The Role of estradiol as a neurotrophomodulatory substance for basal forebrain cholinergic neurons of the female Sprague...

MISSING IMAGE

Material Information

Title:
The Role of estradiol as a neurotrophomodulatory substance for basal forebrain cholinergic neurons of the female Sprague-Dawley rat
Physical Description:
xiii, 123 leaves : ill. ; 29 cm.
Language:
English
Creator:
Singh, Meharvan, 1967-
Publication Date:

Subjects

Subjects / Keywords:
Research   ( mesh )
Estradiol -- pharmacology   ( mesh )
Estradiol -- physiology   ( mesh )
Learning   ( mesh )
Memory   ( mesh )
Brain -- physiology   ( mesh )
Choline -- metabolism   ( mesh )
Choline O-Acetyltransferase -- metabolism   ( mesh )
Nerve Growth Factors -- metabolism   ( mesh )
Brain-Derived Neurotrophic Factor -- metabolism   ( mesh )
Rats, Sprague-Dawley   ( mesh )
Department of Pharmacodynamics thesis Ph.D   ( mesh )
Dissertations, Academic -- College of Pharmacy -- Department of Pharmacodynamics -- UF   ( mesh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1994.
Bibliography:
Bibliography: leaves 110-122.
Statement of Responsibility:
by Meharvan Singh.
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 - 002359197
notis - ALW3702
oclc - 81400150
System ID:
AA00011788:00001

Full Text











THE ROLE OF ESTRADIOL AS A NEUROTROPHOMODULATORY
SUBSTANCE FOR BASAL FOREBRAIN CHOLINERGIC NEURONS OF
THE FEMALE SPRAGUE-DAWLEY RAT












By

MEHARVAN SINGH


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


1994





























I dedicate this dissertation to my parents, Dr. Surjit Singh and Avtar Kaur, and my sister,
Harpreet Kaur, for their continuing support and love that helped make this possible.










ACKNOWLEDGEMENTS


I would like to thank several people, without whom my tenure as a graduate

student would not have been as enjoyable. I would like to thank Dr. James Simpkins,

my mentor and friend, for his immeasurable advice and support for the past 5 years.

His selfless attitudes toward graduate education led to much of my success and

commands my undying respect and gratitude. I would also like to thank the other

equally important members of my committee, Dr. William 'Billie' Millard, Dr. Ralph

Dawson, Dr. Joanna Peris and Dr. Edwin Meyer, for their helpful advice, expertise

and use of their lab facilities. They were always available for consultation or simply a

good chat, which accentuated my positive experience in graduate school.

I would also like to thank Dr. Simpkins' family, Janet, Chris, Gretchen and Lexi

Simpkins, Dr. Millard's family, Beverly, Jason and Justin Millard, and Dr. Dawson's

family, Janice, Alecia and Ralph Jr. for the hospitality and friendship extended to me

over the past years. I am grateful to a number of graduate students, technicians,

faculty and staff who helped in various stages and aspects of my graduate education

and research.

I would like to thank Meri Layden and the late Terry Romano for their assistance

and patience during my early stages of graduate school. I would also like to thank

Jean Bishop, Elizabeth Simpson, Fred Huang, NancyEllen deFiebre, Rhonda Felheim,

Doug Swanson, Manuel Perez, Dr. Tucker Patterson, Dr. Melanie Pecins-Thompson,

Dr. Ping Wu, Dr. Eileen Martin, Dr. Chris deFiebre, Jim D'Arcangelis, Anthony Duva

and Juliet Burry for their technical assistance, scientific input and their friendhsip. I

would also like to thank Dr. Kevin Anderson for his technical assistance in making

the striking slides of the in situ data. A special thanks to Victoria Redd, Roxanne








Federline, B.J. Goins and Linda Panchou for their expert assistance in helping me

deal with the bureaucracy of the system.

I would also like to thank all the other graduate students, faculty, post-docs and

staff in the department of Pharmacodynamics who collectively made my tenure as a

graduate student in the department a very pleasurable and a productive experience. I

would also like to thank my second mom, Cynthia Sweat and her two children,

Allyson and Clifton for their support and caring. Finally, I'd like to give very special

thanks to my parents, Dr. Surjit Singh and Avtar Kaur, and my sister, Harpreet Kaur,

who provided great support throughout my academic career.















TABLE OF CONTENTS


ACKNOWLEDGEMENTS ......................................................................... iii

LIST OF FIGURES.................................................................................... viii

LIST OF TABLES...................................................................................... x

LIST OF PLATES ...................................................................................... xi

ABSTRACT ............................................................................................... xii

CHAPTERS

1 INTRODUCTION 1

Review of Literature........................................................... 2
Estrogen Physiology..................................................... 3
The Cholinergic System............................................... 6
The Cholinergic System and Cognitive Function ......... 8
Neurotrophism and Neuronal Plasticity........................ 9
Mechanism of Action............................................. 10
Distribution of Neurotrophin Receptor Positive
Neurons......................................................................... 12
Importance of Neurotrophins in Development
and Adulthood.............................................................. 14
Estrogens and Neuronal Plasticity................................ 16

2 GENERAL METHODS.............................................................. 19

Animals..................................................................................... 19
Animal Surgical Procedures..................................................... 19
Preparation of Estradiol Pellets................................................ 20
Methods of Blood Sampling ..................................................... 21
Serum Estradiol Assay .............................................................. 21
Vaginal Smears......................................................................... 21
Tissue Dissections.................................................................... 22

3 THE EFFECT OF OVARIECTOMY AND ESTRADIOL
REPLACEMENT ON LEARNING AND MEMORY........... 24

Introduction............................................................................... 25
Methods.................................................................................... 25
Active Avoidance......................................................... 25








Morris Water Task........................................................... 26
Statistical Analysis........................................................... 27
Results.................................................................................... 27
Estradiol Concentrations............................................... 27
Active Avoidance Behavior.......................................... 28
Morris Water Task........................................................ 29
Discussion................................................................................. 37

4 THE EFFECT OF OVARIECTOMY AND ESTRADIOL
REPLACEMENT ON CHOLINERGIC FUNCTION:
HIGH AFFINITY CHOLINE UPTAKE.................................. 41

Introduction.................................................................................. 41
Methods.................................................................................... 42
High Affinity Choline Uptake...................................... 43
Protein Assay................................................................ 44
Statistical Analysis........................................................ 44
Results....................................................................................... 44
Discussion................................................................................. 51

5 THE EFFECTS OF OVARIECTOMY AND ESTRADIOL
REPLACEMENT ON CHOLINERGIC FUNCTION:
CHOLINE ACETYLTRANSFERASE.................................... 54

Introduction............................................................................... 54
Methods.................................................................................... 55
ChAT Assay.................................................................. 55
Statistical Analysis........................................................ 57
Results....................................................................................... 57
Optimization of the Assay............................................ 57
Choline Acetyltransferase Activity............................... 57
Discussion................................................................................. 62

6 THE EFFECT OF OVARIECTOMY AND ESTRADIOL
REPLACEMENT ON NGF PROTEIN
AND mRNA LEVELS ............................................................. 65

Introduction............................................................................... 65
Methods.................................................................................... 66
NGF Protein Measurement........................................... 66
Extraction of NGF Protein............................................ 67
Measurement of NGF mRNA....................................... 69
RNA Isolation............................................................... 69
Preparation of the NGF Probe...................................... 70
Hybridization of the Nylon Membrane......................... 71
Statistical Analysis........................................................ 72
Results....................................................................................... 72
Discussion................................................................................. 81

7 THE EFFECT OF OVARIECTOMY AND ESTRADIOL-
REPLACEMENT ON BDNF mRNA LEVELS: AN IN SITU
HYBRIDIZATION STUDY.................................................... 84

Introduction.............................................................................. 84








M ethods ...................................................................................... 85
Preparation of Tissue.................................................... 86
In Situ H ybridization.................................................... 87
Preparation of BDNF Probe......................................... 88
Quantitation of BDNF mRNA..................................... 89
Statistical A nalysis....................................................... 89
Results...................................................................................... 90
D iscussion................................................................................ 97

8 GENERAL DISCUSSION....................................................... 101
Clinical Perspectives..................................................... 107

REFERENCES.................................................................................. 110

BIOGRAPHICAL SKETCH............................................................. 123
















LIST OF FIGURES


FIGURE

3-1 Learning Curves from Individual Animals following 5
Weeks of Ovariectomy in Intact, Ovariectomized and
Estradiol Replaced Animals ............................................. 34

3-2 Learning Curves from Individual Animals following 28
Weeks of Ovariectomy in Intact, Ovariectomized and
Estradiol Replaced Animals ............................................. 35

3-3 Active Avoidance Performance Following 5 and 28
Weeks of Ovariectomy in Intact, Ovariectomized and
E2 Replaced Animals ....................................................... 36
4-1 The Effect of 5 weeks of Ovariectomy and Two Different
Estradiol Replacement Regimens on High Affinity Choline
Uptake in Behaviorally Tested Rats ................................. 46

4-2 The Effect of Learning Performance on High Affinity
Choline Uptake in the Frontal Cortex and the Hippocampus
of 5 Week Treated Rats .................................................... 47

4-3 The Effect of 28 Weeks of Ovariectomy and Estradiol
Replacement on High Affinity Choline Uptake in the
Frontal Cortex and Hippocampus of Behaviorally Tested
Rats ................................................................................... 48

4-4 The Effect of Learning Performance on High Affinity
Choline Uptake in the Frontal Cortex and the Hippocampus
of 28 Week Treated Rats .................................................. 49
4-5 The Effect of 5 Week Ovariectomy and Estradiol
Replacement on High Affinity Choline Uptake in
Behaviorally Naive Rats .................................................. 50
5-1 Tissue Optimization of the ChAT assay .......................... 59

6-1 The Effect of Different Dilutions of Secondary Ab on the
NGF ELISA Standard Curve at a 1:30 Dilution of the
Coating Ab ....................................................................... 74
6-2 The Effect of Different Dilutions of Secondary Ab on the
NGF ELISA Standard Curve at a 1:60 Dilution of the
Coating Ab ....................................................................... 75








FIGURE


6-3 The Effect of Different Dilutions of Secondary Ab on the
NGF ELISA Standard Curve at a 1:120 Dilution of the
Coating Ab ....................................................................... 76

6-4 The Effect of Different Dilutions of Secondary Ab on the
NGF ELISA Standard Curve at a 1:240 Dilution of the
Coating Ab ....................................................................... 77

6-5 A Low Melting Temperature Agarose Gel for the
Separation of Plasmid DNA from the NGF Probe ........... 78

6-6 The Effect of 3 Month Ovariectomy and E2 Replacement
on NGF Protein Levels in the Frontal Cortex and
Hippocampus .................................................................... 79

6-7 The Effect of 3 Month Ovariectomy and E2 Replacement
on Frontal Cortical and Hippocampal NGF mRNA
Levels................................................................................ 80
















LIST OF TABLES


TABLE

3-1 Plasma Levels of Estradiol in Intact, Ovariectomized
and Estradiol-Replaced Animals ...................................... 31

3-2 Effect of Short-Term and Long Term Ovariectomy and
Estradiol-Replacement on Learning and Retention ............. 32

3-3 The Effect of Ovariectomy and Estradiol Replacement on
Morris Water Task Performance ...................................... 33
5-1 Effect of Short Term and Long Term Ovariectomy and
Estradiol Replacement on Choline Acetyltransferase
Activity in the Frontal Cortex .......................................... 60

5-2 Effect of Short Term and Long Term Ovariectomy and
Estradiol Replacement on Choline Acetyltransferase
Activity in the Hippocampus ........................................... 61

7-1 The Effect of 9 Weeks of Ovariectomy and Estradiol
Replacement on BDNF mRNA ....................................... 91

7-2 The Effect of 9 Weeks of Ovariectomy and Estradiol
Replacement on BDNF mRNA in the Hippocampus .......... 92

7-3 The Effect of 28 Weeks of Ovariectomy and Estradiol
Replacement on BDNF mRNA in Frontal, Parietal and
Temporal Cortices............................................................. 93

7-4 The Effect of 28 Weeks of Ovariectomy and Estradiol
Replacement on BDNF mRNA in the Hippocampus .......... 94















LIST OF PHOTOGRAPHIC PLATES


PLATE

7-1 Autoradiographic Images of Brain Sections of Intact, OVX
and E2 Replaced Animals Following 9 Weeks of
Ovariectomy ..................................................................... 95

7-2 Autoradiographic Images of Brain Sections of Intact, OVX
and E2 Replaced Animals Following 28 Weeks of
Ovariectomy ..................................................................... 96














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

THE ROLE OF ESTRADIOL AS A NEUROTROPHOMODULATORY SUBSTANCE
FOR BASAL FOREBRAIN CHOLINERGIC NEURONS OF THE FEMALE
SPRAGUE-DAWLEY RAT

by

Meharvan Singh

April, 1994

Chairman: Dr. James W. Simpkins
Major Department: Pharmacodynamics

Estradiol's (E2) role as a neurotrophic substance for basal forebrain cholinergic

neurons was investigated in female rats. The effect of ovariectomy (OVX) and E2

replacement (E2 pellet) was assessed on cognitive function and neurochemical indices of

cholinergic function. Ovary-intact (INTACT) animals served as controls.

Learning/memory was assessed using the two-way active avoidance and Morris water

task (MWT) paradigms. Animals OVXed for 5 weeks were learning impaired relative to

E2 pellet animals. When retested at 28 weeks, active avoidance performance was lower in

OVX animals relative to INTACT animals. E2 pellet animals maintained superior

performance relative to OVX animals and showed accelerated rates of learning. MWT

performance was not significantly affected by estrogen despite a trend towards better

performance in E2 animals.

Neurochemical analyses revealed that 5 weeks of ovariectomy reduced high affinity

choline uptake (HACU) in the frontal cortex (CTX) and hippocampus (HIPP) by 24%

and 34%, respectively. E2 replacement elevated HACU in both regions. Choline

acetyltransferase (ChAT) activity was decreased in HIPP but not CTX of 5 week OVX













animals. E2 replacement reversed this effect. At 28 weeks, an unexpected decrease in

ChAT activity was observed across all treatment groups. Interestingly, E2 animals

demonstrated the least severe decline in ChAT, suggesting a previously unreported

cytoprotective effect of E2.

In order to assess whether E2 exerts its neurotrophic properties through classical

neurotrophic substances, the modulation of nerve growth factor (NGF) and brain-derived

neurotrophic factor (BDNF) by E2 was investigated. Three months of OVX reduced NGF

mRNA levels in CTX while HIPP levels were unaffected. Similarly, 9 weeks of OVX

reduced BDNF mRNA in the frontal and temporal cortices while hippocampal levels

were only minimally affected. At the 28 week time point, cortical reductions persisted

and BDNF levels were significantly reduced in hippocampal CA 3, CA 4 and dentate

gyrus (DG). E2 replacement improved the deficit in the frontal cortex and DG of the

hippocampus.

Collectively, these results demonstrate the importance of estrogen on basal forebrain

cholinergic neurons. Since Alzheimer's disease is characterized by widespread

cholinergic deficits, E2 replacement may be of therapeutic importance, especially in post-

menopausal women.














CHAPTER 1

INTRODUCTION


Alzheimer's disease (AD) is a progressive neurodegenerative disorder whose

etiology is presently unknown. Histologically, it is characterized by the widespread

presence of senile plaques (spherical masses of degenerating neurites and reactive

cells around a core of amyloid) and neurofibrillary tangles (large intraneuronal

accumulations of paired helical filaments). While the presence of senile plaques (SP)

and neurofibrillary tangles (NFr) in the non-demented elderly is not uncommon

(Roth, 1986), the widespread abundance of these neuronal abnormalities are known to

correlate well with the degree of choline acetyl transferase and acetylcholinesterase

decrease and hence, postulated to correlate with the degree of dementia (Mountjoy,

1986; Perry et al., 1978).

