LLLLLLLLLLLLLLLLLOLLLLL LLL LLLLLLLLLLLLLLL Z?

3129

 

LIBRARY
Mlchlgan State
Unlverslty

 

 

 

This is to certify that the

dissertation entitled

Food intake and serum insulin responses
to intraventricular infusions of insulin and IGF-1
in sheep

presented by

L. A. Foster III

has been accepted towards fulfillment
of the requirements for

PhD degree in Agriculture

 

 

fJ/tiw/Ay

Major professU

Date 7/5/75

MSU is an Affinmm'w Action/Equal Opportunity Institution 0- 12771

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE

L

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

fif—

MSU Is An Affirmative Action/Equal Opportunity Inditution
email”:

 

 

 

 

 

FOOD INTAKE AND SERUM INSULIN RESPONSES
TO INTRAVENTRICULAR INFUSIONS OF INSULIN AND IGF-I

IN SHEEP

BY
L.A. Foster III

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Animal Science

1990

@55— 57/705

ABSTRACT

FOOD INTAKE AND SERUM INSULIN RESPONSES TO INTRAVENTRICULAR

INFUSIONS OF INSULIN AND IGF-I IN SHEEP

BY

L.A. Foster

Previous studies reported that intracerebroventricular
(ICV) infusion of insulin decreased food intake in rats and
baboons. Insulin can bind to insulin like growth factor I
(IGF-I) receptors and mimic the response of IGF-I. The
objective was to determine the effects of ICV infused
insulin or IGF-I on food intake in sheep. In the present
study a 6 day ICV infusion of insulin (123ng/kg of body
weight/day) but not of IGF-I (123 ng/kg of body weight/day)
decreased food intake by 40% (p<.003) and body weight
(p<.015) compared with control sheep. In addition, sheep
that received ICV insulin or IGF-I had only half the
concentration of insulin in serum as compared with controls.
These results support the hypothesis that ICV insulin does
not decrease food intake through IGF-I receptors.
Nevertheless apparently both insulin and IGF-I in the brain

can influence the concentration of insulin in blood.

This is dedicated to my grandma
Mary Kiley Foster

1894-1990

My grandma was 92 (still living by herself and still an
excellent cook) when I decided to go back to school after 3
years of practicing veterinary medicine. She thought it was
foolish to leave a good paying job, and admixed with
presenting me an unceasing array of delicious food and
coffee, was in not so uncertain terms telling me exactly
that. Such is the psychology of grandmas. Trying to figure
out why I was getting another doctorate degree, yet still
would not be a "real doctor", she asked what I was going
back to school to learn. Wanting to keep it simple, I said
that I wanted to find out why sheep or other animals eat.
With a look of incredulity, she paused a moment holding her
breath, either at a loss for words or so exasperated she
could not say them, before stating what to her was only the
obvious,

"That don't make any sense,

Everyone knows a sheep's gotta eat to live!"

One day I hope to develope a hypothesis I can state

with the conviction my grandma said hers.

iii

ACKNOWLEDGMENTS

I gratefully acknowledge Dr. Della-Ferra for describing
the surgical technique and equipment used for the LV
cannulation in sheep. Pam Buntz for surgical assistance and
George Good, Michigan State's shepherd.

I want to express my gratitude to Dr. Kent Ames for all
his help and advice on surgeries and for his willingness to
always try just ”one more thing".

I would like to thank Jim Liesman for statistical,
editorial and technical help. Our discussions not only
helped solve my research problems, but also eased East
Lansing student-community relation tensions, trade
imbalances with Japan, world hunger and put a liberal
democratic president in office.

I would like to thank my committee: Dr. Roy Fogwell:
Dr. Tom Herdt: and Dr. Dale Romsos, for their helpful
suggestions and guidance. Their input was always
beneficial.

Finally I would like to thank my major advisor Dr.
Emery: it is because of his advice I have decided to make a
career in the ingestive behavior area: it is because of his
example that I realized the importance of keeping current

with the literature.

iv

TABLE OF CONTENTS

List Of Tables vii
List Of Figures viii
Introduction................... ............. 1.
The Set Point Hypothesis......... ..... 2.
The Insulin Hypothesis. ........ . ...... 3.
Mechanism for Insulin's Effects ....... 6.
Neurotransmitters....... ...... ..... 6.
5-HTOOOOOOOOOOOOOOOOOOOOOOOOOOOO 7O
Norepinephrine.................. 8.
DopamineOOOOOOOOOOOOOO00....OOOO 9.
Location of Insulin's Action
on Food Intake.................. 9.

Insulin's Effects on Meal Patterns. 13.
Insulin's Effects at the

Cellular Level.................. 14.

Insulin Hypothesis Via IGF-I Receptor? 16.

Insulin Hypothesis In Sheep.. ..... .... 17.
Objective... ..... .................... ....... 17.
Materials and Methods ..... ...... ............ 18.

Animals and Housing... ................ 18.

Infusion Solutions.............. ...... 18.

Experimental Design................... 19.

Surgery .......... .... ................ 20.

Blood and CSF Sampling Procedure...... 21.
Assays................................ 21.
Analysis of Data...................... 21.
Criteria for Animals Included

in Analysis......................... 22.

Results......................... ............ 23.
Intake and Body Weight..... ........... 23.
Serum Values..................... ..... 23.
CSF Values ......... . ....... .. ..... .... 23.
Cannula Placement.. ................... 23.

DiscuSSionOOOOOOOOOOOOOOOOOOOOOOOOOIOO0.0...
Alternative Explanations... ....... ....
Insulin and a Fat Diet. 0 O O O O O O O O O O O O O O
serumvalues...OOOOOOOOOOOOOOOOOOOOOO.
CSF concentrations. 0 O O O O O O O O C O O O O O O O O O

swam and ConCIuSionSO O O O O O O O O O O O O O O O O O O 0

Future Directions........... 0 00.... O .0... O 0
Classic Approach.....................
Molecular Biology Approach...........
Determining the "Hardware"... ....... .

Appendix A Manufacture of Equipment.......

Appendix B Surgical Procedure.............

Appendix C Preliminary Methods............

Bibliography.... ...... ........... ...........

vi

29.
31.
32.
33.
35.

36.
36.
38.
38.
40.
43.
47.
50.

63.

LIST OF TABLES
TABLE 1. Artificial Cerebrospinal Fluid

composition 18.

TABLE 2. Serum and Cerebrospinal Fluid (CSF)
concentrations of insulin (INS)

or IGF-I 24.

vii

LIST OF FIGURES

FIGURE 1 Mean food intakes by day within

treatments for sheep receiving

intracerebroventricular (ICV) infusions of:

insulin or artificial CSF (CONTROL) for

trials 1-4 (upper panel); or insulin, Control

or IGF-I for trials 3 and 4 (lower panel). 26

FIGURE 2 Mean food intakes (left panels) and

body weight differences (right panels) for

sheep receiving intracerebroventricular

infusions of insulin or A-CSF for trials 1

through 4 (top panels) or infusions of insulin,

A-CSF or IGF-I for trials 3 and 4 (bottom panels). 28

FIGURE 3 Infusion Apparatus 44

viii

INTRODUCTION

An understanding of the basic mechanisms that control
long term ingestive behavior would enable huge benefits for
human and animal health. One of the difficulties
encountered in attempting to find a resolution to this
problem is that almost everything can affect food intake,
from the animal's surroundings (temperature), food
characteristics (caloric density) to changes within the
animal (gastric distension). However, very little seems
necessary for the regulation of food intake to occur. Food
intake will stop during a meal even if the temperature is
cold and gastric distension does not occur. Perhaps this is
not too surprising if one thinks about the different
circumstances where eating is terminated. For example the
control mechanisms that are utilized to cause a person to
finally stop eating at Thanksgiving dinner would probably
not suffice for the continued survival of an antelope
grazing in Africa as a lion initiated pursuit. Both of
these mechanisms would be for short term control of food
intake, the termination of a meal. The ability to control
food intake over a period of days and thereby regulate body
weight, would have a much larger impact on improving the

health of humans and animals.

The Set Point Hypothesis

There is considerable evidence that animals and people
regulate their caloric intake and body weight, but the
mechanisms are controversial. Adult animals strive to
maintain a particular body weight under a variety of imposed
conditions. After a period of forced feeding (Cohn and
Joseph, 1962: Drewry et al., 1989) or food restriction (Sims
et al., 1968) animals adjust voluntary intake to return body
weight to what it was before the imposed diet (for review
see Hervey, 1988). Observations like these have led to the
"set point" hypothesis (Kennedy, 1953; Mayer, 1953,) which
states an animal recognizes a certain "weight" as normal.

In order to maintain that "weight", animals regulate: the
amount of calories consumed: the efficiency of storing those
calories; and/or their physical activity (output of
calories).

There are two basic questions that make up the set
point hypothesis. First, what are the "messenger(s)" that
inform the brain as to the body's current "weight" in
relation to its "set point"? Second, what aspect of
"weight" is monitored, (adipose cell size, amount of liver
glycogen or amount of some blood metabolite)? Insulin has
been proposed as one of the messengers informing the brain
on the amount of adipose tissue (Porte and Woods, 1981;

Woods et al., 1985).

The Insulin Hypothesis

One attractive "set point" hypothesis suggests that the
amount of fat an animal has determines long term food intake
(Woods et al., 1985). This hypothesis suggests that insulin
in cerebrospinal fluid (CSF) is the ”messenger" to the brain
on the amount of fat in the individual. Changes in the
concentration of insulin in the CSF will cause opposite
changes in food intake. An increased amount of insulin in
CSF would cause a decrease in food intake and a decrease in
CSF insulin would lead to an increase in food intake.

It has been known for a long time that an acute
increase of peripheral insulin will decrease blood glucose
and increase food intake (Banting et al., 1922). Therefore
it seems rather paradoxical that an increase of insulin in
CSF should decrease food intake. In order to explain this
conundrum, the hypothesis that 'increases in insulin
concentration in CSF decreases food intake' will be
dissected into components and examined separately. 1) Body
weight is regulated. 2) Average serum insulin concentration
varies from day to day in the same direction as changes in
amount of adipose tissue. 3) Insulin in CSF is proportional
to average serum insulin concentration. 4) Insulin has a
physiological role in the CNS and S) insulin in the CSF can
affect the brain so as to decrease food intake.

The first component of this hypothesis is that animals
"monitor" and regulate their body weight. As discussed

under "The Set Point Hypothesis", adult animals seem to

maintain a certain weight. A classic study by Adolph (1947)
demonstrated this principal with a series of experiments.
Rats fed diets with increasing amounts of cellulose, kaolin
or water (this diluted the caloric density of the diet) ate
increasing amounts of the diet to compensate. Changing the
flavor of the diet had little or no affect on amount of food
consumed. Food directly infused into the stomach caused a
compensatory decrease in food ingested for the next three
days (rats gavaged with the same amount of water did not
have decreased food intakes).

The second component is that changes in the daily
average serum insulin concentration directly reflect changes
in the amount of adipose tissue. This is supported by the
findings that fatter animals and humans have higher serum
insulin than thinner animals or humans and that when animals
lose or gain weight their basal insulin concentrations
change in the same direction (McCann et al., 1989; Owen et
al., 1974)

The third component is that the insulin concentration
in CSF changes proportionally and in the same direction as
serum insulin concentrations averaged over several hours.
One aspect of this component is whether peripheral insulin
can penetrate into the CSF. Over the last 4 years this once
hotly debated issue has been somewhat quelled by a series of
experiments done by Porte and Woods' laboratory in Seattle
WA. (the same group that proposed the insulin hypothesis).

Initially this group reported that IV insulin would increase

the concentration of insulin in CSF of anesthetized dogs
(Woods and Porte, 1977). This experiment was criticized
because the dogs were anesthetized and anesthesia can affect
the blood brain barrier (Ono et al., 1983). In the next
experiment, awake human subjects were infused with insulin
IV in amounts approximating postprandial concentrations and
CSF sampleswere withdrawn from a lumbar spinal tap. The
insulin infusion increased insulin concentrations in CSF
(Wallum et al., 1987). In the next experiment, permanent
cannulas were implanted into the cisterna magna and jugular
vein in rats. While rats were unrestrained and moving,
insulin or glucose (to stimulate endogenous release of
insulin) was infused IV and samples of CSF withdrawn.
Insulin concentration was increased in CSF of rats that
received IV infusions of insulin or glucose (Steffens et
al., 1988). In their most recent study (Schwartz et al.,
1990) they report on the time course for an increase in
serum insulin to be reflected in CSF. Significant increases
of insulin were not measured in CSF until 90 min after a
steady continuous increase in serum insulin had been
maintained.

