RETURNING MATERIALS:
)V{ESI_J P1ace in book drop to
LJBRARJES remove this checkout from
W your record. FMES win

\.

 

 

be charged if book is
returned after the date
stamped below.

 

 

 

CEREBROSPINAL FLUID (CSF) TRANSIENT RESPONSES
INDUCED BY HYPERCAPNIA

BY

Marye Jill Fisher

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Physiology

1984

 

n.-.

ABSTRACT

CEREBROSPINAL FLUID (CS?) TRANSIENT RESPONSES
INDUCED BY EYPERCAPNIA

3}!
larye Jill Fisher

Intracranial CSF volume and brain blood volume (BBV)
are reciprocally related. This dynamic equilibrium is
transiently disrupted when an animal breathes C02. Both its
brain blood flow and BBV increase to displace competitively
a CSF volume (Vd). ‘Ventriculocisternal perfusion studies
using tracers have shown that not only is there a transient
increase in CSF outflow rate with hypercapnia. but also a
transient decrease in CSF effluent tracer concentration
(Cd). These CSF transient responses (Vd and Cd) during
early respiratory acidosis are attributed to two phenomena.
One explanation assumes a homogeneous distribution of tracer
throughout the CSF spaces and that CSF formation (vf)
episodically increases during C02 breathing. The second
explanation allows heterogeneous dispersion of tracer in CSF,
an unchanging 9f, and is based on a redistribution of
craniospinal fluid volumes with C02 breathing. Both
phenomena may contribute to the short-term displacement of

a dilute CSF during hypercapnia.

Harye Jill Fisher

CSF transient responses to C02 inhalation were measured
before and after facilitated perfusate flow through sub-
arachnoid spaces of anesthetized cats during ventriculocis-
ternal perfusion with artificial CSF containing 14C-dextran.
Convective mixing of perfusate in subarachnoid spaces was
augmented while infusion was constant, either by impeding
cisternal efflux of perfusate by raising the cisternal out-
flow cannula (high CSF pressure), or by preventing CSF
outflow by clamping the cisternal outflow cannula (stop-
flow; S-F). CSF transients were also measured before and
after systemic administration of phenoxybenzamine (P82) in
order to evaluate the contribution of sympatho-adrenergic
activity to craniospinal CSF redistribution and mixing.

Results from high CSF pressure and S-F experiments
indicate that unequilibrated CSF contributes significantly
to the reduced tracer concentration in Vd, since Cd was
decreased after subarachnoid facilitated flow. Further,
results from 8-? studies indicate that at least 50% of Cd is
due to craniospinal fluid redistribution, a process which,
along with CSF outflow transients, was unaffected by PBZ.
Conversely, PBz administration decreased steady state CSF
formation and absorption through alpha-mediated cerebrovas-
cular responses and/or through beta-adrenoceptor inhibition

of metabolism of CSF secretory epithelium.

ACKNOWLEDGEMENTS

I wish to thank my mentors, Dr. S. Richard Heisey, and
Dr. Thomas Adams for their guidance and support throughout
my graduate training. Also, thanks to Deborah Traxinger and
David Manner for their collaborative and technical

contributions to this research project.

ii

TABLE OF CONTENTS
Page
LIST OF TABLES O O O O O O O O O O O O O O O O O O O 0 v i 1

LIST OF FIGURES O O O O O O O O O O O O O O O O O O O Vii 1

CHAPTER

I. INTRODUCTION . . . . . . . . . . . . . . . . 1
II. LITERATURE REVIEW . . . . . . . . . . . . . . 6
A. CEREBROSPINAL MENINGES . . . . . . . . . 6
1. Dura Mater . . . . . . . . . . . . . 6
2. Pia-Arachnoid . . . . . . . . . . . . 7
a. Arachnoid Villi . . . . . . . . . 8
B. CEREBRAL CIRCULATION . . . . . . . . . . 9
1. Arterial Supply of the Brain . . . . 9
a. Cephalic Arteries . . . . . . . . 9
b. Extracerebral (Pial) Arteries . . 11

c. Intracerebral (Parenchymal)
Arteries . . . . . . . . . . . . 11

d. Choroid Plexuses . . . . . . . . 12

2. Venous Drainage of the Brain . . . . 14
a. External Cerebral Veins . . . . . 14
b. Internal Cerebral Veins . . . . . 15
3. Cerebrovascular Innervation . . . . . 16
a. Cephalic Vessels . . . . . . . . 16
b. Pial Vessels . . . . . . . . . . 16
c. Parenchymal Vessels . . . . . . . 19
d. Choroid Plexuses . . . . . . . . 22

iii

cnaprsn Page
II. LITERATURE nsvrsw (Cont.)
c. CEREBROSPINAL FLUID . . . . . . . . . . . 25

1. CSF Composition and Homeostatic
neChanisms O O O O O I O O O O O O O 25

2. CSF Secretory Mechanism . . . . . . . 29
3. Methods of Study . . . . . . . . . . 34

a. Sampling of Choroidal Venous

Blood from in sitn Choroid

Plexus . . . . . . . . . . . . . 35
b. Extracorporeal Perfusion of the

Isolated Choroid Plexus . . . . . 37
c. Collection of Nascent Fluid from

the Surface of Exposed Choroidal

Tissue I I O O O O I O O O I O O 39

d. Ventriculocisternal Perfusion . . 40

4. Factors Affecting CSF Formation Rate 42
a. Effects of Temperature . . . . . 43

b. Effects of Osmotic Gradients . . 45
c. Effects of Blood Flow . . . . . . 49
d. Effects of Hydrostatic Pressure

Gradients . . . . . . . . . . . . 53
e. Effects of C02 Tension . . . . . 54
f. Effects of Carbonic Anhydrase

Inhibitors . . . . . . . . . . . 57
9. Effects of Cardiac Glycosides . . 60
h. Effects of the Autonomic Nervous

System . . . . . . . . . . . . . 61

5. Factors Affecting CSF Absorption Rate 67

C. CRANIOSPINAL VOLUME PRESSURE RELATIONSHIP 68

III. HATERIALS AND NETBODS . . . . . . . . . . . . 73
A. ANESTHESIA . . . . . . . . . . . . . . . 73

B. SURGICAL AND EXPERIMENTAL PROCEDURES . . 73

C. BRAIN VENTRICULAR AND CISTERNAL PUNCTURES 75

iv

CHAPTER Page
III. NATERIALS AND IETEODS (Canto)

D. VENTRICULOCISTERNAL PERFUSION EXPERIMENTS 76

1. Protocol A: High Pressure Perfusions 77

2. Protocol B: StOp-Flow Perfusions . . 80

3. Protocol C: Phenoxybenzamine
PerfUSions O O O I O O O O O O O O O 80

E. DETERMINATION OF PERFUSION FLOW
CONCENTRATIONS AND RATES . . . . . . . . 81

F. MATHEMATICAL ANALYSES AND CALCULATIONS . 81
l. Steady State Analysis . . . . . . . . 81
2. Transient Response Analysis . . . . . 83
G. STATISTICAL ANALYSES . . . . . . . . . . 87
Iv. RESULTS 0 O O O O O O O O O O O I O O O O O 0 8 9
A. PROTOCOL A: EFFECTS OF HIGH CSF PRESSURE
ON CSF STEADY STATE AND TRANSIENT
RESPONSES TO HYPERCAPNIA . . . . . . . . 93
B. PROTOCOL B: EFFECTS OF STOP-FLOW ON CSF
STEADY STATE AND TRANSIENT RESPONSES TO
HYPERCAPNIA O O O O O O O O O O O O O O O 9 9
C. PROTOCOL C: EFFECTS OF PHENOXYBENZAMINE
ON CSF STEADY STATE AND TRANSIENT
RESPONSES TO HYPERCAPNIA . . . . . . . . 106
v. DISCUSSION 0 O O O O O O O O O O O O O O O O 1 1 5

VI 0 meUSIONS O O O O O O O O O O O O O O O O O l 4 0

APPENDICES
A. DEFINITION OF TERMS O O O O O O O O O O O O
B. PREPARATION OF DIAL-URETHANE ANESTHETIC
SOLUTION O O C O C O O O O O O O C O O O O
C. ARTERIAL pH AND PC02 MEASUREMENTS . . . . .
D. RESPIRATORY MEASUREMENTS AND CALCULATIONS .
E. ANALYSIS OF CARBON DIOXIDE IN ALVEOLAR AND
MIXED EXPIRED GAS C C I O O O C C C O O O O
F. COMPOSITION AND PREPARATION OF ARTIFICIAL
CAT CEREBROSPINAL FLUID (CSF) . . . . . . .
G. DETERMINATION OF INFLOW AND OUTFLOW
PERFUS ION RATES O O O O O O C O O O O O O O
H. BETA RADIATION COUNTING . . . . . . . . . .

BIBLImRAPBY O O O O O O O O O O O O O O O O O O O 0

vi

Page

141

143
144

147

150

154

156

158

163

LIST OF TABLES

Steady State Respiratory and Cardiovascular
Data (Protocol A) . . . . . . . . . . . . .

Steady State CSF Data (Protocol A) . . . .
Transient CSF Outflow Data (Protocol A) . .

Steady State Respiratory and Cardiovascular
Data (Protocol B) . . . . . . . . . . . . .

Steady State CSF Data (Protocol B) . . . .
Transient CSF Outflow Data (Protocol B) . .

Steady State Respiratory and Cardiovascular
Data (Protocol C) . . . . . . . . . . . . .

Steady State CSF Data (Protocol C) . . . .

Transient CSF Outflow Data (Protocol C) . .

vii

Page

94
95

100

101
103

107

108
110

114

LIST OF FIGURES

FIGURE Page
1. Schematic of Ventriculocisternal Perfusion
Experimental Apparatus . . . . . . . . . . . . . 79
2. Representative Time Courses for CSF Transient
Responses to Hypercapnia . . . . . . . . . . . . 85
3. Data From a Representative Ventriculocisternal
PerfUSion O O O O O O O O O O O O O O O O O O O 92
4. Results of a Representative High CSF Pressure
Ventriculocisternal Perfusion (Protocol A) . . . 98
5. Results of a Representative StOp-Flow Perfusion
(PrOtOCOI B) o o o o o o o o o o o o o o o o o o 105
6. Results of a Representative Phenoxybenzamine
Ventriculocisternal Perfusion (Protocol C) . . . 112
7. CSF Flow Patterns 0 O O O O O O I O O O O O O O 123
8. Data From Prolonged Step-Flow Perfusion . . . . 126

viii

INTRODUCTION

Cerebrospinal fluid (CSF) comprises all the fluid in
the brain's ventricles and in the subarachnoid spaces
surrounding both the brain and spinal cord. Nearly 70% of
this fluid is formed from blood flowing through the
ventricular choroid plexuses (Katzman and Pappius, 1973)
which are a single layer of epithelial cells on a basement
membrane and a highly vascularized stroma of connective
tissue. The remaining 30% of CSF is formed by less well-
defined intracranial extrachoroidal sites (Bering, 1959;
Welch, 1963; Pollay and Curl, 1967). The primary process
for CSF synthesis is the active movement of sodium from the
interstitium across the epithelium into the ventricle
(Wright, 1978; Davson, 1967; Pollay, 1975) which occurs in
two steps; First, sodium is filtered through the Choroidal
capillaries and diffuses across the basolateral membranes of
the Choroidal epithelial cells down its electrochemical
gradient. Second, sodium is pumped out into the ventricle
by an energy dependent sodium-potassium exchange pump on the
apical surface of the epithelial cells. Chloride and
bicarbonate are also secreted into the ventricles and
potassium is absorbed from CSF back into the blood. Water

movement is coupled to the active salt secretion by osmotic

2
obligation (Wright, 1978). This newly formed Choroidal
fluid is slightly alkaline compared to blood, and has a
higher sodium concentration and a lower potassium
concentration than expected in a simple ultrafiltrate of
serum (Davson, 1956; De Rougemont et a1., 1960). It mixes
with the CSF within the ventricles and circulates along a
tortuous path through the ventricular cavities and into the
subarachnoid spaces. During its transit, CSF is continually
exposed to energy-requiring transport processes. Further
modifications are made in its composition (de Rougemont et
a1., 1960), which is insensitive to wide fluctuations in
electrolyte concentrations in blood (Goldstein et a1., 1979;
Reed and Yen, 1978).

