate (upregulate or downregulate) the expression of certain mRNAs of the bcl-2

family and thereby alter susceptibility to ethanol neurotoxicity. Thus, any ethanol-

induced change in bcl-2 family gene expression might disrupt the normal balance of these

proteins in developing neurons and decrease or increase their chance of cell death.

Recent evidence, mentioned above, in addition to the aforementioned expression patterns,

implicates bcl-2 family members in maintaining cerebellar neurons. Purkinje-cell-

degeneration mouse mutants lose nearly all of their Purkinje cells between P22-28.

Gillardon et al. (1995) investigated bcl-2 and bax expression in the cerebella of these

mice and found that bcl-2 mRNA levels decreased while bax mRNA levels remained

unchanged beginning on P22. There was a concomitant reduction of Bcl-2 expressing

Purkinje cells in the mutants compared with wild-types, suggesting a down-regulation of

bcl-2 in Purkinje cells destined to die (Gillardon et al., 1995).

In addition, thyroid hormone-induced upregulation of bcl-2 protects early-

differentiating cerebellar granule cells from apoptosis in vitro (Muller et al., 1995).

Moreover, transgenic mice overexpressing bcl-2 contain more cerebellar Purkinje and

granule cells than controls whether transgene expression was induced in the embryonic or

postnatal periods (Zanjani et al., 1996; Zanjani et al., 1997). Therefore, another goal of








the present work was to test following hypothesis: ethanol-induced alterations in the

expression levels of bcl-2 family PCD genes in the cerebellum contribute to the

cerebellum's relative temporal susceptibility to ethanol neurotoxicity.

The chapters that follow provide a detailed description of methods utilized to test

these hypotheses, and the data collected are presented and fully discussed. Following

these chapters is a concluding chapter that presents interpretations derived from these

data and discusses specific methodological considerations and future directions which the

current data suggest.















CHAPTER 2
EFFECTS OF PRENATAL ETHANOL EXPOSURE ON PARVALBUMIN-
IMMUNOREACTIVE GABAERGIC NEURONS IN THE ADULT RAT MEDIAL
SEPTUM AND ANTERIOR CINGULATE CORTEX


Summary


Exposure of human fetuses to ethanol often results in the fetal alcohol syndrome

(FAS). Animal models of FAS have been developed and have been utilized to examine

the consequences of prenatal ethanol exposure on the central nervous system. While cell

loss in the hippocampus has been documented following prenatal ethanol exposure

(Barnes and Walker, 1981), cellular changes in the cholinergic basal forebrain component

of the septohippocampal (SH) system have not been found after a similar pattern of

ethanol exposure (Swanson et al., 1996). The objective of this study was to determine the

long-term effects of prenatal ethanol exposure on parvalbumin-expressing (PA+)

GABAergic neurons of the rat medial septum and anterior cingulate cortex.

Pregnant Long-Evans rats were maintained on one of three diets throughout

gestation: an ethanol-containing liquid diet in which ethanol accounted for 35% of the

total calories, a similar diet with the isocaloric substitution of sucrose for ethanol, or a lab

chow control diet. Offspring were sacrificed at postnatal-day 60 and their brains were

prepared for parvalbumin immunocytochemistry. Female rats exposed to the ethanol-

containing diet during gestation had 42 % fewer total PA+ neurons in the medial septum

and reduced PA+ cell density when compared to female rats exposed to the sucrose diet.








Ethanol-exposed females also had fewer PA+ neurons per section than sucrose-control

females. Male rats exposed to ethanol did not display a similar change in PA+ neurons or

density. No effect of prenatal diet was found on the area or volume of the medial septum,

or the size of the PA+ neurons therein. As such, prenatal exposure to ethanol appears to

permanently reduce the number of PA+ neurons in the female rat medial septum without

affecting the size of the structure or the size of the neurons.

Due to the observed behavioral problems in children with FAS (Streissguth et al.,

1991) and the fact that alcoholics show decreased glucose utilization in the cingulate

cortex (Joyce et al., 1994), the effect of prenatal ethanol exposure on the number of PA+

GABAergic neurons in the adult rat anterior cingulate cortex was also determined. This

represents the first attempt to document developmental alterations in the cingulate cortex

following ethanol exposure. Rats exposed to the ethanol-containing diet contained 45%

fewer total PA+ neurons in the anterior cingulate cortex, and fewer PA+ neurons per

section compared with sucrose and chow controls. No gender differences were found in

measures of the anterior cingulate. The reduction in PA+ neurons occurred in the absence

of changes in structure area or volume, and occurred in the absence of changes in PA+

neuronal size. Functional implications and possible relations to the fetal alcohol

syndrome are discussed.

Introduction


Exposure of human fetuses to ethanol often results in a constellation of

developmental anomalies which make up FAS. Such developmental perturbations

include pre- and postnatal growth deficiencies, morphological (e.g. craniofacial)








abnormalities and central nervous system (CNS) deficits (Abel, 1995) and these typically

persist into adulthood (Streissguth et al., 1991). The incidence of FAS in the general

obstetric population of the United States is estimated at 1.95 per 1,000 live births (Abel

and Hannigan, 1995) and FAS is thought to be the leading nongenetic cause of mental

retardation (Abel and Sokol, 1986). Experimental studies, utilizing rodent models of

FAS, have been successful in reproducing many of the behavioral and morphological

changes found in FAS, and have described numerous CNS alterations resulting from in

utero ethanol exposure. Alterations in CNS development include changes in neuronal

proliferation and migration (Miller, 1986; Miller, 1995b; Miller, 1996), altered neuronal

and cortical morphology (Burrows et al., 1995; Davies and Smith, 1981; Kotkoskie and

Norton, 1989), changes in receptor density and enzyme ontogeny (Black et al., 1995;

Swanson et al., 1995), delays in synaptogenesis (Hoff, 1988), alterations in neuronal

number (Barnes and Walker, 1981), and long-lasting deficits in learning (Clausing et al.,

1995) and behavior (Riley, 1990). The present study sought to determine the

vulnerability of GABAergic intemeurons expressing parvalbumin in the medial septum

and anterior cingulate cortex following prenatal ethanol exposure.

Medial Septum

Numerous studies investigating the effects of prenatal ethanol exposure on rodents

have determined that the CNS is not uniform in its susceptibility to ethanol. In addition

to differential temporal susceptibility, the CNS exhibits differential regional vulnerability.

The SH system of basal forebrain afferents and hippocampal targets is one region that has

been shown to be sensitive to developmental ethanol exposure. Permanent reductions of

CAl pyramidal neurons (Bames and Walker, 1981; Wigal and Amsel, 1990), alterations








in hippocampal neuronal circuitry (West et al., 1981) and dendritic arborization (Davies

and Smith, 1981; Smith and Davies, 1990), delayed synaptogenesis in the hippocampal

dentate gyrus (Hoff, 1988), changes in hippocampal muscarinic receptors (Black et al.,

1995), and delayed choline acetyltransferase (ChAT) enzyme ontogeny in the basal

forebrain (Swanson et al., 1995) have all been demonstrated in the rodent SH system

following exposure to ethanol in utero. Because of the SH system's role in generating

and maintaining electrical activity in the hippocampus (Dutar et al., 1995), and the

importance of the hippocampus in learning and memory, the possibility exists that

ethanol-induced changes in this system can have detrimental effects on offspring exposed

to ethanol prenatally.

Intemeurons, pyramidal neurons, and granule cells of the hippocampus all receive

synapses from cholinergic and GABAergic fibers originating in the medial septum (MS)

and the horizontal and vertical limbs of the nucleus of the diagonal band of Broca (DBB)

(Dutar et al., 1995; Freund and Antal, 1988). Evidence exists for an ethanol-induced

alteration of the cholinergic component of the SH pathway. Sulik et al. (1984), for

example, noted severe abnormalities in the basal forebrain of fetal mice following an

acute ethanol dose at gestational day 7 (G7). Moreover, Schambra et al. (1990) found a

reduction in the number of ChAT-immunoreactive neurons in fetal mice following an

acute ethanol administration on G7. Similar findings in adult rats were reported

following chronic ethanol treatment (Arendt et al., 1988).

In contrast, studies from our laboratory utilizing a rat model of chronic prenatal

ethanol treatment (CPET) have shown an ethanol-induced disruption of the normal ChAT

ontogeny in the SH pathway, but have not revealed an effect on ChAT-immunoreactive








neuronal number (Swanson et al., 1995; Swanson et al., 1996). The present study sought

to determine whether the GABAergic component of the SH pathway was susceptible to

CPET.

Although not as extensively characterized as the cholinergic component, the

GABAergic component of the SH projection is an important part of this system.

Parvalbumin (PA), an 11.8 kDa member of a Ca2+ binding superfamily of proteins (Kiss

et al., 1990a; Kiss et al., 1990b; McPhalen et al., 1994), is expressed in approximately

33% of the neurons in the MS/DBB region. PA is commonly found in fast-firing

GABAergic neurons where it influences the activity of Ca2+-dependent K+ channels

(Plogmann and Celio, 1993). PA ontogeny begins in the MS of rats at G21 and coincides

with the beginning of physiological activity such as spontaneous firing and excitatory

synaptic input (Lauder et al., 1986; Solbach and Celio, 1991), and the PA-expressing MS

neurons innervate inhibitory intemeurons in the hippocampus (Freund and Antal, 1988).

CNS regions differ in the extent to which PA and GABA co-localize (Alonso et al., 1990;

Brauer et al., 1991; Kiss et al., 1990a), but within the MS nucleus most, if not all, of the

hippocampal-projecting GABAergic neurons are PA+ (Freund, 1989; Krzywkowski et al.,

1995). The present study sought to determine the long-term effects of prenatal exposure

to ethanol on a PA+ subpopulation of GABAergic projection neurons in the SH pathway.

For this determination we performed counts of PA+ neurons in the adult rat MS nucleus

following CPET.

Cingulate Cortex

The present study also sought to determine the vulnerability of GABAergic

intemeurons expressing parvalbumin in the anterior cingulate cortex following CPET.








The ontogeny of PA mRNA and protein in the rat cingulate cortex, and its functional role,

is similar to that noted previously for the medial septum. PA expression begins in the

first postnatal week, and coincides with the functional maturation of cerebral interneurons

in the cingulate (Alcantara et al., 1993; de-Lecea et al., 1995). The cingulate cortex was

chosen for analysis because it is a major relay center of the limbic lobe and is involved in

motor control, attention, emotion, and memory (Kupfermann, 1991; Muir et al., 1996;

Paus et al., 1993; Picard and Strick, 1996). The well documented cognitive and

behavioral impairments in children with FAS, including poor judgment, distractibility,

and hyperkinetic and emotional disorders (Steinhausen et al., 1993; Streissguth et al.,

1991), led us to investigate whether anatomical alterations in the cingulate cortex may

underlie these behavioral defects. Thus, the current study also sought to determine the

long-term effects of prenatal exposure to ethanol on PA+ interneurons of the cingulate

cortex.