A fourth component is whether insulin normally has a
physiological affect on the CNS. Insulin and insulin
receptors have been measured throughout the brain and the
densest population of receptors were found in areas involved
with food intake (Baskin et al., 1983; Havrankova and Roth,

1978: Hill et al., 1986). This would suggest that insulin

normally has some function within the brain. Insulin has
been shown to have a postreceptor effect of increasing
inositol incorporation within the hippocampus (Figlewicz et
al., 1990).

A fifth component is whether an increase of insulin
within the CSF decreases food intake and a decrease of
insulin within CSF increases food intake. Infusions of
insulin into the CSF of rats and baboons has decreased food
intake (Ikeda et al., 1986: Woods et al., 1979). Infusions
of insulin antibodies into the ventromedial hypothalamus
increased food intake (Strubbe and Mein, 1977).

These individual components of the insulin hypothesis
have been addressed and experiments done that support them.
However, combinations have not been tested, such as: does
the concentration of insulin in CSF change with changes in
body weight and will infusions of that amount of insulin
into CSF affect food intake; do the infusions of insulin
into CSF act specifically on insulin receptors to cause the
decrease in food intake: will blocking the insulin receptors

block the effects of infusions of insulin on food intake?

Mechanism for Insulin's Effects

Neurotransmitters

One mechanism by which insulin could affect food intake
is by altering the effects of neurotransmitters. One of the
functions of a neurotransmitter is to carry or transfer

information from one neuron to another. The 'classical'

neurotransmitters are acetylcholine, 5-HT (serotonin),
dopamine (DA), norepinephrine (NE) and epinephrine (Epi).
All except for acetylcholine have an extensive array of
studies that suggest their involvement with food intake
regulation.

S-HT

An infusion of 5-HT into the PVN decreased food intake
by decreasing meal size. There were no affects on meal
frequency (Shor-Posner et al., 1986). The S-HT injections
caused the largest decrease in fat and carbohydrate
ingestion as compared with protein consumption (Shor-Posner
et al., 1986). Experiments of this type have contributed to
the framework of a hypothesis on how 5-HT is involved in
food intake regulation (for review see Wurtman and Wurtman,
1986). Briefly, S-HT is synthesized from tryptophan.
Tryptophan shares a common transport system with other large
neutral amino acids to cross the blood brain barrier. A
high-carbohydrate meal stimulates insulin secretion; this
decreases the amount of large neutral amino acids in blood.
More tryptophan crosses the blood brain barrier.
Availability of tryptophan is the rate limiting step for 5-
HT synthesis in the brain. Therefore S-HT concentrations
within the brain increase and carbohydrate intake is
suppressed. A high-protein meal works the opposite, by
increasing the amount of large neutral amino acids in
plasma, less tryptophan crosses into the brain, less S-HT is

synthesized and food intake, in particular carbohydrate

intake, is increased.

The S-HT hypothesis complements the insulin hypothesis.
Both hypotheses have an increased peripheral insulin
component, and in both cases the peripheral increase of
insulin leads to a decrease in food intake. A major
difference, is the increase in 5-HT is suggested to limit
the carbohydrate intake of the next meal (and not effect
protein consumption) (Shor-Posner et al., 1986). Whereas
central insulin probably takes much longer to have an effect
on food intake (see Discussion sect). Still there does not
seem to be any proof to suggest that one hypothesis is
exclusive of the other.

Norepinephrine

It has been suggested that NE increases carbohydrate
consumption by acting on a2 adrenergic receptors to inhibit
satiety neurons within the PVN (for review see Leibowitz,
1988). This hypothesis is based on these findings. 1)
Infusions of NE into the PVN stimulates consumption of CHO
(Leibowitz et al., 1988). 2) There is an increased release
of NE from the PVN with food deprivation (Stanley et al.,
1989). 3) There is a decreased release of NE from the PVN
after a meal (Stanley et al., 1989). 4) Lesions of the PVN
cause hyperphagia (Sakaquchi et al., 1988). 5) Infusion of
a specific adrenergic toxin into the PVN decreased CHO
intake (Shor-Posner et al., 1986).

. An infusion of insulin into the PVN had no effect on NB

release within the PVN (Minano et al., 1989). An injection

of insulin into the VMH decreased sympathetic activity and
thermogenesis of brown adipose tissue (Amir et al., 1989).
A decrease in lipolysis and thermogenesis could fit in with
a hypothesis for decreasing food intake: however, these
effects of insulin on sympathetic activity are more likely
involved with peripheral homeostasis rather than a central
control of food intake.

Dopamine

The primary focus on DA has been with Parkinson's and
schizophrenia. Much of what is known about schizophrenia
has been learned from an amphetamine induced psychosis that
has been used as a model for schizophrenia (for review see
Chiodo, 1988). Amphetamines cause a large release of DA. A
side effect of amphetamines is decreased food intake (Carr
and White, 1986). Injections of DA agonists decreased food
intake (Rusk and Cooper, 1989). During feeding, endogenous
concentrations of DA increased in the same locations in the
brain where injections of DA decreased food intake
(Hernandez and Hoebel, 1988). DA concentration increased
within the medial hypothalamus (includes PVN, DMN, and VMH)
during insulin perfusion. Therefore insulin may be
affecting food intake via dopaminergic neurons.

Location of Insulin's Action on Food Intake

The primary emphasis of research on the neuroanatomical
location of food intake regulation has focused on the
hypothalamus. More than 20 years ago it was demonstrated

that lesions of the VMH caused hyperphagia and lesions of

10

the LH made animals anorexic (Hoebel and Teitelbaum, 1962).
Other lesions in the hypothalamus such as the PVN
(Kirchgessner and Sclafani, 1988), increase intake: the
dorsomedial hypothalamic nuclei (DMN), decrease intake
(Bernardis et al., 1987). Lesions in higher structures of
the brain, such as the amygdala (Fonberg, 1989) caused
active food rejection in dogs.

Of perhaps greater relevance than inducing a lesion and
examining whether a behavior is changed or lost, is inducing
a lesion and examining which behaviors stay the same. For
example, if a lesion affected motor activity so that an
animal could not walk, the intake of that animal would drop
to zero. Even though, the animal would eat normally if food
was within reach.

Chronic decerebrate rats have an intact spinal cord and
brainstem (no hypothalamic, thalamic or cortical functions).
Chronic decerebrate rats and intact rats ingest or refuse
the same solutions at the same concentrations (Grill and
Norgren, 1978a). Chronic decerebrate rats and intact rats
will stop drinking a solution after they have consumed the
same amount and are sated (Grill and Norgren, 1978b). This
demonstrates that integration between sensory information
from visceral structures (information on satiation), taste
(type of solution in oral cavity) and motor output
(rejection of a substance) must occur someplace caudal to
the hypothalamus.

There has been some type of interaction reported for

11

most structures within the brain and insulin. Injection of
a physiological concentration of insulin into the jugular
vein decreased the activity of taste neurons within the
brainstem (Giza and Scott, 1987). These authors suggested
that the decreased sensitivity of taste neurons to glucose,
after an IV insulin injection, could be a mechanism by which
insulin reduces feeding (Giza and Scott, 1987). An
alternative interpretation, is that the changes in neuronal
activity to insulin, contribute to the increased food intake
that occurs after a peripheral injection of insulin that
induces hypoglycemia. The decreased activity within the
taste neuron to glucose, might cause the animal to prefer a
higher concentrated glucose solution. If a certain activity
within the neuron was related to a preferred amount of
sweetness: then after the insulin injection, a higher
glucose concentration would be necessary to obtain that same
activity. In support of this, it has been reported that in
sodium-deprived rats (sodium deprived rats display an avid
sodium appetite (Fregley et al., 1965)), the activity of
taste neurons to sodium decreases (Jacobs et al., 1988). In
either case, the neuronal response was to exogenous insulin.
The insulin concentration within blood will vary
considerably in response to meals. Mechanisms that are
directly responsive to plasma insulin concentration would
probably be involved with short term food intake regulation.
It has not been reported if similar changes in neuronal

activity occur from an increased concentration of insulin in

12

CSF. If this occurred, then altered taste might be a
mechanism where insulin affects long term food intake.

Infusions of insulin into the LH increased the activity
of glucose sensitive neurons (Oomura and Rita, 1981). Most
other factors, such as glucose, free fatty acids, amino
acids or serotonin that might indicate that a meal had or
was occurring, decreased the activity of glucose sensitive
neurons (reviewed by Oomura, 1988). In this instance the
effect of insulin is opposite to the expected result.
Specific infusions of insulin into the LH did not have any
affect on food intake (McGowan et al., 1990). This would
suggest that insulin is not affecting food intake via the
LH.

Infusions of glucose into the VMH increase the activity
of glucose responsive neurons. Infusions of insulin tend to
inhibit activity, but infusions of glucose and insulin cause
a greater stimulation than glucose infusions alone (Oomura
and Rita, 1981). The same factors that decreased the
activity of glucose sensitive neurons in the LH increase the
activity of glucose responsive neurons in the VMH (reviewed
by Oomura, 1988). Infusions of insulin specifically into
the VMH decrease food intake (McGowan et al., 1990).

Lesions of the VMH cause obesity and hyper-insulinemia (King
et al., 1988). Stimulation of the VMH decreases food intake
(Hoebel and Teitelbaum, 1962). The evidence strongly

supports the inclusion of the VMH within the neuroanatomical

pathway by which insulin affects food intake.

13

Insulin's Effects on Meal Patterns

An experiment has not been reported that was designed
to specifically examine what alterations occur in meal
patterns to account for the decreased intake in response to
insulin infusions. When reported, authors describe a
general decrease in intake with no exclusive effect limited
to meal size, duration or intermeal interval. In an
experiment with baboons ICV insulin reduced feeding
comparably at every interval. The authors did not state
what the length of the interval was, only that they were
examining meal patterns (Woods et al., 1980). Insulin
infused ICV in rats decreased both day and night food
intake, but the authors stated that since rats normally eat
at night, there was only a significant decrease of food
intake during the night (Brief and Davis, 1984). Injections
of insulin antibodies into the VMH only increased intake
during the night time (Strubbe and Mein, 1977). Which again
is not surprising since rats do not normally eat much during
the day. Infusions of small amounts of insulin directly
into the medial hypothalamus, that did not affect total food
intake, increased daytime food intake (McGowan et al.,
1990). The authors suggested that this could have been a
compensatory increase in food intake due to insulin's
effects of decreasing food intake at night. Although the
evidence is far from conclusive, it seems that insulin
decreases meal size, frequency and duration during times

when an animal normally eats.

14

Insulin's Effects at the Cellular Level

Insulin has been reported to have no affect on slices
from the hypothalamus or cortical gray matter for glucose
uptake, oxidation or lipogenesis in brain (Goodner and
Berrie, 1977). In cell cultures derived from the whole
brains of 1 day old rats, insulin inhibited NE uptake more
than maprotiline, an antidepressant and specific NE uptake
inhibitor (Boyd et al., 1986). In cells taken from the
hippocampus, insulin stimulated inositol incorporation into
inositol phosphate and inositol lipids (Figlewicz et al.,
1990).

As stated in a review by Recio-Pinto and Ishii (1988)
there are many complications in trying to discern the
specific functions of insulin on brain neurons. Since there
are more than 10 times as many glial cells as neurons in the
brain, many preparations will be contaminated with
endothelial cells, fibroblasts or glial cells which may
contain their own insulin receptors. But it should not be
ruled out that insulin may be having major affects on
neurons via glial cells. Hexokinase, a rate limiting enzyme
in glucose metabolism, is absent in some neurons: yet
glucose is almost the sole source of energy in the brain
(Katoh-Semba et al., 1988). It has been suggested that some
neurons depend on glucose metabolites such as pyruvate
supplied by glial cells (Katoh-Semba et al., 1988). The
other major point Recio-Pinto and Ishii (1988) made is that

insulin and IGF's are survival factors and therefore results

15

from dose response curves need to be interpreted with
caution, since cells in the control and low dose cultures
may just be dying. Another problem with relating these in
vitro experiments to ingestive behavior is that insulin is
involved with other functions and behaviors in the brain.
There has been a lot of work determining insulin's
involvement with neuron growth or differentiation (for
review see Recio-Pinto and Ishii, 1988). Intraperitoneal
insulin injections changed the electrical activity of 27% of
neurons in the olfactory bulb and 21% of the neurons in the
amygdala to odors (Cain, D.P., 1975). Sleep deficits in
streptozotocin treated rats were normalized by ICV insulin
(Danguir, 1984). Therefore many intracellular changes in
some brain areas, caused by insulin, are probably not
involved with food intake.