Once in the subarachnoid space, bulk absorption of CSF
into venous blood occurs by the cranial and spinal arachnoid
villi (Pollay and Welch, 1962; Hammerstad et a1., 1969;
Davson et a1., 1970; Marmarou et a1., 1975). These absorp-
tive ports are evaginations of the subarachnoid space into
the lumen of the large dural sinuses, and operate like one-
way valves. When CSF pressure exceeds that in the venous
sinuses, the channels are distended, particles up to 7 u.in
diameter pass freely (Welch and Friedman, 1960), and CSF is
absorbed into the blood. When CSF pressure falls below
dural sinus pressure, however, the pressure gradient is
reversed, the villi collapse and CSF absorption ceases.
Colloid osmotic pressure gradients between CSF and blood,

unlike hydrostatic pressure gradients, have little effect on

3
CSF absorption (Pollay and Welch, 1962; Davson et a1.,
1970).

How CSF secretion is controlled remains unknown despite
years of study. Steady state CSF formation rate is reported
to increase (Ames et a1., 1965), decrease (Martins et a1.,
1976), or remain unchanged (Oppelt et a1., 1963) during
bouts of respiratory acidosis. Also, there are large dif-
ferences in CSF production rate within a species despite
control of such influential factors as blood pressure (Carey
and Vela, 1974), body temperature (Snodgrass and Lorenzo,
1972), and osmotic gradients between blood, brain and CSF
(DiMattio et a1., 1975; Wald et a1., 1976). Moreover,
autonomic neural activity influences CSF synthesis
(Edvinsson et a1., 1971; Edvinsson et a1., 1972; Lindvall et
a1., 1978a; Lindvall et a1., 1978b; Lindvall et a1., 1979).
Also, Choroidal epithelium is innervated by cholinergic
fibers (Edvinsson et a1., 1973; Lindvall et a1., 1978) and
by adrenergic nerve fibers originating from the superior
cervical ganglia (Edvinsson, 1977) and from other less well-
defined sources (Lindvall et a1., 1978). These sympathetic
nerve terminals also innervate vascular resistance vessels
(Edvinsson et a1., 1975; Nakamura and Milhorat, 1978), so
that reported changes in CSF formation rate after denerva-
tion or electrical stimulation of cervical sympathetic
ganglia (Lindvall et a1., 1978) may be the result of either
direct effects on the secretory epithelium (Edvinsson et

a1., 1975; Lindvall et a1., 1982) and/or changes in blood

4
perfusion of the choroid plexuses (Lindvall and mean,
1981).

Other sources of variation for CSF production were
first identified by Martins and co-workers (1977). They are
related to the method by which CSF formation rate is
measured, and to the anatomy and physiology of the CSF
spaces. Ventriculocisternal perfusion is a precise tech-
nique for estimating rates of bulk fluid production and
reabsorption of CSF, yet it is not error-free. Multi-
chambered and convoluted CSF spaces inhibit rapid and
complete mixing of perfusion fluid with the entire CSF
volume. Since craniospinal CSF and blood volumes are in a
dynamic equilibrium, pools of CSF unequilibrated with per-
fusion fluid may be mobilized when craniospinal blood volume
changes. This may explain why CSF formation rate appears to
increase transiently during acute respiratory acidosis
(Tschirgi et a1., 1954; Songer et a1., 1980; Fisher et a1.,
1983). CSF formation rate may also increase in response to
C02 breathing. Equations which assess CSF dynamics during
ventriculocisternal perfusion are accurate only if cranio-
spinal fluid volumes are in a steady state. (Heisey et a1.,
1983; Fisher et a1., 1983).

The present study was designed to define phenomena
underlying the apparent transient increase in CSF formation
rate during acute hypercapnia using a modified ventriculo-
cisternal perfusion method. It was also designed to

investigate the influence of alpha adrenergic blockade on

5
CSF transient phenomena and on steady state rates of CSF

formation and absorption.

LITERATURE REVIEW

The soft, gelatinous brain and the slightly firmer
spinal cord are rigidly confined within the skull and
vertebral column, while cushioned in a jacket of secreted
clear cerebrospinal fluid (CSF). The central nervous system
(CNS) is further supported and protected from mechanical
trauma by two connective tissue layers which lie between the
honey encasements and the nervous tissue mass; the firm dura
mater externally, and the delicate, thin pia arachnoid

layer which adheres to the brain and spinal cord.

CEBEBBQSRINAL_MENIN§ES

Dura_nater

The dura mater is a dense, thick layer of collagenous
connective tissue. The spinal dura is tubular, pierced by
spinal nerve roots, and, in humans, extends from the foramen
magnum to the second sacral segment. It is separated from
the wall of the spinal canal by an epidural space which
contains adipose tissue and a venous plexus. The veins
which drain the spinal cord empty into the epidural venous
plexus, which in turn drains into an external vertebral
plexus via conduits in the intervertebral foramina. The
cranial dura mater is firmly affixed to the periosteum of

6

7
the cranial vault, except where it is reflected to form
dural septae. The spaces formed between the septa and the
periosteum.accommodate»the endothelial-lined dural venous
sinuses. The external and internal cerebral veins, which
lie in the subarachnoid space on all surfaces of the
hemispheres and in the central core of the cerebrum, respec-
tively, empty into the venous sinuses through which blood
reaches the internal jugular veins. In some non-human
species, the internal jugular veins are small and venous
drainage from the brain mixes with blood from other

structures.

Pia-:Arachnoid

The delicate arachnoid layer is held to the dura by the
surface tension of a thin layer of fluid between the two
membranes, while the thin pia mater adheres to the surface
of the brain and spinal cord following all their contours.
These pia-arachnoid layers constitute the leptomeninges
("slender membranes"). They originate from a single layer
of mesodermal tissue, as attested by the rich trabecular
meshwork passing between them; both surfaces of the
arachnoid and the outer surface of the pia are covered by
mesenchymal epithelium. The subarachnoid spaces within
these two layers are filled with CSF and are narrow over the
vertices of the gyri, but wide at the base of the brain
(called the cisterna magna) and in the lumbar region of the

spinal canal.

8

While the arachnoid is avascular, the pia mater
contains many fine blood vessels. Large blood vessels go to
the brain surface through the subarachnoid spaces before
penetrating into the parenchyma, the bulk of the brain
tissue, taking with them a sleeve of pia mater. Within the
sleeve, the subarachnoid space and its contained CSF extend
into the Virchow-Robin spaces, becoming progressively
attenuated as the vessels become increasingly smaller, until
finally the perivascular spaces are only a fraction of a
micron long around the capillaries. The subarachnoid space
also extends into and communicates with the subdural space
through the arachnoid villi, which are labyrinths of coapted
tubes that connect with each other and allow open
communication between the subarachnoid and venous channels

of the dura.

W11. The arachnoid villi, which may be the
main site of CSF absorption into venous blood (Welch, 1975;
Marmarou et a1., 1975), are composed of a thin cellular
layer which is derived from arachnoid mesenchymal epithelium
and the endothelium of the sinus. Its cells have a sparse
amount of granular cytOplasm which surrounds round or oval
nuclei. As the cytoplasm extends from the nucleus it
becomes flattened and forms a delicate tubular wall. The
tubular projections enclose an extension of the subarachnoid
space containing trabeculae. The arachnoid tubes form a

communication channel between the subarachnoid space and

9
venous sinus (Welch and Friedman, 1960; Pollay and Welch,
1962).

Descriptions of the architecture of the arachnoid villi
vary depending on how it was examined. ‘Welch and Friedman
(1960) fixed arachnoid villi from the African green monkey
by arterial or venous perfusion of formalin, and then
performed serial sections after the tissue had been
dissected and embedded in paraffin. They found that if the
fixation occurred during low cerebrospinal fluid pressure,
brought about by cisternal drainage, the normal tubular
structure of the arachnoid villi collapsed. The diameters
of the tubular villi decreased, and the villi cells were
less flattened and bulged into the tubular lumen.
Conversely, they found that if the villi.were fixed by
perfusion of formalin into the subarachnoid space under
unphysiologically high pressures, the villi distended and
deformed, while the cytoplasm of the cells was greatly
attenuated. Under normal CSF and subarachnoid pressures,
the villi were found to be tubular, i.e., "open“ but not
distended, and communication was free between the

subarachnoid and dural spaces.

CEBEBRAL.QIBQULAIIQN

WW
Cephalig_brt§rigs, Except for a minor contribution
from the anterior spinal artery to the medulla, blood is

carried to the brain via two routes: the common carotid and

10

the vertebral arteries. The common carotid artery divides
into the external and internal carotid arteries. 'The
external carotid artery gives off several branches to supply
blood to the extracranial tissues. The distal part of the
internal carotid artery or the beginning of the middle
cerebral artery (a collateral branch of the internal
carotid) gives rise to the anterior Choroidal artery, which
has a rather long subarachnoid course and which supplies
blood to the lateral ventricular choroid plexuses as well as
some other cerebral structures. In the adult cat, however,
the internal carotid artery is vestigial and the brain is
supplied with blood through the external carotid, vertebral
and pharyngeal arteries by way of the internal maxillary
artery and the carotid plexus (Crouch, 1949).

The vertebral arteries unite to form the basilar artery
on the ventral side of the spinal cord. This artery gives
off branches, including the posterior cerebral artery and
the posterior choroidal artery, which supplies the choroid
plexuses of a small portion of the rostral end of the
lateral ventricle and the third ventricle. The posterior
cerebral artery finally joins the posterior part of the
Circle of Willis, which is a network of communicating
arteries that provides blood to the ventral side of the
brain in the region of the pituitary. The blood is further
dispersed from these arteries to the other extracerebral bed

(pial vessels), which forms a reticulation on the outer

11
surface of the brain, and then proceeds to the intracerebral

vessels (in the brain parenchyma).

W. The pial vessels, which

lie in the subarachnoid space, are bounded on their outer
surface by a thin adventitial layer which is composed of a
layer of loose connective tissue intermingled with a dense
plexus of adrenergic and cholinergic nerves. A basement
membrane surrounds this layer and confines the adventitial
space from the perivascular space. Toward the center, the
adventitial space is adjoined to a muscular cell layer, the
tunica media, which, in turn, adjoins an innermost single
layer of endothelial cells, the tunica intima. The pial
vessels distribute blood to the vessels which penetrate the

cerebral cortex, the intracerebral or parenchymal vessels.

W. The parenchymal

vessels, which have a short extracerebral course before
penetrating the brain parenchyma, become smaller as they
move inward from the brain surface. There are three main
types: arterioles, capillaries, and venules (Cervos-Navarro
and Matakas, 1973). These vessels have a number of unique
anatomic features which distinguish them from the peripheral
microvasculature and which are considered to constitute the
morphological blood-brain barrier (mean and Hardebo, 1982).
In most regions of the brain, the vascular endothelial cells
are bounded by a basement membrane, are non-fenestrated, are

fused to each other by tight junctions (i.e., intercellular

12
clefts are sealed shut), and demonstrate a paucity of
pinocytotic vesicles in their cytoplasm (Oldendorf, 1975;
Edvinsson, 1975); vesicular transport is presumed to be
virtually absent (Oldendorf, 1975). Further, the brain
capillaries are sometimes ensheathed by the membrane of
astrocytic end-feet (pericytes) which are apposed to the
capillary basement membrane. Whether or not a molecule in
solution in plasma will penetrate the capillary cell wall
(traverse the blood-brain barrier) depends on its oil/water
partition coefficient, and its relative affinity for plasma
and membrane carrier proteins (Oldendorf, 1975).

The integrity of the blood-brain barrier is enhanced by
special enzymatic systems found in the cytoplasm of brain
capillary endothelial cells (mean and Hardebo, 1982).
Specifically, monoamine oxidase (MAO), catechol-o-
methyltransferase (CONT), DOPA-decarboxylase, as well as
other enzymes, show considerable activity in brain
capillaries (Hardebo et a1., 1980). Monoamines such as
dopamine, norepinephrine, epinephrine and S-hydroxytrypta-
mine, and other less well-known neurotransmitters may be
deaminated in the walls of brain vessels (Hardebo and mean,

1982).

Choroid_£lexnses, The choroid plexuses, a special part
of the pial vascular system, also have MAO activity, although
the bulk of the enzyme activity is present in choroidal
epithelial cells rather than in the endothelial lining of
the choroidal capillaries (Lindvall and mean, 1980). ‘The

l3
plexuses in each of the lateral ventricles and in both the
third and fourth ventricles are richly vascularized
epithelial tissue formed by an invagination of the pia mater
(tela choroidea). The lateral ventricle is formed by an
invagination of pia mater on the medial surface of the
cerebral hemisphere, while the plexuses of the third and
fourth ventricles are similarly formed by invagination of
pia mater attached to the roofs of these ventricles. During
the invagination process, the connective tissue acquires a
covering of epithelium from the ependymal lining of the
ventricles. Three distinct elements comprise the choroid
plexus: (l) epithelium of ependymal origin which has
secretory characteristics; (2) a layer of leptomeningeal
tissue which is ectodermal in origin and is characterized by
large aqueous spaces, consistent with the suggestion that
these spaces are fundamentally extentions of the
subarachnoid space (Maxwell and Pease, 1956); and (3) a
dense vascular stroma.

The choroidal vascular stroma contains capillaries
which are anatomically'distinct from.most intracerebral
capillaries. 'The thin endothelium, which rests on basement
membrane like that of the peritubular capillaries of the
kidney (Maxwell and Pease, 1956), has "leaky-tight“
junctions and fenestrations or pores closed by thin
diaphragms (Wright, 1979). These increase the permeability
in these capillaries more than is normally found in

capillaries which supply nervous tissue. In fact, there is

14
free exchange of elements between the blood and the
interstitial fluid bathing the basolateral surface of the
choroidal epithelium. The morphological blood-brain barrier
is lacking (Wright, 1979).