Clinical presentation of this syndrome consists of a deterioration in intellectual

function, particularly in cognition and memory, and secondarily, in social aptitude

and emotional processes. Unlike 'normal aging', the deterioration of memory in AD is

significantly accelerated. Several hypotheses have been put forth to explain the

combined behavioral deficits and neuropathology seen in AD. Many of these make

parallelisms between the aging process and AD. In attempts to understand

degeneration in the aging process, researchers propose one of two general etiologies.

The first makes the assumption that deterioration of the central nervous system (CNS)

leads to subsequent disruption of function in the periphery (so called "downstream

dysfunction"). The second possibility is the converse of the above statement. That is,

disruption in peripheral organ function (such as decreased hormonal release or








synthesis from the organ of origin) leads to improper function and subsequent death

of the CNS (so called "upstream dysfunction"). Unfortunately, either CNS or

physiological dysfunction could contribute to a vicious cycle wherein degeneration is

seen in either aged individuals and those afflicted with AD.

It is also apparent that AD affects a multitude of neurotransmitter systems.

However, probably the most widespread and consistent deficit seen is that of the basal

forebrain cholinergic system; a population of neurons that have been implicated in the

regulation of memory function and cognition. Even so, the neurochemical substrate

for the cognitive dysfunction in AD is only partially understood. Even less well

characterized is the hormonal involvement in this neurodegenerative disease.

Cursory evidence does exist in the literature suggesting the role of estrogens in

cognitive function. It has been documented that women have an approximately 3-fold

higher incidence of AD than men, even after accounting for age, educational level and

socioeconomic status (Aronson et al., 1990). Taken together with some basic science

data that describes the effect of estrogen on nerve cell function, it was postulated that

estrogens may play an important role in the maintenance and preservation of basal

forebrain cholinergic neurons, and as such, may provide significant therapeutic merit

for certain neurodegenerative diseases such as AD.




Review of Literature


Estrogens are known to regulate several neuroendocrine functions of the

hypothalamus relating to reproduction. Recently, however, much focus has been

placed on extrahypothalamic roles of estrogens on both male and female brains.

Furthermore, the ability of the brain to synthesize steroids independently of the

adrenals or gonads is suggestive of the general importance of steroid hormones on








central nervous system function (Robel et al., 1987; Schlinger and Arnold, 1992). In

this thesis, I describe the effects of estradiol on learning and memory and correlate

these findings to estradiol's importance in maintaining the functional integrity of basal

forebrain cholinergic neurons, known to be involved in cognitive processes. In

elucidating the mechanism of estradiol's cytoprotective action, I propose that estradiol

may act either directly, or indirectly, through the modulation of certain compounds

classified broadly as belonging to the family of neurotrophins, examples of which are

nerve growth factor (NGF) and brain derived neurotrophic factor (BDNF).

The menopause, a process that represents the end of the reproductive phase in a

woman's life, is characterized by a precipitous decline in circulating levels of

estrogen. To date, estrogen replacement therapy is used for the treatment of such

conditions as osteoporosis and symptoms associated with the menopause (such as hot

flashes). Given that the basal forebrain cholinergic system of Alzheimer's patients is

severely compromised, I suggest that estrogen replacement in post menopausal

women may have therapeutic merit for neurodegenerative diseases such as AD.


Estrogen Physiology

Both males and females have circulating estrogens, although the levels found in

the female are considerably higher than that found in the male. The three major forms

of estrogen in the female are estradiol (E2), estrone (El) and estriol, of which,

premenopausally, estradiol predominates (Lipsett, 1986). Following the menopause,

however, estrone predominates, a result of the peripheral or extraglandular conversion

of androstenediol (made by the adrenal glands). Nevertheless, both circulating levels

of E2 and El are decreased following the menopause (Jaffe, 1986).

The primary sources of estradiol are the granulosa cells and theca internal cells of

the ovary. While the adrenal gland does produce some estrogens, its contribution to

circulatory levels is relatively minor at any phase of the adult woman's life (Lipsett,








1986). Levels of estradiol in a young healthy woman fluctuate in a predictable cyclic

fashion over a 28 day period (the menstrual cycle). Mean E2 levels during the early

follicular phase of the menstrual cycle (7 to 14 days before the preovulatory LH peak)

are about 44 pg/ml and increase to about 78 pg/ml (3 to 6 days prior to the LH peak.)
Peak estradiol levels are approximately 235 pg/ml (ranging in one study from 150 -

350 pg/ml) occurring at or just prior to the preovulatory LH surge (Mishell, et al.,

1971). Following the LH peak, mean mid luteal levels are approximately 112 pg/ml

after which serum estradiol levels drop sharply at the onset of menses and the cycle

restarts. The menopause, a process that represents the end of the reproductive phase in

a woman's life, occurs at an average age of 54 years. It is characterized by a
precipitous drop in estradiol to undetectable levels.

In the rat, the fluctuations in estradiol occur over a 4 to 5 day estrous cycle.

Circulating levels of estradiol are lowest during estrus (approximately 7 pg/ml) and

gradually increase over diestrus reaching peak levels on proestrus afternoon, just prior

to the preovulatory LH surge (Smith et al., 1975) The range of estradiol levels are

typically between 5 pg/ml and 100 pg/ml.

The major routes by which estrogens exert a biological effect is through an

interaction at the genomic level and through membrane associated receptors. It has

been clearly established that estrogens cause an increase in RNA and protein

synthesis and increase DNA synthesis in association with mitosis in estrogen-

responsive tissue. Most of these effects are blocked by both a translational inhibitor

(puromycin) and a transcription inhibitor actinomycinn D) lending support to the idea

that estrogen exerts its effects at the genomic level (Tepperman and Tepperman,
1987). Until recently, it was believed that unoccupied receptors in the extranuclear
part of the cell bound to the steroid and translocated the complex into the nucleus

where the steroid was allowed to bind to chromatin receptors. However, it is currently

believed that practically all unoccupied steroid receptors are bound in the nucleus








before the steroid encounters the cell and that phosphorylation of the receptor is

necessary prior to hormone binding (Tepperman and Tepperman, 1987).

While it has classically been assumed that steroid hormones, like estradiol,

produce their effects through a genomic mechanism, the time course of certain effects

modulated by estrogens are inconsistent with genomic interaction and thus cannot be

explained by the induction of protein synthesis. For example, iontophoretically

applied estradiol was found to modulate the rate of glutamate-evoked firing of

cerebellar Purkinje neurons (Smith et al., 1988). The time course for this action is too

rapid (less than 100 msec) to be explained by the induction of protein synthesis, a

process that can take hours. Thus, in order to explain these rapid effects of estradiol,

the presence of membrane bound estrogen receptors was postulated and subsequently

characterized (Towle and Sze, 1983).

Estrogen receptor-containing neurons or estrogen-concentrating neurons are found

in several areas of the brain in both male and female brains (Simerly et al., 1990).

Furthermore, the CNS distribution of estrogen receptors in either sex is not very

different. Nevertheless, subtle quantitative differences cannot be ruled out since

sexual dimorphism in the size of certain nuclei have been documented (such as that of

the medial preoptic area of the hypothalamus; Gorski et al., 1978). Therefore,

differences in receptor number may be due to differences in the number of cells

expressing the estrogen receptor. Estrogen receptors, as determined by in situ

hybridization of the estrogen receptor mRNA, have been elucidated to be in highest

concentrations in hypothalamic areas (Simerly et al., 1990). The receptors in this

region are known to have a regulatory and modulatory function on the gonadal-

hypothalamic-pituitary axis. It is important to look at both estrogen receptor mRNA

distribution along with the distribution of estrogen binding in the central nervous

system (CNS) because, at least developmentally, not all estrogen receptor mRNA-

containing neurons exhibit estrogen binding (Toran Allerand et al., 1992). The








presence of estradiol receptor containing neurons are also found in the medial septum

(Simerly et al., 1990; Pfaff and Keiner, 1973; Toran-Allerand et al., 1992), nucleus

basalis/substantia inominata (Simerly et al., 1990; Pfaff and Keiner, 1973; Toran-

Allerand et al., 1992), nucleus of the diagonal band of Broca (Toran-Allerand et al.,

1992), and the hippocampus (Simerly et al., 1990; Pfaff and Keiner, 1973; Loy et al.,

1988; Bettini et al., 1992), thus lending support for estrogen's actions on these

populations of neurons.


The Cholinergic System

A variety of cholinergic neurons in the central nervous system have been

identified (Wainer et al., 1984; Wenk et al., 1980; Johnston et al., 1981). Broadly,

they can be divided into intrinsic and projection systems. Intrinsic or local circuit

neurons include those of the striatum, caudate and putamen, nucleus accumbens and

olfactory tubercle (Perry, 1986). Projection neurons, with predominantly ipsilateral

axons, are more widely distributed and have been identified in the basal forebrain,

projecting to the cortical areas; midbrain, pons, the thalamus; various cranial nerves

and somata in the spinal cord (Perry, 1986). Neocortical regions have both intrinsic as

well as extrinsic sources of cholinergic innervation. It has been estimated that

approximately 70% of the cholinergic innervation of the fronto-parietal cortex is

derived from acetylcholinesterase positive neurons in the peripallidal nucleus basalis

while the remainder appears to be made up of cortical intrinsic neurons (Johnston et

al., 1981).

The major nuclei of the cholinergic system in the rat brain are comprised of 6

regions and are referred to as Chl to Ch6. These correspond to the medial septum

(MS), the nucleus of the vertical limb (NVL), the nucleus of the horizontal limb

(NHL), the nucleus basalis magnocellularis (NBM), the pedunculopontine nucleus

and the laterodorsal tegmental nucleus, respectively.








The principle neurotransmitter for cholinergic neurons, acetylcholine (Ach), is
synthesized at the nerve ending in a one step enzymatic reaction from its precursors,

choline and acetyl coenzyme A (Acetyl CoA) by the enzyme choline acetyltransferase

(ChAT). Acetyl CoA is primarily synthesized in the mitochondria and may be derived

from several metabolic sources (Cooper et al., 1986). Choline, on the other hand, may

be derived from phosphatidyl choline, a constituent of the plasma membrane as well

as from dietary sources. Following the release of Ach into the synapse, it is

hydrolyzed by the enzyme acetylcholinesterase (AchE) into choline and acetate.

About 35 50% of the liberated choline is transported back into the synapse by a

saturable, sodium-dependent, high affinity uptake system for subsequent reutilization

for Ach synthesis. Two mechanisms exist for choline transport: high affinity choline

uptake (HACU) and low affinity choline uptake (LACU). HACU is a saturable,

temperature- and sodium-dependent, carrier-mediated process specific for cholinergic

terminals and thought to be the rate limiting step for Ach synthesis (Simon et al.,

1976). LACU, on the other hand, is found in cell bodies and in some non-neuronal

tissue and is thought to function in the synthesis of choline-containing phospholipids
(Cooper et al.. 1986). Furthermore, HACU can be selectively blocked by

hemicholinium-3 (HC-3) with high affinity (Km of 0.05 1 jiM) while the effects of

this drug on LACU are less potent (Km = 10 120 IM). Simon (1975; 1976) and his

coworkers showed that activated impulse flow increases the Vmax of high affinity

choline transport, while agents that decrease neuronal activity (such as lesions of a

cholinergic tract or administration of pentobarbital), decreased the Vmax. This uptake

process can therefore be regarded as reflecting the moment to moment activity of that
cholinergic population being investigated. ChAT activity, on the other hand, reflects
the relative degree of innervation or alternatively, the relative number of viable
cholinergic neurons.








The Cholinergic System and Cognitive Function

The basal forebrain cholinergic system, consisting of the nucleus basalis and

medial septal/nucleus of the diagonal band, and their projections to cortical and

hippocampal regions, respectively, have been established as having an important role

in cognitive function. Evidence for this comes from behavioral (learning and

memory) deficits that result from lesions of basal forebrain nuclei or transaction of

their pathways. Rats that received either unilateral (Sara et al., 1992) or bilateral (Sara

et al., 1992; Will and Hefti, 1985) lesions of the fimbria demonstrated a learning

deficit in the eight arm radial maze task. Furthermore, this learning/memory

impairment was correlated with a 50% reduction in ChAT activity (Sara et al., 1992).

Chemically-induced lesions of the nucleus basalis have also resulted in deficits in

spatial learning/memory as assessed by the Morris water task (Dekker et al., 1992) as

well as by non-spatial tasks (Flicker et al., 1983; Mouton et al., 1988). These lesions

resulted in a reduction in neocortical ChAT activity that was correlated with the

learning deficits (Sara et al., 1992; Dekker et al., 1992; Flicker et al., 1983).

In humans, it has been documented that the muscarinic receptor antagonist,

scopolamine induces an impairment in memory storage and retrieval (Drachman and

Leavitt, 1974). The lack of behavioral impairment in subjects administered

methscopolamine (a peripherally acting muscarinic antagonist) supported the role for

central cholinergic neurons in cognitive function. Peterson (1977) has also described

the effect of muscarinic blockade on leaning impairment where the effect of

scopolamine had its primary effect on the acquisition of new material. Further

evidence for the cholinergic involvement in cognition comes from the observation

that in AD, a progressive neurodegenerative disorder characterized by a dramatic loss

in memory, a consistent and widespread deficit in cholinergic activity is observed

(Perry, 1986). Cholinergic impairment, as demonstrated by a reduction in ChAT

activity, was also correlated with the extent of intellectual impairment as assessed by








a memory information test in patients with senile dementia (Perry et al., 1978). While

cognition is not exclusively a function of cholinergic activity, and several

neurotransmitter systems are thought to be involved (Wenk et al., 1987), it is clear

that cholinergic activity plays an important role in learning and memory.


Neurotrophism and Neuronal Plasticity

The term neurotrophism implies the ability of a substance to maintain the

functional and/or structural integrity of a specific set of neurons. Lately however,

neurotrophic substances have also been associated with the ability to induce neuronal

plasticity. Neuronal plasticity is that condition in which neuronal connections can be

modified. Both neuronal development and response of the CNS to injury have been

proposed to be at least partially mediated by the release of neurotrophic substances

(Hefti, 1989; Stewart et al., 1988). It has also been postulated that the loss of

cholinergic neurons seen in neurodegenerative disorders such as AD reflects a relative

or absolute deficiency of trophic support (Stewart et al., 1988).

Neurotrophic factors or neurotrophins are a family of polypeptide growth factors

that demonstrate a significant degree of homology with the prototypical neurotrophin,

nerve growth factor (NGF). Other notable examples of neurotrophins that have been

identified are brain-derived neurotrophic factor (BDNF; Barde et al., 1982; Leibrock

et al., 1989) neurotrophin-3 (NT-3; Ernfors et al., 1990; Hohn et al., 1990; Jones

and Reichardt, 1990; Maisonpierre et al., 1990; Rosenthal et al., 1990) and

neurotrophin-4 (NT-4; Hallbo6k et al., 1991).
NGF is a multi-subunit complex consisting of a, 5 and y subunits. Its molecular

weight is approximately 140 kD (Lapchak et al., 1992). The biologically active
component is the j3 subunit, consisting of 118 amino acids and with a molecular

weight of approximately 13.2 kD. NGF is a target-derived growth factor and is

believed to act through a series of steps involving first, the synthesis and release of








NGF from target tissue where it can then bind to its receptor complex, be internalized,

and subsequently transported retrogradely toward the cell body (DiStefano et al.,

1992). The precise mechanism elucidating the second messengers) involved has yet

to be determined although it has been demonstrated that central cholinergic neurons

internalize and transport the receptor itself to the cell body (Johnson et al., 1987;

Springer, 1988), implicating the receptor-ligand complex as a potential second

messenger.

BDNF is a small, basic protein of 112 amino acids having a molecular weight of

about 13.5 kD. BDNF has an approximately 50% amino acid sequence homology

with NGF (Leibrock et al., 1989) and is found in major cholinergic pathways of the

CNS. Regions of the brain that synthesize BDNF include the cerebellum, cerebral

cortex, hippocampus, hypothalamus, septum, striatum and thalamus (Lapchak et al.,

1992). In comparison to NGF concentrations, BDNF is generally found in larger

quantities in the brain. For example, in the hippocampus, BDNF mRNA is present in

50-fold higher concentrations than NGF (Hofer et al., 1990).