Materials and Methods


Subjects and ethanol treatment

Long-Evans hooded rats, purchased from Charles River Co. (Portage, MI), were

housed individually, or in pairs, under controlled temperature and humidity conditions,

and were maintained on a 07:00-19:00 hour light cycle. Nulliparous females were placed

individually with a male overnight until vaginal smear the following morning was

indicative of insemination. This was defined as GO. At this time, females were matched

according to age and weight and assigned to one of three treatment groups from GO-G21:

ethanol, sucrose, or chow. During this time the ethanol group was given free access to an








ethanol-containing liquid diet in which ethanol comprised 35% of the total calories. The

sucrose group was pair-fed the same volume of a similar liquid diet which lacked ethanol

and contained an isocaloric substitution of sucrose for ethanol. The liquid diet was

prepared by mixing a stock ethanol or sucrose solution with Sustacal (Mead Johnson).

Diets were additionally enriched with Vitamin Diet Fortification Mixture and Salt

Mixture (ICN Nutritional Biochemicals). The diets contained 1.3 kcal/ml and have been

shown to provide several times the daily requirement of all essential vitamins and

nutrients (Walker and Freund, 1971). The chow group was given Purina Rodent Chow

and water ad libitum and served as a control for non-specific effects of the liquid diet.

Upon birth (postnatal-day 0 [PO]) pups from an ethanol or sucrose dam were

fostered to chow dams which had given birth at the same time. Litters were then

randomly culled to ten pups, with approximately equal numbers of males and females.

Pups were weaned at P21 and individually housed until perfusion at P60. In order to

avoid litter bias, animals used in this study are representatives from at least eight different

litters, and no more than one individual from each gender was used from a single litter. A

total of 39 animals was used in this study, with approximately equal numbers of males

and females in each diet treatment group (Ethanol, N=13; Sucrose, N=12; Chow, N=14).

Immunocvtochemical procedures

Animals were randomly selected for immunocytochemical staining at P60, an age

which represents adulthood in rats. This age was selected in order to examine the long

term effects of chronic prenatal ethanol treatment on parvalbumin expressing neurons in

the medial septum and cingulate cortex and corresponds to an age used in previous

analyses of prenatal ethanol influences on cholinergic neurons in the medial septum








(Swanson et al., 1996). At P60, animals were euthanized by pentobarbital overdose prior

to transcardial perfusion with phosphate buffered saline (PBS; 0.1 M, with 0.9% sodium

chloride) followed by 10% formalin (in the same PBS). Brains were removed and

equilibrated overnight in a cryoprotectant solution (PBS with 30% sucrose and 15%

ethylene glycol) and frozen at -700C until processing.

Animals of each gender from each diet group were processed for

immunocytochemistry at a given time. This insured against staining differences between

groups resulting from slight procedural differences. Brains were thawed and equilibrated

in a 30% sucrose-PBS solution, and mounted on the frozen stage of a sliding microtome.

Serial coronal sections were cut throughout the basal forebrain at a thickness of 40 pm.

Free-floating sections were immunostained for PA using a monoclonal antibody

(Accurate # 6092). Immunoreactivity was visualized as a blue-black reaction product

using an avidin-biotin conjugate/nickel intensified staining (see below).

Primary incubation with the monoclonal PA antibody (1:2500 in PBS, 0.1%

normal goat serum [NGS], 0.1% Triton X-100, and 0.005% sodium azide) was carried

out at 40C overnight. Sections were then washed with PBS and incubated with

biotinylated anti-mouse IgG (Sigma, B0529; 1:10000 in PBS, 0.1% NGS, 0.1% Triton X-

100, and 0.005% sodium azide) overnight at 4C. Sections were then washed and

incubated with Extravidin-horseradish peroxidase conjugate (Sigma; 1:1000 in PBS)

overnight at 4C. Sections were washed with 0.1 M sodium acetate (pH 7.2) to eliminate

phosphate which can precipitate divalent cations. Sections were then reacted for 3

minutes at room temperature using a developing buffer with 0.8 M sodium acetate, 8 mM

imidazole, 0.5% nickel (II) sulfate, 0.04% 3,3 diaminobenzidine tetrahydrochloride, and








0.005% hydrogen peroxide. Following development, sections were again washed in

sodium acetate buffer, mounted onto slides in PBS, air dried, dehydrated and

coverslipped. Control sections omitting the primary antibody were routinely developed to

ensure that any observed staining was due to PA. Slides from animals of each gender

from the three groups were randomized and coded such that all subsequent analyses were

carried out blind.

MS cell count, area, and volume analysis

PA+ cell counts were conducted bilaterally on alternate sections throughout the

entire rostral-caudal extent of the MS nucleus. The packing density of the DBB nucleus

was too great for accurate cell counts to be made by the image software. Thus, for the

purposes of this study the MS alone was examined and was defined rostrally by the

ventral fusion of the hemispheres (at the level of the genu of the corpus callosum) and

caudally by the decussation of the anterior commissure. The mean number of PA+

neurons per section was derived. Recent data have demonstrated the utility of manual

cell counts (Clarke and Oppenheim, 1995) and show a direct correlation between cell

counts performed manually and cell counts performed with the optical dissector (Hagg et

al., 1997).

Images of PA immunostained sections were captured and digitized using a

RasterOps 24STV video capture board and software on a Macintosh IIvx computer. Low

magnification images (2.5x objective; effective scale 1.59 imn/pixel) were captured in

order to obtain the entire MS region in one image. When capturing each image, the

lighting and contrast enhancement were optimized for identification of individual cells.

Images were digitally processed using the image analysis program NIH Image (freeware








from NIH). To reduce background variation across the image, a digitally defocused

image was created by passing the primary image through a mean filter. The resultant

image was then subtracted from the primary digitized image. This processed image was

then passed through a Laplace filter to enhance edges and separation between cells.

Analysis of individual images was initiated by outlining the region of the MS

nucleus. The area of a section outlined varied depending on the rostral-caudal location of

the individual section. The MS nucleus in rostral sections was defined as the mediodorsal

group of neurons which were separated from the ventrally-located DBB nucleus. In

intermediate sections, where the demarcation between the MS and DBB nuclei is

ambiguous, the MS was defined ventrally by a line perpendicular to midline at the level

of the anterior commissure. In caudal sections, the MS was defined as the medially

located cells dorsal to the anterior commissure. Cells were highlighted interactively by

adjusting the grayscale threshold level to include only objects which were considered

cells. A previous study from our lab has demonstrated that there is a close correlation

(r2= 0.956) between computer-automated and manual counts performed with a

microscope and a drawing tube (Swanson et al., 1996). Given the thickness of each

section (40 pm), the fact that alternate sections were analyzed, and the fact that PA+

neurons in the MS nucleus range from 6-26 pm in diameter (see below), it was not

deemed necessary to perform a split-cell correction on these counts. The program

counted the highlighted objects and measured the area of the outlined region (mm2).

Tissue volume (mm3) was calculated using a modification of the Cavelieri method

(Michel and Cruz, 1988). These measurements were taken in order to determine whether

ethanol treatment changed the size of the structure of interest, and whether changes in the








number of PA+ neurons per section were due to concomitant changes in the size of the

area examined (Peterson et al., 1997). This was done by multiplying the number of 40

p.m sections analyzed for a given animal by the mean area per section for that animal and

section thickness.

Anterior cingulate cell count, area, and volume analysis

The packing density of PA+ neurons in the anterior cingulate cortex was too great

for accurate cell counts to be made by automated imaging software at the magnification

necessary to include the entire region. Therefore, manual counts of PA+ neurons were

performed on every sixth section through the left side from the genu of the corpus

callosum caudally until decussation of the anterior commissure. The anterior cingulate

gyrus was bordered laterally by the cingulum, and ventrally by the corpus callosum

(Paxinos and Watson, 1982). The mean number of PA+ neurons per section was derived

through manual counting of neurons at 400x.

Because our cell count data are expressed as the mean number of cells per section,

it is necessary to demonstrate that any noted change in cells per section is not due to

changes in the volume of the structure being examined (Peterson et al., 1997). Therefore,

the volume of the anterior cingulate gyrus was calculated by a modification of the

Cavaleri method (Michel and Cruz, 1988) to ensure that changes in mean number of PA+

neurons per section as a result of ethanol treatment were not due to changes in the size of

the gyrus. For the volume determination, the mean area of the anterior cingulate gyrus

was determined for each animal by measuring the distance from midline to the cingulum

and the distance from the corpus callosum to the dorsal brain surface at 25x on three

anatomically matched sections with an eyepiece micrometer. These distances were








multiplied and mean area per section was calculated. This area was then multiplied by

the number of sections through the anterior cingulate gyrus and the section thickness to

determine mean cingulate gyrus volume.

Other analyses

The blood ethanol concentration (BEC) in the ethanol group was determined

between 00:00 and 02:00 hours on Gl 8 utilizing the Sigma 333-UV kit. PA+ cell

diameters were determined at 400x utilizing an eyepiece micrometer. Thirty cells from

identical anatomical locations (in the respective structures) for each gender in all groups

were measured. Density of PA+ neurons was determined by dividing the mean number

of cells per section by the mean area (mm2) per section for the respective structures.

Statistical analysis

Statistical analysis was performed with the program StatView (Abacus Concepts,

Berkeley Ca). For each parameter examined, one-way Analysis of Variance (ANOVA)

was used to test for the main effects of diet and gender separately. When appropriate, the

data were further analyzed with the Fisher's Protected Least Significant Difference

(PLSD) post hoc test to determine individual group differences. Because gender

differences were noted for MS measures (following an initial two-way ANOVA for

effects of diet and gender), males and females were analyzed separately. No gender

differences were noted in cingulate measures (following an initial two-way ANOVA for

effects of diet and gender), and so these data were not split by gender for analysis, and

one-way ANOVA was used to test for effects of diet, followed by the PLSD when

appropriate.








Results


BEC, body and brain weight, and brain to body weight ratio measures

BECs in the ethanol group ranged from 150-175 mg/dl (mean = 161 + 18 mg/dl).

This BEC represents a peak or near-peak level, as the nocturnal rats consumed the

majority of the diet after the twelve hour light cycle and blood samples were taken

between 00:00 and 02:00 hours on G18. P60 body weights were taken prior to perfusion

and P60 brain weights were taken following perfusion and are presented in table 2-1,

along with brain to body weight ratios (all tables are located at the end of the chapter).

The ANOVA for P60 body weight revealed no significant effects of diet on P60 body

weight of males or females. A gender difference was noted in P60 body weight for

ethanol [F(1,11) = 24.7; p< 0.0001], sucrose [F(l,10) = 31.4; p< 0.0001], and chow

[F(1,12) = 39.6; p< 0.0001] animals which was expected given the larger size of males at

P60.

The ANOVA for P60 brain weight demonstrated no effect of diet treatment in

males or in females. Likewise, no significant gender differences were noted for brain

weight within ethanol, sucrose, or chow animals.

The ANOVA for brain to body weight ratio showed no effect of diet in males or

females. Gender differences were noted in P60 brain to body weight ratio for ethanol

[F(l,11) = 37.5; p< 0.0001], sucrose [F(1,10) = 41.2; p< 0.0001], and chow [F(1,12) =

38.5; p< 0.0001] animals, which was not surprising given the larger size of male animals

at this age.