Cerebrospinal fluid (CSF) is ultimately derived from
blood flowing through these choroidal capillaries.
Electrolytes and water cross the capillary endothelial
fenestrations and enter the extracellular fluid space of the
choroidal epithelial cells where they are selectively'and
actively secreted into the ventricular cavities (Wright,“
1979). However, choroidal blood flow, of which 25% of the
plasma volume becomes CSF (Cserr, 1971), is nearly twelve
times higher than cortical blood flow in normoxic,
normotensive sheep (Page et a1., 1980) and can be reduced as
much as 50% without affecting CSF formation rate (Ff)

(Nakamura and Hochwald, 1983).

W

The cerebral hemispheres are drained by both an
external and an internal venous system. The external
cerebral veins course through the subarachnoid space and lie
on all surfaces of the cerebral hemispheres. The internal
cerebral veins which are located beneath the corpus callosum
in the transverse fissure drain the central core of the

cerebrum.

W. The superior cerebral veins

and the superficial middle cerebral vein, which are external

15
cerebral veins, course upward over the dorsolateral surface
of the hemispheres, and downward and forward along the
sylvian fissure, respectively. They pierce the arachnoid
and dura mater, and empty into the superior sagittal,
cavernous, and transverse sinuses, and the lateral lacunae
adjacent to the superior sagittal sinus.

The basal vein, formed by the union of the deep middle
and anterior cerebral veins, runs backward along the base of
the brain and eventually empties into the Great Cerebral
Vein of Galen. This latter vein, which is considered to be
an internal cerebral vein, drains into the straight sinus.
Many more external veins, in addition to the few mentioned,
drain circumscribed regions of the brain and channel into

adjacent dural sinuses.

W. The internal venous system
forms beneath the floor of the lateral ventricles and
continues along under the corpus callosum. The choroidal
vein, a rather tortuous vessel, passes along the choroid
plexus of the lateral ventricle and drains the plexus as
well as a few neighboring structures. This vein unites with
the terminal vein, which drains the corpus striatum,
internal capsule, fornix, and septum pellucidum, to form the
internal cerebral vein. The paired internal cerebral veins
course posteriorly in the tela choroidea and unite beneath
the corpus callosum to form the Great Cerebral Vein of

Galen.

16

The cerebral venous system has larger radii and thinner
walls, and is more compliant than cerebral arteries
(Marmarou et a1., 1975). The cerebral blood volume,
determined primarily by the volume of blood contained in the
compliant cerebral venous system (Rich et a1., 1953), exists
in a dynamic equilibrium with the cranial CSF volume.
Sudden changes in one of these volumes elicits a rapid
reciprocal change in the other fluid volume (Ryder et a1.,
1953). Since the cerebral veins, in particular the external
cerebral veins contained in the CSF—filled subarachnoid
space, are highly compliant compared to the cerebral
arteries (Marmarou et a1., 1975) and brain (Sklar et a1.,
1977), intracranial pressure changes induced by sudden
changes in cranial CSF or blood volume are damped primarily
by the rapid reciprocal interaction of these two cranial
fluid volumes (Marmarou et a1., 1975; Weed et a1., 1932;,

Weed et a1., 1933; Ryder et a1., 1953).

W
Cephalie_!£ssels. Nerve fibers were first described on

the anterior and posterior cerebral arteries by Willis in
1664 (cited by Edvinsson, 1975) using silver-impregnated
tissue sections. Confirming studies have traced the origin
of many of these fibers to the superior cervical ganglion

(Edvinsson and MacKenzie, 1975).

W. Pial vessels of many species of animals

(Nielsen et a1., 1971), including the cat (Edvinsson et a1.,

17
1972), have well-developed plexi of sympathetic adrenergic
nerves and cholinergic nerves, as well as some
tryptaminergic and peptidergic fibers (mean and Edvinsson,
1977). The sympathetic innervation, which arises from the
superior cervical ganglia (Edvinsson, 1975) is comparable in
density, distribution, and organization to that found in
other richly innervated peripheral vascular beds (Nielsen et
a1., 1971; Edvinsson and MacKenzie, 1977; Edvinsson, 1975).
Electron.microscopy studies have revealed that the
sympathetic nerves are less dense in cerebral veins,
compared to cerebral arteries (Edvinsson and MacKenzie,
1977), and in the phylogenetically older regions of the
brain (the posterior part of the cerebrum and brain stem),
as compared to the younger parts of the brain (neocortex)
(Edvinsson, 1975).

The extent and distribution of cholinergic nerves to
the pial vessels is similar to that of the adrenergic supply
(Edvinsson et a1., 1972). In fact, cholinergic and
adrenergic nerves frequently run together surrounded by the
same Schwann cell, terminate in juxtaposition, and form a
typical ground plexus in the adventitia immediately adjacent
to the muscular layer of the media. This provides an
opportunity for an interactive mechanism between the two
types of axon terminals (Nielsen et a1., 1971). The origin
of the cholinergic fibers, however, is unknown. It has been
suggested that the majority of cholinergic fibers track in

the facial nerve, course off in the region of the geniculate

18
ganglion, and continue in the greater superficial petrosal
nerve to the plexus of the internal carotid artery where
they form ganglia and project to the pial vessels
(Edvinsson, 1975).

The predominant adrenergic receptor on pial vessels is
the alpha-adrenoceptor, mediating vasoconstriction.
Noradrenaline induces dose-dependent vasoconstrictor
responses in isolated pial vessels which can be reduced or
blocked by specific alpha-receptor blockers, phentolamine or
phenoxybenzamine (Edvinsson and mean, 1974). The cerebral
vessel alpha-adrenoceptors, however, may be distinct from
the alpha-adrenoceptors associated with peripheral arteries.
The cerebral receptors are only weakly sensitive to
noradrenaline and phenylephrine, and demonstrate an atypical
biphasic response in the dose-response curve for
noradrenaline (Duckles and Bevan, 1976). The weak
sensitivity of pial vessels to noradrenaline may be due
either to a limited capacity to introduce extracellular
calcium and/or to mobilize intracellular calcium, or to an
increased activity of the neurotransmitter uptake system in
the sympathetic nerve endings (Marin and Rivilla, 1982).

Beta-adrenoceptors have also been identified in pial
vessels and have been implicated in the vasodilation seen
when isoproterenol is introduced into isolated cerebral
arteries pre-treated with serotonin or prostaglandins;
propranolol (10"8 - 10'7 M) blocks the vasodilation

(Edvinsson and mean, 1974; Sercombe et a1., 1975).

19

25W. Like pial vessels, parenchymal
vessels, including intracerebral capillaries and their
associated pericytes (Rennels and Nelson, 1975), are supplied
with adrenergic nerves (Kawamurs et a1., 1972; Cervos-
Navarro and Matakas, 1973). These nerves, however, may
arise from a source other than the peripheral autonomic
ganglia.

In 1972, Hartman and Udenfriend suggested that small
blood vessels are innervated by a central adrenergic system,
which is part of an intra-axial autonomic system that is
isolated by the blood-brain barrier (Raichle et a1., 1975)
and distinct from the peripheral sympathetic system, but
having analogous functions. Their hypothesis was based on
the immunohistochemical demonstration that bilateral
cervical ganglionectomy in rats did not disrupt the close
association between dOpaminejd-hydroxylase (DBH)-containing
fibers and capillaries found in normal rats. They suggested
that since there are no (DBH)-containing cell bodies rostral
to the pons, the vascular innervation in the rostral brain
arrives via ascending noradrenergic fiber pathways from the
locus coeruleus and other clusters of noradrenergic cell
bodies in the lower brainstem.

Subsequently, several studies investigating the
influence of the central adrenergic system on cerebral
circulation have been conducted. Raichle et a1. chemically'
(1975) and electrically (1976) stimulated the locus

coeruleus, a major cell group of the central adrenergic

20
system (Hartman and Udenfriend, 1972), and measured the
effects on capillary permeability and regional cerebral
blood flow. When stimulating the locus coeruleus either
electrically'(l976) or by microinjections of carbachol
(1975), there was a marked and rapid increase in brain
capillary permeability to water and a reduction in blood
flow; intraventricular administration of phentolamine, a
rapidly acting alpha adrenergic blocker, produced the
opposite effect (1975).

Preskorn and associates (1978) demonstrated similar
phenomena by measuring the extraction fraction of a
diffusion-limited test tracer both before and after the
acute intraperitoneal administration of amitriptyline (AMI)
to rodents and monkeys. AMI, a tricyclic antidepressant,
acts indirectly as an adrenergic agonist by blocking
norepinephrine reuptake. They found that after AMI
treatment, the extraction fraction of both water and ethanol
increased; pre-treatment with phenoxybenzamine or
intraventricular administration of 6-hydroxyd0pamine blocked
the effects. In a later study, Preskorn et a1. (1982)
provided further evidence that the AMI-induced increase in
the extraction fraction of water is mediated by the effect
of the drug on central adrenergic neurons rather than
through the known serotonergic, anticholinergic, and
antihistaminergic actions of AMI, by themselves: central

serotonergic abalation by p-chloroamphetamine, and treatment

21
with atropine and hydroxyzine separately did not alter the
extraction fraction of water.

These data imply a neural control of the blood-brain
barrier at the level of the capillary (Preskorn et a1.,
1978; Grubb and Raichle, 1981; Preskorn et a1., 1981;
Preskorn et a1., 1982). Further evidence for central
adrenergic effects on the cerebral microcirculation was
provided by immunohistochemical demonstration of actomyosin
in pericytes and endothelial cells in the brain capillary
wall (mean et a1., 1977).

The integrity of the blood-brain barrier may also be
affected by the peripheral sympathetic nervous system.
Sadoshima and co-workers (1982) showed that unilateral
superior cervical ganglionectomy in stroke-prone
spontaneously hypertensive rats (SHRSP) increases the
permeability of the blood-brain barrier to the Evans blue-
albumin complex and/or increases the incidence of cerebral
infarction or hemorrhage in the denervated hemisphere. They
concluded that the peripheral sympathetic nervous system
protects the blood-brain barrier during chronic
hypertension, and that disruption of the blood-brain barrier
may play a role in the pathogenesis of cerebral infarction
by allowing vasoactive substances to enter the brain and
produce focal areas of cerebral ischemia. Sadoshima et a1.
speculated that the mechanisms of this protective effect may
be through a trophic effect of sympathetic nerves which may

accentuate cerebral vascular wall thickening (hypertrOphy)

22
during chronic hypertension. Hypertrophy of large cerebral
arteries may augment the structural contribution to cerebral
resistance, reduce pressure in small cerebral vessels, and
therefore protect the cerebral microcirculation. Further,
hypertrOphy of small resistance vessels may reduce the wall
stress of these vessels.

Hardebo et al. (1977a) extended the postulated role of
the peripheral sympathetic nervous system in protecting the
cerebral circulation to include the modification of cerebral
autoregulatory limits during hypertensive episodes. They
showed, as have others (Eklof et a1., 1971; Hernandez et a1.,
1971/72; Hernandez-Perez et a1., 1975; Waltz et a1., 1971;
Stone et alr, 1974), that superior cervical ganglionectomy
does not abolish autoregulation, but causes the denervated

hemisphere to autoregulate at a higher blood flow level.

QhQLQid_£1£anes, .Morphological evidence suggests

extrinsic neural regulation of choroidal secretion of CSF.
The choroid plexuses, like the pial vascular system, are
well innervated with autonomic nerves (Edvinsson et a1.,
1973; Edvinsson et a1., 1974; Edvinsson et al., 1975b;
Lindvall et al., 1978b; Edvinsson and Lindvall, 1978;
Nakmura and Milhorat, 1978). The axons enclose not only
the choroidal arteries, but also form a reticulation in the
choroid plexus and around the veins which drain the tissue
(Edvinsson and Lindvall, 1978L. Using a histo-fluorescence
method in combination with superior cervical ganglionectomy,

Edvinsson and associates (1974) demonstrated the existence

23

of nerve fibers showing a fluorescence typical for that of
catecholamine fluor0phores in the choroid plexuses of
various mammalian species. The nerve fibers formed networks
around both the choroidal arterial and venous vessels, while
others coursed between the base of the choroidal epithelial
cells and the underlying vascular wall. Within one week
after bilateral cervical sympathectomy, the nerve
fluorescence disappeared from the choroid plexuses of the
lateral and the third ventricles, while a small number of
fibers were still visible in the plexus of the fourth
ventricle (Lindvall et al., 1978b; Edvinsson et al., 1975a).