NT-3 is the third homolog of the NGF family of neurotrophins. It is a 119 amino

acid peptide with a molecular weight of approximately 13.6 kD and has an

approximately 57% sequence homology with NGF and BDNF (Ernfors et al., 1990).

NT-3 mRNA has been found to be most abundant in the cerebellum and hippocampal

formation (Maisonpierre et al., 1990).


Mechanism of Action:

Two types of NGF receptors have been postulated: a low affinity (p75) NGF

receptor (Kd 10-9 M) and a high affinity (p145) NGF receptor (Kd 10-11 M; Stach

and Perez-Polo, 1987; Meakin and Shooter, 1992). The nomenclature of these two

classes of NGF receptors reflect the molecular weights of the two species. The

molecular weights of these two receptors are 75 kD and 140 kD for the low affinity








and high affinity nerve growth factor receptors, respectively (Meakin and Shooter,

1992). The p75 NGF receptor (p75NGFR) is a single peptide chain of approximately

400 amino acid residues with a single membrane spanning domain, separating a

slightly longer extracellular domain from a shorter cytoplasmic domain (Meakin and

Shooter, 1992). The cytoplasmic C-terminal region of this receptor contains a single

mastoparan-like domain, a consensus sequence for the binding of G-protein (Feinstein

and Larhammar, 1990), suggesting the possibility of a yet undetermined G-protein

linked signal transduction mechanism.

The high affinity (p140) nerve growth factor receptor (p140NGFR), on the other

hand, was identified as the protein product of the trkA proto-oncogene which encodes

for a tyrosine kinase receptor. NGF was recognized as a ligand for trkA through

studies that showed phosphorylation of trkA on tyrosine and increased tyrosine kinase

activity in response to NGF in PC12 cells (Kaplan et al., 1991). The involvement of

the high affinity pl40NGFR seems to be critical in mediating the biological effects of

NGF while the precise role of the p75NGFR in NGF's actions remain unclear. The

ability of the pl40NGFR to mediate the biological activity of NGF in the absence of

the low affinity receptor supports the importance of the high affinity receptor.

Furthermore, immunological blockade of the low affinity receptor prevents the

binding of NGF to one class of receptor, but its ability to bind to the high affinity

receptor remains unaltered. Despite the prevention of NGF to bind to the low affinity

receptor, NGF's ability to stimulate neuronal survival or neurite outgrowth is

unaffected (Weskamp and Reichardt, 1991). There is recent evidence, however, to

implicate the role of the p75NGFR in innervation of peripheral sympathetic targets

(Lee et al., 1994)One possible role that the low affinity receptor might have, however,

is that the p75NGFR can influence the binding to the pl40NGFR. Supporting

evidence for this hypothesis comes from studies of a PC12 cell line variant, the NR18

cell line. This cell line lacks the p75NGFR and shows only one component of NGF








binding as determined by Scatchard analysis. Transfection of the p75NGFR into this

cell line, however, results in the generation of a second component of binding that is

of higher affinity than the original binding. Furthermore, the addition of this second

component also correlates with the ability of these cells to now induce c-fos

expression, an indication of an NGF-induced biological response (Hempstead et al.,

1989).

The p75NGFR has also been shown to bind BDNF (Rodriguez-T6bar et al.,

1990), NT-3 (Ernfors et al., 1990) and NT-4 (Halb66k et al., 1991) with

approximately the same affinity as NGF. These neurotrophins, however, bind to

different high affinity or trk receptors suggesting that the selectivity of the binding

and thus specificity of effect is imparted by the high affinity tyrosine kinase (trk)

receptors. BDNF and NT-3 interact with the trkB receptor with an affinity of

approximately 1.8 x 10-9 M and 1.3 x 10-9 M, respectively in 3T3 cells (Soppet et al.,

1991). In survival or proliferation assays, NT-3, however was found to be about 10-

fold less active (Glass et al., 1991) and in a transformation assay, about 100-fold less

active than BDNF (Klein et al., 1992). Furthermore, NT-4 also seems to interact with

the trkB receptor with only a five-fold difference in activity than BDNF (Klein et al.,

1992).

Another class of high affinity receptor belonging to the tyrosine kinase family is

the trkC receptor. NT-3 seems to be the major ligand for this receptor with an affinity

of about 2 4 x 10-9 M in 3T3 cells (Lamballe et al., 1991).


Distribution of Neurotrophin Receptor-Positive Neurons

The identification of neurotrophin receptor positive neurons in the CNS have been
elucidated through binding studies utilizing 125I-labeled NGF (Altar et al., 1991,

Richardson et al., 1986) as well as by quantitation of the respective mRNAs for the

low affinity and high affinity NGF receptors (Miranda et al., 1993; Buck et al., 1988).








Binding studies have revealed the presence of the high affinity NGF receptor in

several regions of the brain. Notable examples of such regions are the hippocampus,

basal forebrain, amygdala, paraventricular thalamus, frontal, parietal, occipital and

cingulate cortices and the striatum (Altar et al., 1991; Richardson et al., 1986).

Analysis of mRNA for the NGF receptors have revealed a similar distribution

(Miranda et al., 1993; Buck et al., 1988). Interestingly, the topographical distribution

of NGF receptor positive neurons of the basal forebrain is similar to the distribution

of cholinergic markers such as ChAT activity and [3H]-vesamicol binding (a marker

for the vesicular Ach transporter) suggesting that a large proportion of NGF sensitive

neurons are cholinergic.

BDNF receptor positive neurons also have a wide distribution in the CNS.

Retrograde labeling studies have provided evidence for both overlapping and distinct

populations of BDNF transporting neurons (DiStefano et al., 1992). For example, the

entorhinal cortex retrogradely transports BDNF but not NGF when the neurotrophin

is injected into the hippocampus. Furthermore, while BDNF was widely transported

within the hippocampal formation itself, intrahippocampal transport of [125I]-NGF

was not observed (DiStefano et al., 1992).

The distribution of [125I]-NT-3 binding sites are distinct, more dense and more

widely distributed than that of [125I]-NGF (Altar et al., 1993). The highest levels of

NT-3 binding were in the anterior olfactory nucleus, lateral olfactory tract,

hippocampus, amygdala, anterior thalamus and the neocortex. Lower levels of

binding were also prevalent in the medial septum, hypothalamus and cerebellum

(Altar et al., 1993). NT-3 binding was absent in cell bodies and predominated in

nerve terminal rich regions such as the molecular layer of the hippocampus and layer

1 of the neocortex (Altar et al., 1993), unlike the binding of NGF which demonstrates

dense binding in cell body regions (Altar et al., 1991).








It has recently been demonstrated that certain populations of neurons that express

the neurotrophin receptor also synthesize mRNAs for its cognate ligand (Miranda et

al., 1993). This colocalization of receptor and ligand suggest the potential for

autocrine interactions and regulation of neurotrophin release. This pattern of

colocalization is characteristic of neurons in the cerebral cortex and hippocampus. In

nuclei of the basal forebrain, however, such colocalization is not seen suggesting the

importance of paracrine or more classically associated target-derived trophic

mechanisms that exist for this population of neurons.


Importance of Neurotrophins in Development and Adulthood

During the early or so called critical period of CNS development, neurons and

their connections are sensitive, or plastic, to environmental influences. Over the

course of development, the CNS passes through a phase whereby an excess of

neurons and pathways are created. As development proceeds, redundant projections

are 'pruned' in order to achieve the stable set of afferents and efferents we know as the

mature CNS. It is believed that neurotrophic factors may play a role in this pruning

(Varon and Adler, 1980). Competition for neurotrophic substances leads some

neurons to survive and others to die. In effect, it is the classic Darwinian theory of

"survival of the fittest" at work on the cellular level. These processes that lead to

permanent modifications in morphology and function, and subsequently give rise to

the mature CNS are termed organizational effects.

This neuroplastic property of neurotrophic factors is not, however, restricted to

early development but can also be seen in the mature CNS. Neurotrophic substances

are target-derived factors which act on the neuron in a retrograde fashion. If these

target-derived factors are required for the maintenance of neurons, then interference

with this system will result in death of the neuron. For example, the absence of a

neurotrophic substance for a given set of neurons may result in hypofunctional








neurons. In following, subsequent replacement with the depleted neurotrophic

substance would then lead to normalization of function in these neurons. These

reversible and usually temporary effects are termed activational effects. While the

organizational effects occur primarily during early development (that is, before

maturation of the CNS is complete), activational effects can occur in the adult brain

as well.

Cholinergic neurons of the basal forebrain have revealed the presence of NGF

receptors suggesting that these neurons are a preferential target for NGF action in the

brain (Taniuchi et al., 1986; Piorro and Cuello, 1990). Furthermore, it has been noted

that fimbrial transaction leads to a transient accumulation of hippocampal NGF

(Korsching et al., 1986) suggesting that cholinergic neurons constantly reduce and

utilize hippocampal NGF though retrograde axonal transport.

The potential organizational effects of NGF are reflected by the increase in the

levels of NGF protein and mRNA during development of the rat brain (Whittemore et

al., 1986), followed by a decrease during adult life and senescence (Larkfors et al.,

1987). The activational/organizational potential for NGF to alleviate neuronal

dysfunction seen in such neurodegenerative disorders such as AD has led to

numerous reports. For example, exogenously applied NGF has been shown to

increase high affinity choline uptake (HACU) and choline acetyl transferase (ChAT)

activity in adult and aged rats (Williams and Rylett, 1990). In lesion studies, NGF has

been shown to promote survival of septal cholinergic neurons following fimbrial

transaction (Hefti, 1986). As a behavioral correlate to the neurochemical changes, it

has been demonstrated that NGF improves spatial memory in aged animals (Fischer

et al., 1991) and in animals with bilateral lesions of the fimbria (Will and Hefti,

1985).

Evidence from brains of patients afflicted with AD demonstrate a more complex

picture of the involvement of NGF in the etiology of neurodegenerative disorders.








Some reports document a decrease in NGF receptors (Higgins and Mufson, 1989) and

NGF binding (Strada et al., 1992) which would be consistent with the lack of trophic

support' hypothesis, while others demonstrate no change in NGF protein levels in

varying brain regions (Allen et al., 1991). There has even been a report documenting

a nearly two-fold increase in NGF immunoreactivity in the brains of AD patients

(Crutcher et al., 1993). The discrepancies in NGF levels found in AD brains may

have been due to differences in post-mortem time interval or measurement in different

regions, that is, different brain regions may respond differently to neurodegenerative

changes. While the precise role for NGF in the survival or maintenance of neurons in

AD may still be equivocal, the role of other neurotrophins such as BDNF may be

more critical. In accordance with the establishment of basal forebrain cholinergic

neurons as targets for neurotrophin action along with existing evidence from the

literature that documents the widespread and consistent deficits seen in the basal

forebrain cholinergic system of AD patients, the therapeutic potential for the

activational effects of neurotrophins can be realized.


Estrogens and Neuronal Plasticity

The brain is thought to be a sexually dimorphic structure. In the development of

the CNS, it is believed that the perinatal hormonal milieu is very important in the

development of certain sexually dimorphic nuclei. An example of such a nucleus is

the medial preoptic nucleus (MPN; Gorski et al., 1978). This nucleus is found to be

larger in the male rat than in the female rat. Following manipulation of the perinatal

hormonal environment, it was found that the size of the MPN could be altered from

that of the typical female to the male and vice-versa (Gorski et al., 1978). Castration

of the male neonate, resulted in the marked reduction in nuclear volume of the MPN.

Conversely, injection of testosterone propionate into the neonatal female resulted in

the formation of a significantly larger MPN. The treatment of adult rats with the








gonadal steroid, that is, past the critical period of organizational development or

sexual differentiation of the CNS, was without effect. Although the reported effects

were that of an androgen, it is conceivable that the organizational effects were
mediated through the conversion of testosterone to estradiol by the P450 enzyme

aromatase.

There are several examples of activational effects of ovarian steroids. The peak in

serum LH luteinizingg hormone) that immediately precedes ovulation is under the

influence of circulating estradiol (Tepperman and Tepperman, 1987). Just prior to the

LH peak, E2 levels rise, which in turn, increase the sensitivity to gonadotropin

secretion. This increased sensitivity to GnRH (or LHRH) leads to an eventual burst in

LH secretion. Thus, the rhythmic pattern of E2 secretion modulates the secretion of

LH. E2 has also been shown to affect neurochemical indices in brain regions which

contain putative receptor sites, particularly in the hypothalamus. E2-induced increases

in muscarinic receptors (Rainbow et al., 1980) and ChAT activity (Luine et al., 1975;

Luine et al., 1980) have been reported in hypothalamic nuclei and the basal forebrain.

Morphological changes associated with estrogenic milieu have also been documented.

In ovariectomized female rats, in which complete deafferentations of the medial basal

hypothalamus were performed, estradiol treatment produced marked increases in

axodendritic shafts and spine synapses suggesting that E2 stimulated axonal sprouting

as well as dendritic spine formation (Matsumoto and Arai, 1981). Furthermore,

fluctuations in dendritic spine densities in chronically ovariectomized or estrogen-

replaced animals, as well as across the cyclic day to day variations in estradiol have

been documented (Wooley et al., 1990; Wooley and McEwen, 1992). In view of these

activational effects of estradiol on neuronal function in the CNS, it was hypothesized

that E2 may play a role not only in the development of the mature CNS but also serve

an important function in the maintenance of certain populations of neurons by acting

as a neurotrophomodulatory substance for basal forebrain cholinergic neurons. The






18

term neurotrophomodulatory reflects the ability of E2 to exert its effects either

directly, or indirectly, through the modulation of another classical neurotrophic

substance.














CHAPTER 2

GENERAL METHODS


Animals
Young adult female (3-4 months old), Crl:CD-Sprague-Dawley rats were obtained

from Charles River Breeding Laboratories, Wilmington, MA. Animals were housed

two per cage in an environmentally controlled room on a 14-h light, 10-h dark cycle

(lights on at 0700). The animals were maintained ad libitum on water and rodent

laboratory chow (Ralston Purina Co., St Louis, MO). All procedures performed on

animals were reviewed and approved by the Institutional Animal Care and Use

Committee (IACUC) at the University of Florida prior to initiation of the study.


Animal Surgical Procedures
For all surgical procedures, animals were anesthetized with methoxyflurane

(Metofane, Pitman-Moore, Washington Crossing, NJ). In general terms, three

treatment groups were generated: ovary-intact (INTACT), ovariectomized (OVX) and

estradiol-replaced (E2 pellet). Based on different experimental paradigms and the

measurement of different parameters, the length of time to which an animals was

ovariectomized or estradiol-replaced varied accordingly.

Two-thirds of rats underwent bilateral ovariectomy using a dorsal approach. Three
weeks following ovariectomy, a subset of the ovariectomized animals (the E2-

replaced group) received a 5 mm Silastic (Dow Corning, Midland, MI) pellet

containing a 1:1 mixture of cholesterol (Steraloids, Inc., Wilton, NH) and 17-13

estradiol (Pharmos, Inc., Alachua, Fl.) that was implanted subcutaneously. Estradiol








delivery through Silastic tubing results from diffusion down a large concentration

gradient and the fibrosis, which occurs over time around the Silastic pellet, reduces

diffusion (Smith et al., 1977). In the behavioral studies, neurochemical analyses and

the measurements of different neurotrophin proteins and their respective mRNA, the

effects of short term as well as long term ovariectomy and replacement periods were

investigated. For any treatment regimen that required the maintenance of estradiol

replacement for more than 3 weeks, the Silastic pellets were removed and

repositioned every 2 to 3 weeks to maintain E2 diffusion from the Silastic. The

ovariectomized group received sham pellets that were similarly re-positioned every 2

to 3 weeks.