Another important aspect involved in interpreting the
results of acute experiments, is that insulin's affects on
food intake are chronic. This is a point that has generally
been overlooked when the results from acute experiments are
used to explain how insulin affects food intake (Myers et
al., 1986: Oomura and Rita, 1981). Electrophysiological and
microdialysis measurements need to be made over several days
of insulin infusion in order to make more confident
suggestions on changes caused by insulin that affect food

intake.

16

Insulin Hypothesis Via IGF-I Receptor ?

There is a possibility that all of the pharmacological
manipulations of insulin that have affected food intake have
caused a response via the IGF-I receptor. High
concentrations of insulin can bind to the IGF-I receptor and
elicit a similar response to IGF-I (Rechler and Nissley,
1986). There are IGF-I receptors throughout the brain and
there are more IGF-I receptors in the choroid plexuses of
all ventricles than insulin receptors (Davidson et al.,
1990: Shemer et al., 1987). The choroid plexuses are
located in the ventricles and are responsible for the
synthesis of most of the CSF (Wright, 1982). Peptide
infusions into the ventricles would first contact the
choroid plexus.

Insulin and IGF-I are structurally similar
(Rinderknecht and Humbel, 1978). Receptors for IGF—I and
insulin: both contain a and 8 subunits (McElduff et al.,
1988); both receptors can bind either peptide; both
receptors can be down regulated by either peptide; the 8-
subunit of both receptors have intrinsic tyrosine-specific
kinase activity that is activated by autophosphorylation
(Kasuga et al., 1982: Rechler and Nissley, 1986; Sasaki, et
al., 1985). Central insulin and IGF-I receptors appear to
maintain these homologies, except insulin receptors do not
appear to be down regulated by increased amounts of central
insulin (Havrankova et al., 1981; Manin et al., 1988) and

both types of receptors are smaller in the brain than

17

peripherally (McElduff et al., 1988). It is unknown if IGF—
I receptors in the brain are down regulated in the presence
of increased IGF-I concentrations. In an earlier study a
large bolus of IGF-I did not affect food intake in rats
(Lauterio et al., 1987). However, insulin does not affect
food intake when bolused ICV either (Lauterio et al., 1987:
Tsujii and Bray, 1990). The question "are the affects of
chronic insulin infusions on food intake via IGF-I
receptors?" needs further study.

Insulin Hypothesis In Sheep

In sheep the liver has a net output of glucose in both
fed and fasted states (Brockman and Laarveld, 1986).
Monogastric species, when fed a typical high carbohydrate
diet, only have a net output of glucose in the fasted state.
Short chain fatty acids are the primary energy source
absorbed from the gut in sheep (Bergman et al., 1966). The
concentration of serum insulin is 50% lower than in
monogastrics and serum insulin varies less in response to
meals, compared to monogastrics (Hove and Blom, 1973).
These features make the sheep a unique model to test the
hypothesis that increasing concentrations of insulin in CSF

lead to decreased food intake.

OBJECTIVE
The objective was to determine the effects of ICV

infused insulin or IGF-I on food intake in sheep.

18

MATERIALS AND METHODS

Animals and Housing

Adult castrated male Dorset crossbred sheep (40-60 kg)
were kept in individual metabolism cages. The building was
continuously lighted and maintained at a temperature of
1818°C. Animals were fed a pelleted diet (89% DM, 66.4%
total digestible nutrients (TDN), 12% crude protein, 18%
crude fiber, 2% fat; Countrymark: Columbus, OH) and water ad
libitum. Animals were given new food once daily at which
time refusals were weighed.

Infusion Solutions

Artificial CSF (A-CSF) was made according to methods
reported earlier (Cserr, 1965: Pappenheimer et al., 1962,)
except 100 mg/liter bovine serum albumin was added and the
solutions were sterilized by filtering rather than

autoclaving.
TABLE 1

E lLE' . J ; l . J E] .3 ; m .!.

‘ Compounds Amount of
Went mg. _a_dUe 9.9M on mil

Na 150.0 NaCl 7310.
K 3.0 KCl 224.
Ca 2.3 NaHCO 2100.
Mg 1.6 NazHPg4 23.
Cl 135.0

HCO3 25.0 CaC12 128.
P04 0.5 MgC12°6H20 163.

Albumin 100.

19

Porcine insulin was generously donated by Eli Lilly
(Indianapolis, IN.) and IGF-I was generously donated by
Monsanto (St. Louis, MO.). Each powdered peptide was
reconstituted with 0.01 N HCl to make stock solutions with
concentrations of 1 mg/ml. Each stock solution was diluted
with A-CSF and infused at a dose of 5.6 ug, in 280 ul per
day, which amounts to 123 ng/kg of body weight/day for both
peptides or 3.4 mU/kg/day for insulin. This insulin dose is
greater than the dose used in baboons (0.1 mU/kg/day) (Woods
et al., 1979) and less than the dose used in rats (6
mU/kg/day) (Ikeda et al., 1986). The same mg dose of IGF-I
was chosen so there would not be a difference in amount of
protein added. Since molecular weights are similar (”6000
for insulin and .7000 for IGF-I) molar infusions were also
similar.

Experimental Design

Sheep were adapted to diet, housing and handling for at
least 10 days prior to beginning the baseline period. Four
trials were completed; the first starting in October and the
last ending in March. Trials 1 and 2 used eight sheep each,
four infused with A-CSF and four infused with insulin in A-
CSF. Trials 3 and 4 were similar to trials 1 and 2 except
nine sheep were used in each trial, three were infused with
A-CSF (Controls), three with insulin in A-CSF (INS) and

three with IGF-I in A-CSF (IGF-I).

20

Surgery

Ventricular cannulas were permanently implanted under
isoflurane inhalation anesthesia. A 1.5 cm long, 16-gauge
guide cannula was cemented to the skull with dental acrylic
(Columbus Dental: St. Louis, M0) 0.5 cm lateral to the
midline and 0.5 cm below the bregma suture line. Stainless
steel screws were placed in the skull to anchor the acrylic.
Polyethylene tubing (PE) was filled with sterile artificial
CSF, and attached to a 20 gauge, short beveled needle. The
tubing was held higher than the sheep's head while the 20
gauge needle was lowered through the guide cannula into the
brain. The level of A-CSF dropped in the tubing when the
needle entered the ventricle. The tubing was then held
below the sheep's head and CSF was siphoned out of the
ventricle. A visibly steady drip of CSF out of the tubing
was required before the needle was secured to the guide
cannula. An osmotic pump (Alzet, Palo Alto, CA.) was
inserted subcutaneously in the dorsal cervical area and
connected to the 20 gauge needle via PE-60 tubing filled
with the appropriate test solution.

The criteria used to verify the patency of the system
included: 1. observing a consistent drip of CSF out of the
PE tubing at time of surgery: 2. infusing Bromophenol blue
dye (Sigma: St. Louis, MO) into the PE tubing just prior to
euthanasia and finding blue stained ventricles at necropsy:
3. a two fold increase of insulin and IGF-I in the CSF of

insulin and IGF-I-infused sheep, respectively, as compared

21

with controls. Of the sheep infused with either insulin or
IGF-I, there was only one sheep that had dye stained
ventricles yet did not have elevated insulin in the CSF
sample. This sheep was not included in the analysis.

Blood and CSF Sampling Procedure

After 6 days of continuous ICV infusion, animals were
weighed and a 5 ml CSF sample was obtained aseptically from
a spinal tap into the cisterna magna of anesthetized sheep.
For trials 3 and 4, blood was collected from the jugular
vein and another CSF sample was taken approximately 4 hr
later. Blood and CSF samples were stored in siliconized,
sterile serum tubes on ice until centrifugation. The
supernatant was divided into 2 ml aliquots and frozen until
chemical determinations.

Assays

Insulin was measured by radioimmunoassay utilizing

commercial kits; ImmoPhase (Corning, Medfield, MA.) for
trials 1 and 2: and Micromedic (ICN, Costa Mesa, CA) for
trials 3 and 4. IGF-I was determined using a heterologous
radioimmunoassay (Della-Ferra et al., 1986).

Analysis of Data

For the analysis of food intake, the baseline period
consisted of 6 days prior to surgery. Food intake for days
3 through 6 of the infusion period were averaged for
analysis after determining no treatment*DAY interaction in a
split plot analysis removing sheep within treatment sum of

squares from the residual. The first model compared INS and

22

A-CSF using data from all four trials with fixed factors of
treatment, trial, and treatment*trial interaction. The
second model compared IGF-I, INS and A-CSF using data from
trials 3 and 4. There was no significant trial effect for
the second model, therefore trial was not included in the
analysis. Average food intake of the baseline period was
used as a covariate for both models.

For the analysis of body weight, the difference in body
weight from the end of the baseline period to the end of the
infusion period for each treatment were analyzed in a model
with fixed factors of treatment, trial, and treatment*trial
interaction.

For the analysis of serum and CSF values of insulin and
IGF-I, the same model as body weight was used. Due to the
large range of values for insulin in CSF, there was
heterogeneous variance. Natural logarithms of the CSF
insulin values had homogeneous variance. Therefore the
statistical analysis was done on the natural log
transformation, but actual values are reported.

Criteria for Animals Included in Analysis

Two sheep died during this experiment; one was
hydrocephalic and the other never recovered from anesthesia.
'Ten sheep were removed from the experiment due to failure of
fulfilling criteria for patency of the system. Three sheep
‘were removed from trial one, two from trial two, three from
'trial three and four from trial four, leaving a total of 10

of 14 controls, 9 of 14 INS and 3 of 6 IGF-I infused sheep.

23

RESULTS

Intake and Body Weight

Insulin infused sheep ate 40% less (p<.003) and lost
more weight (p<.015) than controls (FIGURES l and 2). There
were no differences in intake (p<.4) or body weight (p<.5)
between IGF-I infused sheep and controls (FIGURE 2). There
was no difference in food intake for A-CSF infused sheep
between
baseline period and the infusion period (2.4 $0.1 kg/day:
vs. 2.1 10.1 kg/day; p<.08) (FIGURE 1).

Serum values

Sheep infused ICV with INS or IGF-I had less than half
the serum insulin concentration measured in sheep infused
with A-CSF (p<.01; Table 2). There were no differences in
serum IGF-I caused by treatments.

CSF values .

By design, animals infused with INS or IGF-I had to
have two times the concentration of the infused peptide as
compared with controls. The range of values for insulin
infused sheep varied from 11 to 1100 uU/ml, with an average
of 255 uU/ml.

Cannula Placement

At the time of surgery the cannulas were angled
slightly rostral and medial. This caused most of the
cannula tips to be in the medial ventral portion of the

lateral ventricle (LV) where the LV join together to

form the third ventricle. At necropsy all cannula tips were

24

.Aaoo.vae mmou< no HusoH no“:

census“ moose can» mmo :« mzH no mcofiucuucoucoo momma: cc: mzH nu“: census“ doocm w

..~o.va. smote no mzH no“:

caucus“ mound can» mmo ca HIEOH mo mcoHumuucoocoo song“: on: HIEUH cud: consucu moocm m

.Aac.vav hauls nag: census“ gamma 0» coucasoo

no HumoH no mzH cud: cousuca noozm ea uozou xcm who: macaumuucoocoo cfiasmcw asuom a

mos~c> sou umooxo .xmmu mcoofi ucomounou camp on?

.cmoumcfi co>am ma omcou ecu muons .mmo ca mzH no

.HnmoH no mzH .Ammouav emu anaoauauum

«0 seasons“ nooscaucoo amp c o ocfisouuou csouczuqz one: noamsmm mmu use osmoam

 

 

eaum e mm.~« ma «an em o.m« as
cmaduvs unun a.nu e as“ sea 5.6“ mm
eats a s.~« e «an dos se.ea em
:35 £333
mzH Humua HumoH mzH
mmu asuom

c 024 n mfldeB

 

 

 

HImUH
NmVIWA Inna m mzH
old n m mmUI<
«mmmmlllldda\=a. :
mzH msouu
mmo ucoauoous

N 024 H mudeB

 

HumUH no .mzHV cqasmcd no ncofiuouucoocoo Amway macaw ”scammounouoo cco aspen

N ”and?