Number of sections, MS area per section, and PA+ neuronal density in the adult rat
medial septum

Table 2-2 presents the number of sections, MS area per section, and PA+ neuronal

density of the P60 MS. The number of alternate 40 uim sections through the MS was

determined by counting the number of sections analyzed between the ventral fusion of the

hemispheres (at the level of the genu of the corpus callosum) and the anterior

commissure. The ANOVA showed no effect of diet for either males or females on the

mean number of alternate sections through the MS, nor were there gender differences for

ethanol or chow animals. The sucrose group, however, did show a gender effect on the

mean number of sections with females containing significantly fewer than males

(F[1,10]= 4.511; p< 0.05).

The area of the MS outlined on a given section was computed by the image

software. No significant differences were noted between any group for males or females

in mean area per section nor were there gender differences for ethanol, sucrose, or chow

animals. Density of PA+ neurons (mean number of cells per section/mean area [mm2]

per section) was also calculated, and the ANOVA showed an effect of diet on cell density

in females (F[2,16]= 7.347; p< 0.01) but not in males. The PLSD post-hoc test

determined that cell density in ethanol females was significantly reduced compared to that

of sucrose females (p< 0.01). Neuronal density in chow females was also significantly

reduced compared to sucrose females (p< 0.05). The greater density of PA+ neurons

noted in sucrose females is perhaps due to a non-specific effect of liquid diet (see below

for discussion). A gender difference in the ethanol group was also noted for neuronal

density: neuronal density of females was reduced from that of males (F[1,1 1]= 5.075; p<








0.05). This difference is best described by the observed reduction in total PA+ cell

number, and not by differences in MS area. No gender differences in neuronal density

were noted for the sucrose or chow groups.

Total PA+ neurons in the MS

Figure 2-1 (all figures are located at the end of chapter) presents the mean total

number of PA+ cells counted on alternate sections throughout the P60 MS for animals of

both genders from each diet group. The ANOVA for total number of PA+ cells in the

MS determined a significant effect of diet treatment in females (F[2,16]= 4.351; p< 0.05)

but not in males. The PLSD post-hoc test further revealed that ethanol females had 42 %

fewer total PA+ cells than sucrose females (p< 0.01). A qualitative reduction in cell

number in the MS of ethanol females as compared to sucrose females can be observed in

Figure 2-2. Ethanol females were not different from chow females, nor were chow

females different from sucrose females. Gender differences were noted in the ethanol

group and the ANOVA for total number of PA+ cells in the MS indicated a reduction in

the number of PA+ cells in ethanol females compared to ethanol males (F[1,l 1]= 8.421;

p< 0.05). Gender differences were not noted in sucrose or chow groups.

Number of PA+ neurons per section in the MS

Figure 2-3 displays the mean number of PA+ neurons detected per section for

animals of both genders from each group in the P60 MS. The ANOVA for mean number

of cells per section revealed a significant effect of diet treatment in females (F[2,16]=

7.342; p< 0.01) but not males. The PLSD post-hoc test further determined that ethanol

females had 44 % fewer PA+ cells per section than sucrose females (p< 0.01). Chow

females also displayed a significant reduction in mean PA+ cells per section compared








with sucrose females (p< 0.05), while ethanol and chow females were not different from

each other. The difference between ethanol females and sucrose females is due to a

reduction in total cell number and not to a difference in tissue volume. The difference

noted between chow females and sucrose females is perhaps due to a non-specific effect

of liquid diet and may be related to the aforementioned increase in PA+ neuronal density

noted in sucrose females (see below for discussion). Gender differences were also noted.

The ANOVA showed a reduced mean number of PA+ cells per section in ethanol

females, compared to ethanol males (F[1,11]= 5.063; p< 0.05). This difference was again

due to a reduction in total PA+ cell number, and not to a difference in volume. No gender

differences were noted for the sucrose or chow groups.

Diameter of PA+ MS neurons

PA+ cell diameters were measured for both genders in all groups and are

presented in figure 2-4. No differences were noted between any treatment group for

males or females; nor were gender differences noted for the ethanol, sucrose, or chow

groups.

MS volume determination

Figure 2-5 presents MS volume (mm 3) for the analyzed sections. Volume was

calculated by multiplying the number of 40 pm sections analyzed for a given animal by

the mean area per section for each animal and the section thickness. The resulting

volume was then compared across groups for each gender and between both genders for

each diet treatment. No significant differences were found between diet groups for males

or females. Gender specific effects of tissue volume, however, were noted in sucrose

animals. The ANOVA showed a significant reduction of tissue volume in the sucrose








females compared to sucrose males (F[1,10]= 6.102; p< 0.05). This was likely due to a

decrease in the number of sections for female sucrose animals compared to male sucrose

animals (see table 2-2; p< 0.05). No gender differences in MS volume were noted for the

ethanol or chow groups.

Number of sections, mean area per section, and PA+ neuronal density in the adult rat
anterior cingulate cortex

Table 2-3 presents the number of sections, mean area per section, and PA+

neuronal density in the adult rat cingulate cortex. The number of 40 pm sections through

the anterior cingulate gyrus was determined by counting the number of sections between

the genu of the corpus callosum and the decussation of the anterior commissure. The

ANOVA showed no effect of diet on the mean number of sections through this region.

The area of the gyrus on a given section was determined by measuring the

distance from midline to the cingulum and the distance from the corpus callosum to the

dorsal brain surface at a magnification of 25x on three anatomically matched sections

with an eyepiece micrometer. No significant differences were noted in mean area per

section. The Density of PA+ neurons in the anterior cingulate (mean number of cells per

section/mean area [mm2] per section) was also calculated, and the ANOVA showed an

effect of diet on cell density (F[2,30]= 7.041; p< 0.01). The PLSD post-hoc test revealed

that cell density in the ethanol group was significantly reduced compared to that of the

sucrose (p< 0.01) and chow groups (p< 0.01).

Total PA+ neurons in anterior cingulate

Figure 2-6 presents the total number of PA+ neurons counted throughout the

anterior cingulate cortex. The ANOVA for total number of PA+ neurons in the anterior








cingulate showed a significant effect of diet treatment (F[2,30]= 12.314; p< 0.0001). The

PLSD post hoc test further revealed that ethanol animals had 45% fewer PA+ neurons

than sucrose (p< 0.0001) or chow (p< 0.0001). Figure 2-7 demonstrates a qualitative

reduction of PA+ neurons in the anterior cingulate cortex of adult rats exposed to ethanol

prenatally when compared with sucrose controls.

Number of PA+ neurons per section in the anterior cingulate

Figure 2-8 presents the mean number of PA+ neurons per section in the anterior

cingulate. The ANOVA revealed a significant effect of diet treatment on the mean

number of PA+ neurons per section (F[2,30]= 12.80; p< 0.0001). The PLSD further

determined that ethanol treatment reduced the mean number of PA+ neurons per section

when compared with sucrose (p< 0.001) and chow (p< 0.0001) controls.

PA+ neuronal diameter in the anterior cingulate cortex

PA+ cell diameters were determined with an eyepiece micrometer at 400x. Ten

whole cells from identical anatomical location for each gender in all groups were

measured on three sections for each brain region. The mean PA+ neuronal diameter for

each group was determined and is presented in figure 2-9. No significant differences in

the mean PA+ neuronal diameter were noted.

Anterior cingulate gyrus volume

The mean gyrus volume for each group was determined, and is presented in figure

2-10. No significant effect of treatment on mean volume was noted.








Discussion


The long-term effects of chronic prenatal ethanol exposure on PA+ neurons were

examined in the medial septum and anterior cingulate cortex of adult rats. This pattern of

exposure produced no significant alterations in P60 body weight (although males were

larger than females for all groups examined). Similarly, no significant differences were

noted in P60 brain weight, or brain to body weight ratio as a function of ethanol treatment

(although brain to body weight ratios were higher in females due to the larger size of

males at P60). The lack of effect of chronic prenatal ethanol treatment (CPET) on the

long-term growth of these animals is consistent with other reports from our laboratory

(Swanson et al., 1995; Swanson et al., 1996). This exposure paradigm did, however,

produce significant alterations in PA+ neurons in both regions examined, although a

sexually dimorphic effect was noted for the medial septum. The results for each brain

region are discussed separately below.

Medial Septum

A major conclusion drawn by the current study is that CPET affects the MS in a

sexually dimorphic manner. This study has demonstrated anatomical changes in the

expression of PA in the female rat MS following prenatal ethanol exposure. The

evidence from this study suggests that a reduction in the number of MS neurons

expressing PA occurs in the absence of area, volume, or PA+ cell size effects. Whether

this reduction is due to lower PA levels in existing cells or to a loss of PA+ neurons is

impossible to conclude. Regardless, the data suggest that alterations in PA expression








occur in female rats following CPET and that PA expression in female rats may be

particularly susceptible to the long-term consequences of CPET.

Although male offspring exposed to ethanol in utero were unaffected as adults, a

number of differences were noted in female ethanol-treated animals when compared to

females in the sucrose group. The mean total number of PA+ neurons in the MS nucleus

of adult females was reduced by 42 % following prenatal exposure to ethanol, PA+

neuronal density was reduced in ethanol females, and ethanol females had fewer PA+

neurons per section when compared to sucrose females. There were also gender

differences in the ethanol groups for a number of measures. Female rats exposed to

ethanol in utero contained fewer total PA+ neurons, reduced density of MS PA+ neurons,

and fewer PA+ neurons per section when compared to males. Gender differences in MS

volume in the sucrose group were also noted, with the MS of females significantly

reduced compared to that of sucrose males. As mentioned previously, this difference was

due to a lower number of 40 gpm sections from the MS of female sucrose animals when

compared to sucrose males. This is not surprising since P60 female rats are smaller than

males. This difference was apparently enhanced in the sucrose animals, since neither the

ethanol nor the chow groups displayed MS volume differences.

Differential gender susceptibility of PA+ neurons in the MS following CPET

The finding that PA expression was affected by prenatal ethanol exposure only in

female animals suggests that differences in the hormonal environment of the males and

females influenced PA-immunoreactivity and susceptibility to ethanol. In fact, the SH

system itself is highly sexually dimorphic (Loy, 1986). Specifically, morphological

differences in hippocampal asymmetry (Diamond et al., 1982) and differences in binding








capacity of hippocampal glucocorticoid receptors (Turner and Weaver, 1985) have been

noted in rats. Additionally, Loy and Milner (Loy and Milner, 1980) have found

differences in lesion-induced hippocampal sprouting between male and female rats.

Whereas sprouting in female rats is uniform in the hippocampus following injury,

sprouting in males occurs predominately in the dentate molecular layer. These results

suggest that the CNS response to damage may be dictated by hormonal environment, and

that this response is sexually dimorphic. Perhaps similar differences in hormonal

environment determine sensitivity to ethanol in this region and account for the noted

sexual dimorphism.