Preparation of Estradiol Pellets

Silastic medical grade tubing (inner diameter: 0.062 in outer diameter: 0.125

in) was sectioned into several 5 mm portions. Each section, called a pellet, was sealed

at one end with Silastic medical adhesive (Silicone Type A) and allowed to dry

overnight. At this point, the pellets were separated into two groups and stored in

different rooms to avoid possible contamination of sham pellets with estradiol. Sham

pellets were stored in a 20 ml glass scintillation vial. The other subset of pellets were

filled with a 1:1 mixture of 17-0 estradiol (E2) and cholesterol. Preparation of the E2

pellets were done under a biological hood. The E2 and cholesterol were weighed out

and mixed to homogeneity. Each pellet was then filled with this mixture and sealed at

the other end and left to dry overnight under the hood. On the following day, the

silastic pellets were washed twice with 100% ethanol to remove any excess E2

present on the outsides of the pellet so as to avoid presenting the animal with an

initial high dose or spike of E2 following implantation. After the second wash, the

ethanol was decanted and the pellets were allowed to dry. Beginning 48 hours prior to








implantation, both the sham and E2 pellets were incubated in phosphate buffered

saline (PBS) at room temperature.


Methods of Blood Sampling

Blood samples were obtained either by cardiac puncture or trunk blood collection

following decapitation. For the cardiac puncture technique, the animals were first

anesthetized with Metofane. Once anesthetized, a 27 gauge needle attached to a sterile

1 ml syringe was used to draw up between 0.75 and 1 ml of blood. The animals were

placed on their side and the needle was inserted between the ribs at the point where

the heart beat felt the strongest. Trunk blood collection was the technique more

frequently employed. Upon decapitation, a funnel was used to collect the trunk blood

into an Eppendorf tube.


Serum Estradiol Assay

Blood samples obtained by the above mentioned methods were processed and

used for the determination of serum estradiol. The blood was centrifuged at 13,500 x

g for 1.5 minutes and resulting serum was aliquoted into a separate tube, frozen and

stored at -30C for steroid level assessment at a later date. Serum concentrations of

E2 were assayed by radioimmunoassay (RIA) using commercial kits supplied by

Diagnostic Products Corp. (Los Angeles, CA). The range of assay detectability was

20-3600 pg/ml.


Vaginal Smears

In order to ascertain the effectiveness of the steroid level manipulation, selected

groups of rats were used to perform vaginal smears. An eye dropper containing a

single drop of normal tap water was inserted into the vagina. Cells were obtained by

flushing the vagina with the small volume of water and allowing fluid to flow back








into the dropper without actually aspirating cells from the endometrial lining. The

cells were then spread out on a glass slide, allowed to dry on a hot plate set on low

heat for approximately 2 min. and stained with 0.1% Toluidine Blue. The relative

proportion of nucleated epithelia, cornified epithelia, and leukocytes were determined

under medium power light microscopy and recorded. Vaginal smears were usually

taken over a 5 day period so as to determine cyclicity patterns in intact rats. A high

proportion of nucleated epithelia relative to any other cell group signified that the rats

were in proestrus (the period of highest circulating estradiol). Relatively large number

of cornified epithelia indicated a prolonged exposure to estradiol and implied intact

rats were in estrus and confirmed that E2 pelleted animals were receiving a sustained

level of estradiol. Ovariectomized animals typically showed very few cells on the

slide indicating the lack of trophic support by estradiol on the uterine wall. A

predominance of intensely staining leukocytes implied that the animal was in diestrus,

a period that correlates with relatively low levels of estradiol and coincides with the

shedding of the uterine wall cells associated with the withdrawal of the uterine trophic

hormones E2 and progesterone.


Tissue Dissections

The regions of the brain investigated for cholinergic markers (high affinity

choline uptake, HACU; and choline acetyltransferase, ChAT) and molecular markers

(NGF and BDNF) were the frontal cortex and the hippocampus, target regions of the

basal forebrain cholinergic neurons.

Following decapitation, the brain was carefully removed from the skull, and

immediately placed on a wet paper towel directly over a layer of ice on which the

dissections were carried out. The frontal cortex was cut away with a pair of dissection

scissors. The dissection of the frontal cortex was made carefully so as not to invade

any of the temporal or parietal cortex. Any colossal fibers that were taken in the initial








cut were trimmed off. As a result of this precise dissection, only about 120 mg of

frontal cortex tissue was obtained per brain.

In order to gain access to the hippocampi, the entire cortex was peeled away.

Careful removal of the dura facilitated this process. Once the cortex had been

separated from the diencephalon, the hippocampus could be seen on the inner side of

the cortex. Using a pair of forceps, the myelinated fibers that hold the hippocampus in

place were severed. Subsequently, the hippocampus was pulled away from the rest of

the cortex. Following this dissection procedure, the amount of hippocampal tissue

obtained per brain was approximately 120-140 mg. The tissues were then immersed

in the appropriate buffer required in further processing of the tissue.















CHAPTER 3

THE EFFECT OF OVARIECTOMY AND ESTRADIOL-REPLACEMENT ON

LEARNING AND MEMORY




Classically, gonadal hormones have been thought to have their major influence on

reproductive function by interacting primarily with the hypothalamic-pituitary-

gonadal axis. Recently, however, there has been evidence to suggest that estrogen

may have an impact on learning and memory. One example is the effect of E2

replacement therapy on cognition in surgically menopausal women (Sherwin, 1988).

A statistical decline in cognitive performance was coincident with the decrement in

circulating sex steroids in oophorectomized women. Patients who had hysterectomies

but had their ovaries retained, showed stability in cognitive performance as well as in

circulating sex steroids (Sherwin, 1988; Phillips and Sherwin, 1992). Another

example is that in patients with senile dementia of the Alzheimer's type (SDAT),

estradiol replacement over a 6 week period produced significant improvements in

attention, orientation, mood and social interaction (Fillit et al., 1986). Other

preliminary studies have pointed toward a similar effect of estrogen on cognition in

demented subjects (Honjo et al., 1989). An important caveat to this observation is that

one cannot ascribe the beneficial effects of estrogen on the improvement in

psychological test performance without acknowledging the effect of estrogen on

mood (Gerdes et al., 1982; Sherwin and Gelfand, 1985).








In pursuit of a novel animal model to study neurodegenerative disorders

associated with dementia, such as AD, we addressed the involvement of estradiol in

learning and memory performance in the rat. To test our hypothesis, we assessed

learning behavior using the 2-way active avoidance and the Morris water task

paradigms in three groups of animals: INTACT, OVX, and estradiol replaced

ovariectomized animals (E2 Pellet).




Methods


Animals were maintained and treated as described in the General Methods

section. Rats were tested using the active avoidance paradigm at two time points

following ovariectomy: 5 weeks and 28 weeks. E2 replaced animals received the

pellet implants after 3 weeks of ovariectomy and were maintained on this treatment

regimen for the next 2 weeks prior to testing. These same animals were retested on

their ability to learn at the 28 week time point. E2 pellet animals were maintained on

their treatment regimen throughout this time and were therefore exposed to constant

levels of estrogen for 25 weeks.

Plasma estradiol levels were measured in 5, 12 and 28 week ovariectomized

animals. At these time points, the corresponding E2 pellet animals had undergone E2

replacement for 2, 9 and 25 weeks, respectively.


Active Avoidance

To assess learning, the 2-way active avoidance paradigm was employed following

the procedure of Mouton et. al. (1988). All three groups of animals were tested for 14

consecutive days, each day consisting of 15 trials. Each trial lasted for 1 minute and

was comprised of the presentation of the conditional stimulus (a light and sound cue)








for the first 5 seconds and a 7 second interval followed by an electrical foot-shock of

1.4 mA for a 2 second duration. Successful learning was determined by the number of

correct responses or 'avoidances' and was defined as transferring from one side of the

shuttle-box to the other within the first 12 seconds of each trial, before the onset of

the footshock. If the animal transferred only upon the administration of the brief

footshock, an 'escape' was recorded. In order to assess potential motivational

differences between animals in different treatment groups, the number of "no

transfers" was also recorded. This parameter describes the number of trials in which

the animal did not transfer from one side of the shuttle box to the other upon

stimulation with the electrical shock.


Morris Water Task

Our model of the Morris Water task is a circular pool 43 inches in diameter and

23 inches deep. The pool was filled with water to 16 inches in depth and maintained

at a constant temperature of 25C. Blue Crayola powdered paint was mixed with the

water in order to produce an opaque media. The principle of this paradigm, described

by Morris (1981) and modified from that of Gage et al.. (1989) and Paylor et al..

(1991), is to train animals to escape from the water onto a submerged platform based

on conspicuously positioned external visual cues (paintings) around the pool. The

animals are then tested for their ability to remember that position of the platform.

Each training trial consisted of placing animals randomly at four sites, 90 degrees

apart, along the tanks periphery. The animals were then allowed 60 seconds to find

the submerged platform. Once the 60 seconds had elapsed, the animal was physically

removed from the platform and prompted to find the platform from four separate

approach angles. If the animal did not find the platform within the allotted 60

seconds, they were physically guided to the platform by the experimenter. Once

animals had found the platform, they were allowed 30 seconds to orient themselves








according to the external visual cues. Training lasted for 3 days, each day consisting

of 12 trials.

Testing was videotaped and was reviewed at a later time for analysis of individual

animal performance. The pool was divided into 4 quadrants and designated as target,

adjacent 1, adjacent 2 and opposite quadrants. The amount of time spent in each

quadrant as well as the number of target platform crossings were recorded.

Furthermore, distance swum and time to initially reach the target quadrant were also

recorded.


Statistical Analyses

Behavioral data as well as serum E2 levels were analyzed non-parametrically

using the Kruskal-Wallis one-way analysis of variance followed by the Mann

Whitney U test for assessment of group differences.




Results


Estradiol Concentrations

Serum E2 levels in each of the three treatment groups at three different periods

following E2 pellet implantation are presented in Table 3-1. The data are from

animals that were used for behavioral studies as well from animals used for

assessment of neurochemical markers (HACU and ChAT) and molecular markers

(NGF and BDNF).

Ovariectomized animals produced serum E2 levels that were consistently below

the sensitivity of the radioimmunoassay employed (20 pg/ml). At the 5 week time

point, intact animals produced serum E2 levels that were within the expected range of

a normal cycling rat (43.0 10.1 pg/ml). At the 28 week time point, the average








levels of serum E2 in both the intact and E2 Pellet groups were not statistically

different from those seen at the 5 week time point. It should be noted, however, that

behaviorally-tested intact animals showed lower E2 levels (26.4 4.3 pg/ml) than

non-behaviorally tested animals (59.6 17.9). This observation was consistent with

the finding that the behavioral testing using the active avoidance paradigm rendered

approximately 50% of the intact animal population acyclic. When behaviorally tested

intact animals were excluded from the calculation, the plasma estradiol levels in the 5

week intact animals were on average 15 pg/ml higher than levels found in the

behaviorally naive, long term (28-week) intact group (although this difference was

not statistically significant due to the high variance in the short term group). While

serum E2 levels in the E2 pellet group were not statistically different across all time

points tested, the levels of E2 in 28 week intact animals were significantly higher than

those observed in the 28 week E2 pellet group.


Active Avoidance Behavior,

INTACT and E2-pellet animals performed better on the 2 way active avoidance

paradigm relative to OVX animals. By plotting the number of avoidances made by

each animal over the 14 day testing period, learning curves were generated (Figures

3-1 and 3-2). Figure 3-1 depicts individual learning curves from each of the three

treatment groups. Logarithmic learning curves were observed in which animals

performed poorly early in the experiment, but as the experiment approached

completion, a higher level of performance was achieved. This typical pattern of

learning was seen primarily in INTACT and E2 pellet animals. OVX animals on

average, were not as successful in this learning paradigm. Learning curves from

individual animals that were retested on their ability to learn at the 28 week time point

are shown in Figure 3-2. At this time point, the INTACT and E2 pellet animals start

out the testing period at a relatively higher level of performance.








Figure 3-3 shows the average of the total number of avoidances made by each

animal over the 14 day testing period in the three groups of animals tested at 5 weeks

post-ovariectomy and 28 weeks post-ovariectomy. The short term OVX rats showed a

59% decrease in the number of avoidances achieved relative to the intact group, but

this difference was not statistically significant (Figure 3-3). E2 replacement of OVX

rats caused a 8.5 fold increase in the number of avoidances relative to the OVX

group. When these same animals were maintained on their respective treatments and

were behaviorally tested at 28 weeks, the number of total avoidances was increased in

all groups relative to the 5 week testing period (Figure 3-3). However, at this 28 week

testing point, OVX significantly reduced the total number of avoidances by 61% and

E2 replacement continued to increase avoidances by 4.5-fold versus OVX rats.

Furthermore, E2-pellet rats showed a marked acceleration in their rate of learning at

28 weeks, achieving the criteria of performing correctly 11 out of a possible 15 times

in a given day by 1.3 0.3 days of testing (Table 3-2). INTACT rats did not show this

acceleration in learning requiring 9 2.8 days to reach criteria. The OVX animals

maintained their inability to learn the task in the allotted 14 days and were therefore

assigned a value of 15 days (Table 3-2).


Morris Water Task

Analysis of the rate of learning in the Morris water task showed no significant

treatment effects although a marked trend was observed. OVX animals had greater

overall mean training times over the 3 day training period (Table 3-3). There were no

significant differences observed in either time spent in the target quadrant or number

of target platform crossings among all three treatment groups (Table 3-3). However,

in E2-pellet rats, there was a 28% increase in the time spent in the target quadrant and

a 75% increase in the number of target platform crossings. The time it took to initially






30

locate the platform or to reach the target quadrant was also not different across

treatment groups.




















8
I
00
p4













8 C
-
I








'3 -
I I
M -
6

















H


N
CO
9-4


- I


\
v e


C'

+1
e II
5
en











-
C14






-II
54





| II




MS
0-










S4

w
g
m




."


> 4

d0 0
VI VI
&0 4
" ." 9


-HII
0.


elI


S0

I~ i









Table 3 2. Effect of Short-Term and Long-Term Ovariectomy and
Estradiol Replacement on Learning and Retention.


Treatment Group 5 wks 28 wks


Days to reach criteria

INTACT 14 1.0 9 + 2.8

OVARIECTOMIZED 15 0.0 15 0.0

E2 PELLET 9.5 2.1* 1.3 0.3 t

t p < 0.05 vs. Ovariectomized and Intact animals
* p 0.05 vs Ovariectomized animals using the Mann Whitney U nonparametric
statistic





























C0; 00
-H -H -H









q- ell in(




N 0 0V
+H +H -H
e'I. i


O






00





iv
I





U,
4- 60
b.-










o







4-


*S

Hi'


4-

4- 0 0

U 4-


t- 0

+1 +1
06 vi
1-4 M


4)
O
U,

0


U
6)











'5!
U,


















0^
UG<
I8^


00

+1

00
v-.









a 0


INTACT


I-I I I E I


OVX


" "" "r "" I IF


M a


U H
a B


E2 PELLET


12 14


Figure 3-1.


Learning Curves from Individual Animals following 5 weeks of
Ovariectomy in Intact, Ovariectomized and Estradiol Replaced
Animals. Following 5 weeks of ovariectomy, animals were tested for
their ability to learn in an active avoidance task. Avoidances made by
individual animals each day over the 14 day testing period were
plotted thereby generating learning curves.


15:
12-
9-
6-
3-


0~


0s
0


15
12-
9:
6-
3:
O


15-
12-
9-
6-
3-
0
(


1 3


0 2


4
4


I I 1 1
6 8 10
DAY











INTACT
U a


a M


0 2 4 6 8 10 12 14 I I
0 2 4 6 8 10 12 14


ovx


a


a E


E U


0 2 4 6 8 10 12 14-.
0 2 4 6 8 10 12 14


a B


0 2 4 6 8
DAY


Figure 3-2.


E2 PELLET


10 12 14


Learning Curves from Individual Animals following 28 weeks of
Ovariectomy in Intact, Ovariectomized and Estradiol Replaced
Animals. Animals that were tested at the 5 week period were retested
for their ability to learn in an active avoidance task at the 28 week time
period. Avoidances made by individual animals each day over the 14
day testing period were plotted thereby generating learning curves.