In a subsequent study, Edvinsson and co-workers (1975b)
found corroborative ultrastructural evidence of autonomic
innervation of the choroid plexus from fluorescent and
electron microscopy of cat and rabbit plexuses after
administration of S-hydroxydopamine. They found thick
bundles of nonmyelinated axons in the proximal portions of
the plexuses where they accompanied small arterioles, and
smaller axon bundles located contiguous to the walls of
small arterioles as well as to therbase of the epithelial
cells, between adjacent cells, or within cellular
invaginations. The distance between the membrane of the
perivascular nerves and the membrane of the smooth muscle
cells of the arterioles was less than 100 nm. ‘Axonal-
epithelial cell distances were sometimes as small as 30 nm.
The nerve terminals had no neurilemmal sheath but had an

abundance of electron-dense vesicles (indicative of

24
adrenergic terminals) or of electron-lucent vesicles
(suggestive of cholinergic terminals) of the same size.
They concluded that the sympathetic vasomotor innervation of
the choroid plexus resembles that of the pial and
intracerebral vascular beds, and that the close apposition
of nerve cells to plexus epithelia provides morphological
evidence for true autonomic innervation.

Histochemical studies of the choroid plexuses of
superior cervical ganglionectomized rabbits and cats
confirmed the presence of numerous acetylcholinesterase-
containing nerve terminals (Edvinsson et a1., 1973). The
conditions under which the cholinesterase staining developed
in the plexus preparations under study (i.e., preincubated
in 4 uM solution of the cholinesterase inhibitor Mipalox and
incubated for 4 to 6 hours in the presence of
acetylthiocholine) suggested that the cholinergic nerves
were parasympathetic rather than sensory. The cholinergic
nerve fibers were found to form a wide-meshed network around
most vessels supplying and draining the plexus, and to
ramify among epithelial cells without obvious relation to
the subjacent vascular system. Edvinsson and co-workers
concluded that the distribution of the cholinergic nerves
resembled that reported for the adrenergic system
(Edvinsson et a1., 1974) and that, in fact, the overall
density of nerve contribution to plexus tissue was even

greater for the cholinergic system.

25
CEBEBBQSEINAL_ELHID_1£5£1.

CSF is known to be produced by the choroid plexuses of
all four ventricles (de Rougemont et a1., 1960; Welch, 1963;
Ames et a1., 1965; Wright, 1978), the ependymal surface of
the brain (Pollay and Curl, 1967), and perhaps in the cranial
(Sato and Bering, 1967) but not the spinal subarachnoid
spaces (Hamer and Sahar, 1978; Lux and Fenstermacher, 1975).
In experimental animals, 55 to 65% of the CSF formed within
the ventricles is reportedly produced by the choroid plexus
in the fourth ventricle (Wright, 1978).

CSF, which is slightly hypertonic to plasma (Wright,
1979 ; Cserr, 1971), flows presumably along regional
pressure gradients established by fluid secretion from the
lateral ventricles through the third ventricle to the fourth
ventricle where it enters subarachnoid cisterns surrounding
the brain via the foramina of Luschka and Magendie. It then
flows over the brain convexities and down the spinal cord
where it is absorbed through the cranial arachnoid villi and
spinal nerve roots, respectively'(Pollay and Welch, 1962;
Hammerstad et a1., 1969; Davson et a1., 1970; Marmarou et
a1., 1975).

ESE: 'I‘ in III
Mechanism

CSF closely resembles an ultrafiltrate of plasma,
however, both anatomical and functional evidence clearly
indicate that CSF production is an active mechanism and not

a simple ultrafiltration process. Several characteristics

26
of the choroid plexus epithelial cell seen in electron
micrographs suggest that the production of CSF is at least
in part an energy-requiring active process. There is a
large spherical nucleus, numerous and randomly arranged
mitochondria, and a slightly polarized system of vesicles
(representing end0plasmic reticulum) in an abundant
cytoplasm; the vesicles are both more numerous and often
large near the apex of the cell (Maxwell and Pease, 1956).
Further, the choroidal epithelial plasma membrane, which has
a low hydraulic conductivity (Wright, 1978), has the lateral
aspects of its apical surface tightly adjoined to adjacent
cells (Wright, 1979) forming a barrier to solutes between
the CSF and the cellular interstitial spaces; filtration
cannot play a significant role in the passage of fluid
across this layer. The apical surfaces of the epithelial
cells are lavishly folded, as.are the basal surfaces, and
specialized into a fairly well-developed brush border
(Maxwell and Pease, 1956). ‘This folding vastly increases
surface area, and is common to epithelial cells involved in
active secretory processes.

Chemical analyses of CSF support the idea that CSF
elaboration involves active processes. In 1960, de
Rougemont et a1. published the first of a series of papers
in which they analyzed newly formed choroidal fluid, and
compared it to that of serum and to that of CSF obtained
from sites progressively further along its circulation

course (ldh, cisternal magna and cisterna pericallosa).

27
Their technique for sampling from the choroid plexus
involved the surgical exposure and subsequent filling of the
lateral ventricles of a cat with PantOpaque warmed to 37°.
Since the choroid plexus and its secretions have lower
specific gravities than PantOpaque, they rose in the pool of
oil; permitting their collection of newly formed CSF
droplets with a fine glass pipette. This fluid had a lower
potassium and a higher concentration of sodium and chloride
than expected in a simple ultrafiltrate of serum. Further,
although there are no detectable differences in Na+
concentration among samples from the ventricular and
cisternal collection sites, fluid collected from the
cisternae contained even higher chloride than plexus fluid,
and potassium concentration progressively declines from the
choroid plexus to cisterna magna to the cisterna
pericallosa. de Rougemont and associates interpreted these
findings to indicate that CSF is actively elaborated from
serum by transport processes located at the site of its bulk
formation, and that CSF continues to be exposed to the action
of energy-requiring transport processes as it circulates
along a course between the ventricles and the subarachnoid
spaces. These active mechanisms produce further
modifications in CSF in the same direction as those
initiated during its choroidal formation process. They
postulated that extrachoroidal transport mechanisms reside
in cells that have a similar embryological origin to that of

the choroid plexus.

28

Ames et a1. (1964) also examined the composition of CSF
collected from the exposed ventricular choroid plexus to
that which had contacted the ependyma and pia mater. Their
data agreed with those published by de Rougemont: CSF
collected from the choroid plexus and the cisterna magna
have concentration differences in four ions (chloride,
magnesium, calcium and potassium) while sodium concentration
remained constant between the two sites. Chloride,
magnesium, and calcium concentrations increase as CSF
traverses the ventricular system, while the potassium
concentration declines. Ames et al. also estimated
bicarbonate concentrations based on charge balance and
determined that choroidal bicarbonate is greater than
cisternal bicarbonate concentration.

These regional differences in concentrations of
potassium, magnesium, and calcium in CSF stay constant
despite wide fluctuations in plasma electrolyte composition
(Goldstein et a1., 1979; Reed and Yen, 1978). Ames et a1.
(1965) varied plasma potassium concentration from 1.59 to
11.08 mM/kg H20 and found that CSF potassium changed less
than 2 mM. Reed and Yen (1978) investigated the regulation
of CSF magnesium concentration using the cat choroid plexus
isolated in a chamber in_sitn, and varied the magnesium
concentration from 0 to 4.8 mEq/liter in the plasma and/or
the solution bathing the plexus. They observed that chamber
magnesium concentrations were rapidly returned to the

control value of 1.8 mEq/liter, and concluded that the

29
choroid plexus is directly involved in maintaining the
constancy of CSF magnesium concentration by sensing changes
in the normal CSF magnesium concentration and altering the
rate of active magnesium secretion.

Subsequently, Goldstein et a1. (1979) arrived at
similar conclusions about the role of the choroid plexus in
CSF calcium regulation. They simultaneously measured total
and ionized calcium and inorganic phosphorous in blood and
CSF in human subjects and dogs with and without parathyroid
glands, in dogs which had been treated three days with
either parathyroid extract or phosphate infusion, or in dogs
which were made and maintained hypercalcemic for 3 months
with oral administration of vitamin D. IDespite marked
variations in the blood concentrations of total calcium,
ionized calcium, inorganic phosphorous, and of the calcium-
phosphorous product, these concentrations and products
varied only slightly in CSF under all conditions tested.

These studies show that CSF is at least partly actively
secreted, and suggest that CSF is contiguous with the
immediate environment of neural cells, the brain
extracellular fluid. CSF electrolyte levels are actively
maintained constant providing a critial step in the overall
process which prevents changes in brain composition (Wright,

1978).

W

Considering the architecture of the choroidal

capillaries and the overlying epithelium, filtration may be

30
an important process in moving fluid from capillary lumen to
interstitial space of the plexus and a necessary prelude to
an active transepithelial secretion (Welch, 1975).

Pollay (1975) develOped a model which provided a basis
for the current understanding of the CSF production process.
He applied the standing gradient hypothesis of solute and
water transport (Diamond and Bossert, 1967) to the choroidal
ependyma. Considering the tight junctions connecting
adjacent choroidal ependymal cells as well as the large
fenestrated capillaries in the plexus, he proposed that the
first step in the formation of CSF is the transcapillary
movement of an ultrafiltrate of plasma under the influence
of a hydrostatic gradient. The filtrate, which is isotonic
to blood, would then traverse the basal foldings of the
choroidal cell, or between choroidal cells proximal to the
apical tight junctions; the large particles are filtered by
the basement membrane. Active transport of sodium from
these channels renders the local interstitial fluid
hypotonic as it approaches the blind end of the channel, and
a favorable osmotic gradient is established for water flow
into the cell. A similar mechanism would couple solute and
solvent.movement into the ventricular system through the
apical microvilli channels.

Pollay proposed that CSF synthesis involves both filtration
and the active transport of sodium as primary processes in
CSF secretion (Wright, 1979). The problem with his theory,

however, rests in the fact that it predicts basolateral

31
sodium pumping and explains water flux upon the osmotic
gradient induced by these pumps. Wright (1978) using
autoradiographic techniques studied the binding of 3H-
ouabain to its natural substrate, sodium—potassium adenosine
triphosphatase (Na/K-ATPase) which is now recognized as the
sodium pumping mechanism (Wright, 1978). He determined that
Na/K-ATPase, which represents one-fourth to one-third of the
total ATPase activity in the plexus (Lindvall et a1., 1982),
was largely confined to the apical cell membrane, and not to
the basolateral membrane as Pollay“s theory would require.
In fact, the ouabain-sensitive influx of sodium across the
serosal surface of the plexus has been demonstrated to be
only 7% of that across the apical plasma membrane (Wright,
1978). .Adenylate cyclase, an enzyme which converts
adenosine triphosphate (ATP) to adenosine 3', 5“-
monophosphate (cyclic AMP), is involved in the mediation of
both alpha— and beta-adrenergic effects (Exton, 1982;
Nathanson, 1979; Epstein et a1., 1977), and alkaline
phosphatase (Lindvall et a1., 1982) has also been localized
on the apical plasma membrane of choroidal epithelial cells.
The basolateral membranes, though relatively devoid of
sodium pumps, are not inactive. Guanylate cyclase, an
enzyme that converts guanine triphosphate (GTP) to guanosine
3', 5'-monophosphate (cyclic GMP), a compound which has been
shown to decrease the binding of agonists to cortical
alphaz—receptors (Exton, 1982), and glutamyl transpeptidase,

an enzyme active in transmembrane amino acid transport,

32
have been localized on the basolateral membrane CLindvall et
a1., 1982).

The entry of sodium into the epithelium across the
basolateral membrane does not need to be active. The
electrochemical potential gradient for sodium influx, in
fact, favors its passive movement into plexus cells (i.e.,
.‘fiua+ is approximately 150 mv) (Wright, 1978; Wright, 1979;
Welch, 19750. Once within the epithelium, however, sodium
is pumped out in the ventricle by Na/K-ATPase (Wright,
1979). The Na/K pump is ouabain-sensitive (Wright, 1978),
and is electrogenic. It contributes approximately 10 mv to
the electrical potential difference across the apical
membrane of the plexus; 3 sodium ions are pumped out of the
cell into the ventricles in exchange for two potassium ions.
Potassium accumulation within the plexus is against an
electrochemical gradient (approximately 57 mv) (Wright,
1979), and its leakage across the basolateral membrane
accounts for the observed net absorption of potassium from
CSF to blood (Wright, 1978; Wright, 1979).

The rate of sodium-potassium pumping is limited by the
entry of sodium into the epithelium across the basolateral
membrane. Sodium influx is accelerated by the presence of
an increased supply of intracellular protons which couple
with sodium in an antiport process, and of intracellular
buffer anions, especially'bicarbonate (Wright, 1978), and
chloride to accompany sodium transport out of the cell

across the apical membrane to maintain electroneutrality

33

(Curl and Pollay, 1967; Wright, 1978; Wright, 1979; Welch,
1975). Bicarbonate can enter the cell in the form of C02
where it is hydrated, a process catalyzed by carbonic
anhydrase, to form carbonic acid. Carbonic acid is an
unstable compound and forms bicarbonate and hydrogen ions.
Acetazolamide, a carbonic anhydrase inhibitor, causes a
decrease in CSF formation rate (Ames et a1., 1965), possibly
by limiting the number of intracellular protons to exchange
for sodium at the apical epithelial membrane (Wright, 1979).