25

FIGURE 1. Mean food intakes by day within treatments for
sheep receiving intracerebroventricular (ICV) infusions of:
artificial CSF (CONTROL) or insulin for trials 1-4 (upper
panel); CONTROL, insulin, or IGF-I for trials 3 and 4 (lower

panel).

roan INTAKE (g)

roan INTAKE (g)

26

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SURGERY
aooo, BASELINE PERIOD I l INFUSION PEIRIOD
2500
-‘V‘ CV“--
I “v“ V
2000-
1500 - “
. V"‘V‘~-v
-O -O CONTROLS ‘.
1000- ‘V
-V-‘VINSULIN
500-
0 J— l L L l L l l l l 1 ]
-6-5—4-3-2-1 o 1 2 3 4 5 6
DAY
SURGERY
WOW L, Imusxon pmon]
—-O—O CONTROLS
1°°°' “fir-v INSULIN
--v---v- IGF-I
500-
o 1 L 1 J l l I l L l l l
—6-5-4-3-2-1 o 1 2 3 4 5 6
DAY

FIGURE 1

27

FIGURE 2. Mean food intakes for days 3-6 of infusion (left
jpanels) and body weight differences (right panels) for sheep
recieving intracerebroventricular infusions of A-CSF or
insulin for trials 1 through 4 (top panels) or infusions of
.A-CSF, insulin, or IGF-I for trials 3 and 4 (bottom panels).
**p<.003 insulin infused sheep ate less than controls.
*p<.02 insulin infused sheep ate less than A-CSF and IGF-I
infused sheep. ..p<.015 insulin infused sheep lost more

‘weight than controls. .p<.03 insulin infused sheep lost

:more weight than A-CSF and IGF-I infused sheep.

FOOD INTAKE (g\doy)

FOODINTAKE(g/day)

2500

2000

1500

1000

500

2000

1500

1000

500

 

A-CSF

 

A-CSF

 

INSULIN

 

lNSULlN

28

TRIALS l - 4

2000 -

r1310

1000 -

0

-lOOO -

  

-2000 -

~3000 -

.1.

A-CSF INSULIN

BODY WEIGHT CHANGE (g)

-4000 -

 

 

TRIALS 3 AND 4

 

 

 

2000
1000

E

g 0

z

<

r

o -1000

.—

I

o

:3 —2000

3

E

o -3000

a:
-4000

[OF-l A-CSF INSULIN Icr—I

FIGURE 2

29

the lateral ventricles for the INS and A-CSF infused
sheep. The tips of cannulas for the IGF-I infused sheep were
actually in the third ventricle.

DISCUSSION

When insulin is infused ICV, food intake is often
(Brief and Davis, 1984; Ikeda et al., 1986; Woods et al.,
1979) but not always (Brief and Davis, 1984: Lauterio et
al., 1987; Manin et al., 1988) decreased. A generalization
that may explain these conflicting studies is: ICV insulin
reduces food intake only when chronically infused at a
relatively large and steady dose. That insulin does not
have a short term affect on food intake is supported by the
experiments in which insulin bolused into rat ventricles had
no short term affect on food intake (Lauterio et al., 1987;
Plata-Salaman and Oomura, 1986; Tsujii and Bray, 1990).
Insulin was infused into baboons for three days before a
significant decrease in food intake occurred (Woods et al.,
1979). In this experiment no significant decreases in food
intake were measured until after two days of infusion.

This necessity for a chronic ICV infusion of insulin in
order to negatively affect food intake might be due to the
distance from the infusion site to the active site, or the
changes initiated by insulin might take time. That higher
doses of insulin decrease the time lag between the start of
an ICV infusion and a significant decrease in food intake

(Brief and Davis, 1984), support the idea that insulin must

30

diffuse to the active site. However, ICV infusions of
insulin can affect pancreatic function almost immediately
(Woods and Porte, 1975), which suggests that at least some
active sites are available to insulin in CSF. The idea that
insulin might cause slow changes, also fits in well with the
proposal that insulin is a component of the ”set point"
hypothesis. Animals normally take at least a day to
compensate for sudden changes made in the energy density of
their food or in their energy expenditure (Adolph, 1947;
Edholm et al., 1955).

It also appears that a large dose of insulin, relative
to endogenous CSF insulin concentrations, must be infused
into the CSF to decrease food intake. The endogenous
insulin concentration in CSF of baboons is 3 uU/ml (Woods et
al., 1981), 150 -1500 uU/day of insulin were required to
decrease food intake when infused into the ventricles of
baboons (Woods et al., 1979). In rats an ICV infusion of 5
mU/day of insulin did not alter food intake whereas 7.5 and
10 mU/day caused a dose dependent decrease in food intake
(Brief and Davis, 1984). The normal concentration of
insulin in CSF of rats is 6 uU/ml (Manin et al., 1988). In
this experiment the endogenous concentration of insulin was
increased in CSF 10 to 100 times.

There is no question that the doses infused for both
INS and IGF-I were pharmacological; however since the active

site is unknown, the concentration of insulin at the active

31

site might be in the physiological range. Duffy and
Pardridge (1987) have identified an insulin transporter
between blood and brain parenchyma, suggesting that insulin
goes from blood to brain to CSF. This would mean that the
infusion site was "downstream" of the active site and it
might be expected that much higher doses would be required.

Alternative Explanations

The decreased food intake with insulin in this trial
might have been due to a non-specific protein effect. This
is not likely since the same dose of IGF-I was also infused,
plus the large amount of albumin (.1mg/ml) that was included
in the A-CSF should have masked any nonspecific protein
effect that insulin might have had. In earlier reports
glucagon (Woods et al., 1979) or denatured insulin (Plata-
Salaman et al., 1986) were infused ICV with no effect of
food intake.

Insulin infused ICV might cross the blood brain barrier
and have an effect peripherally. This is not likely since
the amount of insulin infused into the CSF was very minor
compared to peripheral insulin concentrations. If the
amount of insulin infused ICV (2.3uU/kg/min) for 15 min (two
times the half life of insulin in blood) (Grizard et al.,
1988) was given in a single dose into the blood, the insulin
concentration in blood would increase by approximately 0.7
uU/ml or less than 2%. In sheep, intravenous jugular

infusions ranging from 20 to 2000 uU/kg/min for 24 hr had no

32

effect on food intake (Deetz et al., 1980).

High concentrations of insulin bind to IGF-I receptors
and elicit a response similar to IGF-I (Rechler and Nissley,
1986). It is possible that large doses of insulin (relative
to endogenous concentrations of insulin in CSF) need to be
infused into the CSF to decrease food intake because insulin
is acting via IGF-I receptors. These results would suggest
that this was not the case since there was a trend for IGF-I
to increase, not decrease, food intake (FIGURE 2).

It is possible that IGF-I did not cause an effect since
it was infused into the third ventricle; however, insulin
decreases food intake when infused into the third ventricle
(Brief and Davis, 1984; Ikeda et al., 1986)., There is
mixing of CSF between the ventricles (Wright, 1982); insulin.
infused into the right lateral ventricle will spread into
the third ventricle and on to the left lateral ventricle
(Manin et al., 1990). This would suggest that IGF-I infused
into the third ventricle would diffuse into the lateral
ventricles.

Insulin and a Fat Diet

Arase et a1. (1988) reported that ICV insulin decreased
intake in rats fed a high-carbohydrate diet, but not in rats
fed a high-fat diet. Sheep can be said to be on a high-fat
diet, based on what is absorbed from the gut into the blood,
but sheep primarily absorb short chain fatty acids rather

than long chain fatty acids (Bergman et al., 1966). These

33

results suggest that long chain fatty acids -not short chain
fatty acids- prevent the effects of ICV insulin on food
intake in high fat diets.

Rats need to be tested with a high-fat diet of short
chain fatty acids to determine if they respond similar to
sheep. This would have interesting implications into the
mechanism of insulin's actions on food intake. Short chain
fatty acids are absorbed directly across the gut into the
blood stream whereas long chain fatty acids are absorbed
into the lymph; long chain fatty acids require carnitine
acyl transferase to cross the mitochrondrial membrane to be
oxidized whereas short chain fatty acids do not (McGarry and
Foster, 1980).

Serum Values

There was a large and significant decrease in serum
insulin concentration for sheep infused with either INS or
IGF-I as compared with A-CSF (p<.01). Decreased intake
might explain the reduced serum insulin in the INS infused
sheep. Animals with decreased food intake will have
decreased serum insulin (de Boer et al., 1985; Heitmann et
al., 1986). However the trend was for IGF-I infused sheep
to eat more than controls, suggesting that IGF-I either
decreased pancreatic insulin secretion or increased the rate
of insulin turnover. '

A subcutaneous infusion of IGF-I in humans decreased

insulin secretion, based on C-peptide concentration (Guler

34

et al., 1989). Plasma insulin concentrations were not
affected which suggests that insulin turnover was decreased
also. During the subcutaneous IGF-I infusion, serum IGF-I
concentrations were increased four times higher than normal.
In this experiment there were no differences in plasma IGF-
I. It is not likely that the IGF-I infused in sheep was
acting peripherally. It is possible the high concentration
of IGF-I obtained in humans might have crossed the blood
brain barrier and affected insulin secretion. This
experiment supports the proposal that IGF-I lowers serum
insulin via a central mechanism.

Pancreatic insulin secretion is altered by stimulating
the hypothalamus (de Jong et al., 1977). Therefore, neural
connections exist to the pancreas that might be activated by
IGF-I. Electrical stimulation of the hypothalamus can alter
enzyme kinetics in the liver (Shimazu and Ogasawara, 1975).
Therefore, it is possible that IGF-I might affect insulin
turnover via the CNS. An alternative explanation is that
IGF-I induces the release of another peptide or hormone that
is then responsible for the peripheral changes in serum
insulin concentrations. ICV infusion of IGF can decrease
pituitary release of GH by increasing hypothalamic
somatostatin concentrations (Berelowitz et al., 1981:
Tannenbaum et al., 1983).

A significant decrease in serum insulin due to an ICV

infusion of insulin or IGF-I has not been reported

35

previously. Most studies have measured a nonsignificant
decrease in serum insulin in response to an ICV infusion of
insulin (Arase et al., 1988; Manin et al., 1988; Steffens et
al., 1985; Woods et al., 1979); except in an experiment with
anesthetized dogs, an ICV infusion of insulin caused an
increase in serum concentration of insulin (Woods and Porte,
1975). Possible reasons for this discrepancy might be due to
factors such as: the dogs had been fasted prior to insulin
infusion; the dogs were anesthetized; or the experiment only
had a 3 hr duration.

CSF Concentrations

Reasons for the huge variation in insulin concentration
within the CSF, even though the same amount of insulin was
being infused, is unknown, but has been found by other
workers (Buonomo, personal communication). In the last two
trials, two CSF samples were obtained per sheep
approximately 4 hr apart. There were large differences in
insulin concentrations within, as well as among sheep (data
not presented). Within the sheep infused with insulin, the
concentration of insulin in CSF was not correlated with food
intake. The production and outflow of CSF is very
responsive to hydrostatic pressure (Pappenheimer et al.,
1962: Wright, 1982), so perhaps the variation was due to the
way the sheep had been holding his head just prior to

sampling.

36

SUMMARY AND CONCLUSIONS

I think that insulin has a specific effect on food
intake within the central nervous system (CNS). Infusion of
insulin into the ventricle of the brain decreased food
intake. Infusion of insulin antibody into ventromedial
hypothalamus caused an increase in food intake (Strubbe and
Mein, 1977). The dose of insulin I infused was too small to
cross the blood brain barrier and have an effect
peripherally. Insulin binding sites have been identified
within the brain in areas associated with food intake
(Havrankova and Roth, 1978; van Houten and Posner, 1981;
Wilcox et al., 1989). Insulin has been shown to have a
postreceptor effect of increasing inositol incorporation
within the hippocampus (Figlewicz, et al., 1990). In
conclusion, ICV infusion of insulin significantly decreased
food intake and body weight as compared to IGF-I or A-CSF
infused sheep. A surprising, albeit unique, finding was
that IGF-I and INS infused sheep had significantly lower
serum insulin concentrations than A-CSF infused sheep.
Further experiments will be necessary to elucidate the
mechanism(s) by which this occurred.

FUTURE DIRECTIONS

One of the most exciting results found, was that
central administration of IGF-I and insulin decreased plasma
insulin. This experiment needs to be repeated with more

frequent withdrawal of blood samples. The minimal analysis

37

on the blood samples would include insulin, glucose, and C-
peptide concentrations. With this information, the time
course for the decrease in plasma insulin could be charted.
A decreasing amount of C peptide would support the
hypothesis that there was a decreased secretion of insulin
(versus an increased turnover of insulin).