^s
^
pfi
&
3
iz


Ci2


1










T


INTACT


INTACT


OVX E2 PELLET


OVX E2 PELLET


Figure 3-3. Active Avoidance Performance Following 5 and 28 Weeks of
Ovariectomy in Intact, Ovariectomized and E2 Replaced Animals.
Total number of avoidances made over the 14 day testing period were
recorded for each animal group. The top panel depicts the total
avoidances made at 5 weeks post ovariectomy while the bottom panel
represents performance following 28 weeks of ovariectomy. The sample
size was 6 in all groups except the 28 week OVX and E2 pellet groups
in which case the sample size was reduced to 3 due to death during
surgery. Overall group difference was assessed using the Kruskal-Wallis
one way ANOVA followed by Mann Whitney U test for individual
group comparisons. ( : p < 0.05 vs. OVX; t : p < 0.05 vs. Intact)


80-
O-
60
40-
20-
0-


200


100
so-
0-

0-


1201
1 00


1T-








Discussion


These data demonstrate that estrogen deprivation of young adult female rats

results in an impairment of non-spatial learning which is reversible through

replacement of E2 to physiological levels. Our observation that non-spatial learning is

affected by estrogen deprivation and replacement is in agreement with some but not

all reports on steroid modulation of avoidance behavior in adult female rats. Three

studies have reported a reduction in avoidance learning in response to OVX

(Cannizzaro et al., 1970; Davis et al., 1976; Milner, 1976), while one report observed

no effect of OVX on lever press avoidance behavior (Van Oyen et al., 1981), and two

reports observed an increase in avoidance in response to OVX (Diaz-Veliz et al.,

1989; Diaz-Veliz et al., 1991). Similarly, E2 replacement in OVX rats has been

reported to increase active avoidance behavior (Cannizzaro et al., 1970), to extend the

extinction of avoidance behavior in another study (Telegdy and Stark, 1973) but also

to decrease avoidance behavior in yet another study (Diaz-Veliz et al., 1991). It

should be noted, however, that in the latter study, while low levels of E2 impaired

avoidance behavior, higher doses of estradiol facilitated this behavior. Furthermore,

the stage of the estrous cycle has also been implicated in avoidance behavior

performance (Sfikakis et al., 1978). During periods of relatively high levels of

circulating E2 (proestrus), avoidance behavior was significantly facilitated while

during low periods of serum E2, avoidance behavior was impaired (Sfikakis et al.,

1978). While on the surface, the literature on the effects of estrogen on

learning/memory seems complex, alot of the differences could be attributable to dose

of E2 as well as the type of experimental paradigm utilized. In our experimental

paradigm, E2 has a positive effect on non-spatial learning.

Studies which address sex differences in avoidance behavior have shown that

females outperform males in avoidance paradigms (Van Haaren et al., 1990). It has








also been shown that when animals are ovariectomized during adulthood, their

performance tends to resemble that of males (Davis et al., 1976), while castration

results in female-like performance (Davis et al., 1976; Van Oyen et al., 1981).

Furthermore, McCord et. al. (1979) have demonstrated that estradiol-treated, castrate

male rats show a tendency toward superior acquisition of Sidman avoidance behavior.

Taken together, these results suggest the importance of estradiol in the learning of

avoidance behaviors under at least some experimental conditions.

At both observation times learning in the OVX group was about 60% less than the

INTACT group, but at only the 28 week time point was the difference significant.

This was due in part to the high variance of the INTACT group, which exhibited

subpopulations of learners and non-learners. Vaginal lavages revealed that the

animals which learned exhibited estrous cycles while those that failed to learn in this

avoidance paradigm were in constant estrus, a period of relatively low estrogen levels.

Subsequently, we observed that the avoidance paradigm induces constant estrus in

about half of all intact rats subjected to the 14 days of testing. This is supported by the

E2 levels seen in our behaviorally tested INTACT animals versus our non

behaviorally tested INTACT animals. Behaviorally tested INTACT animals

demonstrated plasma E2 levels that were 56% lower than those seen in behaviorally

naive INTACT rats. This apparent test-related reduction in performance of the

positive controls may have obscured the extent of the OVX effect and could explain

the apparent superior performance of the E2-pellet when compared to the INTACT

group. Between 5 and 28 weeks, rats in each group maintained there relative order of

proficiency in active avoidance and all animals performed better in the second trial

than first. This enhanced performance during the second exposure to the paradigm

likely reflects recall of the behavior learned during the first test. This long-term

memory is particularly apparent in the E2-pellet group, which at the 28 week time-

point, reached the performance criteria in 1.3 0.3 days. It appears that chronic








exposure to low doses of E2 may enhance long-term memory in addition to its

stimulation of learning of this active avoidance paradigm.

Estrogens have well documented effects on mood and affect in human subjects

(Gerdes et al., 1982; Sherwin and Gelfand, 1985). Along with documentation that

memory impairments are associated with depression (Sternberg and Jarvik, 1976), it

is conceivable then, that the observed differences in performance were due to

motivational differences among rats. These motivational differences could have arisen

from steroid-induced changes in sensitivity to the negative stimulus or to differences

in spontaneous locomotor activity. We observed that all animals, irrespective of

treatment, vocalized briefly upon presentation of the shock stimulus, suggesting that

steroid environment did not affect sensitivity to the electrical shock level used.

Further evidence against steroid-induced differences in shock sensitivity was

presented by Beatty and Beatty (1970) and Beatty and Fessler (1976), who

demonstrated that ovariectomy did not affect flinch-jump thresholds to a shock

stimulus. To address further the possibility of motivational differences related to

steroid treatment, the active avoidance data were analyzed for the number of "no

transfers" among the treatment groups. This parameter reflects the number of trials in

which the animal did not transfer to the other side of the shuttle box upon stimulation

with the electrical shock, suggesting a decreased sensitivity to the shock or decreased

locomotor activity. No significant differences were observed between the treatment

groups, suggesting that animals from each of the treatment groups were similarly

likely to transfer. With respect to locomotor activity, Stewart and Cygan (1978) found

decreased open field activity in perinatally gonadectomized female rats; however, in

adult rats, as is the case in this study, gonadectomy was without effect (Bengelloun et

al., 1976; Blizard and Denef, 1973). In these experiments, no obvious difference in

locomotor activity was observed.








Contrary to the learning deficits seen in the non-spatial task, the results showed

that estrogenic milieu did not markedly affect spatial memory. However, there was a

trend, albeit non-significant, in that E2 reduced mean total training time and increased

the means of parameters related to the identification of the target in the Morris water

task. Nonetheless, active avoidance learning seems to be more sensitive to estrogenic

milieu than is spatial memory. It should be noted that in this study, and other studies

done in this laboratory, sample sizes up to 6 per group did not alter the lack of

significance in the data. It is conceivable, however, that the effects of estrogenic

milieu on spatial memory is very subtle and that only with very large sample sizes can

significant differences be seen. While data on the effect of gonadal hormone

manipulation on spatial memory tasks seems relatively scarce, Van Hest et al.. (1988)

showed that in an operant delayed spatial alternation task, ovariectomized animals

reached similar levels of accuracy as did intact control animals, supporting our

observation that gonadal hormone manipulation does not appreciably affect spatial

memory/learning.















CHAPTER 4
THE EFFECT OF OVARIECTOMY AND ESTRADIOL REPLACEMENT

ON CHOLINERGIC FUNCTION: HIGH AFFINITY CHOLINE UPTAKE




In rats, lesions of certain basal forebrain nuclei (medial septal area and nucleus

basalis magnocellularis) lead to disruptions in learning and memory function (Flicker

et al., 1991) implicating these nuclei and their respective projections (to the

hippocampus and cortex) in cognitive performance. Furthermore, lesions of the

fimbria/fornix, the pathway that connects the hippocampus with the medial septum,

lead to similar impairments in learning and memory (Will and Hefti, 1985). In

memory formation and consolidation, one of the predominant neurotransmitters

involved in these projection systems is ACh, further implicating the basal forebrain

cholinergic systems in memory and learning functions (Perry, 1987).

Estradiol (E2), the major form of circulating estrogen in the body, has been shown

to affect the cholinergic system in a variety of brain regions. In several nuclei of the

hypothalamus, ovarian steroid deprivation (via ovariectomy) and replacement has

been shown to alter ChAT activity and cholinergic receptor density (Luine et al.,

1975; Luine et al., 1980). Further, O'Malley et. al. (1987) have shown that

ovariectomy reduces and E2 replacement normalizes HACU in the frontal cortex of

rats. In view of these activational effects of E2 on cholinergic function in the CNS, I

hypothesized that the previously described effect of E2 on learning and memory may

occur through the modulation of basal forebrain cholinergic function. In order to








address this question, HACU was measured in two target regions of the basal

forebrain cholinergic system, the frontal cortex and hippocampus of intact,

ovariectomized and E2-replaced animals.




Methods


Animals were maintained and treated as described in the General Methods

section. HACU was assessed in two groups of animals, behaviorally tested and

behaviorally untested (naive) animals. HACU was determined in the frontal cortex

and hippocampus of behaviorally tested and behaviorally naive intact, OVX and E2-

replaced OVX animals. In one set of behaviorally tested animals, OVX animals had

been ovariectomized for 3 weeks. The other 2 groups of animals in this experimental

set were both E2 replaced, but for different durations. Following 3 weeks of

ovariectomy, one of the E2 replaced groups received an E2 pellet for 1 day and was

subsequently behaviorally tested. The second E2 replaced group was maintained on

the E2 pellet replacement regimen for 2 weeks prior to behavioral testing. The

resulting 3 groups in this experimental set were: OVX, 1 day E2 replaced, and 2 week

E2 replaced. The E2 replacement regimens were maintained throughout the

behavioral testing period. On the day following the 2 weeks of behavioral testing,

animals were sacrificed by decapitation and were processed for HACU determination.

Thus, at the point of measuring neurochemical markers, the OVX animals had been

E2 deprived for 5 weeks, while the 1 day E2 and 2 week E2 groups had received the

E2 replacement for 2 weeks and 4 weeks, respectively.

A second group of behaviorally tested animals underwent a longer treatment

regimen in which OVX animals had been ovariectomized for 28 week prior to

behavioral testing. E2 pellet animals had been ovariectomized for 3 weeks followed


I








by 25 weeks of E2 replacement. Following behavioral testing at the 28 week time

point, animals were sacrificed on the following day for the determination of HACU.

In both behaviorally tested groups, the data were analyzed according to hormonal

environment as well as according to learning performance, that is, all animals,

regardless of hormonal treatment, were segregated into 2 groups: learners and non-

learners.

The behaviorally naive group was maintained on the treatment regimen for 5

weeks, after which HACU was assessed. This was done in order to address the effect

of hormonal environment alone on cholinergic function.


High Affinity Choline Uptake (HACU)

Following decapitation, brains were removed from the skull and placed on an ice-

cooled surface. The hippocampus and frontal cortex were then dissected and

immediately placed into ice cold 0.32 M sucrose buffer (0.32 M sucrose, 1.0 mM

EGTA, 100 gM TRIS-HCI, pH = 7.4 at 40C). Average wet weights were 180 mg and

110 mg for the frontal cortex and hippocampus, respectively. Tissue samples were

then homogenized with a dounce homogenizer at 400 rpm. Homogenized samples

were subsequently centrifuged at 1000 x g for 8 minutes at 40C. The supernatant (Sl

fraction) was centrifuged at 30000 x g for 15 minutes at 4C. Following this spin, the

supernatant was discarded and the resulting pellet (P2 fraction) was resuspended in 2

ml of cold, oxygenated Krebs buffer (139 mM NaCI, 5 mM KCI, 13 mM NaHCO3, 1

mM MgCI2, 1 mM NaH2PO4, 10 mM glucose, 1 mM CaCl2; oxygenated for 15

minutes with 95% 02/5% CO2). Each reaction tube contained 200 tl of P2

suspension. High affinity choline uptake was determined in triplicate in the presence
of 1 piM [3H]-choline (final specific activity: 4.5 Ci/mmol, New England Nuclear,

Cambridge, MA) at 370C for 3 min. Non-specific uptake was estimated by adding 5

pM hemicholinium-3 (Sigma Chemical Corp., St. Louis, MO); these hemicholinium-








3 values were subtracted to obtain high affinity values. Unused tissue was stored at

-300C for subsequent determination of ChAT activity and protein levels. Final

specific uptake values were normalized for protein concentration in the P2 fraction.


Protein Assay

Analysis of protein content in the P2 preparation was conducted according to the

method of Bradford using Coomassie Blue dye (Bradford, 1976). 100 Wl aliquots of

sonicated P2 sample were placed into clean 13 x 100 mm borosilicate tubes. Freshly

made Coomassie blue dye (5 ml) was added to each tube, vortexed and read at 595

nm on a spectrophotometer. Once the dye was added, samples were run within an

hour in order to avoid the protein-dye complex from becoming unstable and

dissociating. All samples were run in duplicate and in parallel with a standard curve

ranging from 0 to 100 gg protein per tube.


Statistical Analysis

HACU differences for the behaviorally naive animals were analyzed using a t-test

since Intact vs. OVX and OVX vs. E2-treated were evaluated only once each in two

separate studies. HACU values in other experiments underwent ln(x) transformation

prior to analysis. Statistical analysis was carried out using an analysis of variance

(ANOVA). Multiple comparisons among groups were performed using Scheffe's

post-hoc test.




Results


In the behaviorally tested, short-term treatment group, animals in the 1 day E2

group demonstrated the highest level of HACU, while the levels for the OVX and 2








week E2 groups were approximately identical (Figure 4-1). However, in spite of the
nearly 2-fold difference between the 1 day E2 group and the other two groups, this

elevation was not statistically significant due to the relatively high variability in the

data. When the animals were categorized into learners and non-learners irrespective

of treatment group, frontal cortical HACU was found to be 2-fold higher than that for

non-learners (Figure 4-2). Hippocampal HACU was not different between learners

and non-learners (Figure 4-2).

At the 28 week time point, frontal cortical HACU was significantly elevated in

the E2 pellet animals relative to the OVX and Intact groups by 49% and 46%,

respectively, while hippocampal HACU was elevated by 59% and 47%, respectively

(Figure 4-3). When animals were categorized according to learning performance

(learners vs. non-learners), learners showed similarly elevated levels of HACU in the

frontal cortex relative to non-learners similar to those seen in the short term group,

although at this time point, the difference was not as great (15%; Figure 4-4). As in

the short term group, no significant differences in HACU were observed in the

hippocampus between learners and non-learners (Figure 4-4). Within each treatment

group, it was also observed that 83% of the E2-pellet animals learned, while only

16% of the OVX animals learned (based on the criteria of an animal performing

correctly in at least 11 out of 15 trials).

In behaviorally naive animals, however, ovariectomy significantly reduced

HACU by 24 % in the frontal cortex and by 34 % in the hippocampus (Figure 4-5).

E2-replacement resulted in a reversal of this effect of ovariectomy, increasing HACU

by 82 % in the frontal cortex and by 46 % in the hippocampus (Figure 4-5).







Frontal Cortex


T


-F


-T-


OVX E2 (1 Day) E2 (2 Wks)

Hippocampus


-F


OVX


--


T


E2 (1 Day) E2 (2 Wks)


Figure 4 1.


The effect of 5 weeks of Ovariectomy and Two different Estradiol
Replacement Regimens on High Affinity Choline Uptake in
Behaviorally Tested Rats. High affinity choline uptake (HACU) was
assessed in animals that had been ovariectomized for 3 weeks prior to
behavioral testing and in animals that had been E2 replaced for 1 day
and 2 weeks prior to behavioral testing. At the time of HACU
assessment, OVX animals had been estrogen deprived for 5 weeks.
Sample sizes were 6 for both OVX and 1 day E2 animals, and 5 for 2
week E2 animals.








Frontal Cortex


*
~~1~~


Learner


Non-learner


Hippocampus


-r


0--


Learner


Figure 4'-2.