Gender differences have been reported in the alcohol literature as well: female

rats are more affected than males in measures of radial arm maze performance following

chronic adult ethanol treatment (Maier and Pohorecky, 1986); Witt et al. (1986) found an

increase in muscarinic receptor binding sites in the hippocampus of ethanol-treated

females but not ethanol-treated males following chronic adult ethanol treatment; Kelly et

al. (Kelly et al., 1988) reported impaired spatial navigation in adult female rats but not

adult male rats following neonatal ethanol exposure. As in the current study, these

reports suggest an increased susceptibility to ethanol's effects in female animals and it

has been suggested that females may be more sensitive than males to ethanol in animal

models (West et al., 1989). However, not all studies have agreed with an increased

susceptibility for females. Goodlett and Peterson (Goodlett and Peterson, 1995), for

example, found increased susceptibility for spatial learning deficits in male rats following

time-limited binge ethanol exposure. Sexually dimorphic effects of prenatal ethanol

treatment have also been noted on daily water consumption, with males (but not females)








consuming more water than pair-fed controls as adults (McGivem et al., 1998). Clearly a

greater focus on potential gender differences in animal models of FAS is warranted in

order to clarify sexual dimorphism in relation to ethanol exposure. Moreover, an

investigation into gender differences in human FAS is warranted, as no sexual

dimorphisms have been reported in human offspring of mothers who drink during

pregnancy.

What specific differences in hormonal environment might account for the

observed sexually dimorphic effect of CPET on PA+ neurons in the female medial

septum? Estradiol, for example, is known to influence the functioning of the cholinergic

component of the SH system through its influence on choline re-uptake in the

hippocampus (Singh et al., 1994). Indeed, removal of estradiol and other ovarian steroids

by ovariectomy decreases cholinergic neurotransmission in the female hippocampus, and

this is reversible by estradiol treatment (Singh et al., 1994). Moreover, estradiol infusion

ameliorates fimbria-fomix lesion-induced decline of ChAT+ neurons in the rat medial

septum (Rabbani et al., 1997). While estradiol is not known to influence the functioning

of the GABAergic component of the SH system, estrogen is known to increase glutamic

acid decarboxylase (the rate limiting enzyme in the synthesis of GABA) mRNA levels in

the female rat brain (McCarthy et al., 1995). Thus, it is possible that an ethanol-induced

decrease in ovarian steroids might account for some of the effects noted in the present

study on GABAergic neurons. Effects of CPET on ovarian steroids would be a useful

avenue of future research and may shed light on this interesting sexually dimorphic effect.








Liquid diet effects on the MS

In addition to the effects of ethanol on the MS of female rats, a non-specific effect

of the liquid diet on females was noted for some measurements. Previous studies

utilizing rodent models of prenatal ethanol exposure have noted an effect of liquid diet

treatment. For example, Swanson et al. (1995) found a stimulatory effect of liquid diet on

ChAT enzymatic activity during the first postnatal week. Additionally, studies

quantifying ChAT+ cell number in the P 14 rat MS found a liquid diet-induced increase in

ChAT+ neuronal number for female sucrose animals, but not ethanol females or male

sucrose animals (Swanson et al., 1996). It was speculated that a possible sucrose diet-

induced stimulation of MS cholinergic development occurred in female rat pups at P14.

The seemingly protective and perhaps stimulatory nature of the liquid diet is probably due

to its high vitamin and mineral content.

The present study indicates similar diet effects for sucrose females. There seems

to be a liquid diet-induced increase in both the density of PA+ neurons in the MS and the

number of PA+ neurons per section in sucrose females. We speculate that a liquid diet-

induced increase in cell density for the sucrose group, similar to the increase in ChAT+

neurons noted by Swanson et al. (Swanson et al., 1996), raised the base level of PA+

neurons in the liquid diet animals to a level greater than that seen in chows. Although

there was not also a significant increase in cell number, the possibility exists that the

period of naturally occurring cell death served to reduce this overall number while density

remained high. The ethanol-treated group may also display this effect of liquid diet, but,

nonetheless, a significant ethanol effect was noted as ethanol females had fewer PA+

neurons. Apparently the effect of prenatal ethanol treatment is sufficient to produce a








significant cellular reduction in ethanol females. The raised baseline due to the liquid

diet, however, may mask a difference between ethanol and chow females.

Functional considerations for the MS

Although CNS regions differ in the extent to which PA and GABA co-localize,

the MS is a region where most, if not all, GABAergic neurons are PA+. This statement is

supported by studies of Freund (Freund, 1989) utilizing anterograde transport from the

septum to the hippocampus. Of the projection neurons identified by the anterograde

transport, all of the axons shown to be immunoreactive for GABA were immunoreactive

for PA. Thus, PA-immunoreactivity serves as a reliable marker for GABAergic

hippocampal projection neurons in the MS nucleus and identifies a subpopulation of the

total GABAergic neuronal pool within the basal forebrain.

As noted previously, in the MS nucleus PA expression begins at G21, after the

appearance of GABAergic neurons at G16 (Lauder et al., 1986; Solbach and Celio, 1991).

The ethanol exposure regimen utilized in the present study encompasses the entire

prenatal period, and thus ethanol was present in CNS tissue during the initiation of PA

expression in MS neurons and the physiological maturation of these cells. Exposure of

MS neurons to ethanol resulted in a long-term reduction in the number of PA+ neurons in

adult females. How might these neurons have been affected by changes in the normal

expression of PA?

The intracellular concentration of PA in neurons (6-45 gM) and the fact that it is

commonly found in fast-firing neurons (Plogmann and Celio, 1993) implicate PA in

buffering excess Ca2+ at presynaptic nerve terminals following rapid trains of action

potentials (Heizmann, 1984). Since excessively high levels of intracellular Ca2+ are








known to initiate cell death, the possibility exists that an ethanol-induced reduction of PA

expression resulted in fewer cells in the females by reducing those neurons' ability to

buffer calcium and suppress the process of cell death. Though neurons expressing normal

levels of PA can survive experimental cerebral ischemia (Nitsch et al., 1989) and fimbria-

fornix transaction (Kermer et al., 1995), any reduction in PA expression due to ethanol

may impede a protective effect. Solodkin et al. (1996) have shown a decrease in PA

immunostaining in parts of the entorhinal cortex which showed Alzheimer's pathologies.

Since PA expression is reduced in the basal forebrain of aged rats (Krzywkowski et al.,

1995) the possibility exists that the vulnerability of the septohippocampal pathway to

degenerative diseases such as Alzheimer's is due to a reduction in PA expression. The

observed reduction in PA+ neurons found following CPET in the current study may be

due to direct loss of neurons by the process of cell death or a reduction in the level of PA

in MS neurons which would preclude detection by immunostaining due to subthreshold

amounts of antigen/epitope. In fact, Kermer et al. (1995) have detected reduced PA

immunoreactivity in the medial septum following fimbria-fornix transaction in the

absence of PA+ neuronal cell death. However, whether cells are lost in our paradigm or

not, the ability of these neurons to function adequately would likely be diminished in

affected animals.

Altered MS function and FAS

A loss of effective GABAergic MS neurons, whether by a reduction in cell

number or a reduction of PA expression, may impede the normal function of the

septohippocampal system in learning and memory. GABAergic neurons in the basal

forebrain are known to control activity of cholinergic neurons in this region (Dudchenko








and Sarter, 1991) and working memory in the rat is disrupted by antagonism of

GABAergic transmission in the septum (Chrobak and Napier, 1992). Smythe et al.

(1992) have shown that both the cholinergic and GABAergic components of the MS are

necessary to influence the electrical activity in the hippocampus. Mechanistically, the

influence of the basal forebrain on hippocampal electrical activity is known to be on theta

cell rhythm (Bland and Bland, 1986) and the GABAergic component of the MS plays a

part in this influence (Allen and Crawford, 1984; Smythe et al., 1992). PA+ MS neurons

are known to depress the activity of inhibitory interneurons in the hippocampus (Freund

and Antal, 1988). This would serve to increase the excitability of the hippocampus by

removing local inhibitory potentials; theta rhythm would be initiated and the excitability

of the principal hippocampal cells involved in long-term potentiation would be increased

(Freund and Antal, 1988). The current study has demonstrated that CPET permanently

reduces the number of PA+ GABAergic neurons in the female rat MS by 42 % in the

absence of area or volume effects. Such a reduction in the number of PA+ GABAergic

neurons could potentially affect the functioning of the septum and hippocampus. Indeed,

some of the observed phenotypes of FAS, including mental retardation and spatial

learning deficits (see Kelly et al., (1988) Goodlett and Peterson (1995) for animal

studies), may be partially explained by a reduced septohippocampal efficacy. This

possibility is supported by the findings of Miettinen et al. (1993) that the decline in aged

rats in spatial learning correlates with a reduction in PA-containing neurons in the

entorhinal cortex. A similar decline in learning in FAS children may also relate to PA

expression in the SH system.








Cingulate Cortex

This investigation represents, to our knowledge, the first demonstration of

neuroanatomical alterations in the cingulate cortex as a result of developmental ethanol

exposure. In fact, this appears to be the first examination of teratogen-induced alterations

in cingulate cortex neuroanatomy, and is a significant contribution to the literature since

there exists a relative paucity of data on the cingulate cortex in normal or pathological

states.

A conclusion of the present study is that CPET permanently reduced the number

of PA+ neurons in the anterior cingulate cortex without altering size of the structure.

These data are similar in their significance to those reported previously for the MS. One

critical difference, however, between the effect of CPET on the anterior cingulate and the

effect of CPET on the MS is the lack of sexually dimorphic effects in the former. The

factors that are responsible for the lack of gender specific effects of CPET on the anterior

cingulate remain unknown. It is possible that the medial septum and anterior cingulate

respond differentially to different hormonal environments. While the medial septum is

known as a sexually dimorphic region (see above), no data exist which point to a similar

dimorphism for the anterior cingulate. Regardless, the overall pattern of ethanol-induced

abnormalities is similar, with the volume of the anterior cingulate unaffected by prenatal

ethanol exposure even though mean total PA+ neuronal number and mean PA+ neurons

per section was affected. Additionally, the size of the PA+ neurons was determined, and

was found to be unaltered by this pattern of in utero exposure. Our results are similar to

those of Kril and Homewood (1993), who found that PA+ neuronal number was








decreased in the frontal cortex of adult rats following chronic ethanol treatment and

thiamin deficiency.

Since the volume of the anterior cingulate was unaffected by treatment, and

because the cells examined showed no decrease in size, it is unlikely that the decline in

mean PA+ neurons per section is a result of an ethanol effect on the size of this area.

Because our cell counts are derived from PA immunostained sections, we cannot

definitively conclude that the reduced number of PA+ neurons is a result of ethanol-

induced cell death. While this is a possibility, it is equally likely that the reduction of

PA+ neurons noted after prenatal ethanol treatment is due to a decreased expression of

PA in living neurons (perhaps as a result of ethanol interfering with PA ontogeny) which

would preclude detection by immunostaining. Krzywkowski et al. (1995) examined PA

immunoreactivity in the septum of aged rats and found a decrease in the number of PA+

neurons without loss of GABAergic neurons, suggesting that PA+ expression levels were

decreased without cell loss. Similar findings have been reported in patients with

Parkinson's disease (Hardman et al., 1996).