Another important relationship to determine is the
change in serum insulin over time as compared to the change
in food intake. If serum insulin changed before the change
in food intake, this would suggest that the decreased
peripheral insulin was not due to decreased intake.

If C-peptide concentration decreased, then it would be
worthwhile to sever the vagal branches to the pancreas and
repeat the infusion experiment with insulin and IGF-I. If
the decrease in serum insulin was eliminated, this would
suggest that the peptides were decreasing insulin secretion
via a neural mechanism. If C-peptide was not decreased,
than a selective vagotomy to the liver might be indicated to
determine if the decrease in serum insulin was due to a
change in insulin turnover within the liver that was
neurally mediated. If a hepatic vagotomy had no effect,
than measuring other hormones or peptides, such as
somatostatin, might be indicated.

Three directions that could be pursued from the insulin
and food intake part of this research would include: 1) a

"classic" approach that has typified previous research

38

investigating the role of a peptide with food intake
regulation. 2) An approach that would examine the "basic"
mechanisms of insulin in the brain, with molecular biology
techniques. 3) An approach that would work on elucidating
the "central hardware" for food intake regulation.

Classic Approach

Typically after an ICV infusion of a peptide has been
shown to affect food intake, micro-infusions are done in
representative sites throughout the brain. In the most
sensitive areas, several different amounts of that peptide
are infused to determine: a dose response curve, affects on
macronutrient selection and affects on meal patterns. The
most sensitive area can be lesioned to determine if this
will attenuate the response to an ICV infusion. Then
various combinations of manipulations such as: lesions,
pharmacological agonists or antagonists, and diets are
tested.

Molecular Biology Approach

Since: insulin receptors are found throughout the
brain: insulin is involved with different behaviors and
functions in the brain: and the location of insulin's
affects on food intake are unknown: it is likely that work
on post receptor effects of insulin would lead to other
fields of research than food intake. However, a mechanism
can be postulated on how insulin could affect long term food

intake at the cellular level. Since insulin is thought to

39

be involved with the regulation of body weight and long term
food intake regulation, it is plausible that the short term
regulators of food intake (initiation and termination of a
meal) are functioning normally and that only the sensitivity
to these short term regulators is changed. Two possible
mechanisms for modifying the sensitivity, would be to alter
the extracellular ion concentration or change the number of
synapses between neurons.

Small changes in the extracellular potassium can cause
large changes in the excitability of a neuron. As an action
potential travels down a neuron, potassium moves from inside
the neuron to the outside. One of the functions of glial
cells, the structural cells in the brain, is maintaining a
constant extracellular potassium concentration (Kandel,
1985). Postreceptor effects of insulin could affect ion
channels in the glial cells to either increase or decrease
the amount of potassium taken up by the glial cell.

Glial cells in the pituitary can insert and retract
processes between neurons and thereby change the number of
synapses between neurons (Hatton et al., 1988). There are
more than 10 times the number of glial cells as compared to
neurons in the brain (Kandel, 1985). Different peptides
have been reported to alter cAMP or inositol phospholipid
turnover in glial cells in culture, although none of these
peptides were insulin (Cholewinski and Wilkin, 1988).

Insulin receptors have been identified on glial cells

40

(Bouhaddi et al., 1988). Insulin has been shown to increase
phospholipid incorporation in membranes derived from the
whole brain (Figlewicz et al., 1990) (considering the
proportion of neurons to glial cells, it is a good
possibility that this finding occurred in glial cells). It
is possible, insulin is causing post-receptor effects on
glial cells to alter the number of synapses between neurons.

Therefore there is evidence to suggest that insulin
could be causing long term effects on food intake by
interacting with glial cells. Increasing or decreasing
neuron sensitivity could either increase or decrease the
amount of information making it through the various synaptic
relays or checkpoints in the brain. This could either
increase or decrease the probability of a "satiety peptide"
terminating a meal.

Determining the "Hardware"

An area that is crucial to determine the
'physiological' mechanisms for food intake, are the neural
pathways involved with food intake. By using different dyes
and nerve tract staining techniques, it can be determined
where all the nerve fibers from a particular structure go.
For example, injection of a dye into the stomach wall will
label neurons in the dorsal motor nucleus of the vagus
(Yoshida et al., 1988). Injecting label into the VMH,
(Fahrbach, et al., 1989) identified over 12 major areas of

the brain that have neurons that terminate in the VMH. But

41

just knowing that neurons connect two structures, does not
explain what kind of information is carried by those
neurons.

Starting with structures where it is relatively easy to
determine what type of information is being carried, like
the vagus, single unit recording (to monitor electrical
activity) and microdialysis (to monitor chemical activity)
need to be done in the awake animal. The pathways must be
carefully mapped out, one step at a time, moving up into
higher structures of the brain.

In the not to distant future, the technology will
probably be available where recording from several neurons
simultaneously, in different connecting areas of the brain,
in an awake animal will be possible. With this type of
information, associating behavior with specific neuronal
tracts will be greatly facilitated.

Once the normal activity within the neuronal pathway
has been described over the duration of a meal: the neural
activity elicited by various infused peptides could be
compared to what normally happens during a meal. This would
be stronger support that a peptide is or is not involved in
physiological hunger or satiety, than any type of evidence
available to date.

Comparing the effects of an infused substance on neural
activity, to the normal activity that occurs during

ingestive behavior, could work as a physiological screening

42

method for identifying novel compounds that might be useful

in modifying intake.

APPENDICES

APPENDIX A

MANUFACTURE OF EQUIPMENT
Luer Plug (see Figure 3)

The luer plug (perfectum adapter: Popper and Sons) is
solid stainless steel (a). The luer plug was made to fit
into the hub of a needle. A small hole is drilled through
the luer plug so that the ventricular needle can just pass
through easily. Another hole was drilled perpendicular (b)
to accommodate a stainless steel screw. When the screw was
tightened, the ventricular needle could not be moved.

Guide Cannulas

Guide cannulas (c) were made from 16 gauge stainless
steel hypodermic needles. A rotary grindstone was used to
cut the needle at 1.5 cm from the bottom of the hub of the
needle. The ventricular needles needed to be passed through
the guide cannula to test for patency prior to surgery. In
about 30% of the 16 gauge needles, there was a stricture
where the needle part was attached to the hub. To open this
stricture, a 1" safety pin was used. The bottom of the
safety pin was held in a vice grips. The hub end of the 16
gauge needle was twisted on and off of the pointed end of
the safety pin. This was enough to allow the ventricular

needle to pass easily.

43

44

 

 

 

 

L

FIGURE 3 INFUSION APPARATUS

45

Ventricular Needle

The ventricular needle (d) was put into the luer plug.
The luer plug was put snugly into the hub of the guide
cannula. The ventricular needle was pushed down through the
guide cannula until the distance from the bottom of the hub
of the guide cannula to the point of the ventricular needle
was 4 cm. A black line was marked on the ventricular needle
just on top of the luer plug as a guide for depth during
surgery.

The top of the ventricular needle sticking out of the
luer plug was wrapped 3/4 times around a pencil, taking care
not to crimp the needle. This left a loop at the top (e),
and the blunt end sticking out perpendicular to the pointed
end. The loop allowed the cap to fit on, and acted like a
spring so that small movements were less likely to put a
hole in the tubing or pull the ventricular needle out.

Stylets

Stylets were made by soldering 1.7 cm long, 0.635 mm
diameter, stainless steel rod to a luer plug.

Protective Caps

The protective caps were made from 8 oz plastic
containers. The container was thick but non brittle. The
bottom was cut off with a hack saw; a bunsen burner was used
to smooth the bottom where it was jagged from being cut and
to roughly shape it for the sheep's head. A solder iron was

used to make small holes for stainless screws in the bottom

46

so that the plastic cap would stick to the dental acrylic.
A larger (‘1 cm) hole was made in the back of the cap for
the poly ethylene (PE) -tubing to go through to connect to

the osmotic minipump.

APPENDIX B

SURGICAL PROCEDURE: Step by step and

helpful hints

The sheep were anesthetized with injectable sodium
thiopental, intubated and anesthesia was maintained with
isoflurane gas. A stereotaxic apparatus was used to
stabilize and fix the position of the sheep's head. In
order to make the proper size and shape of incision, the
protective cap was positioned 2 cm cranial to the ligamentum
nuchae crest (where the neck muscles attach to the back of
the skull) on the midline. Three stab incisions were made
just inside of the cap, one on each side and one at the
caudal end. The stab incisions were guides to form a U
shaped skin flap. The initial incision was made smaller
than the cap because the skin would stretch once the hole
was made. The skin and underlying tissues were scraped off
the skull with a periosteal elevator and pulled rostrally
until an area of the top of the skull had been cleared that
extended about 2 cm cranial to the bregma suture line. An 8
mm hole was trephined 1 cm caudal to the bregma suture line
and 1 cm lateral to the midline.

While drilling a hole through the skull, the bone
seemed to be made up of 3 layers based on the ease of

drilling. A thin hard layer on the top and bottom and a

47

48

soft wider layer in between. Therefore a lot of pressure
could be put on the trephine at the start, but when the
drilling became hard again for the second time, care and
time needed to be taken so that the bone did not break off
and the trephine pith the sheep. Typically the skull was
approx. 0.8 cm thick. The soft part of the skull usually
did not end till after 0.5 cm. White faced sheep had softer
skulls than black faced sheep.

After the hole was drilled, four 1/2" stainless steel
screws were placed in the skull around the hole. I used a
1.5 mm in diameter intra- medullary pin to put small holes
in the skull to start the screws. The screws were only
tightened to approx. 9 mm (3/8") so that the dental acrylic
could flow under the screw heads.

The ventricular needle in the luer plug was inserted
into the guide cannula so that just the point of the
ventricular needle stuck out past the end of the guide
cannula. The needle point would puncture the dura mater
when I inserted the guide cannula through the hole into the
brain parenchyma.

Once the guide cannula was in the brain parenchyma, I
would withdraw the ventricular needle (being careful not to
move the guide cannula) and replace the ventricular needle
with a stylet.

The protective cap was put over the screws and guide
cannula and dental acrylic was poured in the protective cap.

Enough acrylic was poured in to cover the screws in the

49

skull and protective cap, but not cover the guide cannula.

As the acrylic hardened, PE-60 tubing was filled with
the appropriate test solution. The syringe was left
attached to the tubing, to keep the solution from running
out.

Once the acrylic was hard, the ventricular needle was
inserted into the guide cannula. As soon as the tip of the
ventricular needle entered the brain parenchyma, the syringe
was removed from the tubing and the free end of tubing was
held up above the sheep's head. When the ventricular needle
entered the ventricle, the fluid level would drop in the
tubing, then the tubing was lowered below the sheep's head.
If the ventricular needle was actually in a ventricle, there
was a steady drip of CSF siphoned out of the ventricle.

An osmotic minipump was inserted subcutaneously in the
neck. The PE tubing, that was connected to the ventricular
needle, was cut and attached to the osmotic minipump so
there was enough slack to allow the sheep free movement. A
helmet and halter was placed on the sheep after the sheep
was disconnected from the anesthetic gas machine but before
the sheep woke up. After recovery the sheep would be
returned to the metabolism cage.

During the days that followed, any air bubbles in
tubing would be an indication of a hole or interruption in

the line between the pump and the ventricular needle.

APPENDIX C

PRELIMINARY METHODS: What Worked

And What Did Not

In the process of finding a method that worked, I found
a lot of things that did not work. When I realized that a
method was not working, I would try and come up with all the
possibilities why. Then I would change as many of the
potential problems as I could all at once. I was not really
interested in finding out why the technique was not working,
I merely wanted a technique that worked. This section is an
overview of the different things I tried and why I thought
they did not work.

Housing

Initially, 4 ft by 8 ft by 3.5 ft high pens for the
sheep were made out of woven wire hog panels. In order that
each pen could be opened individually, scrap hog panels were
cut 4 ft wide and used as a gate for the front of the pen.
Wood shavings were used for bedding. The sheep pushed the
shavings out of the pens through the hog panels. Therefore
holes were drilled in scrap pieces of wood, mostly 2" X 6"s,
and were wired to the bottom of the panels to keep the
shavings from coming out. The 2 X 6's also made the panels
more stable.