Non-learner


The Effect of Learning Performance on High Affinity Choline
Uptake in the Frontal Cortex and the Hippocampus of Short term
Treated Rats. Animals that were ovariectomized for 3 weeks and E2
replaced for 1 day and 2 weeks prior to behavioral testing were
segregated into learners and non-learners. The level of HACU was
higher in the frontal cortex of learners while HACU was unaffected by
learning status in the hippocampus. Statistical significance was
determined by ln(x) transforming the data followed by an ANOVA (* :
p 5 0.05 vs learner).








Frontal Cortex


0-4-


OVX E2 Pellet


Hippocampus


0 4-


OVX E2 Pellet


Figure 4-3.


The Effect of 28 Weeks of Ovariectomy and Estradiol
Replacement on High Affinity Choline Uptake in the Frontal
Cortex and Hippocampus of Behaviorally Tested Rats. E2 pellet
animals demonstrated significantly higher HACU levels than either
Intact or OVX animals in both the frontal cortex and hippocampus.
Sample sizes were 8 per group. Statistical significance was determined
by first ln(x) transforming the data followed by an ANOVA. Post hoc
determination of group differences was done using the Scheffe's test (*
: p < 0.05 vs Intact and OVX).


Intact


Intact








Frontal Cortex


*


0 4-


Learner


Non-learner


Hippocampus


!T


Learner


Non-learner


Figure 4 4.The Effect of Learning Performance on High Affinity Choline
Uptake in the Frontal Cortex and the Hippocampus of 28 Week
Treated Rats. Animals that were ovariectomized for 28 weeks and E2
replaced prior to behavioral testing were segregated into learners and
non-learners. The level of HACU was higher in the frontal cortex of
learners while HACU was unaffected by learning status in the
hippocampus. Statistical significance was determined by ln(x)
transforming the data followed by an ANOVA (* : p < 0.05 vs
learner).


"-- -









Frontal Cortex


-r


Intact OVX


OVX E2


Hippocampus


-1-


Intact OVX


Figure 4 5.


SOVX E2


The effect of 5 week Ovariectomy and Estradiol Replacement on
High Affinity Choline Uptake in Behaviorally Naive Rats.
Ovariectomy resulted in a significant reduction in HACU relative to
Intact animals. E2 replacement resulted in a significant elevation of
HACU relative to OVX. Sample size was 6 per group. Statistical
evaluation of the data was performed using the students t-test since the
differences between Intact vs. OVX and OVX vs. E2 were evaluated
from different experiments (* : p 5 0.05 vs. Intact; t : P 5 0.05 vs
OVX).








Discussion


High affinity choline uptake (HACU) reflects moment to moment activity of the

cholinergic neurons (Simon et al., 1976; Kuhar and Murin, 1978). In this

experimental paradigm, HACU was assessed in both behaviorally tested and

behaviorally naive animals. In the short term treatment group, HACU did not

correlate well with estrogenic milieu in behaviorally tested animals. Instead, it was

observed that HACU correlated much more closely with successful learning. In

support of this conclusion, it has been observed that the effects of training (Decker et

al., 1988) and learning (Wenk et al., 1984) alters the high affinity uptake of [3H]-

choline in the frontal cortex and hippocampus of young adult rats. Decker et al..

(1988) described a reduction in hippocampal HACU 4 days after place-training in a

Morris water task but observed no effect on hippocampal HACU in cue-trained rats.
Wenk et al.. (1994) described a similar alteration in hippocampal HACU, but in the

opposite direction using different tasks (Radial arm maze and the T-maze). Thus it

seems apparent that the effect of training on HACU is task dependent, presumably

employing different anatomical pathways for learning and memory formation. In our

studies, the primary region that was affected by active avoidance training was the

frontal cortex.

It is conceivable that altered levels of stress during the testing procedure may

affect learning and consequently HACU. In the hippocampus, short term restraint

stress has been documented to increase HACU (Finkelstein et al., 1985). However,

when the stress is prolonged or repeated, decreased levels of hippocampal HACU
have been observed (Finkelstein et al., 1985; Lai et al., 1986). Typically, the
immobilization stress that produces the above mentioned changes in hippocampal

HACU parallels changes in the frontal cortex (Lai et al., 1986). Our results document

changes in frontal cortical HACU that are distinct from those in the hippocampus.








While the potential effect of stress from behavioral testing in the active avoidance

paradigm cannot be overlooked, the regional difference in HACU in response to

learning suggests that the level of successful learning itself may alter HACU.

In the 28 week behaviorally tested animals, a similar correlation between learning

and HACU was observed although at this time point, the magnitude of the difference

was not as great (15%). The animals at this time point are about 11.5 months of age.

It is possible that at this age, the basal forebrain cholinergic system of female rats

undergoes an age related diminution in responsiveness to learning or behavior. This

hypothesis is consistent with the observation that aged male rats do not demonstrate

the training induced HACU changes as do young adult rats when place-trained in the

Morris water task (Decker et al., 1988). Furthermore, while 12 months of age in a

male rat is typically not considered to represent an aged animal, in female rats, a

different pattern may hold true. Evidence for this comes from the observed

occurrence of reproductive acyclicity, suggesting reproductive senescence, at about

12 months of age (Lu et al., 1979). Consistent with this theory is the observation that

in the 28 week Intact animals, the levels of serum estradiol are statistically equivalent,

but demonstrate a smaller standard error. This reflects the probability that

approximately half of the animals became acyclic, thus producing a constant level of

estrogen.

Following the observation that HACU reflected successful learning more closely

than estrogenic milieu, HACU was assessed in behaviorally naive animals as well in

order to dissect out the effect of hormonal environment alone on the function of basal

forebrain cholinergic neurons. Five weeks of ovariectomy and two weeks of E2

treatment were sufficient to modify cholinergic activity as assessed by HACU.

O'Malley et al.. (1987) reported that 4 weeks of ovariectomy reduced, and E2-

replacement by way of 3 subcutaneous injections given every other day, normalized

HACU in the frontal cortex of rats. This reversible decline in cholinergic activity






53

indicates that E2 environment of the animal helps to define the activity of the

cholinergic projections to the cortex and the hippocampus. As such, the reversible

learning deficit induced by ovariectomy may be dependent upon the activational

effects of physiological levels of E2 on these cholinergic neurons.















CHAPTER 5
THE EFFECT OF OVARIECTOMY AND ESTRADIOL-REPLACEMENT

ON CHOLINERGIC FUNCTION: CHOLINE ACETYLTRANSFERASE



The involvement of the basal forebrain cholinergic system in cognitive

performance has been proposed by several authors (see Dekker et al., 1991; Gray and

McNaughton, 1982). Some of this evidence comes from animal models in which

discrete lesions of forebrain nuclei (Flicker et al., 1983; Mouton et al., 1988; Decker

et al., 1992) or their projections (Bresnahan et al., 1992) lead to deficits in learning

and/or memory. These behavioral deficits are found to correlate well with cholinergic

neurochemical deficits also induced by these lesions. Studies in post-mortem tissue

that have correlated behavioral deficits with cholinergic pathology have also been a

useful tool in understanding the neurochemical substrates for learning and memory.

For example, in AD, perhaps the most consistent and widespread neurochemical

deficit seen is that of the basal forebrain cholinergic system. Furthermore, the

pathology associated with this disease has been associated with a decrease in ChAT

and the hydrolytic enzyme, acetylcholinesterase (AchE; Mountjoy, 1986). As in the

animals studies, this decline in cholinergic neurochemical markers has been

correlated with the degree of dementia (Perry et al., 1978).

Given the aforementioned behavioral deficits associated with E2 deprivation, and

the impairment in cholinergic function (as supported by the previously observed

decline in HACU), it was postulated that these deficits may be the result of the

withdrawal of a trophic influence on basal forebrain cholinergic neurons.








Accordingly, the functional deficit described by the decline in HACU may be a

consequence of neuronal degeneration. In order to assess the potential trophic

influence of estradiol on cholinergic neurons, ChAT was measured in the frontal

cortex and hippocampus of OVX and E2-replaced OVX rats.




Methods


Animals were maintained and treated as described in the General Methods

section. ChAT was measured in the frontal cortex and hippocampus of 5 week- and

28 week OVX and E2 replaced rats. Intact controls were run in parallel to these two

groups of animals. The short- and long-term E2 replaced animals had both been

OVXed for 3 weeks. The replacement regimen lasted for 2 weeks for the 5 week

group and 25 weeks for the 28 week group.


ChAT Assay

ChAT activity was determined in both 5 week group as well as the 28 week group

of animals. ChAT was assayed following a modified version of Fonnum (1975). The

ChAT assay measures the amount of enzymatic activity in a given tissue sample by

quantifying the amount of [3H]-Ach formed from its precursors, [3HJ-acetyl

coenzyme A (ACoA) and choline.

Since this assay had not been previously run in our laboratory, it was necessary to

evaluate the appropriate concentration of tissue extract to use in order to optimize the

assay. The tissue samples used were frozen crude synaptosomal preparations (P2

samples) previously prepared for the determination of HACU. P2 samples were

already in a volume of KRB. An aliquot of 100 Wl of the P2 sample was diluted 1:1

with KRB containing 2% 1-butanol giving a final concentration of 1% 1-butanol in








the mixture. This mixture was then sonicated for 15 seconds followed by

centrifugation at 13000 x g for 5 min. Different volumes (thus reflecting different

concentrations) of the resulting supernatant (tissue extract containing both membrane

bound and cytosolic sources of ChAT) were then evaluated for ChAT activity. The

volumes used were 0 gl, 5 pl, 10 gl, 15 pl and 20 pl of the supernatant. The

appropriate amount of KRB was added to each tube in order to bring the total volume
of the sample up to 20 pl. The goal of this procedure was to select a tissue

concentration that did not make the amount of substrate for the reaction, [3H]-ACoA,

the limiting reagent in the assay. If the enzymatic activity in the tissue sample was too

high (a reflection of tissue concentration), then the ACoA would always be exhausted

and potential group differences would not be detected. ChAT levels that corresponded

to the varying tissue levels are presented in Figure 5 1.

20 pl of the tissue extract was incubated with the reaction mixture which

contained 0.28 mM [3H]-ACoA (specific activity: 45 gCi/gmol, New England

Nuclear, Cambridge, MA), 7.8 mM choline chloride and 0.2 mM physostigmine
(Sigma Chemical Corp., St. Louis, MO). Incubation with the [3H]-ACoA was carried

out for 30 min at 37C. The reaction was terminated by the addition of ice cold

glycyl-glycine buffer (pH = 8.6). Following a 10 minute incubation at 4C,

tetraphenyl boron dissolved in butyronitrile (10 mg/ml) was added to the reaction

tube allowing for liquid cation exchange extraction of acetylcholine (ACh). Samples

were vortexed and centrifuged in a bucket centrifuge at low speed (185 x g) for 5

minutes to allow settling and separation of the organic and aqueous phases. 100 pl of

the organic phase was then aliquoted into 7 ml scintillation vials and 4 ml of
scintillation fluid (LiquiscintT, National Diagnostics, Atlanta, GA) were added.

Vials were then counted in a Hewlett-Packard scintillation counter for 5 minutes, and

the dpm were converted to pmoles and values were normalized for protein content in

the P2 preparation.








Statistical Analysis

Statistical evaluation of the data was performed using an ANOVA followed by

Scheffe's test for post-hoc determination of group differences. Evaluation of treatment

effects as well as the effect of length of ovariectomy were analyzed.




Results


Optimization of the Assay:

In order to optimize the assay, an appropriate concentration of sample to be used

was evaluated. The concentration selected was 20 pl of the supernatant derived from

the ChAT extraction. This concentration was on the linear portion of the

concentration-activity curve that was generated (Figure 5-1). The correlation

coefficients for the curves generated from frontal cortical and hippocampal tissue

were 0.998 and 0.999, respectively. The correlation coefficients were derived based

on the assumption that the dose-response curve followed a 3rd degree polynomial fit.


Choline Acetyltransferase Activity

In the frontal cortex, no significant differences in ChAT activity due to hormonal

manipulation were detected within 5 weeks (Table 5-1). Interestingly, at the 28 week

time point, ChAT levels in the frontal cortex were reduced in both the INTACT and

OVX groups by 61% and 56 % respectively (Table 5-1). In the E2-pellet group,

however, this reduction was only 16% (Table 5-1).
In the hippocampus, five weeks of ovariectomy was a sufficient time period to

induce a significant reduction in ChAT activity, and 3 weeks of E2-replacement

reversed this effect (Table 5-2). The reductions in ChAT activities in INTACT and

OVX animals between 5 and 28 weeks were comparable with those seen in the frontal






58

cortex. In E2-pellet animals, loss of ChAT activity was larger than that seen in the

frontal cortex but was less than that seen in OVX or INTACT animals (Table 5-2).











1000

800O

600

400-

200-


I .


Frontal Cortex


R2 = 0.998








30


Hippocampus


R2- 0.999


0 10 20
Volume of supernatant (pl)


Figure 5 1. Tissue Optimization of the ChAT assay. Varying concentrations of
tissue extracts containing the enzyme, ChAT, were assayed in order to
determine the optimal concentration for the assay used. In order to
determine if different concentrations were required for the two tissue
types used, this preliminary determination of optimal extract
concentration was determined in both the hippocampus and the frontal
cortex. Correlation coefficients (R2 values) are provided based on a
3rd degree polynomial curve fit.


1000

800

600

400

200

0


re'/










Table 5 1. Effect of Short Term and Long Term Ovariectomy and
Estradiol Replacement on Choline Acetyltransferase
Activity in the Frontal Cortex.


Treatment Group 5 wks 28 wks

ChAT activity (nmol/30 min/mg protein)

INTACT 10.2 0.5 4.0 0.1

OVARIECTOMIZED 9.2 0.6 4.0 0.2

E2 PELLET 9.8 0.6 8.2 0.8*

n = 6 for Ovariectomized and E2 Pellet groups and n = 5 for Intact group for the 5
week time period. For the 28 week time period, n = 6 for all treatment groups.
* p < 0.05 vs Intact and OVX












Table 5 2.


Effect of Short Term and Long Term Ovariectomy and
Estradiol Replacement on Choline Acetyltransferase
Activity in the Hippocampus.


Treatment Group 5 wks 28 wks

ChAT activity (nmol/30 min/mg protein)

INTACT 13.2 0.8 5.7 0.3

OVARIECTOMIZED 10.3 0.3 6.2 + 1.1

E2 PELLET 12.7 0.5 8.0 1.1

n = 6 for Ovariectomized and E2 Pellet groups and n = 5 for Intact group for the 5
week time period. For the 28 week time period, n = 6 for all treatment groups.
* p < 0.05 vs. Intact and E2 Pellet








Discussion


ChAT activity provides an estimate of the relative extent of cholinergic

innervation of target tissues. As such, ChAT activity was measured in an attempt to

define the effect of ovariectomy on the cholinergic innervation of the cortex and

hippocampus. ChAT activity was decreased in the hippocampus, but not in the frontal

cortex, of animals ovariectomized for 5 weeks. The magnitude of the decline in ChAT

activity in the hippocampus (22%) was similar to the observed decline in HACU in

this region (34%), suggesting that this enzymatic change may have contributed to the

decline in HACU in the hippocampus. While our experiments did not include the

measurement of Ach formation or release, O'Malley et al.. (1987) have described an

ovariectomy-induced decrease in HACU without a concomitant decrease in

spontaneous or potassium-induced Ach release. The length of ovariectomy, however,

in the O'Malley study was only 3 weeks. It is conceivable that longer durations of

estrogen deprivation may lead to compromised Ach formation or release.

The observed decline in neocortical HACU in the absence of an associated decline

in ChAT activity at the 5 week time point indicates that the two neurochemical events

may be affected independently by E2. The fact that the neocortex contains both

intrinsic populations of cholinergic neurons as well as terminals arising from nucleus

basalis and thalamic inputs (Emson and Lindvall, 1986; Perry, 1986) may explain, at

least in part, the region specific effects of E2. It is conceivable that declines in ChAT

activity in the frontal cortex may have been masked by an insensitivity of either

intrinsic neurons or the projection neuron terminals to E2. Additionally, the apparent

lack of effect of ovariectomy on cortical ChAT activity may also reflect a decreased

sensitivity to estrogenic milieu by virtue of its lower estrogen receptor density (Pfaff

and Keiner, 1973, Simerly et al., 1990).