Yet another possibility is that ethanol interfered with the generation of neurons

within the anterior cingulate cortex. Evidence from Miller (1986) shows that prenatal

ethanol exposure delays and extends the period of cortical neuronal generation while

reducing the number of neurons and altering the distribution of neurons in the mature

cerebral cortex. It is conceivable that a similar reduction in the generation of cortical

neurons is responsible for our observed decrease in PA+ neuronal number as a result of

ethanol treatment.








Functional considerations for the cingulate

As mentioned previously, the present study examined PA+ neuronal number in

the cingulate cortex following prenatal ethanol treatment because of earlier observations

related to cerebral cortex functioning following ethanol exposure. These include severe

behavioral problems in children with FAS, including poor judgment, distractibility, and

difficulty perceiving social cues (Streissguth et al., 1991). Additionally, many alcoholics

who develop Korsakoff's syndrome have deficiencies in glucose utilization within the

cingulate cortex, potentially contributing to learning and memory defects due to

interruption of Papez' circuitry (Joyce et al., 1994). Papez' circuit is a neuroanatomical

pathway which originates in the hippocampus, proceeds through the fornix to the

mammillary bodies, and then to the anterior thalamic nucleus, cingulate gyrus, entorhinal

cortex, and then back to the hippocampus (Nolte, 1993). Therefore, it is conceivable that

alterations in PA expression patterns might contribute to behavioral anomalies and/or

learning and memory deficiencies.

Because the intracellular concentration of PA in neurons is in the range necessary

for calcium buffering (6-45 jpM) and the fact that it is commonly found in fast-firing

neurons (Plogmann and Celio, 1993), PA appears to be involved in buffering excess Ca2+

at presynaptic nerve terminals following rapid trains of action potentials (Heizmann,

1984). Since excessively high levels of intracellular Ca2+ are known to initiate cell death,

the possibility exists that an ethanol-induced reduction of PA expression resulted in fewer

cells in the ethanol animals by reducing those neurons' ability to buffer calcium and

suppress the process of cell death. Though neurons expressing normal levels of PA can

survive experimental cerebral ischemia (Nitsch et al., 1989) and fimbria-fomix








transaction (Kermer et al., 1995) any reduction in PA expression due to ethanol may

impede a protective effect. Solodkin et al. (1996) have shown a decrease in PA

immunostaining in parts of the entorhinal cortex which showed Alzheimer's pathologies.

Similarly, the number of PA+ neurons is reduced in the prefrontal cortex of

schizophrenics (Beasley and Reynolds, 1997).

While we cannot determine from the present study whether cell death or reduced

PA antigen is responsible for the noted reduction in PA+ neurons, it should be noted that

whatever the mechanism of the ethanol-induced reduction of PA+ neurons, inhibitory

neurotransmission in the cingulate cortex is likely to be altered. For example, PA+

neurons are lost at epileptic foci in animal models of epilepsy (De Felipe et al., 1993).

Moreover, Jacobs et al. (1996) showed that decreased PA immunoreactivity in cortical

freeze lesions of neonatal rats leads to hyperexcitability in adjacent cortex. Such

alterations in inhibitory and excitatory processes have clear functional consequences to

the organism, and likely result in an imbalance in excitation and inhibition in these

cortical regions (Wang et al., 1996). The current findings of decreased number of PA+

neurons in the anterior cingulate cortex of adult rats as a result of prenatal ethanol

exposure may provide a neuroanatomical basis for the well described cognitive and

behavioral impairments in offspring exposed to ethanol in utero.

Conclusions

The major findings of this study are that PA+ neurons in the medial septum and

anterior cingulate cortex are susceptible to prenatal ethanol exposure. In the medial

septum, a sexually dimorphic effect of ethanol was noted, with females more susceptible

than males. No gender specific effects were noted for the anterior cingulate. These






55

reductions in PA+ neurons occurred in the absence of significant ethanol-induced

alterations in the size of the structures examined, the size of the PA+ neurons within these

structures, or the growth of these animals. This represents the first description of ethanol-

induced alterations in PA+ GABAergic neurons in animal models of FAS. Further

research will help to further define developmental windows of GABAergic vulnerability

(for example, neonatal exposure models).









Table 2-1. Postnatal day 60 (P60) body and brain weight and brain to body weight ratio
(br/bd)*
Diet group N P60 body weight (g) P60 brain weight (g) P60 br/bd (%)
Females
Ethanol 8 220 + 2.92 1.32 + 0.02 0.600 + 0.010"
Sucrose 6 216+5.58 1.31 + 0.02 0.606+ 0.021t
Chow 6 224 + 5.06 1.36 + 0.02 0.607 + 0.022t
Males
Ethanol 5 341 + 13.5t 1.34 + 0.05 0.392 + 0.032
Sucrose 6 351 + 5.81t 1.35 + 0.06 0.384 + 0.019
Chow 8 348 + 13.6t 1.39 + 0.02 0.399 + 0.030
*All measures are expressed as mean + S.E.M. No significant differences were noted.
tSignificantly increased, compared with other gender of same group (p< 0.0001).


Table 2-2. Number of sections, medial septum (MS) area per section, and parvalbumin-
immunoreactive (PA+) neuronal density in the adult rat MS*
Diet group N Number of 40 ipm MS area/section Density (PA+
sections (mm2) cells/mm2)
Females
Ethanol 8 15.75 + 0.453 0.722 + 0.066 49.55 + 6.716t'f
Sucrose 6 15.50 + 0.428 0.655 + 0.061 89.38 + 9.700
Chow 6 17.67 + 1.054 0.670 + 0.065 59.87 + 6.237
Males
Ethanol 5 16.20 + 0.374 0.733 + 0.103 78.031 + 12.01
Sucrose 6 17.50 + 0.428 0.778 + 0.061 66.16 + 10.98
Chow 8 16.87 + 0.811 0.733 + 0.049 63.49 + 5.957
*All measures are expressed as mean + S.E.M. Area and volume measures are
representative of approximately 1/2 of the total MS (alternate sections were analyzed; see
materials and methods section for detail).
tSignificantly reduced from sucrose females (p< 0.01).
ISignificantly reduced from ethanol males (p< 0.05).
Significantly reduced from sucrose females (p< 0.05).
TSignificantly reduced from sucrose males (p< 0.05).






57



Table 2-3. Number of sections, mean area per section, and parvalbumin-immunoreactive
(PA+) neuronal density in the adult rat anterior cingulate cortex*
Diet group N Number of 40 pm Mean area/section Density (PA+
sections (mm2) cells/mm2)
Ethanol 13 41.5 +2.56 2.135 + 0.069 395.5 + 67.85f
Sucrose 12 39.8 + 2.56 2.262 + 0.060 669.6 + 54.56
Chow 14 35.8+ 1.34 2.385 + 0.080 697.4 + 61.88
*All measures are expressed as mean + S.E.M.
tSignificantly reduced compared with sucrose (p< 0.01) and chow (p< 0.01).









1200-

1000-

800-


Male


=Z Ethanol
M Sucrose
- Chow


Female


Figure 2-1. The mean total number of parvalbumin-expressing neurons detected on
alternate sections through the medial septum of postnatal-day 60 rats exposed in utero to
one of three diets, ethanol, sucrose, or chow (see materials and methods section for
detail). Data are expressed as mean + SEM. a: significantly reduced compared with
sucrose females (p< 0.01). b: significantly reduced compared with ethanol males (p<
0.05).


a,b

T


600-

400-

200-


0--































Figure 2-2. Photomicrographs of 40 ptm coronal sections through the medial septum of
ethanol-treated (A) and sucrose-treated (B) postnatal-day 60 female rats. Sections are
matched for anatomical location and are representative of their respective group. A
qualitative reduction in cell number for the ethanol females can be noted, as can an
increase in cell density for sucrose females. Scale bar equals 125 tm.
















A



.4 ',
p


'1



t


0..'
a


'I 4

~41 *j~

I,
I *~
I I
dl fl
4. 1'

r
fr ? *



~ Ii

p.
U


)
~


I ft
*4


I


4
6


I,






fr,1

ft


4/



.* *,


^*;(
'1
>. t^ a









3d e
S .
..- .
f :-\.
*; a
*'
'1 i


"'U
A


r
IVA


S

I!. %


4



Ga
q'~* I
1~


*


Ir
'S.
'119
'I
,

*1
6*


-d I


fr


4


4


e


'I


9
V.
* I


.Ip..Y
4W
h ~

I ~


B'
I~ '

es I
I



S! ^


#
fc


dL











=-- Ethanol
M Sucrose
M Chow


75-



50-



Z 25-



0-


Figure 2-3. The mean number of parvalbumin-expressing neurons per section in alternate
sections through the medial septum of postnatal-day 60 rats exposed in utero to one of
three diets, ethanol, sucrose, or chow (see materials and methods section for detail). Data
are expressed as mean + S.E.M. a: significantly reduced compared with sucrose females
(p< 0.01). b: significantly reduced compared with sucrose females (p< 0.05). c:
significantly reduced compared with ethanol males (p< 0.05).


Female


Male












---1 Ethanol
M Sucrose
M Chow


Female


Figure 2-4. The mean diameter of parvalbumin-expressing neurons on alternate sections
through the medial septum of postnatal-day 60 rats exposed in utero to one of three diets,
ethanol, sucrose, or chow (see materials and methods section for detail). Data are
expressed as mean + S.E.M. No significant differences were noted.


p.i
h

II-0


+


0-


Male















a T


=C1 Ethanol
m Sucrose
M Chow


Female


Figure 2-5. The mean medial septum (MS) volume of postnatal-day 60 rats exposed
throughout gestation to an ethanol-containing liquid diet, sucrose-containing liquid diet,
or lab chow and water. Data are expressed as mean + SEM. a: significantly reduced
compared with sucrose males (p< 0.05).


0.75-


2
2
2
0U


0.50-



0.25-


0.00 -


Male











2000-


1000-


__Tr


-IT_


-7-


Ethanol


Sucrose


Chow


Figure 2-6. The mean total number of parvalbumin-expressing (PA+) neurons detected
on alternate sections through the anterior cingulate cortex of postnatal-day 60 rats
exposed in utero to one of three diets, ethanol, sucrose, or chow (see materials and
methods section for detail). Data are expressed as mean + SEM. a: significantly reduced
compared with sucrose (p< 0.0001) and chow (p< 0.0001).
































Figure 2-7. Photomicrographs of 40 p.m coronal sections through the anterior cingulate
cortex of ethanol-treated (A) and sucrose-treated (B) postnatal-day 60 rats. Sections are
matched for anatomical location and are representative of their respective group. A
qualitative reduction in parvalbumin-immunoreactive neuronal number is seen in the
ethanol-treated animal compared with the sucrose control. Scale bar equals 110 upm.






66






A B





67



300-



S200
a


100-



0
Ethanol Sucrose Chow


Figure 2-8. The mean number of parvalbumin-immunoreactive (PA+) neurons per
section counted in the anterior cingulate cortex of postnatal-day 60 rats exposed
throughout gestation to an ethanol-containing liquid diet, sucrose-containing liquid diet,
or lab chow and water. Data are expressed as mean + SEM. a: significantly reduced
compared with sucrose (p< 0.001) and chow animals (p< 0.0001).