The sheep were fed and watered in separate 8 gal black

rubber buckets with metal o-rings on top. Both buckets were

50

51

wired in the corners of the pen so that the sheep could not
tip them over. The sheep stood in and on the buckets and
pawed their feed: either spilling or adding bedding material
to their feed.

To stop the sheep from spilling feed, wooden gates were
made from 1" x 6" commercial grade boards for the front of
the pens. Two 1'X 1' windows were cut out of each gate 1 ft
from the bottom. Wooden boxes were made from commercial
grade boards and secured to the gate below the windows. The
food and water buckets were fastened to the boxes with
rubber straps with S hooks on each end. The sheep stuck
their heads through the window into the buckets to eat or
drink. This set up worked well for keeping the sheep from
climbing in their feed. The wood was cheap and a few times
a sheep jumped through the hole and broke the gate and the
box.

There were no major problems with this setup. Minor
problems include it was not convenient to clean out: the
sheep were in the way and there were to many corners to
clean easily with a scoop shovel. Shavings would get
between the back panel and the wall. Once in a while a
sheep would jump over the panels. The sheep would
occasionally butt the sides of the pen, especially during
and after cleaning. A few sheep damaged their infusion set—
up in this manner.

Due primarily to changes in the method of infusing (see

infusion) I needed to restrict the movement of the sheep.

52

Ten calf metabolism crates were obtained from: the veterinary
research building: ULAR: and the blacksmith at farm
maintenance. In 2 crates the floor was replaced with 1/2"
cold rolled steel. This looked good but the surface was
rough: it was heavy: and I was concerned that it might be
abrasive on the sheep's legs. Several coats of "rubber
paint" were brushed on the steel to make it smoother and
softer, but it came off in long strips after the sheep had
been in the crate for a few weeks. I ended up just painting
them with four coats of metal paint. The floor still felt
rather rough but none of the sheep became lame or sore.

In the front of the metabolism cages, one wooden box
was made for each metabolism crate to hold both buckets for
feed and water. There was no support for the front of the
boxes and they did not hold up well.

Because of the problem with the boxes, wire bucket
holders (from Dairy supply) which fit on a two inch wide
board were purchased to hold buckets for water and feed. A
2 X 6 board was attached with 2 hinges to the metabolism
crate. Rubber straps with an S hook on each end were again
used to keep the sheep from pulling the buckets out of the
bucket holders.

This setup worked well, however when some of the sheep
ate, they would take a large mouthful of feed, than lift
their heads out of the bucket and look around while they
ate. Some food would spill out of the sides of their

mouths. It was difficult to quantify this amount. Therefore

53

12" diameter PVC pipe was cut in 20" lengths, with a 7" wide
slot cut out along the length of pipe. The PVC pipe was
bolted in the bucket so that the slot faced the front. This
raised the sides of the bucket 18", with a slot in front for
the sheep to stick their heads in. Now when the sheep were
eating, the sheep could not move their heads from side to
side unless they backed up first. This worked well for
avoiding spillage.
Peed

In the beginning I bought grain from Mason elevator and
mixed a 2 wk supply with chopped hay. The mix was not very
uniform and settling would occur. The feed was switched to
a maintenance pellet (feed # 1730, Countrymark, Columbus,
OH) and that worked well.

Infusion

I started using auto syringe pumps and infused 2.7
ml/day. Sheep ate less when infused with artificial CSF (A-
CSF). The amount infused was decreased to 2 ml/day. The
sheep still did not eat as well as they had before surgery.

A 3 ml syringe is the smallest syringe that will fit
securely in the auto syringe pump. In order to decrease the
infusion rate further osmotic minipumps (2m12, Alzet, Palo
Alto, CA) that delivered 140 ul/day for 2 wk were tried. At
necropsy it was found that what appeared to be the choroid
plexus, had completely walled off the tip of the cannula
from the ventricle. The dye, that was infused just prior to

euthanizing the sheep, went out of the cannula: stained the

54

choroid plexus: then was shunted back up along the outside
of the cannula to the top of the brain. So I decided I
needed a faster infusion rate, and went back to the Auto-
syringe's.

Polyvinyl (tygon) tubing was placed around tuberculin
syringes (1 ml capacity), and then the tuberculin syringes
fit securely in the Auto-syringe. Using tuberculin
syringes, I infused 0.7 ml/day. Controls still went off
feed. I went back to the minipumps, but to the 2mll which
infuse at a rate of 280 ul/day for 1 wk.

Angiotensin II (Ang II)

When Ang II is injected into the lateral ventricle an
abrupt drinking response is elicited. This effect of Ang II
has been used by other investigators as support that their
solutions were going into the ventricle (Woods et al.,
1979). However, the response of Ang II in other areas of
the brain has not been reported: Ang II in other areas of
the brain might elicit water intake also. I found it
difficult to be objective on what constituted a response.
Sometimes immediately after infusing the Ang II a sheep
would go to the water bucket and drink a 1/2 kg of water:
another sheep would just stand there for a minute and then
drink 4 kg of water all at once. There seemed to be a
variation in response. Plus I was concerned about what
effects Ang II might have on food intake, and if Ang II
might change the response to insulin. In the final method

I did not use a response to Ang II as a criteria for patency

55

of the infusion system.
Surgical Techniques

I first used a method described in Pappenheimer et al.,
(1962) to implant a permanent cannula into the cisterna
magna. Later I learned from Dr. Heisey (a co-author in the
paper by Pappenheimer et al., 1962: and a professor here at
Michigan State Univ. in the physiology department), that
there were some major differences between their technique
and ours (Foster et al.,1987).

The cannulas used to access the cisterna magna were
made from 14 gauge hypodermic needles cut to a length of 5
cm. A 1 cm crossbar was made with 16 gauge baling wire.
Copper wire was used to tie the crossbar 1.7 cm from the tip
of the hypodermic needle. A 90% metal epoxy was used to
cement the crossbar to the hypodermic needle.

These cannulas were placed in holes drilled through the
occipital bone directly above the cisterna magna. Two metal
screws were placed into the occipital bone, one above and
one below the crossbar. Then dental acrylic was used to
cement the cannula to the skull. The screws held the dental
acrylic to the skull. This type of cannula was used on 12
sheep. All sheep recovered for 2 wk before any experiments
started. When I went to collect CSF I found that the
cannulas had come loose on seven of the sheep. What I think
happened is an electric current developed because of all the
different metals that I used to make the cannulas. This

weakened the metal epoxy and allowed the crossbar to come

56

loose from the 16 gauge needle. I tried doing three of the
sheep over again, but the occipital bone seemed to have
become restructured also. These changes in the skull
supported the idea that I might have set up an electric
current.

In the sheep that did not lose their cannulas, there
was a cellular infiltration around the obturator up the
guide cannula.

The second model of cisterna magna cannulas were made
from stainless steel hypodermic needles, the crossbars were
made from stainless steel wire. The crossbars were soldered
on the hypodermic needle. The solder was coated with rubber
paint. An obturator was made of stainless steel wire, and
coated with rubber paint for a tighter fit in the cannula to
stop the cellular infiltration.

The second model was used on 6 sheep. I was able to
get a CSF sample on all these sheep. However the day after
collecting several samples from each sheep, 4 of the 6 sheep
were showing abnormal clinical signs, like the spinal cord
had been damaged. The first sample of CSF was easy to get,
but regardless of the amount of CSF I removed, or even the
length of time between samples on the same day, each
successive sample was considerably more difficult to
collect. When I had trouble collecting a sample of CSF I
would move the needle around. I used 18 gauge 4" spinal
needles to puncture the dura and collect the CSF. The

diameter of the spinal needles was smaller than the guide

57

cannula. The up and down movement probably did not cause a
problem. The side to side movement, that I was not even
aware of, could easily have damaged the spinal cord.

At the time I thought that this technique would not
work. If I had used the luer plug with a side screw to hold
the ventricular needle solid that I used for the last
experiments, this technique might have worked.

The second technique I tried was cannulating the
subarachnoid space of the spinal cord. A 14 gauge 1" needle
was inserted into the subarachnoid space between the lumbar
vertebrae. Polyethylene (PE) tubing was pushed through the
needle and subarachnoid space to the first cervical
vertebrae. The tubing went easy until the thoracic cervical
junction. A fine wire (initially I used a guitar wire) that
would fit inside the PE tubing was used as a stylet. I
could verify the location of the tip of the tubing by
leaving the wire in the tubing and taking an x-ray.

(A subjective observation was that this method
increased the respiratory rate of sheep without affecting
their temperature.)

The subarachnoid cannulation was performed on two sheep
initially. When A-CSF (control solution) was infused into
the subarachnoid cannulas, their food intake immediately
dropped by 50%. The infusion was stopped, and the sheep
returned to eating the same amount they had been eating
before the infusion. New A-CSF was made and the infusion

rate was decreased. Three days elapsed before the infusion

58

began again. This time food intakes did not decrease.
After two weeks 400 uU/kg/day of insulin was added to the
infusion. There was no response in food intake after 2
weeks, so the dose of insulin was doubled. Three weeks
later the dose was increased 10 times. None of the doses
affected food intake. The sheep were anesthetized and a
radio-opaque dye was injected into the tubing under a
fluoro-scope. The dye seemed to leak out of the tube
everywhere the tubing rubbed against the vertebrae of the
spinal cord (three places). No dye was observed in the
ventricle system.

At necropsy I found that connective tissue had been
laid down over the entire PE tubing I was infusing through.
I do not know how long it took for this connective tissue
growth to occur, but I thought that once it was formed that
it would be unlikely that the infusion would reach the
ventricular system. This also reinforced the importance of
being able to verify where the infusion was actually going.
That the mere physical presence of the tubing in the proper
location at necropsy did not prove the infusion had been
going in the right place. Since it seemed that glial cells
or some type of fibroblast (connective tissue forming cells)
were so active within the spinal cord and ventricular
system. I wanted a method of verifying the patency_of the
tubing that could be used throughout the experimental period
and not just at necropsy.

The third technique came from Breuhaus et al., (1989).

59

A 0.5 cm hole was drilled in the skull over the lateral
ventricle. Polyvinyl tubing (.085" 00) was lowered by hand
into the brain parenchyma. A manometer filled with A-CSF
was connected to the tubing: once the tubing entered the
ventricle the level of A-CSF in the manometer would drop.
Sometimes CSF would drip from the end of the tubing. I felt
this was a good indication that I was actually in the
ventricle. For the final method a consistent drip of CSF
out of the tube was one of the criteria for patency of the
infusion system.

Sometimes the fluid did not drop while the tubing was
pushed through the brain because of either: a tissue plug in
the end of the tubing: or the tubing did not penetrate a
ventricle.

When the tube was pulled back out, sometimes there was
an obvious plug, other times it was hard to tell. Since the
tube was lowered by hand, the second time the tubing was
lowered, it was impossible to ensure the tube went down the
same tract. Another concern was that the tubing made a
large enough tract to act like a false ventricle or to leave
a tract to the ventricle that would close up later. In
either case the A-CSF would drop in the manometer even
though the end of the tubing was not actually in the
ventricle.

Another potential problem area was when the tube was
cemented in. While one person held the tube, another person

would press the dental acrylic around the hole and tubing.

60

The person holding the tube could pull the tube out of the
ventricle and not realize it.

Another potential problem with this method was the
tubing that stuck out of the hole was bent down and cemented
flat against the skull. Over the next several days the
portion of the tube that was in the brain may have partially
straightened out, removing the tube from the ventricle.

A vacuum pac (Olympia Washington) was used to support
the head. The vacuum pac feels like a small bean bag chair.
When vacuum is applied to the vacuum pac it becomes rigid.
The vacuum pac worked well for support, but if the sheep
moved during surgery, the vacuum pac would not hold the head
in the same position. The movement itself was a very minor
problem but a small rotation of the head could easily go
unnoticed because of the drapes. Just a small rotation
could make a big difference in where the tip of the cannula
ended up. This problem was eliminated by the use of a
stereotaxic instrument.

I do not know which of these potential problems were
actually problems. I do know that sometimes I saw a drop in
the level of A-CSF in the manometer at the time of surgery
and sometimes I saw a dramatic drinking response to Ang II
at the beginning of the experiment but a poor or no
drinking response at the end of the experiment and/or the
cannula out of the ventricle at necropsy.

To get around the problem of the tube straightening out

after the tube had been cemented to the skull, and the

61

problem of it being so easy to move the tube while cementing
it in, I switched from polyvinyl to silicone tubing
(silastic). A 14 gauge 2" needle and obturator were used to
make a hole. The proper depth of the hole was determined by
connecting a manometer filled with A-CSF to the needle.