While the results demonstrate a decrease in enzyme activity, they do not address

whether this decline in activity is due to a decrease in the Vmax or Km of the enzyme

for either substrate. It has been reported the decreases in ChAT activity observed in

post mortem tissue of AD brains can be explained by either decreased Vmax or

increases in Km. In spite of this lack of information, decreases in ChAT activity

(without dissecting out the effect of Vmax vs. Kin) have been correlated well with

neurodegenerative changes in Alzheimers disease.

In other neurodegenerative disorders such as Parkinson's disease (PD), it has been

noted that only after the levels of the neurotransmitter dopamine decrease to below

80% of control levels does one see symptomatology of PD. Thus, smaller decreases in

neurotransmitter levels may not be biologically relevant. I argue against this

hypothesis, in that, smaller decreases in either neurotransmitter levels or enzyme

levels may suggest the early stages of the progressive disorder. Accordingly, the

level of decrease that was observed in the short term group may suggest the beginning

of degeneration, while in the long term Intact and OVX groups, the decline in ChAT

may be severe enough to affect the ability of the enzyme to make Ach, the native

neurotransmitter necessary for the basal forebrain cholinergic neurons.

The neocortex and hippocampus also differed in their response to chronic E2

exposure over time. At the 28 week time point, we found that neocortical ChAT

activity was reduced by 61% and 56 % in the INTACT and OVX groups,

respectively. However, E2-replaced animals showed decreases in ChAT activity of

only 16% over the same time period. This suggests that E2 may have a

neuroprotective effect on basal forebrain cholinergic projections to the cortex. In the

hippocampus, the protective effects of chronic E2 exposure were less apparent, that

is, the magnitude of ChAT decreases were similar to those seen in the frontal cortex

in Intact and OVX animals, while E2 replacement was not as efficacious in protecting

the hippocampal ChAT activity. The decrease in ChAT activity in the INTACT








female rats over the 28 week course of this study was unexpected given the reported

lack of decline in ChAT activity in male rats over this same time period (Williams

and Rylett, 1990). It is possible that female rats undergo an age related diminution in

ChAT activity as early as 12 months of age in association with the observed

occurrence of reproductive acyclicity at about 12 months of age (Lu et al., 1979).

Recently, it has been shown that female Fisher 344 rats undergo a 30% and 40%

decrease in ChAT and AchE, respectively in the ventral globus pallidus (Luine et al.,

1986). It should be noted, however, that the effects of age on cholinergic markers in

the brain are strain specific (Michalek et al., 1989; Michalek et al., 1990). Therefore,

while the effects of age on cholinergic markers in the male Fischer 344 rat are well

documented, the age-associated changes in the female Sprague-Dawley rat are

relatively less known.

The preservation of ChAT activity in the hippocampus and frontal cortex of E2

treated animals may suggest the prevention of neuronal degeneration in these areas.

This neurotrophic or cytoprotective role of E2 may have therapeutic merit for such

neurodegenerative diseases such as AD.
















CHAPTER 6
THE EFFECT OF OVARIECTOMY AND ESTRADIOL-REPLACEMENT ON

NGF PROTEIN AND mRNA LEVELS




The preservation of ChAT activity in the frontal cortex and hippocampus of E2

replaced rats has led to the suggestion that estrogen may serve a cytoprotective or

neurotrophic function for basal forebrain cholinergic neurons. The neurotrophic

actions of E2 may occur either through a direct interaction with the target cell or

through a more indirect mechanism, specifically, by modulating the synthesis and/or

release of a putative neurotrophin such as NGF. The discovery that certain neurons of

the basal forebrain and their target regions possess both estrogen receptors and

neurotrophin receptors (Toran-Allerand et al., 1992; Miranda et al., 1993) has set a

precedence for the potential interaction of these two substances.

The regulation of cholinergic function and potentially the viability of cholinergic

neurons have been documented to be affected by estrogenic milieu (Chapters 4 and

5). This modulation of cholinergic activity has also been shown to be correlated to

learning and memory function (Chapter 3). Coincidentally, neurochemical cholinergic

markers such as ChAT, HACU and Ach synthesis and release have also been shown

to be regulated by NGF (Williams and Rylett, 1990; Rylett et al., 1993). In lesion

studies, NGF has been shown to rescue the damaged neurons (Montero and Hefti,

1988). Furthermore, in animal lesion models that impair cognitive function in

animals, NGF has also been shown to improve the behavioral deficits (Dekker et al.,








1992). Taken together, these effects of NGF seem to be closely related to the

documented effects of estrogen. As such, the potential neurotrophomodulatory role of

E2 was investigated by quantifying the level of NGF protein and its message in

ovariectomized and estradiol-treated rats.




Methods


Animals were maintained and treated as described in the General Methods

section. Both NGF protein and NGF mRNA were measured in the frontal cortex and

hippocampus of 3 month ovariectomized and E2 replaced animals. E2 replaced

animals were ovariectomized for 3 weeks followed by E2 replacement for 9 weeks.

Intact controls were run in parallel with these animals. However, due to a technical

error, the NGF protein determination in the frontal cortex of 3 month INTACT

animals was not performed. As a result, the only comparison made for frontal cortical

NGF protein levels was between OVX and E2 pellet animals.


NGF Protein Measurement:

The method employed in the measurement of nerve growth factor (NGF) protein

levels was the double antibody enzyme-linked immunosorbent assay (ELISA).

Following extraction of NGF from tissue, the extract was subjected to 2 monoclonal

antibodies against NGF (clone 27/21, obtained from Boehringer Mannheim,

Indianapolis, IN). The 2 antibodies are applied at different stages of the ELISA

procedure in order to facilitate a sandwich effect. Simplistically, one could envisage

the 2 antibodies on either side of the NGF molecule. The second Ab is conjugated

with an enzyme which converts an added substrate to a colored product. The intensity

of color generated is a function of NGF concentration. Thus, when experimental








samples are run in parallel with a group of known concentration standards, NGF

protein content can be estimated.


Extraction of NGF Protein

Following decapitation, brains were removed from the skull and placed on an ice-

cooled surface. The frontal cortex and hippocampus were dissected out and weighed.

Based on this wet weight, 2 volumes of extraction buffer (100 mM Tris-HCl, 400 mM

NaC1, 2% (w/v) bovine serum albumin, 0.05 % (w/v) Na Azide, Protease inhibitors:

EDTA 4 mM, 7 tg/ml Aprotinin and 1 mM PMSF) were added and the tissue was

homogenized in a dounce homogenizer set at 400 rpm. Following homogenization,

the tissue samples were centrifuged in a Beckman high speed centrifuge at 48,400 x g

at 4C. The resulting supernatant was diluted 1:1 with the supernatant buffer (0.2%

Triton x-100, 20 mM CaCI2). This resulting extract was applied to the appropriate

well of the 96-well microtiter plate as the sample.

The assay itself consisted of first coating the wells of the 96-well plate with the

primary Ab. Following wash off of excess primary antibody, either the unknown

processed tissue, or standard NGF (obtained from Boehringer Mannheim,

Indianapolis, IN) in a fixed volume, was added and incubated with the primary

antibody. The next step involved incubating a second Ab (the anti-NGF p-

galactosidase conjugate, Boehringer Mannheim, Indianapolis, IN) with the already

bound primary Ab-NGF complex. The amount of p-galactosidase conjugate bound

is directly proportional to the NGF content of the sample. The amount of conjugated

Ab (anti NGF-p-galactosidase) bound was determined by adding the substrate for the
enzyme (Chlorophenol red-p-D-galactopyranoside, Boehringer Mannheim,
Indianapolis, IN) in excess quantity. The substrate is broken down into

galactopyranose and chlorophenol red. The latter imparts a red color to the media.

Higher concentrations of NGF correspond to a more intense color which is








quantitatively measured as the absorbance read at 575 nm. By running a standard

curve along with the samples, the concentration of NGF present in a given tissue

sample was determined. Due to the fact that some NGF is lost during the processing

of tissue, a single tissue sample (usually run in triplicate) was spiked with a known

amount of NGF and assayed. By comparing the concentration of NGF in the unspiked
tissue versus that of the spiked tissue, the percent NGF recovered over the tissue

processing steps can be calculated.

Tissue concentration, primary and secondary Ab concentrations (or dilutions) and
varying lengths of exposure of the conjugate Ab to the substrate were tested for

conditions that would provide optimal conditions for this assay. Initially, a dilution

factor of 1:4 was selected for frontal cortical and hippocampal tissue so as to provide
enough sample at the end of the processing to run the extract in triplicate. However, it

was determined that tissue from one animal that was diluted in this manner produced
NGF levels that were below the sensitivity of the assay (6.25 pg/ml). As a result,

frontal cortical and hippocampal tissue from 2 animals were pooled and referred to as

a single sample. An initial dilution of 1:2 with extraction buffer was used for the

pooled tissue. The subsequent dilution with the supernatant buffer resulted in a final

dilution of 1:4 per pooled tissue sample.

The concentration of primary and secondary (conjugated) antibodies were also

evaluated for that which would optimize the assay without wasting costly Ab. The

primary or coating Ab was tested at four dilutions (1:30, 1:60, 1:120 and 1:240). The

secondary Ab was tested at a 1:10, 1:20, 1:40, 1:80 and 1:160 dilution. The resulting

standard curves generated are presented in Figure 6 1 through 6 4.
The effect of varying lengths of time to which the secondary conjugated Ab was

incubated with the substrate was also evaluated. The optimal incubation time was 1.5

hr as assessed by the time at which the standard curve had the largest range of NGF

values on the linear portion of the curve before it started to plateau.








Measurement of NGF mRNA:

NGF mRNA was quantified by the use of the dot blot technique. Total RNA was

first extracted from fresh tissue samples, after which they were blotted onto a nylon

membrane and U.V. crosslinked to avoid further degradation of RNA. The nylon blot

was then hybridized with a [32P]-labeled NGF probe (771 bp) recognizing the entire
pre-pro sequence of the NGF mRNA. The amount of hybridization was then

quantified with the use of a Betagen that converted radioactive signal to blots of

different intensity. The NGF signal was normalized for the amount of DNA loaded

onto the nylon membrane by stripping the NGF probe and subsequently rehybridizing

the blot with an actin probe.


RNA Isolation:

Total RNA was isolated using the acid guanidinium isothiocyanate method
followed by the phenol-chloroform method of purification and isopropanol

precipitation (Sambrook et al., 1989). All solutions used in the isolation and

purification of RNA were diethylpyrrocarbonate (DEPC) treated. Freshly dissected

tissue was homogenized in Lysis solution (4M guanidinium thiocyanate, 25 mM Na

citrate, 0.5% N-lauryl sarcosine, 0.1M 2-mercaptoethanol) using a Polytron

homogenizer set at setting 6. Following this procedure, buffer-saturated phenol

(Gibco-BRL) and chloroform/isoamyl alcohol (49:1) were used to purify the RNA

from protein contaminants. The phenol aids to remove the protein contaminants while

the chloroform/isoamyl purification steps are used to remove the phenol. 2M NaAc,

TRIS buffered phenol and 49:1 chloroform/isoamyl alcohol was added to the

homogenate and vortexed for about 1 min. This suspension was centrifuged at 10000

x g at 4C for 15 min. after which the aqueous (top) layer was aliquoted into a second

tube. Phenol and the chloroformn/isoamyl solution were re-added to the new tube,

vortexed and recentrifuged at 10000 x g for 15 min. at 4C. The upper phase was








again removed and aliquoted into yet another fresh tube. An equal volume of

isopropanol was added and vortexed. This mixture was stored overnight at -200C. The

following day, isopropanol containing solution was centrifuged at 10000 x g at 4C.

At this point a pellet had formed. The supernatant was discarded and the pellet was

dissolved in a small volume of lysis solution. The isopropanol precipitation was

repeated. Once this was accomplished, an ice-cold 75% ethanol wash was performed.

The ethanol was evaporated off and the remaining pellet was resuspended in 0.5%

SDS. The concentration of the RNA was evaluated spectrophotometrically at a

wavelength of 260 nm. The purity of RNA was assessed by calculating the ratio of
absorbance at 260/280 nm. A ratio of 2 is considered pure RNA. Samples that had

ratios below 1.6 were not used as they were deemed to be too contaminated by

protein.

Total RNA samples were then blotted onto a nylon membrane and U.V.

crosslinked. The blot was subsequently stored in a dessicator until the time of

hybridization.


Preparation of the NGF Probe

The NGF probe was a gift from Dr. Scott Whittemore (Miami, Fl) and Dr. Paul
Isackson (Mayo Clinic, Jacksonville). The probe was received incorporated in pBS

(Bluescript). Before the probe could be labeled with the radionuclide, 32P, it had to be

excised from the plasmid. Consequently, a restriction enzyme digestion was

performed using the enzymes BAMH I (4 U/tg DNA) and ECOR I (5 U/pg DNA).

The restriction enzyme digestion was carried out at 370C for 3 hrs.

At this point the excised probe was separated from the plasmid by running the
sample on a 1% low melting temperature agarose gel. After letting the gel run for 3

hrs at 60V in a cold room, the bands were photographed (Figure 6 5) and the band

corresponding to the NGF probe was cut out of the gel and placed into an eppendorf








tube. The following day, the probe was extracted from the gel and purified using the

phenol/chloroform/isoamyl alcohol method previously described. The resulting pellet

following the ethanol wash was resuspended in 1 x TE (pH = 8). The concentration

and purity of the DNA probe was assessed spectrophotometrically. A 260/280 ratio of

1.8 is considered pure DNA.


Hybridization of Nylon Membrane

The nylon membrane was placed in a Seal-a-Meal plastic pouch in which

prehybridization buffer (5 x SSPE, 5 x Denhardts solution, 50% formamide, 0.5%

SDS) was added. The pouch was sealed and incubated in a shaking water bath at

420C for approximately 1 hr. During this prehybridization time period, the NGF

probe was labeled with [32P]-CTP (Specific activity: 3000 Ci/mM, New England

Nuclear, Cambridge, MA) using the Random primer probe labeling kit (Gibco-BRL).

Once the labeling procedure was complete, the labeled probe was separated from the
free [32P]-CTP by passing the reaction solution through a Sephadex-50 column. 1 gl

of the labeled probe fraction was counted in 4 ml of scintillation cocktail on a

scintillation counter. Based on a precalculated volume of hybridization buffer to be

used, the amount of labeled probe that would give a final activity of 1-2 x 106

cpm/ml of hybridization buffer was added to the hybridization buffer. The

hybridization buffer was maintained at 420C in a warm water bath until the time of

use.

Following prehybridization, the prehybridization buffer was discarded. The

hybridization buffer was then carefully added and the pouch resealed. The pouch was

cleaned on its outer surface and placed into a second pouch to avoid contamination of

the water bath in which it was to be placed. Hybridization was carried out at 420C in a

warm shaking water bath for 22 hours.








The following day, the hybridization buffer was removed and the nylon
membrane was subjected to 4 washes (wash 1: 2 X SSC/0.1% SDS at room

temperature, wash 2: 2 X SSC/0.1% SDS at 600C, wash 3 and 4: 0.2 X SSC/0.1%
SDS at 600C). Following the washes, the membrane was blotted on Whatmans filter

paper. While still damp (but not wet), the nylon membrane was placed in a Betagen
densitometer overnight for the quantitation of NGF mRNA signal.
After the membrane had been quantified for NGF signal, the NGF probe was
stripped by incubating the membrane in a solution containing Na2P207 (5%), 2M
TRIS (pH = 8), 0.5M EDTA (pH = 8), 0.1 x Denhardts solution) at 650C for 3 hrs.

After this procedure, the blot was then ready for rehybridization with the actin probe.


Statistical Analysis

Each blot corresponded to tissue from a single animal. Levels of radioactivity
from each blot within a given treatment group and brain region were ln(x)
transformed prior to analysis. Analysis was performed using the ANOVA followed

by Scheffe's post hoc test for determinations of group differences.