20-






o 10-






Ethanol Sucrose Chow



Figure 2-9. The mean diameter of parvalbumin-immunoreactive (PA+) neurons in the
anterior cingulate cortex of postnatal-day 60 rats exposed throughout gestation to an
ethanol-containing liquid diet, sucrose-containing liquid diet, or lab chow and water.
Data are expressed as mean + SEM. No significant differences were noted.






69




4-





1-







0
Ethanol Sucrose Chow


Figure 2-10. The mean cingulate gyms volume of postnatal-day 60 rats exposed
throughout gestation to an ethanol-containing liquid diet, sucrose-containing liquid diet,
or lab chow and water. Data are expressed as mean + SEM. No significant differences
were noted.















CHAPTER 3
EFFECTS OF NEONATAL ETHANOL EXPOSURE ON CHOLINERGIC NEURONS
OF THE RAT MEDIAL SEPTUM


Summary


Developmental ethanol exposure has been known to affect the normal

development of the central nervous system. Studies in animal models have determined

that chronic prenatal ethanol exposure has no effect on the number of cholinergic neurons

in the rat medial septum (Swanson et al., 1996). Since many brain regions exhibit tight

temporal windows of vulnerability to ethanol, the objective of this study was to determine

the long-term effects of chronic neonatal ethanol exposure on the cholinergic neurons in

the medial septum (MS) of the rat. On postnatal day 4 (P4) pups were assigned to one of

three groups: an ethanol-receiving, gastrostomized group (EtOH); a pair-fed,

gastrostomized control group (GC); and a dam-reared suckle control group (SC).

Gastrostomized pups were infused with ethanol-containing or control diet as a 9.1% v/v

solution for two feedings on each day from P4-10. Choline acetyltransferase (ChAT)

immunocytochemistry was analyzed at P60.

Ethanol treatment resulted in long-lasting microencephaly in P60 animals.

Ethanol exposure did not directly reduce mean total ChAT-expressing (ChAT+) neuronal

number, or the mean number of ChAT+ neurons per section. Neither were changes noted

in MS volume, mean area section, or cell density as a result of ethanol treatment.








However, ethanol exposure significantly reduced ChAT+ neuronal size in males

compared with GC males but not SC males. No differences in ChAT+ neuronal size

were noted in females. Thus neonatal ethanol exposure, while producing long-lived

microencephaly and small changes in ChAT+ neuronal size, has no effect on the number

of cholinergic neurons in the adult rat MS, and has no effect on the size of the MS.

Introduction


Fetal alcohol syndrome (FAS) has been well characterized since its description in

1973 (Jones and Smith, 1973). Perinatal growth deficiencies, craniofacial abnormalities,

as well as central nervous system (CNS) dysfunction have been noted in human offspring

exposed to ethanol prenatally (West et al., 1994). In attempts to thoroughly examine and

accurately define the consequences of developmental ethanol exposure, rodent models of

FAS have been developed and extensively utilized. These animal models have identified

susceptible populations of neurons within the CNS and have described periods of

development during which particular populations exhibit heightened vulnerability to the

effects of ethanol (West et al., 1994). Studies in rats have demonstrated that the early

postnatal period, the so-called brain growth spurt and the equivalent of the human third

trimester (Dobbing and Sands, 1979), is a time of vulnerability for particular CNS

regions.

The septohippocampal (SH) system of basal forebrain afferents and hippocampal

targets is one such region in rats that exhibits susceptibility to ethanol during neonatal

development. The SH system is a pathway ofcholinergic and GABAergic fibers

originating from the MS nucleus and the horizontal and vertical limbs of the diagonal








band of Broca (DBB) that influences electrical activity in the hippocampus (Dutar et al.,

1995). These fibers originate from cell bodies in the basal forebrain and, beginning at

embryonic day 20, innervate the hippocampus where they form synapses on hippocampal

pyramidal neurons, granule cells, and intemeurons (Freund and Antal, 1988; Milner et al.,

1983). Cholinergic neurons comprise the major projection of the SH system and first

express the catalytic enzyme choline acetyltransferase (ChAT) on embryonic day 17

(Armstrong et al., 1987; Dutar et al., 1995).

While numerous studies have described the effects of neonatal ethanol exposure

on neurons within the hippocampus (Bonthius and West, 1990; Bonthius and West, 1991;

Greene et al., 1992; Pierce and West, 1987; West and Pierce, 1986) and have documented

deficits in spatial learning following neonatal ethanol exposure (Goodlett and Peterson,

1995; Kelly et al., 1988), the effect of ethanol exposure during the brain growth spurt on

the cholinergic component of the SH system has yet to be characterized. However,

studies examining the effect of chronic adult ethanol exposure on the neurons of the SH

system have been reported and have suggested a decrease in the number of cholinergic

neurons in the MS, though pair-fed controls were not examined (Arendt et al., 1995;

Arendt et al., 1988). Moreover, studies examining the effect of prenatal ethanol exposure

on the SH system in the rat (Swanson et al., 1995; Swanson et al., 1996) and mouse

(Schambra et al., 1990; Sulik et al., 1984) have been completed. When rats were exposed

to ethanol throughout gestation, a reduction in ChAT enzymatic activity was noted in the

MS during the second postnatal week compared with pair-fed controls (Swanson et al.,

1995).








In contrast, no change in the number or morphology of ChAT+ neurons in the MS

was found (Swanson et al., 1996), though earlier experiments in mice had shown the

cholinergic neurons of the basal forebrain to be particularly sensitive to acute prenatal

ethanol exposure (Schambra et al., 1990; Sulik et al., 1984). The objective of the present

study was to determine the long-term effects of ethanol exposure on the cholinergic

neurons in the rat MS when ethanol was delivered during the brain growth spurt, a time

when the cholinergic neurons are more mature and potentially more susceptible than at

the time points examined in the previous prenatal studies (Bonthius and West, 1991).

Ethanol was delivered from P4-10, a period of differentiation and synaptogenesis in the

SH system (Dutar et al., 1995). ChAT+ neuronal number and morphology was examined

in the adult MS of artificially reared ethanol-treated and pair-fed controls as well as in

dam-raised suckle controls.

Methods


Subjects and artificial rearing

Long-Evans hooded rat pups were obtained from nine timed pregnant dams

ordered from Charles River Co. (Portage, MI). Animals were housed with a 07:00-19:00

light cycle under controlled temperature and humidity conditions. At the time of birth

(PO), litters were culled to 12 pups and the pups were randomly assigned to one of three

groups: ethanol-receiving gastrostomized pups (EtOH); pair-fed gastrostomy controls

(GC); and dam-reared suckle controls (SC). Artificial rearing was performed after West

et al. (West et al., 1984). On P4 a gastrostomy feeding tube was surgically implanted into

the EtOH and GC pups. Pups were placed under methoxyfluorane anesthetic while








gastrostomy tubes were inserted down the esophagus into the stomach. The tube was

pulled through a small hole in the stomach and abdominal wall and secured on the outside

by a small plastic washer. Pups were reared individually in plastic cups filled with

bedding and a fur-like material. Cups were floated in a covered, heated aquarium (400C)

and pups were maintained on a 07:00-19:00 light cycle. SC pups were weighed daily

from P4-10. Dams with SC pups always had a total of eight pups to minimize weight

differences between groups.

Gastrostomized pups were infused with a liquid diet containing evaporated milk,

sterile water, soy protein, L-methionine, L-tryptophan, calcium phosphate, deoxycholic

acid, a vitamin mixture, and a mineral mixture (West et al., 1984). Pups received the

milk formula from P4-10 in 12 feeding periods of 20 minutes each. EtOH pups received

ethanol-supplemented formula as a 9.1% v/v solution for the first two feedings on each

day for a total of 4 g/kg/day. GC pups received an isocaloric amount of maltose-dextrin-

supplemented formula for the first two feedings on each day. The remaining 10 daily

feedings consisted of milk formula alone. Gastrostomized pups were weighed daily and

the daily diet consumption volume (in mls) was equivalent to 33% of the mean litter body

weight. The liquid diet was administered to the EtOH and GC pups by connecting the

gastrostomy tubes to a feeding line connected to diet-filled syringes held within a

Stoelting (Wood Dale, IL) programmable infusion pump. On the morning of P11, pups

were disconnected from the pump, feeding tubes were sealed, and they were returned to

their original dam. EtOH and GC pups were accepted by the dams and began nursing

soon after re-introduction to the home cage. At P21, pups were weaned and housed








individually until sacrifice and perfusion at P60. A total of 38 animals were used in this

analysis (EtOH, n = 12; GC, n = 10; SC, n = 16).

Tissue preparation and immunocytochemistry

Animals were selected for immunocytochemical staining at P60, an age which

represents adulthood in rats. This age was selected in order to examine the long-term

effects of neonatal ethanol treatment on ChAT expression in cholinergic neurons of the

medial septum and corresponds to an age used in previous analyses of prenatal ethanol

influences on cholinergic and GABAergic neurons in the medial septum (Moore et al.,

1997; Swanson et al., 1996). On P60, rats were anesthetized by pentobarbital overdose

just prior to perfusion. Animals were perfused transcardially with phosphate buffered

saline (PBS; 0.1M containing 0.9% sodium chloride) followed by 10% formalin (in the

same PBS). The brains were removed and post-fixed overnight in the same fixative. The

brain tissue was equilibrated overnight (or until it sank) in a cryoprotectant solution (PBS

containing 30% sucrose and 15% ethylene glycol) and stored frozen at -700C until

processed for immunocytochemistry.

Brains were prepared for frozen sectioning by thawing and equilibrating with a

30% sucrose-PBS solution. Frozen serial coronal sections throughout the forebrain were

cut at a thickness of 40 pm and processed free-floating. Alternate sections were

immunostained for choline acetyltransferase using a polyclonal antibody (rabbit anti-

human placental ChAT, Chemicon, Temecula, CA). Immunoreactivity was visualized as

a blue-black reaction product using an avidin-biotin conjugate/nickel intensified staining

(see below). Animals from each diet group were processed for immunocytochemistry at a








given time. This ensured against staining differences between groups resulting from

procedural differences.

Endogenous peroxidase activity was quenched by initially treating tissue sections

for 60 min at room temperature (RT) with PBS containing 3% hydrogen peroxide (H202)

and 10% methanol. Sections were pretreated with 0.4% Triton-X100 (T-X) in PBS for 30

min at RT, then blocked for 60 min with PBS containing 3% normal goat serum (NGS)

and 0.1% T-X. Primary incubation with the polyclonal ChAT antibody (1:750 in PBS +

3% NGS and 0.1% T-X) was carried out for 48 hr at 4C. Sections were then washed in

PBS containing 1% bovine serum albumin (BSA) and incubated with biotinylated goat

anti-rabbit IgG (1:1000 in PBS +1% BSA) overnight at 40C. Tissues were washed and

incubated with extravidin-horseradish peroxidase conjugate (Sigma, St. Louis, MO;

1:1000 in PBS + 1% BSA) overnight at 4C. Tissues were washed with 0.1M sodium

acetate (pH 7.2) to eliminate phosphate which can precipitate divalent cations. Sections

were then reacted for 3 min at RT using a developing buffer, pH 7.2, containing 0.8M

sodium acetate, 8 mM imidazole, 0.5% nickel (II) sulfate, 0.04% 3,3 diaminobenzidine

tetrahydrochloride, and 0.005% H202. The chromogen precipitation reaction was stopped

with 0.1M acetate buffer (pH 7.2) containing 0.1% sodium azide. Sections were mounted

in the same acetate buffer, air dried, dehydrated, and coverslipped. Control for

nonspecific staining was carried out in parallel with stained tissues with omission of the

primary antibody. Slides from individual animals were randomized and coded so that all

subsequent analyses were carried out blind with regard to the diet treatment group and

gender.