When the pressure dropped in the manometer: the manometer
was disconnected: the silastic tubing was pushed into the
needle: and the needle was pulled out. The problem with
this was keeping the silastic tubing at the same depth while
pulling the needle out (ie pulling the silastic tubing out
with the needle). Plus if there was not a flow of CSF out
of the end of the tube, the tube could not be put in deeper.
The problem of a tissue plug that I encountered with the
polyvinyl technique was made worse by this technique.

Next a model of a cannula was obtained from Dr.
Latteman in Seattle WA that was used with baboons. Dr.
Latteman's model had six guide cannulas arranged in 2 rows
of three. The guide cannulas were attached to a 2 cm in
diameter flat round piece of stainless steel (approx. the
size of a quarter) with holes around the perimeter for
screws to be placed into the skull. Engineering said they
could make one for something like 200 ~300$. I also
obtained a model from Dr. Della- Ferra at St. Louis, MO.
that was used with sheep. Dr. Della-Ferra's model consisted
of a single guide cannula, a luer plug with a side screw so
that the depth of the ventricular needle could be changed.

Initially I tried combining Dr. Latteman's and Dr.

62

Della-Ferra's models. I had engineering put a 16 gauge
stainless steel needle through a 2 cm in diameter round flat
piece of stainless steel. The stainless steel disk had
three holes to screw it to the skull. Although the screws
through the disk and into the skull made the guide cannula
very solid, while turning the screws in, the disk would
wobble on the skull, increasing the risk of complications
due to trauma. In the last experiment, I relied on the
dental acrylic for holding the guide cannula.

There were in particular, 3 very expensive lessons
learned during this project about the development of a new
technique. First, keep the experimental turnover time as
short as possible. For the first several experiments I had
four months invested in a group of animals before finding
out if the cannulas were in the ventricles or not. Included
in keeping the experimental turnover time down are: keep the
number of animals to a minimum: keep the design very simple
and short: do the things that you think will cause the
maximum response in the shortest time (once you know your
technique works you can refine it later). Second, go the
"best route first": you can evaluate the performance of
cheaper substitutes once you know how the "right stuff"
works. Third, if anyone is doing anything similar, go watch

them and have them watch you.

BIBLIOGRAPHY

10.

11.

12.

BIBLIOGRAPHY

Adolph, E.F. Urges to eat and drink in rats. Am. J.
Physiol. 151:110-125: 1947.

Amir, S.: Lagiorgia, M.: Pollock, R. Intra-ventromedial
hypothalamic injection of insulin suppresses brown fat
thermogenesis in the anaesthetized rat. Brain Res.
480:340-343; 1989.

Arase, K.: Fisler, J.S.: Shargill, N.S.: York, D.A.: Bray,
G.A. Intracerebroventricular infusions of 3-OHB and insu-
lin in a rat model of dietary obesity. Am. J. Physiol.
255:R974-R981; 1988.

Banting, P.: Best, C.H.: Collip, J.B.: Macleod J.J.R.:
Noble, E.C. The effect of pancreatic extract (insulin) on
normal rabbits. Am. J. Physiol. 62:162-176: 1922.

Baskin, D.G.: Porte, D.: Guest, K.: Dorsa, D.M. Regional
concentrations of insulin in the rat brain. Endocrinology
112:898-903; 1983.

Berelowitz, M.: Szabo, M.: Frohman, L.A.: Firestone, S.:
Chu, L. Somatomedin-C mediates growth hormone negative
feedback by effects on both the hypothalamus and the
pituitary. Science 212:1279-1281: 1981.

Bergman, E.N.: Roe, W.E.: Kon, K. Quantitative aspects of
propionate metabolism and gluconeogenesis in sheep. Am.
J. Physiol. 211:793-799: 1966.

Bernardis, L.L.: McEwen, G.: Kodis, M.: Feldman, M.J.
Pair-feeding of sham-operated controls to rats with
dorsomedial hypothalamic lesions: new evidence for an
'organismic' set point. Behav. Brain Res. 26:99-108: 1987.

Bouhaddi, K.: Thomopoulos, P.: Pages, C.: Khelil, M.:
Tardy, M. Insulin effect on GABA uptake in astroglial
primary cultures. Neurochem. Res. 13:1119-1124: 1988.
Boyd, P.: Clarke, D.W.: Raizada, M.K. Insulin inhibits
specific norepinephrine uptake in neuronal cultures from
rat brain. Brain Res. 398:1-5: 1986.

Breuhaus, B.A.: Demarest, K.T.: Chimoskey, J.E. Comparison

63

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

64

of intracerebroventricular and intracarotid infusions of
PGFE2 in conscious sheep. Am. J. Physiol. 256:R685-R693:
1989.

Brief, D.J.: Davis, J.D. Reduction of food intake and body
weight by chronic intraventricular insulin infusion. Brain
Res. Bull. 12:571-575: 1984.

Brockman, R.P.: Laarveld, B. Hormonal regulation of metab-
olism in ruminants: a review. Livest. Prod. Sci. 14:313-
334; 1986.

Cain, D.P. Effects of insulin injection on responses of
olfactory bulb and amygdala single units to odors. Brain
Res. 99:69-83: 1975.

Carr, G.D.: White, N.M. Contributions of dopamine terminal
areas to amphetamine-induced anorexia and adipsia.
Pharmacol. Biochem. Behav. 25:17-22: 1986.

Chiodo, L.A. Dopamine-containing neurons in the mammalian
central nervous system: electrophysiology and
pharmacology. Neurosci. Biobehav. Rev. 12:49-91: 1988.

Cholewinski, A.J.: Wilkin, G.P. Peptide receptors on
astrocytes:evidence for regional heterogeneity. Biochem.
Soc. Trans. 16:429-432: 1988.

Cohn, C.: Joseph, D. Influence of body weight and body
fat on appetite of "normal" lean and obese rats. Yale J.

Cserr, H. Potassium exchange between cerebrospinal fluid,
plasma, and brain. Am. J. Physiol. 209:1219-1226: 1965.

Danguir, J. Sleep deficits in diabetic rats: restoration
following chronic intravenous or intracerebroventricular
infusions of insulin. Brain Res. Bull. 12:641-645: 1984.

Davidson, D.A.: Bohannon, N.J.: Corp, E.S.: Lattemann,
D.P.: Woods, S.C.: Porte, D.: Dorsa, D.M.: Baskin, D.J.
Evidence for separate receptors for insulin and insulin-
like growth factor-I in choroid plexus of rat brain by
quantitative autoradiography. J. Histochem. Cytochem.
38:1289-1294; 1990.

de Boer, G.: Trenkle, A.: Young, J.W. Glucagon, insulin,
growth hormone, and some blood metabolites during energy
restriction ketonemia of lactating cows. J. Dairy Sci.
68:326-337.: 1985.

de Jong, A.: Strubbe, J.H.: Steffens, A.B. Hypothalamic
influence on insulin and glucagon release in the rat. Am.
J. Physiol. 233:E380-E388: 1977.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

65

Deetz, L.E.: Wangsness, P.J.: Kavanaugh, J.F.: Griel, L.C.
Effect of intraportal and continuous intrajugular adminis-
tration of insulin on feeding in sheep. J. Nutr. 110:1983-
1991; 1980.

Della-Fera, M.A.: Buonomo, E.C.: Baile, C.A. Growth hor-
mone secretory responsiveness to multiple injections of
growth hormone-releasing factor in sheep. Dom. Anim.
Endocrinol. 3:153-164: 1986.

Drewry, M.M.: Harris, R.B.S.: Martin, R.J. The effect of
increased adiposity on food intake of juvenile rats.
Physiol. Behav. 45:381-386: 1989.

Duffy, K.R.: Pardridge, W.M. Blood-brain barrier transcy-
tosis of insulin in developing rabbits. Brain Res. 420:32-
38; 1987.

Edholm, O.G.: Fletcher, J.G.: Widdowson, E.M.: McCance,
R.A. The energy expenditure and food intake of individu-
al men. Brit. J. Nut. 9:286-301: 1955.

Fahrbach, S.F.: Morrell, J.J.: Pfaff, D.W. Studies of
ventromedial hypothalamic afferents in the rat using three
methods of HRP application. Exp. Brain Res. 77:221-233:
1989.

Figlewicz, D.: Szot, P.: Greenwood, M.R.C. Insulin stimu-
lates inositol incorporation in hippocampus of lean but
not obese zucker rats. Physiol. Behav. 47:325-330: 1990.

Fonberg, E. Amygdala, depression and drug treatment. Acta
Physiol. Hungarica 74:103-114: 1989.

Foster, L.A.: Ames, N.K.: Emery, R.S. Technique for
continuous access to cerebrospinal fluid (CSF) in sheep to
quantify peptides related to feed intake. J. Anim. Sci.
65(81):390: 1987.

Fregley, H.J.: Garper, J.M.: Radford, E.P. Regulation of
sodium chloride intake by rats. Am. J. Physiol. 209:287-
292: 1965.

Giza, B.K.: Scott, T.R. Intravenous insulin infusions in
rats decrease gustatory-evoked responses to sugars. Am. J.
Physiol. 252:R994-R1002: 1987.

Goodner, C.J.: Berrie, M.A. The failure of rat hypothalam-
ic tissues to take up labeled insulin in vivo or to re-
spond to insulin in vitro. Endocrinol. 101:605-612: 1977.

Grill, H.J.: Norgren, R. The taste reactivity test. II.
mimetic responses to gustatory stimuli in chronic thalamic

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

66

and chronic decerebrate rats. Brain Res. 143:281-297:
1978a.

Grill, H.J.: Norgren, R. Chronically decerebrate rats
demonstrate satiation but not bait shyness. Science
201:267-269: 1978b.

Grizard, J.: Champredon, C.: Aina, E.: Sornet, C.: Debras,
E. Metabolism and action of insulin and glucagon in goat
during lactating and dry period. Horm. Metabol. Res.
20:71-76: 1988.

Guler, H.: Schmid, C.: Zapf, J.: Froesch, E.R. Effects of
recombinant insulin-like growth factor I on insulin secre-
tion and renal function in normal human subjects. Proc.
Natl. Acad. Sci. 86:2868-2872: 1989.

Hatton. G.I.: Yang, Q.z.: Smithson, K.G. Synaptic inputs
and electrical coupling among magnocellular neuroendocrine
cells. Brain Res. Bull. 20:751-755: 1988.

Havrankova, J.: Brownstein, M.: Roth, J. Insulin and
insulin receptors in rodent brain. Diabetologia 20:268-
273: 1981.

Havrankova, J.: Roth, J. Insulin receptors are widely
distributed in the central nervous system of the rat.
Nature 272:827-829: 1978.

Heitmann, R.N.: Sensenig, S.C.: Reynolds, C.K.: Fernandez
J.M.: Dawes, D.J. Changes in energy metabolite and
regulatory hormone concentrations and net fluxes across
splanchnic and peripheral tissues in fed and progressively
fasted ewes. J. Nutr. 116:2516-2524: 1986.

Hernandez, L.: Hoebel, B.G. Feeding and hypothalamic
stimulation increase dopamine turnover in the accumbens.
Physiol. Behav. 44:599-606: 1988.

Hervey, G.R. Physiological factors involved in long-term
control of food intake. Internat. J. Vit. Nutr. Res.
58:477-490: 1988.

Hill, J.M.: Lesniak, M.A.: Pert, C.B.: Roth, J. Autora-
diographic localization of insulin receptors in rat brain:
prominence in olfactory and limbic areas. Neuroscience
17:1127-1138: 1986.

Hoebel, B.G.: Teitelbaum, P. Hypothalamic control of
feeding and self-stimulation. Science 135:375-377 1962.

Hove, K.: Blom, A.K. Plasma insulin and growth hormone in
dairy cows: diurnal variation and relation to food intake
and plasma sugar and acetoacetate levels. ACTA Endocrinol.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

67

73:289-303: 1973.

Ikeda, H.: West, D.B.: Pustek, J.J.: Figlewicz, D.P.:
Greenwood, M.R.C.: Porte, D.: Woods, S.C. Intraventricu-
lar insulin reduces food intake and body weight of lean
but not obese zucker rats. Appetite 7:381-386: 1986.

Jacobs, E.M.: Mark, G.P.: Scott, T.R. Taste responses in
the nucleus tractus solitarius of sodium-deprived rats. J.
Physiol. 406:393-410: 1988

Handel, E.R. In: Principles of Neural Science. Eds Handel,
E.R. and Schwartz. 2nd ed. Elsevier New York: p17: 1985.