Results


The testing of antibody dilution was performed to see what combination of
primary (coating) and secondary (conjugated) antibodies would result in the best

standard curve (Figures 6-1 through 6-4). From preliminary trials, it was observed
that NGF values from experimental samples fell on the lower part of the standard
curve. As such, it was necessary to optimize the slope of the lower portion of the
standard curve. By assessing what antibody titers produced a curve with greatest

sensitivity (small change in NGF concentration reflecting a relatively larger change in








optical density reading), it was found that a primary Ab dilution of 1:30 and a

secondary Ab dilution of 1:10 generated the best curve. Consequently, these Ab

concentrations were used in all subsequent assays.

In both the frontal cortex and hippocampus, 3 months of ovariectomy did not

result in a significant difference in the levels of NGF protein relative to E2-replaced

animals (Figure 6 6). Similarly, 3 months of ovariectomy and E2 replacement was

without effect in the hippocampus (Figure 6 6). In contrast, 3 months of

ovariectomy resulted in a significant reduction of 45% in frontal cortical NGF mRNA

levels (Figure 6 7). E2-replacement resulted in a partial, albeit non-significant

elevation in NGF mRNA. In the hippocampus, however, 3 months of ovariectomy

was without effect while E2 replacement resulted in a significant elevation of NGF

mRNA (Figure 6 7).










1:30 Dilution of 10 Ab


2 Ab Dilution:
---- 1:10
1:20
-- 1:40
1:80


0.1 1 I I


NGF (pg/ml)


Figure 6 1.The effect of different dilutions of secondary Ab on the NGF
ELISA standard curve at a 1:30 dilution of the coating Ab. 4
different dilutions of the conjugated (secondary) Ab were evaluated on
their ability to produce a standard curve with a relatively steep slope so
as to provide a more accurate optical density reading at the lower
concentrations of the curve.


0.3-




0.2-









1:60 Dilution


of 1 Ab


2 Ab Dilution
----- 1:10
1:20
-- 1:40
1:80


0 10 20


NGF (pg/ml)


Figure 6 -


2.The effect of different dilutions of secondary Ab on the NGF
ELISA standard curve at a 1:60 dilution of the coating Ab. 4
different dilutions of the conjugated (secondary) Ab were evaluated on
their ability to produce a standard curve with a relatively steep slope so
as to provide a more accurate optical density reading at the lower
concentrations of the curve.









1 Ab Dilution


of 1:120


2 Ab Dilution
-- -- 1:10
1:20
-- 1:40
1:80


0 10 20


NGF (pg/ml)


Figure 6 3.The effect of different dilutions of secondary Ab on the NGF
ELISA standard curve at a 1:120 dilution of the coating Ab. 4
different dilutions of the conjugated (secondary) Ab were evaluated on
their ability to produce a standard curve with a relatively steep slope so
as to provide a more accurate optical density reading at the lower
concentrations of the curve.










10 Ab Dilution


of 1:240


20 Ab Dilution
-U-- 1:10
-- 1:20
-- 1:40
0 1:80


0 10 20


NGF (pg/ml)


Figure 6 4.The effect of different dilutions of secondary Ab on the NGF
ELISA standard curve at a 1:240 dilution of the coating Ab. 4
different dilutions of the conjugated (secondary) Ab were evaluated on
their ability to produce a standard curve with a relatively steep slope so
as to provide a more accurate optical density reading at the lower
concentrations of the curve.


0.22


0.20


0.18


0.16


0.14


0.12


0.10







































Figure 6 5A low melting temperature agarose gel for the separation of
plasmid DNA from the NGF probe. Following restriction enzyme
digestion of the pBS plasmid using BAMH I (4 U/gg DNA) and
ECOR I (5 U/p.g DNA), the excised NGF probe was separated on a
low melting temperature agarose gel. Lanes 1 and 6 are DNA ladders.
Lanes 2 through 5 represent replicate samples obtained from the
enzyme digested mixture. By comparison with the DNA ladder, the
separated DNA fragment has an approximate size of 800 b.p. which
corresponds very closely with the actual size of the NGF probe (771
b.p).









Frontal


Cortex


1200-
1000
800-
600-
400
200
0-


Figure 6 6.


The effect of 3 month Ovariectomy and E2 Replacement on NGF
protein levels in the frontal cortex and hippocampus. No
significant differences were seen in frontal cortical NGF protein
levels when comparing OVX vs E2 pellet animals. In hippocampus,
ovariectomy and E2 replacement was also without effect. Sample
sizes were 11 for OVX and 6 for E2 Pellet in the frontal cortex while
sample size was 6 per group in the hippocampus.


rT


X E2 Pellet

Hippocampus












OVX E2 pellet


2000-

1500-

1000-

500-

0-


OV


Intact









Frontal


*


Cortex


-1-


0.0 ---


Intact OVX E2 Pellet

Hippocampus


--r


-I,,-


0.o(-


OVX E2 Pellet


Figure 6 -


7.The effect of 3 month ovariectomy and E2 replacement on
Frontal Cortical and Hippocampal NGF mRNA levels. In the
frontal cortex, 3 months of ovariectomy was sufficient to induce a
marked decline in NGF mRNA. E2 replacement resulted in only a
partial reversal of this deficit. Hippocampal NGF mRNA was
unaffected by ovariectomy. E2 replacement resulted in a modest
elevation of NGF mRNA. Sample sizes for Intact, OVX and E2 pellet
animals were 10, 12 and 8 for the frontal cortex and 6, 12 and 12 for
the hippocampus, respectively. (* : p < 0.05 vs Intact)


z
I --

U.-..
E ''
Z *


Intact








Discussion


NGF protein levels observed were consistent with literature values that report

highest levels of NGF in the hippocampus and slightly lower value in the cerebral

cortical regions (Whittemore et al., 1986). While NGF protein levels in the frontal

cortex and hippocampus seemed relatively unaffected by estrogenic milieu, E2

deprivation through ovariectomy resulted in a markedly significant decline in NGF

mRNA levels in the frontal cortex. Without the NGF protein levels from Intact

animals in the frontal cortex, it is difficult to determine if the mRNA data are in

agreement with the protein data. It may well have been that the Intact NGF protein

levels are higher than the OVX levels, and that chronic E2 replacement was not

successful in normalizing the protein levels.

Alternatively, the lack of effect of OVX on NGF protein (at least in the

hippocampus) may reflect the relative stability of NGF protein levels in this relatively

short term absence of ovarian steroids. At a later time point, (e.g. 7 months post

ovariectomy when the severity of the insult may become more prevalent) the levels of

NGF protein may decrease. This stability of NGF protein at the 3 month time period

may reflect the existence of an exhaustible pool of NGF mRNA. The assumption

made is that not all mRNA transcribed is translated to the mature protein. Thus, while

3 months of ovariectomy may result in a significant decline in mRNA, translation of

the remaining mRNA in the pool can still account for the naturally occurring NGF

protein levels in a given tissue region. Beyond a certain threshold of NGF mRNA

decline, however, protein levels may also fall resulting in neurochemical changes

indicative of hypofunction or degeneration in neurons responsive to NGF. Along the

same lines, the decline in mRNA but not protein levels may be the beginnings of a

series of events that leads to nerve cell death. Unfortunately, without data for NGF

protein levels at the 7 month period, the proposed explanation is purely speculative.








The discrepancy between protein and mRNA levels may also be attributed to an

impairment in the release of NGF from the synthesizing cells of the frontal cortex and

hippocampus. It has been documented that fimbrial transaction results in a transient

accumulation of NGF protein in the hippocampus, suggesting that NGF is constantly

being reduced at the site of production through retrograde transport (Gasser et al.,

1986; Korsching et al., 1986). Thus, in spite of mRNA levels being reduced, the

amount of translated protein present in the hippocampus may be higher than expected

due to an impairment in the transport away from this region. Estradiol has been

documented to increase the rate of axonal transport in motoneurons of the spinal cord

(Frolkis et al., 1985). It follows then that in the presence of estradiol, NGF in the

target regions of the basal forebrain cholinergic system, is continually being reduced

at its normal rate. However, in an estradiol deprived state it is conceivable the

opposite of the above is true, that is, the rate of retrograde axonal transport of NGF

from the frontal cortex and hippocampus is reduced thereby leading to an

accumulation of NGF in those regions. As a result, the levels of NGF would be

artificially high and would not necessarily correlate with the level of reduced NGF

mRNA.

The fact that E2 replacement produced only modest increases in NGF mRNA

could be explained by the fact that the replacement regiment utilized provides chronic

levels of E2. Consequently, prolonged exposure to moderate doses of E2 may result

in down regulation of the estrogen receptor or compromise the estrogen receptor-

effector system.

Taken together, these data document the importance of ovarian steroids in

regulating NGF mRNA levels. The decrease in NGF levels in the frontal cortex may

provide a mechanism by which neocortical cholinergic activity is compromised in the

ovariectomized state.






83

NGF has been proposed to be a potential therapeutic agent in certain

neurodegenerative disorders such as AD (Phelps et al., 1989). Due to the bulky nature

of the peptide, invasive techniques must be used in order to deliver NGF into the

brain. With estradiol, on the other hand, peripheral administration could potentially

increase the levels of endogenous NGF. The cytoprotective and

neurotrophomodulatory effects of estradiol may therefore be of significant therapeutic

or prophylactic benefit in the aging population.















CHAPTER 7
THE EFFECT OF OVARIECTOMY AND ESTRADIOL-REPLACEMENT

ON BDNF mRNA LEVELS :

AN IN SITU HYBRIDIZATION STUDY




The field of neurotrophin research has led to the discovery and characterization of

several neurotrophic factors belonging to the NGF superfamily. Of notable examples

are BDNF and NT-3. While NGF has been ascribed the role of maintaining neurons

and their connection, NT-3, due to its temporal expression pattern in the developing

brain, is believed to be more closely associated with proliferation, migration and

differentiation of immature neurons (Maisonpierre et al., 1990). BDNF has a

neurotrophic profile that resembles NGF more closely. BDNF, like NGF, has been

demonstrated to protect basal forebrain (Widmer et al., 1993) and medial septal

(Morse et al., 1993) cholinergic neurons following axotomy, although the efficacy of

neuronal rescue for BDNF seemed to be less than that for NGF (Widmer et al., 1993;

Morse et al., 1993; Knusel et al., 1992)). Furthermore, while the topographical

distribution of NGF is found to overlap considerably both with cholinergic

innervation (Korsching et al., 1985) and indices of cholinergic function, alterations in

BDNF due to fimbrial lesions are similarly associated with decreases in

acetylcholinesterase (AchE) staining and [3H]-vesamicol binding (Lapchak et al.,

1993). Unlike NGF, however, BDNF seems to have an effect on certain non-

cholinergic neurons as well. For example, embryonic mesencephalic dopaminergic








neurons were shown to be rescued by BDNF following the application of the

neurotoxicant, 2, 4, 5-trihydroxyphenylalanine (TOPA; Skaper et al., 1993). BDNF

has also been demonstrated to act as a neurotrophic factor for dopaminergic neurons

of the substantial nigra (Hyman et al., 1991).

The ability of NGF to rescue fimbrially transected neurons with relatively higher

efficacy (Widmer et al., 1993; Morse et al., 1993) suggests that NGF may play a

major role in neurotrophic responses to injury. BDNF, on the other hand, due to its

temporal expression during development and in adulthood has been proposed as a

maturation or maintenance factor for distinct neuronal populations (Maisonpierre et

al., 1990; Nonomura and Hatanaka, 1992).

The colocalization of the estrogen receptor with low affinity p75 NGF receptors

and high affinity trkB receptors in the basal forebrain (Miranda et al., 1993; Toran-

Allerand, 1992) has also set a precedence for the possible interaction between

estrogens and neurotrophins.

In lieu of the similarities in neurotrophic properties with NGF, the possible

involvement of neurotrophic factors in the mediation of the cytoprotective effect

observed (Chapter 5) was extended to involve BDNF. In order to investigate the role

of estrogenic milieu on BDNF production, mRNA for this neurotrophin was

measured using in situ hybridization in 4 major forebrain target regions, the frontal

cortex, the parietal cortex, the temporal cortex and the hippocampus.




Methods


Animals were maintained and treated as described in the General Methods

section. BDNF mRNA levels were measured in the frontal cortex, parietal cortex,

temporal cortex and hippocampus of 9 and 28 week ovariectomized and E2 replaced








animals. E2 replaced animals were ovariectomized for 3 weeks followed by 6 weeks

and 25 weeks of E2 treatment for the 9 and 28 week groups, respectively. Intact

controls were run in parallel with the OVX and E2 pellet animals at each of the two

time points. Due to a technical error where certain tissues were exposed accidentally

to a lower specific activity of hybridization probe, analysis of BDNF mRNA could

only be performed on 3 brains representing each of the three treatment groups. For the

long term group, however, analysis of BDNF mRNA was averaged across 5 animals

per group.


Preparation of tissue

Animals were perfused transcardially with 100 ml of PBS followed by 400 ml of

phosphate buffered 4% paraformaldehyde solution (made with 0.1M sodium

phosphate buffer). The perfusion process was performed with the aid of a peristaltic

pump that was set at a flow rate of 65 ml/min (the approximate cardiac output of an

adult rat). The perfused animal was kept at 40C overnight after which the brains were

removed and placed into a glass vial containing 10 ml of 4% paraformaldehyde

solution. The next day, the brains were transferred to a vial containing a 20% (w/v)

sucrose in 4% paraformaldehyde solution and maintained in this solution for 2 days.

After this time, the brains were removed, blocked (the olfactory bulbs, cerebellum

and brainstem were removed), and frozen on dry ice. Once frozen, the brains were

placed in foil and stored at -800C until the time of use.

At least 24 hours prior to use of the brains for in situ hybridization, the brains

were sectioned at a thickness of 25 pm in a cryostat and placed in a cell culture

cluster dish well containing approximately 3 ml of 4% paraformaldehyde solution.

These free floating slices were considered suitable for in situ hybridization for 4

weeks. Beyond this point, the signal to noise ratio may become compromised.

Sections were taken at two major sites. The first region was referred to as frontal








cortical slices where the preponderance of sections were taken at the level of the

septal nuclei up to the point where the anterior commissure was seen spanning the

two hemispheres. The second region was referred to as hippocampal/parietal cortical
regions. The anatomical landmarks that served as rostro-caudal boundaries for
sectioning were the rostral most region of the hippocampus (CA 3) up to the point
where CA 3 was found to extend to the ventrolateral portion of the section. The
cortical mantle was separated into parietal and temporal cortices in sections where the
rhinal fissure was visible. From each brain, approximately 25 total tissue sections

were used for hybridization.


In situ Hybridization
The hybridization technique was adapted according to the method of Gall and

Isackson (1989). Tissue sections were used in a free floating hybridization protocol
for BDNF mRNA quantitation. Sections were placed in a steel basket which in turn
was placed in a cluster plate well containing the appropriate solution treatment. Brain

sections were first subjected to a 0.1 M phosphate buffer (PB) wash followed by
glycine buffer (0.75 g glycine/100 ml 0.1 M PB) treatment. Following another PB

wash, the slices were exposed to Proteinase K treatment at 370C (lIg Proteinase
K/ml proteinase K buffer, Proteinase K buffer contained 0.1M Tris HC1, pH = 8 and

50 mM EDTA). Proteinase K permeabilizes tissue to allow access of the probe,

thereby increasing the sensitivity of the hybridization. The sections were then
transferred to 0.25% acetic anhydride in 0.1 M Triethanolamine (pH = 8). This step
helps to reduce non-specific binding by acetylating basic groups in tissue thereby
reducing the electrostatic binding of the probe (Hayashi et al., 1978). Furthermore,

this treatment has also been suggested to inhibit any further proteinase K activity
(Angerer et al., 1987). After this step, the sections were exposed to two 2 X SSC

(0.3M NaCl/0.03M Na-citrate) washes. Sections were then placed into the