MS cell count, volume, and area analysis

ChAT-positive cell counts were conducted on alternate sections throughout the

rostral-caudal extent of the MS nucleus. The packing density of the ChAT+ neurons in

the DBB nucleus was too great for accurate cell counts to be made. Thus, for the

purposes of this study the MS alone was examined and was defined rostrally by the

ventral fusion of the hemispheres (at the level of the genu of the corpus callosum) and

caudally by the decussation of the anterior commissure

Images of ChAT immunostained sections were captured and digitized using a

RasterOps 24STV video capture board and software on a Macintosh computer. Low

magnification images (4x objective; effective scale 1.59 itm/pixel) were captured in order

to obtain the entire MS region in one image. When capturing each image, the lighting

and contrast enhancement were optimized for identification of individual cells. Images

were digitally processed using the image analysis program NIH Image (freeware from

NIH, Bethesda MD). To reduce background variation across the image, a digitally

defocused image (created by passing the primary image through a mean filter) was

subtracted from the primary digitized image. This processed image was then passed

through a Laplace filter to enhance edges and separation between cells.

Analysis of individual images was initiated by outlining the region of the MS

nucleus. The area of a section outlined varied depending on the rostral-caudal location of

the individual section. The MS nucleus in rostral sections was defined as the mediodorsal

group of neurons which were separated from the ventrally-located DBB nucleus. In

intermediate sections, where the demarcation between the MS and DBB nuclei is

ambiguous, the MS was defined ventrally by a line perpendicular to midline at the level








of the anterior commissure. In caudal sections, the MS was defined as the medially

located cells dorsal to the anterior commissure. Cells were highlighted interactively by

adjusting the grayscale threshold level to include only objects which were considered

cells. Given the thickness of each section (40 gm), the fact that alternate sections were

analyzed, the fact that ChAT+ neurons in the MS nucleus range from 10-25 um in

diameter, and the fact that an estimation of the total population was not desired, it was not

deemed necessary to perform a split-cell correction on these counts.

The program counted the highlighted objects and measured the area of the

outlined region (mm2). Because the MS is defined by clusters of ChAT+ cell bodies, the

area outlined was determined by the pattern of ChAT staining on each individual section.

MS volume (mm3) was calculated using a modification of the Cavelieri method (Michel

and Cruz, 1988) to ensure that changes in mean number of ChAT+ neurons per section as

a result of ethanol treatment were not due to changes in the size of the structure of interest

(Peterson et al., 1997). This was done by multiplying the number of 40 pim sections

analyzed for a given animal by the mean area per section for that animal and the section

thickness.

ChAT+ cell size determination

Because a previous study from our laboratory (Swanson et al., 1996) had analyzed

ChAT+ neuronal size using the computer instead of simple measurements of cell

diameters, we also chose to use the computer for these purposes so that comparisons

could be made to the previous work. For morphometric analysis, moderate magnification

images (10x objective; effective scale 0.978 pnm/pixel) were captured in order to

maximize the resolution yet minimize the loss of cells out of the focal plane of each








image. Primary images were not processed further prior to morphometric analysis. Cell

size measurements were performed on ChAT+ neurons in the MS nucleus. Three

sections through the MS were regionally matched in rostro-caudal extent for all animals.

25 non-overlapping cells per section were selected at random and measured. The cross

sectional somatic area of individual cells was highlighted interactively by adjusting the

grayscale threshold and was automatically measured by the computer. The program

computed cross-sectional area for each cell and mean somatic area was determined for

each animal.

Other analyses

The peak blood ethanol concentration (BEC) in the ethanol group was determined

on P7 two hours after the second ethanol infusion (Goodlett et al., 1990) utilizing the

Sigma 333-UV kit. Body and brain weights were recorded for individual animals at P60

following perfusion. Density of ChAT+ neurons was determined by dividing the mean

number of cells per section by the mean area (mm2) per section.

Statistical analyses

Statistical analysis was performed with the program StatView (Abacus Concepts,

Berkeley Ca). For each parameter examined, one-way Analysis of Variance (ANOVA)

was used to test for the main effects of diet and gender separately. When appropriate, the

data were further analyzed with the Fisher's Protected Least Significant Difference

(PLSD) post hoc test to determine individual group differences. Because gender

differences were noted for many measures (following an initial two-way ANOVA for

effects of diet and gender), males and females were analyzed separately.








Results


BEC, body and brain weights, and brain/body weight ratios

The mean BEC in EtOH pups two hours following the last ethanol treatment on

P7 was 269 + 23 mg/dl (n=5). Table 3-1 (all tables are located at the end of the chapter)

presents the mean P60 body and brain weights, and brain/body weight ratios (br/bd) for

all diet groups separated by gender. The ANOVA for P60 body weight revealed no

significant effects of diet on P60 body weight of males or females. A gender difference

was noted in P60 body weight for EtOH [F(1,22) = 45.2; p< 0.0001], GC [F(1,14) = 38.5;

p< 0.0001], and SC [F(l,16) = 40.8; p< 0.0001] animals which was expected given the

larger size of males at P60.

The ANOVA for P60 brain weight demonstrated an effect of diet treatment in

males [F(2,18) = 12.3; p< 0.001] and in females [F(2,25) = 6.69; p< 0.01]. The PLSD

revealed that EtOH male brains weighed less than GC male (p< 0.01), and SC male (p<

0.01) brains. EtOH female brains also weighed less than GC female (p< 0.05), and SC

female (p< 0.01) brains. No significant gender differences were noted for brain weight

within EtOH, GC, or SC animals.

The ANOVA for br/bd showed an effect of diet in males [F(2,18) = 5.16; p< 0.05]

and females [F(2,25) = 11.3; p< 0.001]. The PLSD demonstrated that EtOH males had

reduced br/bd compared with GC males (p< 0.01) while EtOH females had reduced br/bd

compared with both GC females (p< 0.001) and SC females (p< 0.001). Gender

differences were also noted in P60 br/bd for EtOH [F(1,22) = 37.5; p< 0.0001], GC








[F(1,14) = 41.2; p< 0.0001], and SC [F(1,16) = 38.5; p< 0.0001] animals, which was not

surprising given the larger size of male animals at this age.

Number of sections, MS area per section, and ChAT+ neuronal density

Table 3-2 presents number of sections, MS area per section, and ChAT+ neuronal

density measurements for each diet group for either gender. The number of alternate 40

gm sections through the MS was determined by counting the number of sections analyzed

between the ventral fusion of the hemispheres (at the level of the genu of the corpus

callosum) and the anterior commissure. The ANOVA showed no effect of diet for either

males or females on the mean number of alternate sections through the MS. Additionally,

no significant gender differences were noted in the number of alternate sections for EtOH,

GC, or SC animals.

As mentioned above, the computer program determined the area of the outlined

MS on each individual section. The mean MS area per section was determined for each

animal. The ANOVA revealed no significant effect of diet in males or females on the

mean MS area per section. Similarly, no significant gender differences were noted in MS

area per section for EtOH, GC, or SC animals.

ChAT+ neuronal density was calculated by dividing the mean number of ChAT+

neurons per section by the mean area per section. The ANOVA demonstrated no

significant effect of diet in males or females on ChAT+ neuronal density. Also, no

significant gender differences were noted in ChAT+ neuronal density for EtOH, GC, or

SC animals.








Cholinergic cell counts in the MS

The total number of ChAT+ neurons counted on alternate sections through the

P60 MS was determined and is presented in figure 3-1 (all figures are located at the end

of the chapter). The ANOVA for total number of ChAT+ cells demonstrated an effect of

diet in males [F(2,18) = 4.45; p< 0.05] but not females. The PLSD further revealed that

EtOH males contained 24 % fewer MS ChAT+ neurons than SC males (p< 0.05). No

significant gender differences were noted in mean total number of ChAT+ neurons for

EtOH, GC, or SC animals.

In order to normalize for volume, the average number of ChAT+ neurons per

section was determined. Figure 3-2 presents the mean number of ChAT+ neurons per

section. The ANOVA for mean number of ChAT+ neurons per section revealed an effect

of diet on males [F(2,18) = 3.34; p< 0.05] but not females. The PLSD further

demonstrated that EtOH males contained 19 % fewer ChAT+ neurons per section than

SC males (p< 0.05), while the 14 % difference between GC males and SC males was not

significant. No significant gender differences were noted in mean ChAT+ cells per

section for EtOH, GC, or SC animals.

Cholinergic neuronal size

Figure 3-3 presents the mean somatic cross sectional area for ChAT+ neurons

throughout the P60 MS. The ANOVA for mean somatic cross sectional area determined

an effect of diet in males [F(2,18) = 5.76; p< 0.05] but not females. The PLSD further

revealed that GC males contained an increased somatic cross sectional area compared

with SC males (p< 0.05), while EtOH males contained a reduced somatic cross sectional








area compared with GC males (p< 0.01). No significant gender differences were noted in

ChAT+ neuronal size for EtOH, GC, or SC animals.

MS volume

MS volume was determined by multiplying the number of 40 Jpm sections

analyzed for a given animal by the mean area per section for that animal. Figure 3-4

presents the mean MS volume of P60 rats. The ANOVA demonstrated no effect of diet

in males or females on the mean number of alternate sections through the MS. No

significant gender differences in MS volume were noted for EtOH, GC, or SC animals.

Discussion


The major conclusion of the present work is that small reductions in the size of

ChAT+ neurons of the medial septum are produced by neonatal ethanol exposure while

changes in the number of cholinergic neurons or size of the MS are not noted. This study

has also determined that long-lasting microencephaly is found in male and female P60

rats exposed to ethanol via artificial rearing from P4-10.

Because EtOH animals showed no difference when compared with GC animals in

total ChAT+ neuronal number or mean number of cholinergic neurons per section,

ethanol exposure did not directly reduce ChAT+ neuronal number. However, since EtOH

males contained fewer cholinergic neurons when compared with SC males it appears that

ethanol in combination with artificial rearing does affect these neurons. No changes were

noted in mean area per section, cell density, or MS volume as a result of ethanol

treatment in males. Female animals similarly exposed to ethanol did not contain a

reduced number of ChAT+ MS neurons and no changes were noted in MS volume, mean








area per section, or cell density. A permanent reduction in the size of ChAT+ neurons

was noted in EtOH males compared with GC males. However, this was concurrent with

an increase in size of cholinergic neurons in GC compared with SC males. Moreover, no

differences in cell size were found in female rats, making it difficult to determine whether

the reduction in ChAT+ cell size in EtOH males is an effect of ethanol on neuronal size

or merely an effect of diet and artificial rearing on GC males. Thus neonatal ethanol

exposure, while producing long-lived microencephaly, has little effect on the cholinergic

neurons of the adult rat MS, as measured by ChAT+ neuronal number and morphology.