With pumps only on the apical membrane, the link
between active salt transport and fluid secretion remained
an enigma. Wright (1977) isolated the fourth ventricular
choroid plexus of the frog, removed all the solution from
the ventricular surface of the plexus, covered it with oil,
and collected the newly secreted fluid from the tissue/oil
interface at regular intervals of time. He found that when
the serosal side of the frog choroid plexus was exposed to a
Ringer solution containing 25 mM sodium glycodiazine and
equilibrated with 100% 02 (viz., control conditions), the
nascent CSF was hypertonic to the serosal bathing medium,
due primarily to an increase in the sodium concentration of
the newly formed CSF. Further, Wright discovered that when
the osmolarity of the bathing solution was varied from 110
to 290 mOsm/liter by varying the sodium chloride
concentration, the osmolarity of the newly secreted CSF
remained hypertonic to the bathing solution by approximately

30 mOsm.

34

A local osmotic pressure gradient, established by
active sodium transport, may provide the link between solute
and water transport (Wright, 1975). The local gradient may
be perpetuated by the presence of a large “unstirred“ CSF
layer adjacent to the ventricular surface of the tissue
(Wright and Prather, 1970). Active salt transport across
the epithelium, then, increases the osmotic pressure in this
unstirred layer, thereby passively pulling water from the
serosal side of the cell. The exact mechanism of solute-
solvent coupling remains the subject of debate in this and

other epithelia specialized for water transport.

MethodsJLSthdx
Numerous in gitrg and in vigg methods have been used to

study choroidal and brain ependymal transport processes as
well as to estimate rates of CSF formation and absorption.
Principal methods for assessing CSF fluid dynamics include:
1) sampling of choroidal venous blood from in_aitn choroid
plexus (Welch, 1963); 2) extracorporeal perfusion of the
isolated plexus (Pollay et a1., 1972); 3) collection of
nascent fluid from the surface of exposed choroidal tissue
(deRougemont et a1., 1960); and 4) ventriculocisternal
perfusion (Pappenheimer et a1., 1962; Heisey et a1., 1962).
Of these, the ventriculocisternal perfusion system has been
the most widely used means of studying rates of CSF

formation (Vf) and absorption (Va) (Cserr, 1977).

35
1. Sampling of Choroidal Venous Blood from 1n_aitn

Choroid Plexus

In 1963, Welch developed an approach for determining

CSF formation rate which involves the estimation of the
blood flow through the choroid plexus and the loss of plasma
fluid volume during the transit of blood. He observed that
the vascularization of the rabbit}s lateral ventricular
choroid plexus favors the direct measurement of venous blood
flow, since the venous drainage route is chiefly through one
large vein, the internal cerebral vein. He surmised that if
red blood cells do not appreciably change in volume as they
pass through the choroid plexus, then the volume of red
blood cells entering and leaving the plexus during any

period are necessarily constant.

Hcta x Qa = Hctv x Qv (l-l)
where,
Hct = hematocrit of choroidal blood

Q = choroidal blood flow (ul/mg-min'l)

arterial

a

V venous

Welch assumed that the fluid loss from the blood in its
transit through the rabbit plexus, (i.ea, the difference
between choroidal arterial and venous flow)*would represent

CSF formation (Vf; ul/min).

6: = Oa - 0v (1-2)

36
Substituting Eq. 1-1 into Eq. 1—2 and rearranging, CSF
formation rate can be estimated if choroidal venous blood
flow (0v), and choroidal arterial (Beta) and venous (Hctv)
hematocrit are measured.

Hct
Hct

Vf = 0v [ v

 

- 1] (1-3)
a

By the use of an eyepiece micrometer, the diameter and
a length of the main choroidal vein were measured. A
micropipette filled with 1-octanol was introduced into the
vein and spherules of 1-octanol were injected. The linear
velocity of the spherules was determined using a
cinematographic technique from film projections on graph
paper. Blood flow (Ov) in the main choroidal vein was then

determined by:

Qv = (nr2 x l/t) - (ngfir3) (1-4)
where,
wrz = cross-sectional area of the vein (cmz)
l/t = the linear velocity (cm/min)

n a the number of spherules injected during one
unit of time (number/min)

gwr3 = volume of l-octanol spherules injected (cm3)

Venous blood from the plexus and blood from the aorta
were collected in capillary tubes and hematocrits determined
by centrifugation and microscopic inspection.

This method permits the in 111g measurement of

choroidal CSF formation rate, yet may have inherent sources

37
of error in at least one of its assumptions. Welch sampled
from the aorta to determine anterior choroid arterial
hematocrit;(Hcta)*when it is known that small cerebral
arteries have a smaller hematocrit than that of aortic blood
(Cserr, 1971); this would lead to an underestimation of

transependymal volume flow.

2. Extracorporeal Perfusion of the Isolated Choroid

Plexus

In 1972, Pollay and his associates determined that the
choroid plexus of sheep can be maintained satisfactorily in
an extracorporeal perfusion system for up to 7 hours. They
procured heads from freshly decapitated sheep, rapidly
entered the skull and removed the brain, being careful to
section the supraclinoid portion of the internal carotid
artery. 4All branches of the carotid artery were ligated,
except for the branch to the anterior choroidal artery, so
that arterial inflow to the plexus could be controlled by
carotid perfusion. Venous outflow from the plexus was
collected from a cannula introduced into the Great Vein of
Galen. The vascular perfusate consisted of donor sheep
blood which had been diluted with a salt solution containing
dextran (10% solution); osmolality'was adjusted with sucrose
while the hematocrit remained approximately one-half that
of whole blood. The fluid bathing the plexus was artificial

CSF similar in composition to that of sheep CSF.

38

The method of analysis and determination of CSF
formation rate (Vf) is similar to that used by Welch (1963),
and depends on the constancy of red cell volume between
choroidal arterial and venous blood. However, Pollay and
co-workers could measure blood flow into (Oa) and out of
(Ov) the perfused plexus by determining the time required to
fill 100 pl precision volumetric pipets, and measured
choroidal arterial and venous hematocrits. Although it was
possible to measure directly 0a, Pollay chose to compute it
since there was uncertainty as to the quantity of blood lost
to brain from the anterior choroidal artery prior to its
entrance into the inferior tip of the choroid plexus. The
actual blood flow into choroidal tissue from the anterior
choroidal artery can be calculated from the venous outflow
and arterial (Hcta) and venous (Hctv) hematocrits as:

. . Hctv
= v
Qa Q Hcta

 

(2-1)

Further, the volume flow rate of CSF is the inequality
between arterial (Qa) and venous (Qv) inflow and outflow,
respectively. CSF formation rate (Vf;‘u1/min) can be

calculated from:

Hct
Hct

V

 

Vf = Qa - Qv = Qv [ - 1] (2-2)
a
Although this procedure incorporates the direct

measurement of choroidal arterial and venous hematocrit and

39
permits the control and indirect measurement of choroidal
blood flow, errors in collecting these data might occur
since the posterior choroidal artery supplies a portion of
the choroid plexus, yet is not perfused in this preparation.
Further, if Oa is measured instead of calculated, and all
deep branches of the anterior choroidal artery (which supply
brain areas other than the choroid plexus) are not
electrocoagulated, some inflowing blood will be lost through
these vessels, and Vf will be overestimated. These errors,
however, are probably small. Less than 5% of the plexus by
weight receives blood from the posterior choroidal artery,
and less than 5% of the inflowing blood is lost from deep
anterior choroidal arterial branches as determined
volumetrically (Pollay et a1., 1972). Clearly the strength
of this method is that it allows the separation of the role
of the choroid plexus in the formation and modification of
CSF from the complex solvent and solute exchanges that occur

between CSF and brain tissue.

3. Collection of Nascent Fluid from the Surface of

Exposed Choroidal Tissue

The technique of collecting freshly formed CSF from cat
choroid plexus developed by de Rougemont in 1960 has been
previously discussed with regards to its use in determining
the composition of nascent CSF. However, the direct
sampling procedure has proven useful in estimating the rate

of choroid plexus fluid formation (Vf) as well.

40

The determination of choroidal Vf is based on the
measurement of the maximum rate at which newly secreted
fluid can be collected with a standardized micropipette
reproducibly positioned on the surface of the plexus. The
pipettes are made of tubing of uniform bore and shape, and
are held horizontally to avoid the development of‘a
hydrostatic pressure as the fluid is collected; capillary
action provides the force for fluid collection. This
technique is most useful in predicting changes in choroidal
Vf between control and test conditions since the area
sampled represents only a small portion of the whole choroid
plexus. 'These changes, however, may also reflect those in

whole choroid plexus Vf.
4. Ventriculocisternal Perfusion

Perfusion of the ventricular system to the cisterna
magna was originally introduced as a means of recovering
molecules of cerebral origin and as a method for studying
the effects of ions and drugs on behavior (Cserr, 1971;
Welch, 1975). Pappenheimer and his associates (1962)
developed quantitative methods for investigating the rates
of molecular exchange between the perfusion fluid, plasma
and brain, and for estimating rates of bulk CSF production
and reabsorption. This method involves the perfusion of an
artificial CSF, similar in composition to normal CSF, into
one or both lateral ventricles at a known and constant rate

and with a known concentration of a tracer material, and the

41
quantitative collection of perfusate samples from another
site, i.e., usually the cisterna magna. Alternatively,
inflow or outflow cannulae may be placed in the cortical or
lumbar subarachnoid spaces.

Expressions for the steady state rates of bulk CSF
production (Vf; nl/min) and absorption (Va; ul/min) were
developed from the law of conservation of mass, assuming
that all CSF formation occurs upstream from the outflow
cannula. A reference compound of large molecular size, such
that it is unlikely to diffuse from the system, is mixed
with artificial CSF and perfused into the system at a known
rate (Vi; ul/min) and concentration (ci; quantity/ul).
Perfusate outflow rate (V0; ul/min) and concentration (co;
quantity/DI) are the other measured variables. Given that
tracer loss by diffusion is negligible, then the dilution of
the tracer molecule that occurs during its transit through
the CSF system can be used to calculate the steady state
rate at which newly formed (tracer-free) CSF is added to the

perfusate.

Vf 8 Vi (ci - co)
co

 

(4—1)

Steady state CSF absorption rate, ine., the rate at
which fluid is absorbed in the subarachnoid space, can be
estimated by the clearance of the non-diffusible tracer

molecule.

Va 8 Vi ci - Vo co (4_2)
co

 

42
Finally, a mass fluid balance expression must be
satisfied in order to demonstrate definitively the existence

of a perfusion steady state during sampling periods.
6: + vi = 6. + 6. (4-3)

As pointed out by Heisey et a1. (1983), the apprOpri-
ateness and accuracy of these equations for estimating Vf
and Va are rigidly dependent upon the stability of Vi, Vo,
ci and co. 'This is because tracer molecules may not be
uniformly distributed throughout all portions of the CSF
system which are otherwise in communication. Martins et a1.
(1977) suggested that the ventriculocisternal perfusion
method possesses inherent potential sources of error linked
directly with perfusion inhomogeneities which may exist. He
prOposed that erroneous estimates of CSF formation rate may
be a consequence of the mobilization of sequestrated and
incompletely equilibrated CSF when craniospinal blood volume
is suddenly changed by either an increased PaC02 or an
increased central venous pressure. Ventriculocisternal
perfusion equations (Eqs. 4-1 to 4-3) derived for the steady
state analyses of CSF formation and absorption rate are not
reliable indicators of these quantities during transient
changes in the CSF fluid system (Heisey et a1., 1983; Fisher
et a1., 1983).

E | EEE l' ESE E l' E |
CSF production is a two step process that includes a

passive transudation of fluid and electrolytes from brain

43
capillaries and an active secretory process which involves
the expenditure of energy (Wright, 1979). Consequently,
factors that interfere either with the delivery of solutes
and solvent to the secretory epithelium, or‘with local
supplies of metabolic energy will affect the rate at which
CSF is formed. ‘Although results of experiments are
sometimes contradictory and mechanisms underlying changes in
secretory rate are poorly understood, a consistent pattern
emerges which relates various physical, chemical and

biological factors to CSF formation.
1. Effects of Temperature

Several investigations have been conducted on the
effect of temperature on the secretion of CSF. Davson and
Spaziani (1962) compared the rates of penetration of ethyl
alcohol, ethyl thiourea, and Na24 from blood into brain and
CSF in anesthetized rabbits at 37°C (normal body
temperature) and at 25°C (hypothermia). The rate of
penetration of Na24 into CSF is considered to be a measure
of CSF turnover rate. The rates at which lipid-soluble
ethyl alcohol and ethyl thiourea enter CSF, however, are
largely affected by diffusion rates from the capillaries of
the central nervous parenchyma. ‘Davson et a1. discovered
that cooling the rabbit to 25°C significantly slowed the
penetration of Na24 and ethyl thiourea, but not ethyl
alcohol, into CSF and brain tissue. They postulated that the

decreased Na24 and ethyl thiourea turnover rates are due to

44
a decrease in the primary secretion rate of CSF and/or to a
reduction in choroidal capillary circulation. The small
effects of cooling on the penetration of ethyl alcohol, they
proposed, is due to the fact that ethyl alcohol is normally
consumed quite rapidly by nervous tissue; cooling the animal
markedly reduced the consumption and thereby obscured a fall
in penetration rate that had, in fact, occurred.