Kasuga, M.: zick, Y.: Blithe, D.L.: Crettaz, M.: Kahn,
C.R. Insulin stimulates tyrosine phosphorylation of the
insulin receptor in a cell-free system. Nature 298:667-
669; 1982.

Katoh-Semba, R.: Keino, H.: Kashiwamata, S. A possible
contribution by glial cells to neuronal energy production:
enzyme-histochemical studies in the developing rat cere-
bellum. Cell Tissue Res. 252:133-139: 1988.

Kennedy, G.C. The role of depot fat in the hypothalamic
control of food intake in the rat. Proc. Roy. Soc. B.
(London) 140:578-592: 1953.

King, E.M.: Dallman, M.F.: Esquerre, E.R.: Frohman, L.A.
Radio-frequency vs. electrolytic VMH lesions: differential
effects on plasma hormones. Am. J. Physiol. 254:R918-R924:
1988.

Kirchgessner, A.L.: Sclafani, A.: Nilaver, G.
Histochemical identification of a PVN-hindbrain feeding
pathway. Physiol. Behav. 42:529-543: 1988.

Lauterio, T.J.: Marson, L.: Daughaday, W.H.: Baile, C.A.
Evidence for the role of insulin-like growth factor II
(IGF-II) in the control of food intake. Physiol. Behav.
40:755-758; 1987.

Leibowitz, S.F. Hypothalamic paraventricular nucleus:
interaction between az-noradrenergic system and
circulating hormones and nutrients in relation to energy
balance. 12:101-109: 1988.

Manin, M.: Balage, M.: Larue-Achagiotis, C.: Grizard, J.
Chronic intracerebroventricular infusion of insulin failed
to alter brain insulin-binding sites, food intake, and
body weight. J. Neurochem. 51:1689-1695: 1988.

Manin, M.: Broer, Y.: Balage, M.: Rostene, W.: Grizard, J.
Metabolic clearance of insulin from the cerebrospinal

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

68

fluid in the anesthetized rat. Peptides 11:5-12: 1990.

Mayer, J. Glucostatic mechanism of regulation of food
intake. New Engl. J. Med. 249:13-16: 1953.

McCann, J.P.: Bergman, E.N.: Reimers, T.J. Effects of
obesity and ovarian steroids on insulin secretion and
removal in sheep. Am. J. Physiol. 256:E116-E128: 1989.

McElduff, A.: Poronnik, P.: Baxter, R.C.: Williams, P. A
comparison of the insulin and insulin-like growth factor I
receptors from rat brain and liver. Endocrinol. 122:1933-
1939; 1988.

McGarry, J.D.: Foster, D.W. Regulation of hepatic fatty
acid oxidation and ketone body production. Ann. Rev.

*Biochem. 49:395-420: 1980.

McGowan, M.K.: Andrews, K.M.: Kelly, J.: Grossman, S.P.
Effects of chronic intrahypothalamic infusion of insulin
on food intake and diurnal meal patterning in the rat.
Behav. Neurosci. 104:373-385: 1990.

Minano, F.J.: Peinado, J.M.: Myers, R.D. Profile of NE, DA
and 5-HT activity shifts in medial hypothamus perfused by
2-DG and insulin in the sated or fasted rat. Brain Res.
Bull. 22:695-704; 1989.

Myers, R.D.: Peinado, J.M.: Minano, F.J. Monoamine
transmitter activity in lateral hypothalamus during its
perfusion with insulin or 2-DG in sated and fasted rat.
Physiol. Behav. 44:633-643 1988.

Myers, R.D.: Swartzwelder, H.S.: Peinado, J.M.: Lee, T.F.:
Hepler, J.R.: Denbow, D.M.: Ferrer, J.M.R. CCK and other
peptides modulate hypothalamic norepinephrine release in
the rat: dependence on hunger or satiety. Brain Res. Bull.
17:583-597: 1986.

Ono, T.: A.B. Steffens: Sasaki, K. Influence of peripheral
and intracerebroventricular glucose and insulin infusions
on peripheral and cerebrospinal fluid glucose and insulin
levels. Physiol. Behav. 30:301-306: 1983.

Oomura, Y. Chemical and neuronal control of feeding
motivation. Physiol. Behav. 44:555-560: 1988.

Oomura, Y.: Kita, H. Insulin acting as a modulator of
feeding through the hypothalamus. Diabetologia 20:290-298
1981.

Owen, O.E.: Reichard, C.A.: Boden, G.: Shuman, C. Compara-
tive measurements of glucose, beta-hydroxybutyrate, aceto-
acetate, and insulin in blood and cerebrospinal fluid

74.

75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

69

during starvation. Metabolism 23:7-14: 1974.

Pappenheimer, J.R.: Heisey, S.R.: Jordan, E.F.: Downer, J.
deC. Perfusion of the cerebral ventricular system in
unanesthetized goats. Am. J. Physiol. 203:763-774: 1962.

Plata-Salaman, C.R.: Oomura, Y. Effect of intra-third
ventricular administration of insulin on food intake after
food deprivation. Physiol. Behav. 37:735-739: 1986.

Plata-Salaman, C.R.: Oomura, Y: Shimizu, N. Dependence of
food intake on acute and chronic ventricular administra-
tion of insulin. Physiol. Behav. 37:717-734: 1986.

Porte, D.: Woods, S.C. Regulation of food intake and body
weight by insulin. Diabetologia 20:274-280: 1981.

Rechler, M.M.: Nissley, S.P. Insulin-like growth factor
(IGF)/somatomedin receptor subtypes:structure, function,
and relationships to insulin receptors and IGF carrier
proteins. Hormone Res. 24:152-159: 1986.

Recio-Pinto, E.: Ishii, D.N. Insulin and related growth

factors: effects on the nervous system and mechanism for
neurite growth and regeneration. Neurochem. Int. 12:397-
414: 1988.

Rinderknecht, E.: Humbel, R.E. Primary structure of human
insulin-like growth factor II. FEBS Let. 89:283-287: 1978.

Rusk, I.N.: Cooper, S.J. The selective dopamine D1
receptor agonist SK&F 38393: its effects on palatability-
and deprivation-induced feeding, and operant responding
for food. Pharmacol. Biochem. Behav. 34:17-22: 1989.

Sakaguchi, T.: Bray, C.A.: Eddlestone, G. Sympathetic
activity following paraventricular or ventromedial
hypothalamic lesions in rats. Brain Res. Bull. 20:461-465:
1988

Sasaki, N.: Rees-Jones, R.W.: Zick, Y.: Nissley, S.P.:
Rechler, M.M. Characterization of insulin-like growth
factor I-stimulated tyrosine kinase activity associated
with the fl-subunit of type I insulin-like growth factor
receptors of rat liver cells. J. Biol. Chem. 260:9793-
9804: 1985.

Schwartz, M.W.: Sipols, A.: Kahn, S.E.: Lattemann, D.P.:
Taborsky, C.J.: Bergman, R.N.: Woods, S.C.: Porte, D.
Kinetics and Specificity of insulin uptake from plasma
into cerebrospinal fluid. Am. J. Physiol. 259:E378-383:
1990.

Shemer, J.: Raizada, M.K.: Masters, B.A.: Ota, A.: Le-

86.

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

70

Roith, D. Insulin-like growth factor I receptors in neu-
ronal and glial cells. J. Biol. Chem. 262:7693-7699: 1987.

Shimazu,T.: Ogasawara, S. Effects of hypothalamic stimu-
lation on gluconeogenesis and glycolysis in rat liver. Am.
J. Physiol. 228:1787-1793: 1975.

Sims, E.A.H.: Goldman, R.P.: Gluck, C.M.: Horton, E.S.:
Kelleher, P.C.: Rowe, D.W. Experimental obesity in man.
Trans. Assoc. Am. Physicians 81:153-170: 1968.

Shor-Posner, G.: Grinker, J.A.: Marinescu, C.: Brown, 0.:
Leibowitz, S.F. Hypothalamic serotonin in the control of
meal patterns and macronutrient selection. Brain Res.
17:663-671: 1986.

Stanley, B.G.: Schwartz, D.H.: Hernandez, L.: Hoebel,
B.G.: Leibowitz, S.F. Patterns of extracellular
norepinephrine in the paraventricular hypothalamus:
relationship to circadian rhythm and deprivation-induced
eating behavior. Life Sci. 45:275-282: 1989

Steffens, A.B.: Scheurink, A.J.W.: Luiten, P.G.M. Inter-
ference of the nutritional condition of the rat with
peripheral glucose regulation determined by CNS mecha-
nisms. Physiol. Behav. 35:405-410: 1985.

Steffens, A.B.: Scheurink, A.J.W.: Porte, D.: Woods, S.C.
Penetration of peripheral glucose and insulin into cere-
brospinal fluid in rats. Am. J. Physiol. 255:R200- R204:
1988.

Strubbe, J.H.: Mein, C.G. Increased feeding in response
to bilateral injection of insulin antibodies in the VMH.
Physiol. Behav. 19:309-313; 1977.

Tannenbaum, G.S.: Guyda, H.J.: Posner, B.I. Insulin-like
growth factors: a role in growth hormone negative feedback
and body weight regulation via brain. Science 220:77-79:
1983.

Tsujii, 8.: Bray, C.A. Effects of glucose, 2- deoxyglu-
cose, phlorizin, and insulin on food intake of lean and
fatty rats. Am. J. Physiol. 258:E476-E481: 1990.

van Houten, M.: Posner, B.I. Cellular basis of direct
insulin action in the central nervous system. Diabetologia
20:255-267; 1981.

Wallum, B.J.: Taborsky, G.J.: Porte, D. Jr.: Figlewicz,
D.P.: Jacobson, L.: Beard, J.C.: Ward, W.K.: Dorsa, D.
Cerebrospinal fluid insulin levels increase during in-
travenous insulin infusions in man. J. Clin. Endocrinol.
Metab. 64:190-194: 1987.

97.

98.

99.

100.
101.
102.

103.

104.

105.

106.

71

Wilcox, B.J.: Corp, E.S.: Dorsa, D.M.: Figlewicz, D.P.:
Greenwood, M.R.C.: Woods, S.C.: Baskin, D.G. Insulin
binding in the hypothalamus of lean and genetically obese
zucker rats. Peptides 10:1159-1164: 1989.

Woods, S.C.: Lotter, E.C.: McKay, L.D.: Porte, D. Chronic
intracerebroventricular infusion of insulin reduces food
intake and body weight of baboons. Nature 282:503-505:
1979.

Woods, S.C.: McKay, L.D.: Stein, L.J.: West, D.B. Insulin
in the cerebrospinal fluid (CSF) of the baboon: response
to feeding and intravenous (iv) infusions. Diabetes
30:119A: 1981.

Woods, S.C.: McKay, L.D.: Stein, L.J.: West, D.B.: Lotter,
E.C.: Porte, D. Neuroendocrine regulation of food intake
and body weight. Brain Res. Bull. 5(S4):1-5: 1980.

Woods, S.C.: Porte, D. Effect of intacisternal insulin on
plasma glucose and insulin in the dog. Diabetes 24:905-
909; 1975.

Woods, S.C.: Porte, D. Relationship between plasma and
cerebrospinal fluid insulin levels of dogs. Am. J. Physi-
Ol. 233:E331-E334; 1977.

Woods, S.C.: Porte, D.: Bobbioni, E.: Ionescu, E.: Sauter,
J.P.: Rohner-Jeanrenaud, P.: Jeanrenaud, B. Insulin: its
relationship to the central nervous system and to the
control of food intake and body weight. Am. J. Clin. Nutr.
42:1063-1071; 1985.

Wright, E.M. Secretion and circulation of the cerebro-
spinal fluid. In: Rodriquez E.M.: van Wimersma Greidanus,
T.B. eds. Cerebrospinal Fluid (CSF) and Peptide Hormones,
volume 9. Frontiers of hormone research. Basel, New York,
Karger: 1982:4-15.

Wurtman, R.J.: Wurtman, J.J. Carbohydrate craving, obesity
and brain serotonin. Appetite 7:99-103: 1986.

Yoshida, J.: Polley, E.H.: Nyhus, L.M.: Donahue, P.E.
Labeling of nerve cells in the dorsal motor nucleus of the
vagus of rats by retrograde transport of Fluoro-Gold.
Brain Res. 455:1-8: 1988.

HIGRN STRTE UNIV LIBRQR

IIIIIIII IIIIIIIIIIIIIIIIIIII I7IIIIII7IIIIIIIIIILIIIIIIII