Long-term effects of ethanol on brain growth

Microcephaly and microencephaly are hallmarks of FAS as well as neonatal

animal models of FAS (West and Pierce, 1986). Reduced brain size represents one of the

most reproducible features of ethanol-induced alterations in the brain, and the current

study is no exception. Artificial rearing from P4-10 produced no long-lasting deficits in

body weight and by P60 EtOH and GC pups weighed the same as SC pups. Despite the

lack of effect on body weight, EtOH animals of both genders showed reduced brain

weight at P60 compared with GC and SC animals. Brain to body weight ratio (br/bd) was

similarly reduced in both males and females exposed to ethanol. Taken together, the

reduced brain weight of EtOH animals and the reduced br/bd in EtOH animals

demonstrate that neonatal ethanol exposure produced long-lasting microencephaly.

Goodlett et al. (Goodlett et al., 1991) exposed rats from P4-9 in two of the 12 daily

feedings as a 10.2% solution (an exposure paradigm similar to the current experiments).

Exposure to this concentration of ethanol resulted in a peak BEC of 361 mg/dl (compared

with 269 mg/dl in the current study) and produced microencephaly in both genders in








adult animals. The current study demonstrates that even with a lower peak BEC, brain

growth is severely compromised.

The degree of ethanol-induced reduction in brain weight depends directly on the

peak BEC produced. Bonthius and West (1991) provided an illustration of this by

comparing the effects of a 6.6 g/kg/day ethanol treatment spread out over 12 feedings

from P4-11 with the effects of 4.5 g/kg/day ethanol treatment given in just two feedings.

Animals were sacrificed at P90 to observe long-term effects of neonatal ethanol delivery

on brain weight. Interestingly, they found that the lower 4.5 g/kg/day dose produced

profound brain weight and brain to body weight ratio deficiencies, while the 6.6 g/kg/day

group exhibited no such microencephaly. This differential response to varying ethanol

delivery patterns was directly related to peak BEC as the higher dose spread out over 12

feedings produced a peak BEC of only 43 mg/dl while the lower dose condensed into two

feedings produced a peak BEC of 318 mg/dl (Bonthius and West, 1991). Thus, the

pattern of exposure plays a critical role in the severity of ethanol-induced teratogenicity.

Long-term effects of ethanol and artificial rearing on ChAT+ neurons in the MS

Artificial rearing and ethanol exposure produced a decrease in the mean ChAT+

neuronal number per section in EtOH males compared with SC males in the absence of

volume, area, or density changes. Female animals displayed no differences in mean

cholinergic neuronal number per section, volume, area, or density measurements and,

thus, the reduction occurs in a sexually dimorphic manner. Because the reduction in

ChAT+ neuronal number noted in males was between the gastrostomized EtOH group

and the dam-reared SC group (and not the pair-fed GC group) the reduction appears to be

the result of an interaction between ethanol treatment and some characteristic of the








artificial rearing technique rather than the result of ethanol alone. Kelly (1997) has noted

an effect of artificial rearing on the concentration of neurotransmitters in the SH system

of male rats: neurotransmitter concentration was reduced in artificially reared males

independent of ethanol exposure. While the current results do not implicate artificial

rearing in reducing cholinergic neuronal number (the difference between GC and SC

males in mean ChAT+ neurons per section is not significant) they do suggest that ethanol

in combination with artificial rearing can reduce ChAT+ neuronal number. Perhaps an

interaction between ethanol and the stress of the procedure (from invasiveness or

maternal separation) accounts for the reduced cholinergic neuronal number in artificially

reared and ethanol exposed pups.

Developmental ethanol exposure and ChAT+ neuronal number

The current finding that cholinergic neuronal number in the MS is unaffected by

ethanol alone is consistent with previous experiments from our laboratory. Swanson et

al. (1995; 1996) demonstrated reduced ChAT activity in P14 rats chronically exposed to

ethanol in utero. However, analysis of ChAT+ neuronal number in the MS of P14 and

P60 rats revealed no effect of diet treatment on cholinergic cell number. Similarly,

evidence from Heaton et al. (1996) showed that ChAT+ neuronal number is not

significantly affected in the striatum of male or female rats following CPET. A transient

increase in cholinergic neuronal number was noted in the striatum on P14, but this was

reduced to control levels by P60.

While studies from our laboratory (see above) have not indicated an effect of

developmental ethanol exposure on cholinergic neuronal number in the rat MS or

striatum, studies from other groups have demonstrated ethanol-induced alterations in








ChAT+ number, particularly in the MS. For example data suggest that acute prenatal

exposure of mice to ethanol resulted in severe midline anomalies at embryonic day 18,

including loss of basal forebrain neurons (Schambra et al., 1990) as well as loss of septal

neurons and reduced volume at P 15 (Ashwell and Zhang, 1996). However, it should be

noted that these experiments produced high acute BECs (600-700 mg/dl) which were the

result of one or two ethanol exposures early in embryonic development. Thus,

differences in timing and dosage may account for the apparent discrepancy (Swanson et

al., 1996).

Exposure of adult rats to ethanol for 12 or 28 weeks has been reported to reduce

cholinergic neuronal number in the MS (Arendt et al., 1995; Arendt et al., 1988).

However, it should be noted that comparisons were made between rats exposed to ethanol

as a 20% solution of drinking water and rats fed chow and water ad libitum; no pair-fed

controls were examined. Thus, it is possible that the reduction of cholinergic neurons in

the MS is not due to the influence of ethanol alone. For instance, an interaction between

ethanol and malnutrition may explain the noted differences. However, it is conceivable

that chronic ethanol exposure during mature stages may reduce the number of MS

cholinergic neurons, as found by Arendt et al. (1995; 1988). Regardless, the current data

in combination with the Swanson et al. (1996) and Heaton et al. (1996) studies suggests

that cholinergic neuronal number in the MS is not susceptible to moderately-high or

moderate doses of ethanol during early postnatal or prenatal development, respectively.

A future study examining ChAT+ neuronal number after exposure to a higher peak BEC

would determine whether the apparent lack of effect of ethanol on cholinergic neurons is

due to the moderate doses used in these studies.








Developmental ethanol exposure and ChAT+ neuronal size in the MS

Although ChAT+ neuronal counts were not directly affected by ethanol, analysis

of the mean somatic cross sectional area of cholinergic neurons in the P60 MS

demonstrated a diet effect in male animals. Specifically, neuronal size was reduced in

EtOH males compared with GC males. However, GC males contained larger neurons

compared with SC males and no differences were found in females, making it difficult to

determine whether this is an effect of ethanol on cholinergic neuronal size or merely an

effect of diet and artificial rearing on GC males. It is unlikely that the smaller size of

cholinergic neurons in EtOH males is simply because of the noted microencephaly since

the brains of EtOH and SC males were different in size but the mean size of ChAT+

neurons was similar.

It is possible, however, that GC males experienced an increase in somal size in

response to either the stress of the artificial rearing procedure (such as maternal

separation, isolation, or the invasive surgery) or to some component of the maltose-

dextrin containing milk diet. Factors such as nerve growth factor (NGF) can induce

hypertrophy of ChAT+ basal forebrain neurons (Higgins et al., 1989). NGF is also

known to increase the size of cholinergic neurons in cultures of rat septal neurons

(Markova and Isaev, 1992), and is able to reverse axotmy-induced decreases in MS

cholinergic neuronal cell bodies (Hagg et al., 1989). It is conceivable that male GC

animals increased the expression of NGF, as is known to occur following stress (Smith,

1996), resulting in larger cholinergic MS neurons. However, female GC animals were

similarly stressed and did not display hypertrophy of ChAT+ neurons. Perhaps a more

likely explanation is that a constituent of the liquid diet acted in atrophic manner on these








neurons. Whatever the explanation, these data are consistent with data from other

investigators who have reported sexually dimorphic effects on the size of cholinergic

neurons of the MS. For example, neonatal hyperthyroidism is capable of increasing

ChAT+ neuronal size in the MS of male animals while simultaneously decreasing the size

of ChAT+ neurons in females (Westlind-Danielsson et al., 1991).

One other report from our laboratory has examined cholinergic neuronal size at

P60 following chronic prenatal ethanol exposure (Swanson et al., 1996). While the

exposure times were different in the current study, it is useful to compare the present data

with the Swanson et al data. The size of the ChAT+ neurons at P60 is quite similar. For

example, the range of neuronal size in the Swanson et al. study was 90-100 pm2, which is

comparable to the range noted in the current study (80-110 im2). However, in contrast to

the present data, Swanson et al. (1996) found no ethanol-induced alterations in ChAT+

neuronal size, and found no gender differences.

ChAT+ neuronal insusceptibility: possible role of neurotrophic factors

Why is ChAT+ neuronal number apparently unaltered following developmental

ethanol exposure? One possible explanation is the importance of neurotrophic factors in

regulating the development and maintenance of the SH system. The hippocampus is

known to express a rich variety of neurotrophic factors, including NGF, brain-derived

neurotrophic factor (BDNF), neurotrophin-3, and basic fibroblast growth factor (Ernfors

et al., 1990; Maisonpierre et al., 1990). The high affinity receptors for the various

neurotrophins found in the hippocampus, including trk A, trk B, and trk C, are also

present in the developing septum, including the cholinergic neurons of the medial septum

(Ringstedt et al., 1993). Evidence suggests a role for neurotrophins in regulating the








expression of ChAT in the basal forebrain. Intraventricular injection of NGF antibodies

into rat neonates reduces ChAT immunostaining in the septum (Vantini et al., 1989).

Moreover, mice which have been genetically engineered to lack the various neurotrophin

receptors have been generated, and have further defined a role for neurotrophins in

sustaining septal cholinergic neurons in vivo. For example, mice which lack trk A show

reduced acetylcholinesterase activity in septal projection fibers (Smeyne et al., 1994), and

mice which lack BDNF show reduced density of cholinergic neurons in the MS (Jones et

al., 1994).

In vitro studies have also demonstrated a role for neurotrophins in sustaining

septal neurons in normal conditions and in the presence of ethanol. For example, NGF

stimulates ChAT activity, neuritic complexity, increased fiber length, and increased fiber

outgrowth in cultures of septal cholinergic neurons (Hartikka and Hefti, 1988).

Moreover, studies from our laboratory have demonstrated that NGF protects both

cultured dorsal root ganglion neurons and septal neurons from ethanol neurotoxicity

(Heaton et al., 1993; Heaton et al., 1994). It is conceivable that neurotrophic interactions

are preserved in the current study, and that ethanol is unable to produce toxic effects on

cholinergic neurons in the medial septum because these normal interactions are

maintained. A consequence of this maintenance may be growth factor-mediated

induction of protective molecules of the bcl-2 family of cell death molecules which would

prevent ethanol-induced toxicity in this population. Similar mechanisms are known to

operate in cerebellar granule cells during development. For example, Muller et al. (1997)

demonstrated that NGF induces Bcl-2 protein expression and that this induction resulted

in cell survival.