Another study which elucidated the effects of
temperature on CSF formation rate was conducted by
Fenstermacher and colleagues in 1969. They selectively
cooled the brains of anesthetized cats to 15°C by perfusing
cooled blood through the right carotid artery which was
exposed and connected by polyethylene tubing to an
extracorporeal system with a heat exchanger. CSF formation
rate (Vf) was measured using ventriculocisternal perfusion
methods during both normo- and hypothermic conditions. They
found that CSF formation rate of the cat whose brain was
made hypothermic was about one-sixth that of the cat with a
normothermic brain. In fact, in some of their experiments,
CSF formation appeared to cease at 15°C.

Snodgrass and Lorenzo (1972) concurred with these
reports. Like Fenstermacher et a1. (1969), they measured
CSF formation rate in anesthetized cats using
ventriculocisternal perfusion, but varied temperature
between 31 and 41°C by heating the whole animal with a
heating blanket or by allowing it to cool. Cats were

divided into two groups: 1) an ”ascending group' in which

45
each steady state was at a higher temperature than its
predecessor, and 2) a “descending group“ in which the first
temperature was the highest. Within the temperature range
studied, a change in 1°C in rectal temperature in either
direction produced a matching 11% change in CSF formation
rate. They concluded that the thermal effects on CSF
formation were not due to injury, but rather resembled the
relationship between rate of a chemical reaction and

temperature.
2. Effects of Osmotic Gradients

The effects of ventricular fluid osmolality on the bulk
flow of nascent CSF into the cerebral ventricles has been
investigated. Heisey et a1. (1962) used the
ventriculocisternal perfusion method to determine CSF
secretion rates in unanesthetized goats provided with
chronically implanted ventricular and cisternal cannulae.
The osmolarity of the perfusion fluid was varied by altering
sodium chloride concentrations while the concentrations of
all other ions were maintained constant. Osmolality was
determined by freezing point or by calculation from sodium
analyses. ‘When the perfusion fluid was hypotonic to plasma,
CSF production decreased. Hypertonic perfusion fluid caused
CSF synthesis to increase. They concluded that since CSF
secretion continued despite an adverse gradient (chemical
potential of water greater in CSF than in blood), CSF

formation must be dependent upon a mechanism other than

46
osmotic flow secondary to the ependymal cell secretion of a
hypertonic sodium chloride solution. They also concluded
that the ventricular ependyma must significantly contribute
to CSF secretion, at least when the CSF was hypertonic to
blood, since abnormally high rates of formation were
observed; if the choroid plexus were the sole source of CSF,
choroidal blood flow would have to be on the order of 3
ml/min per gram tissue if 20% of the plasma flow were
secreted as CSF. In a later study it was determined that
choroid plexus blood flow in anesthetized sheep is, in fact,
6 ml/min per gram tissue (Page et a1., 1980).

In a later study, Wald et a1. (1976) determined not
only the effects of osmolality gradients between CSF and
blood on CSF production rate, but also attempted to
differentiate between the sources of the nascent fluid
during the osmotic tests. 'They measured the bulk flow rate
of newly formed CSF in anesthetized cats during
ventriculocisternal perfusion with solutions containing
sucrose, urea, and sodium chloride in different
concentrations. Subsequently, they inhibited the active
secretion of the choroid plexus with acetazolamide while
osmotic gradients were induced using urea and sodium
chloride solutions and again measured the bulk flow rate of
nascent CSF. The difference between the normal CSF
formation.(measured while perfusing with mock CSF) and that
measured during perfusion with anisosmotic fluids was

linearly related to corresponding differences in osmolality

_ 47
of the effluent from the ventricles. The slope of this
relationship is the coefficient of osmotic flow. 'The
coefficient for sodium chloride and sucrose were similar,
and these molecules were considered to have equal reflection
coefficients in the CSF compartment. However, a larger
osmotic gradient was required to inhibit bulk flow
completely when the perfusion flow contained sucrose instead
of sodium chloride, and perfusion with isotonic sucrose
solution rather than isotonic sodium chloride increased bulk
flow of nascent fluid. They explained this difference
between sucrose, a nonelectrolyte, and sodium chloride, an
electrolyte normally found in nascent CSF, to the fact that
when a sodium-free solution is perfused through the‘
ventricles (i.e., a sucrose solution), a sodium gradient
between the perfusion fluid and the blood and brain is
established. The sodium gradient results in a net sodium
influx from the blood into the perfusion fluid within the
ventricles, which by coupling between solute and solvent,
increases bulk flow rate.

When perfusing with urea and sodium chloride, the
coefficient of osmotic flow and the reflection coefficient
for urea in the CSF system were smaller than for sodium
chloride, while the y-intercepts (idh, the osmotic gradient
required to inhibit nascent fluid bulk flow completely) for
both urea and sodium chloride perfusions were not
significantly different. The bulk flow rates of nascent CSF

with either of the test solutions were not different from

48
the nascent fluid bulk flow rate when mock CSF was perfused.
After intravenous administration of acetazolamide, however,.
both the slopes and the y-intercepts for the two isotonic
test solutions were equal. Wald and co-workers interpreted
these data to indicate that acetazolamide inhibits a source
of fluid which is responsive to differences in the
reflection coefficient, while another mechanism, related
solely to osmotic gradient, still remains active. Since CSF
production is an active process at the choroid plexus, they
interpreted their data to indicate that the choroid plexuses
contriubte significantly to the increase in CSF production
observed during hypertonic perfusions.

More recently, Nakamura and Hochwald (1983) using
anesthetized cats, measured choroidal and regional cerebral
blood flow with microspheres during ventriculocisternal
perfusion with anisosmotic mannitol solutions. They
confirmed that perfusion of the cerebral ventricles with
anisosmotic solutions alters CSF formation. When the
perfusion fluid was 720 mOsm/liter, CSF formation rate
increased by 174%, but during perfusion with a 60 mOsm/liter
solution of mannitol, there was an 80% reduction. How CSF
volume flow rate increases or decreases during anisosmotic
perfusions remains unknown since neither choroidal nor
regional cerebral blood flow changed in response to the

induced changes in perfusion fluid osmolarity.

49

3. Effects of Blood Flow

Blood flow to choroidal or extrachoroidal sites of CSF
production may be influential in regulating CSF formation
rate (Heisey et a1., 1962; Ames et a1., 1965; Carey and
Vela, 1974; Haywood and Vogh, 1979; Martins et a1., 1976;
Davson and Spaziani, 1962; Snodgrass and Lorenzo, 1972;
Lindvall et a1., 1979; Deane and Segal, 1978). It has been
suggested that blood flow places an upper limit on the rate
of CSF formation because as fluid is extracted from CSF
secretory cells, the concentration of plasma proteins
increase. An increased plasma protein concentraton will
increase colloid osmotic pressure in the capillaries and
slow filtration (Cserr, 1971).

Carey and Vela (1974) provided support for the idea
that blood flow to CSF secretory sites alters CSF production
when they found significant reductions in formation rate in
anesthetized dogs whose systemic blood pressures were
lowered to 60 Torr by exsanguination. 4Although neither
choroidal nor cerebral blood flow was measured in these
experiments, Carey and Vela speculated that the decrease in
Vf with decreased blood pressure was due to a diminished
blood flow to CSF secretory sites.

Weiss and Wertman (1978) found results similar to Carey
and Vela when they measured the rate of CSF production in
anesthetized cats while cerebral perfusion pressure (CPP;
CPP = mean arterial blood pressure - CSF pressure) was

decreased either by decreasing systemic arterial blood

50
pressure, or by raising CSF pressure. Only when CPP was
subsequently reduced (50 mmHg) was there an inverse
relationship between CSF formation rate and CSF pressure.
They suggested that variations in CSF formation rate pertain
to the maintenance of constant blood flow through the
choroid plexus under conditions of normal CPP (i.e., when
cerebral blood flow autoregulation is functional). When the
capacity for compensation by cerebral autoregulation is
exceeded (when CPP falls below 55 mmHg; Drake, 1977, cited
in Weiss and Wertman, 1978), adequate choroidal perfusion is
not.maintained and ultrafiltration at the level of the
choroidal capillary is decreased which, in turn, decreases
CSF production.

In 1980, Sklar and co-workers corroborated that CSF
formation rate varies with CSF pressure below limits of
cerebral blood flow autoregulation. They found that CSF
dynamics were altered when cerebral perfusion pressure (CPP)
transiently fell below the lower limits of autoregulation
(spontaneous hypotensive bouts), and CSF formation rate
decreased with increasing CSF pressure. Sklar hypothesized
that because intact autoregulation maintains constant
cerebral blood flow and perhaps choroidal blood flow, CSF
formation may normally be independent of CPP. With a
disturbance of blood flow autoregulation, CSF formation may
follow cerebral blood flow passively with changes in CPP.

Heisey et a1. (1983) found a direct relationship

between CSF formation rate and CPP over the range of 70 to

51
105 Torr, and also suggested that cerebral blood flow
autoregulatory mechanisms may have been disturbed in their
anesthetized cats, possibly by anesthetization. They
suggested that both changes in CPP and regional brain blood
flow would predictably affect CSF formation rate by altering
the rate of plasma ultrafiltration and the amount of fluid
and solutes presented to the epithelial cells of the choroid
plexuses as predicted by the standing gradient hypothesis
(Diamond and Bossert, 1967).

Pollay (1975) was the first to relate the standing
gradient hypothesis of solute and water transport (Diamond
and Bossert, 1967) to the relationship between choroidal Vf
and cerebral blood flow. This theory predicts that Vf
should be regulated both by the metabolically driven
transport systems found at the apical and basal aspects of
the choroidal epithelium, and by the blood flow supplying
these secretory cells with ultrafiltered fluid. If cerebral
blood flow is restricted, as it would be were systemic blood
pressure below the level of autoregulation (Carey and Vela,
1974; Weiss and Wertman, 1978), then transudation of plasma
ultrafiltrate would be decreased. 'This could limit the
delivery of available solute and water to the choroidal
transport systems and result in a decreased rate of CSF
formation. Even with increased cortical and choridal blood
flow (as found with an increased PaCOz (Page et a1., 1980;
Ames et a1., 1965; Martins et a1., 1976)), and increased

transudation of solutes and water from the choroidal

52

capillaries, Vf may not necessarily increase, because with
adequate blood flow, the limiting process in the formation
of CSF may be metabolic (Sklar et a1., 1980). Unless
metabolism is changed during increased transudation, Vf may
remain stable despite an increase in choroidal blood flow.

Bering (1959) investigated the relationships between
cerebral blood flow and the rate of CSFformation by simul-
taneously measuring in the same animal global cerebral blood
flow and cerebral oxygen consumption using the nitrous-oxide
Kety-Schmidt method, and CSF formation rate (Vf) using
timed-collections of CSF drainage. He found that Vf (ml/min
per 100 mg brain weight) was linearly correlated with cere-
bral oxygen consumption (ml/min per 100 gm brain weight),
and Vf (ml/min per 100 mg choroid plexus tissue) was
correlated with cerebral blood flow'(ml/min per 100 gm brain
weight), and concluded that while both cerebral metabolism
and cerebral blood flow are important influences on the net
rate of CSF formation, there are differences between these
processes. Cerebral oxygen consumption is a primary regula-
tory mechanism for CSF production from brain ependyma,
whereas cerebral blood flow is preeminent in determining the
rate of CSF synthesis from choroidal sources.

Cerebral blood flow and CSF formation rate were
measured later by Nakamura and Hochwald (1983). They used
microspheres to measure blood flow in cerebral cortex and

choroid plexus of anesthetized cats during normocapnia, and

53
after Paco2 was increased by 300% or decreased by 50%.
Blood flow measurements were also made during perfusion of
the ventricular system with mock CSF. They found that blood
flow in the choroid plexus, unlike that in the cortex,
decreased by almost 50% during hypercapnia while CSF forma-
tion rate remained constant. Both choroid plexus and corti-
cal blood flow decreased and CSF formation rate stayed
constant during hypocapnia. Nakamura and Hochwald did not
explain the decreased choroidal blood flow during
hypercapnia, but concluded that choroid plexus blood flow
does not limit or even affect the volume flow rate of CSF

from the choroid plexus.
4. Effects of Hydrostatic Pressure Gradients

Heisey et a1. (1962) studied the influence of acute
changes of hydrostatic pressure on the rate of ventricular
fluid formation in goats during ventriculocisternal
perfusion. CSF pressure was varied over a range of -10 to
+30 cmHZO by lowering and raising the height of the
cisternal outflow cannula. They found that although CSF
steady state absorption rate varies linearly with CSF
pressure, steady state CSF formation rate is independent of
acute changes in CSF pressure over the range tested.
Hochwald and‘Wallenstein (1967), and Sklar et a1. (1980)
found similar results from experiments conducted on

anesthetized cats and dogs.

54

Bering and Sato (1963) investigated the effects of
chronic increases in CSF pressure on CSF formation and
absorption rates by inducing hydrocephalus in anesthetized
dogs and then measuring these rates using either a lateral
ventricle-lateral ventricle or lateral ventricle-fourth
ventricle perfusion systems Effects of acute increases in
CSF pressure were also studied in normal dogs using these
perfusion systems, as well as those for ventriculocisternal,
subarachnoid-cisternal, and ventriculo-aqueductal perfusion.
The formation of CSF was found to be constant, independent
of intracranial pressure, and unaffected by the development
of hydrocephalus. (Absorption, however, was linearly and
directly related to the hydrostatic pressure in the
ventricular and subarachnoid spaces for both the normal and
hydrocephalic dog.

Decreases in CSF formation rate with increasing CSF
pressure (20 to 25 cmHZO) however, have been reported for
anesthetized normal rabbits and cats (Hochwald and Sahar,
1971) and anesthetized cats made hydrocephalic by cisternal
instillation of kaolin (Eisenberg et a1., 1974). Sklar and
his associates (1980) suggested that CSF formation rate is
influenced by CSF pressure through impaired cerebral blood

flow autoregulation.
5. Effects of C02 Tension

Hypercapnia induces cerebral vasodilation with an

increase in cerebral blood flow and volume (Reivich, 1964;

55

Rich et a1., 1953; Adams et a1., 1980; Grubb et a1., 1974)
and hypocapnia, a vasoconstriction with a corresponding
decline in cerebral blood flow (Reivich, 1964; Rich et a1.,
1953).

In 1954, Tschirgi and colleagues administered a 70% 02-.
30% C02 gas mixture to anesthetized cats and rabbits before
and after the administration of acetazolamide, a carbonic
anhydrase inhibitor. CSF formation rate was measured by
timed-collections of CSF draining from a cisternal cannula.
C02 inhalation caused only a transient increase in the rate
of CSF outflow which they attributed to a rapid change in
the volume of the intracranial vascular compartment due to
the increased C02 tension. The inhalation of 30% C02
following acetazolamide (Diamox) administration, which alone
caused CSF flow to decline over a period of an hour to a new
steady rate approximately one-third its previous value,
still produced a transient, but smaller, rise in CSF flow.
CSF flow rate gradually returned to its prehypercapnic
level. Tschirgi concluded that large changes in inspired
C02 do not effect steady state changes in CSF flow and that
acetazolamide must inhibit CSF flow through some mechanism
other than through an increased 002 tension of intracranial
fluids. Others (Oppelt et a1., 1963; Hochwald et a1., 1973;
Martins et a1., 1976; Fisher et a1., 1983) conclude similar-
ly to Tschirgi that although the inhalation of C02 produces
transient alterations in CSF flow rate, it does not produce

consistent changes in steady state CSF formation rate.

56

Conversely, Ames et a1. (1965) reported that both the
inhalation of 10% C02 and changing the PC02 in the
Pantopaque medium covering the exposed lateral ventricular
choroid plexus caused a rise in the maximum rate at which
choroidal fluid formed on the surface of the plexus. They
observed.the choroidal vasodilation during hypercapnia, and
suggested that an increase in blood flow may have been
primarily responsible for the changes in fluid production.

In ventriculocisternal perfusion studies (Oppelt et
a1., 1963; Oppelt et a1., 1964; Hochwald et a1., 1973;
Davson and Segal, 1970), including those in which average
cerebral blood flow changed with increased Paco2 (Martins et
a1., 1976; Nakamura and Hochwald, 1982), 6: was found to be
unaffected by increases in PC02. This apparent
contradiction may be partly explained on the basis of
methodological differences between Ames' in Sitll choroid
plexus preparation and the ventriculocisternal perfusion
method. First, the former experimental technique requires
more extensive surgery than does the ventriculocisternal
pefusion technique, and may damage the blood-brain barrier
and cerebral autoregulation. Secondly, data from in sitn
studies report only choroidal production of CSF, whereas
data from ventriculocisternal perfusions report net changes
in Vf independently of their source, which may be more
realistic, since there are two sources of Vf; the choroid
plexuses and the brain ependyma (Hochwald and Wallenstein,

1967; Bering and Sato, 1963; Welch, 1963; Pollay and Curl,

57
1967; Pollay, 1975; Bering, 1959; Curl and Pollay, 1968;

Davson and Luck, 1957; Davson and Segal, 1970).
6. Effects of Carbonic Anydrase Inhibitors

Common to the kidney, stomach, pancreas, and other
exocrine structures, the choroid plexus has a high
concentration of carbonic anhydrase (Maren, 1972; Vogh,
1980) which is also found in brain glial cells (Vogh, 19801.
This enzyme catalyzes the formation of bicarbonate from
gaseous C02, and is intimately involved with the movement
of fluid across epithelial tissues. In fact, the first
substance found to affect materially the rate of CSF
formation was acetazolamide, which inhibits carbonic
anhydrase. Carbonic anhydrase inhibitors markedly depress
CSF formation rate (Ames et a1., m 1965; Oppelt et a1., 1963;
Pollay, 1975; Cserr, 1971; Welch, 1975; Welch, 1963; Oppelt
et a1., 1964; Knopp et a1., 1957; Davson and Luck, 1957;
Davson and Segal, 1970; Maren, 1972; Pollay et a1., 1972;
Curl and Pollay, 1968), however, the mechanism is unknown.

Tschirgi and co-workers (1954) administered
acetazolamide (150 mg/kg;iv) to anesthetized cats and
rabbits, and compared the rate at which CSF drained freely
from a cisternal cannula before and after the drug
treatment. They observed a 30% decrease in CSF pressure in
the intact ventricular system, (which was later confirmed by
Knopp et a1. (1957)), and a sustained 3- to lS-fold (70%)

decrease in the CSF drainage rate after acetazolamide

58
treatment. They proposed that acetazolamide inhibits CSF
synthesis by decreasing the rate of carbonic acid production
from metabolically produced C02. They postulated also that
choroidal and extrachoroidal CSF production is dependent
upon an ion exchange between plasma and CSF. Transport
processes located on choroidal epithelium and throughout
brain glial processes exchange hydrogen and bicarbonate
derived from the hydration of C02 for sodium and chloride.
One mole of metabolically produced C02 is exchanged for one
mole each of sodium and chloride which is transported from
plasma to CSF. To maintain osmotic equilibrium, it was
proposed that water passively follows transported sodium and
chloride into CSF. Carbonic anhydrase, then decreases CSF
formation rate by decreasing the amount of hydrogen and
bicarbonate available for exchange.

Davson and Luck (1957) refuted Tschirgiis proposal that
acetazolamide inhibits extrachoroidal CSF formation when
they found that Na24 turnover (an index of CSF production)
decreases in CSF but not in brain after intravenous adminis-
tration of acetazolamide. IDavson and Luck agreed that
acetazolamide may be interfering with ion-exchange at the
choroid plexus, an idea supported by Pollay et a1. (1972)
who demonstrated using the isolated perfused choroid plexus
that the addition of acetazolamide to the arterial blood (20
pg/ml) or to CSF (10"3 M) stopped choroidal CSF secretion.
Pollay and his co-workers also measured choroidal blood flow

and found that it was not affected by acetazolamide.

59

If an ion-exchange mechanism were operating at the
choroid plexus, reasoned Welch (1963), the content of C02 in
choroidal venous blood would have to be enormous. At rates
of choroidal CSF formation measured by him, approximately
24 mEq of sodium ions are secreted, an equivalent amount of
C02 would be expected to be recovered in choroidal venous
blood if the ion-exchange mechanism were operating. He
measured the C02 content of venous blood from the plexus and
found it insufficient to account for the sodium chloride
flux into CSF.

A more recent view of the role of carbonic anhydrase in
CSF secretion, and the mechanism whereby acetazolamide
inhibits CSF formation was presented by Maren in 1972. He
discovered that intravenous acetazolamide inhibited the
transependymal flux of bicarbonate to the CSF similarly to
that of sodium and water. He proposed that plexus secretory
cells, through the protolysis of water, produce hydroxyl
ions which react with C02, with or perhaps without carbonic
anhydrase. Bicarbonate and hydrogen ions are formed as a
result of this reaction. The proton would subsequently
enter the bloodstream and be buffered while a bicarbonate
gradient was established. Sodium would then serve as a
counter ion for transport with bicarbonate into the CSF
system. Maren proposed, therefore, that carbonic anhydrase
inhibitors act by interrupting the necessary coupling of
bicarbonate and sodium which, in turn, prevents the

osmotically obligated movement of water into the CSF and CSF

60
formation declines. Further, he hypothesized that since
sodium and bicarbonate are intimately coupled, anything that
would interfere with carbonic anhydrase or ATPase would

produce similar changes in sodium and water transport.
7. Effects of Cardiac Glycosides

The choroid plexuses have transport systems for
inorganic ions as well as for organic acids and bases
(Wright, 1978). Since ouabain can inhibit these processes,
they have been implicated to be dependent on Na/K-ATPase
(Lindvall et a1., 1982; Vates et a1., 1964). Likewise, CSF
formation is dependent upon this enyzme system since ouabain
produces sharp declines in CSF production when administered
intraventricularly (Cserr, 1971; Vates et a1., 1964),
intrathecally (Oppelt et a1., 1964), or when applied
tepically to the choroid plexus (Welch, 1963). In fact, if
ouabain is added to a medium which bathes the plexus, the
elaboration of CSF is almost completely inhibited (Welch,
1963). Unfortunately, ouabain is not a useful in 1119,
research tool because, unlike acetazolamide which is well
tolerated at doses required to achieve maximal effects, it
is extremely toxic producing both marked circulatory and
respiratory disturbances.

The mode of action of cardiac glycosides on CSF
formation rate is better defined than that of carbonic
anhydrase inhibitors (Pollay, 1975; Cserr, 1971; Ames et
a1., 1965; Curl and Pollay, 1968; Davson and Segal, 1970).

61
Ouabain decreases CSF formation by inhibiting the Na/K-
ATPase enzyme system (Vates et a1., 1964). The Na/K pumps
localized on the apical surface of choroidal epithelial
cells (Wright, 1978) and found throughout brain tissue are
fueled by the hydrolysis of intracellular ATP by the enzyme
Na/K-ATPase. Inhibiting the pump prevents the active
secretion of sodium into the ventricle (Vates, 1964), which
is considered by Wright to be the primary process in CSF

secretion.
8. Effects of the Autonomic Nervous System

Possible neural influences on the production rate of
CSF has been suggested by histological studies since 1874,
when Benedikt demonstrated the presence of nerves in the
choroid plexuses. It is now well known that both
sympathetic and cholinergic nerve terminals within the
plexus innervate both the resistance vessels and the
secretory epithelium (Edvinsson et a1., 1973; Edvinsson et
a1., 1974; Edvinsson et al., 1975a; Edvinsson et al., 1975b;
Lindvall et a1., 1982). In 1971, Edvinsson and co-workers
helped to establish that the choroid plexus innervation was
indeed functional. They measured the activity of carbonic
anhydrase from rabbit choroid plexus homogenates before and
after cervical sympathectomy by incubating the plexus
homogenates in phosphate buffer containing NaH14CO3 and the
14C02 formed was collected in Protosol-soaked filter paper

strips. The radioactivity of the strips was measured by

62

liquid scintillation spectrometry. They found that at one
and three days post-sympathectomy, carbonic anhydrase
activity had approximately doubled; at 14 days it was
reduced by 60% compared with unoperated controls.

Subsequently, Edvinsson and colleagues (1975) were able
to confirm their previous observations in homogenates of
blood-free rabbit choroid plexus. They found that carbonic
anhydrase activity was higher in the choroid plexus from the
fourth ventricle than from the lateral ventricles, and that
depletion of norepinephrine from the adrenergic nerves by
reserpine treatment increased carbonic anhydrase activity.
Two weeks after sympathectomy, however, the carbonic
anhydrase activity had returned to normal which they
attributed to the development of denervation supersensi-
tivity of the adrenergic receptors to circulating catechola-
mines. They concluded that the sympathetic nervous system
inhibits carbonic anhydrase activity and CSF production.

Lindvall et al. (1978b) directly measured the influence
of sympathetic nerves on the rate of CSF production. One
group of anesthetized rabbits underwent ventriculocisternal
perfusion one week after superior cervical ganglionectomy.
Another group of animals had their cervical sympathetic
trunks electrically stimulated during ventriculocisternal
perfusion. The first group of animals showed a 33% increase
in the rate of bulk CSF production compared to the second
group before electrical stimulation (control). In the

intact group, sympathetic activation produced a reduction in

63
the rate of CSF formation by 32%, but systemic blood
pressure effects were negligible. Based on the results from
this study and from the carbonic anhydrase study on blood-
free choroidal tissue (Edvinsson et a1., 1975), they
concluded that the sympathetic nerves in the choroid plexus
inhibit CSF production primarily by their effects on the
plexus epithelium and its secretory functions.

Little was known in 1978 about the nature of the local
adrenoceptors mediating the sympathetic-nerve induced effect
on CSF formation. Lindvall and associates (1979)
pharmacologically characterized these receptors by infusing
sympathomimetic agonists and corresponding specific receptor
antagonists via the intraventricular or vascular routes
during ventriculocisternal perfusion of anesthetized
rabbits. They found that intraventricular or intravenous
administration of norepinephrine produced a dose-related
decrease in CSF production up to 50% which could be
counteracted by both phentolamine (an alpha-antagonist) and
intravenously» bur.not intraventricularly administered
prOpranolol.(a beta antagonistL. Systemic arterial blood
pressure remained stable during intraventricular infusions
of norepinephrine, but transiently increased with
intravenous infusions of norepinephrine. Further,
intraventricular administration of a betal-receptor agonist,
HBO/62, reduced CSF formation; practolol (a betal-
antagonist) blocked this effect. Conversely, terbutaline, a

betaz-agonist had little effect on CSF synthesis. They

64
suggested that sympathomimetic reduction of CSF production
is a complex function of both betal-receptor-mediated
inhibition of plexus epithelium secretion as well as a
reduced choroidal blood flow resulting from stimulation of
the vascular alpha-receptors.

Biochemical evidence supporting the suppostion that
beta-receptors exist on choroidal epithelium was provided by
Nathanson in 1979. Binding of beta-adrenergic agonists to
such receptors leads to the intracellular synthesis of 3',
5'-mon0phosphate (cyclic AMP) from adenosine triphosphate
(ATP) (Edvinsson and MacKenzie, 1977), a reaction catalyzed
by adenylate cyclase. Nathanson investigated the possible
existence of beta-adrenoceptors by enzymatically determining
the synthesis of cyclic AMP from ATP in broken cell prepara-
tions of cat, rabbit, calf, and dog choroid plexus. Cell
preparations were exposed to various concentrations of
biogenic amines and cyclic AMP activity was measured. He
discovered beta-adrenergic-sensitive adenylate cyclase in
broken cell preparations of both whole choroid plexus and
choroidal epithelium which had been dissociated from its
subjacent vasculature. A substantial portion of the enzyme
was present in the epithelial cell fractions. Further, he
showed that l-isoproteronol-induced stimulation of adenylate
cyclase was greater in epithelial cell fractions than in
choroidal vessel fractions, and that the cat fourth
ventricular choroid plexus, which contains relatively'more

secretory epithelium than the lateral ventricular plexus,

65
shows a twofold greater response in adenylate cyclase to
isoproteronol. L-norepinephrine, a mixed beta— and alpha-
adrenergic agonist, stimulated enzyme activity nearly as
much as isOproteronol (310% of control levels) but was less
potent (Ka - 3 x 10"5 M); l-phenylephrine, an alpha-
adrenergic agonist, was both less potent (Ra = l x 10" M)
and less effective (185% of control levels).

Likewise, alpha-adrenergic mediated vasomotor
reactivity in the choroid plexus was confirmed from in 11:19,
studies on anterior choroidal artery strips. Edvinsson and
Lindvall (1978) dissected choroidal arteries from freshly
slaughtered cows, and mounted two pieces of the arteries as
cylinders in an organ bath containing a CSF buffer solution.
Circular vasomotor activity was registered isometrically and
reflected by the voltage output from force-displacement
transducers. Measurements were made after various
vasoactive pharmacological agents were added to the organ
bath. They found no spontaneous contractions of the
arterial strips, but found weak contractile responses to
epinephrine, norepinephrine, phenylephrine, and
isOproterenol in decreasing order of potency; Phentolamine
(alpha-antagonist) inhibited these responses. Other agents
such as histamine, prostaglandin F2“, and saturated
potassium chloride produced strong contractions. Dilator
responses could also be elicited, provided the choroidal
arteries were constricted with prostaglandin F2“ and

contractile effects blocked by phenoxybenzamine (an alpha-

66
antagonist). The relative potency for dilator
sympathomimetic agents were, in decreasing order,
isoproterenol, norepinephrine, epinephrine, terbutaline.
The dilator effect was inhibited by propranolol (a beta-
antagonist) which indicated the dilator beta-adrenergic
receptors coexist with alpha-adrenergic receptors on
choroidal arteries.

Cholinergic effects on CSF formation rate have also
been directly studied and appear to have an influence.
Lindvall et a1, (1978a) measured CSF production in
anesthetized rabbits by ventriculocisternal perfusion.
Intraventricular or intravenous infusion of‘carbamylcholine
(a parasympathomimetic) or acetylcholine produced a dose-
related reduction in CSF formation by as much as 55%.
AtrOpine (a muscarinic cholinergic antagonist) antagonized
this response while hexamethonium (a nicotinic cholinergic
blocking agent) did not. Intraventricular carbachol (para-
sympathomimetic) potentiated the reduction in CSF secretion
induced by electrical activation of the cervical sympathetic
nerve trunks. They concluded that muscarinic cholinergic
mechanisms markedly reduce CSF production, though the site
of the parasympathomimetic agents is unknown. Since CSF
responses were observed after the agents were administered
intraventricularly, Lindvall proposed that the effects were
mediated through the secretory epithelium rather than

through the vascular system.

67
W

Although CSF may be absorbed by the brain (Bering and
Sato, 1963; Rosenberg et a1., 1978; Rosenberg et a1., 1982)
and by cranial and spinal nerve roots (Pollay and Welch,
1962; Hammerstand et a1., 1969; Davson et a1., 1970;
Marmarou et a1., 1975), its main absorption into venous
blood is normally'through the arachnoid villi.(Welch, 1975)
which project into dural venous sinuses.

Welch and Friedman (1960) and Pollay and Welch (1962)
investigated the function of the arachnoid villi in monkeys
and dogs. They mounted a disc of dura mater between two
chambers as an occluding membrane. The two chambers were
then filled with saline, and one chamber was attached to a
reservoir containing Ringer lactate solution. The other
chamber was attached to a small polyethylene tube, and flow
was measured by monitoring bubble movement. Flow through
the arachnoid villi, seen when the pressure difference
across the disc was approximately 30 mm of perfusion fluid,
was only from the subarachnoid to the sinus side. ‘Above
30 mm, the critical opening pressure, flow increased non-
linearly with increasing pressure. (At pressures just higher
than the critical opening pressure, the lepe of the
pressure-flow curve was steeper than that for high pres-
sures, presumably due to drainage of CSF through valved
elastic tubes.

Others have confirmed that CSF absorption through the

arachnoid villi is a pressure-dependent phenomenon in many

68
mammals (Welch and Friedman, 1960; Pollay and Welch, 1962;
Heisey et a1., 1962; Bering and Sato, 1963; Davson et a1.,
1970; Hochwald and Sahar, 1971; Marmarou et a1., 1978), and
that colloid osmotic pressure differences between CSF and
venous blood in the dural sinuses are inconsequential to the
rate of CSF absorption through the villi.(Pollay and Welch,
1962; Davson et a1., 1970; Pollay and Davson, 1963). The
‘villi seem to allow an unrestricted passage ofjproteins
through their pores, as corroborated by Pollay and Welch
(1962) who showed that 8 n.diameter polystyrene spheres and
canine erythrocytes pass freely through canine arachnoid
villir Also, Mann and co-workers (1979) demonstrated in the
rat that.0.5 n.polystyrene beads and inulin flow through
arachnoid villi. The arachnoid villi are a major efflux
route for CSF and for macromolecular and particulate

substances.

W

The relationship between the volume of CSF and its
pressure has been investigated clinically and in the
laboratory for many years. Weed et a1. (1932) studied CSF
pressure changes produced by changes in body position and
determined that when an animal was tilted from a horizontal
to a vertical (head-down or tail-down) position, the volume
of CSF displaced from the CSF system through an Open-end
manometer was linearly related to CSF pressure. They sug-
gested that the central nervous system enclosures behaved

like an elastic system which permits relocation of

69
craniospinal fluid. .Although cranial dura mater is less
elastic because of its close apposition to the skull, the
spinal dura and the compressible, distensible spinal and
cranial venous beds permit CSF movement.

weed and his colleagues later investigated whether the
cranial venous system is an elastic component of the central
nervous system. lThey measured subarachnoid and saggital
sinus pressures in etherized dogs during venous compression
and during cranial subarachnoid fluid volume expansion due
to cisternal infusions of CSF. Any intracranial venous
pressure change affected CSF pressure, while saggital sinus
venous pressure was uninfluenced by changes in subarachnoid
pressure. They proposed that the pressure-volume
relationship between cerebral veins and CSF is that of two
fluids of unequal volume but normally'of equal pressure
which are separated by an elastic membrane. The two fluids
have different pressures because the cerebral blood volume
is not confined to the cranium, but CSF volume is confined
in rigid craniospinal membranes, which it can leave only
slowly. They concluded that abrupt changes in CSF volume
are permitted mainly by a concomitant compression of the
cerebral venous bed.

Ryder et a1. (1953) provided a similar assessment of
the volume-pressure relationshiplbetween the vascular and
CSF volumes based on CSF pressure recordings taken from
patients with different intracranial disorders. They

observed in all patients that lateral ventricular or lumbar

70
pressures were stable. Subarachnoid volume additions or
withdrawals of CSF did not affect CSF pressure. When CSF
pressure was recorded during the withdrawal or addition of
several milliliters of fluid from or to the CSF space,
return of the displaced pressure to its pre-disturbance
level occured within minutes. ‘When large'volumes of fluid
were added or withdrawn, it was necessary to withdraw or add
fluid to the CSF space in order to re-establish pressure
within minutes. When fluid was added to the CSF space,
however, less was needed to be withdrawn in order to re-
establish CSF pressure. Likewise, after volume withdrawals,
a lesser CSF was necessary to restore pressure. They con-
cluded that if intracranial volume is constant at a given
pressure, then a change in the volume of the cerebrospinal
fluid space induces an immediate reciprocal change in the
craniospinal vascular bed, as well as a concomitant net
seepage of CSF into or out of the CSF system.

In 1972, Martins and co-workers characterized the
intracranial pressure-volume relationship slightly
differently. Using human volunteers, they observed the
movement of a PantOpaque column in the spinal subarachnoid
space during myelography. Each of the volunteers performed,
in turn, a‘Valsalva.maneuver, hyperventilated for two
minutes, breathed 10% C02 in oxygen for two minutes, and
then repeated the Valsalva maneuver. Finally, the
volunteers underwent fifteen seconds of jugular compression.

During both the forced expiration with glottis closed

71
(Valsalva maneuver) and during hyperventilation, the spinal
dural sac partially collapsed, as indicated by the
Pantopaque column narrowing and elongating. Conversely,
during 10% C02 inhalation and during bilateral jugular
compression, the spinal sac distended, as indicated by the
column widening as it moved caudally. to fill the
lumbosacral nerve roots. They concluded that the spinal
dural sac serves as a dynamic reservoir that can change its
capacity in response to prevailing transmural pressure
gradients. Since the cranial and spinal CSF spaces freely
communicate through the foramen magnum, CSF could rapidly
move between these two compartments, and volume-induced
sudden changes in CSF pressure could be effectively-
buffered. They concluded also that the spinal dural sac
together with its loose fatty tissue and epidural venous
plexus provide the major source of elasticity which
underlies the apparent reciprocal relationship between
cranial blood and CSF volumes.

Marmarou et a1. (1975), however, disagree with Martinfis
hypothesis that the majority of central nervous system
elasticity is in the spinal axis. They conducted a series
of experiments on anesthetized cats designed to apportion
the total central nervous system elasticity into cranial and
spinal components. Cerebral and spinal CSF components were
isolated by an inflated balloon positioned epidurally'at the
C-6 spinal level. CSF volume additions were made, in turn,

into the two compartments, while the change in CSF volume in

72
respect to CSF pressure (dV/dP) as well as the change in CSF
volume per time (dV/dt) were measured. These ratios were
used to define CSF compliance and absorption rate,
respectively. They found that 68% of total compliance is
attributable to the cerebral compartment, and 32% to the
spinal axis. They also found that 84% of CSF absorption is
in the cranial compartment. They showed that central
nervous system elasticity is due to the compressibility, and
distensibility of cranial venous vessels, but recognized
that the spinal axis provides an additional buffering system
against sudden volume-induced changes in CSF pressure.
Cranial, and to a lesser extent spinal, CSF absorptive
mechanisms were more important in buffering “steady-state"
pressure disturbances, such as those associated with

expanding extra- or intracerebral masses.

MATERIALS AND METHODS

ANESIHESIA
Adult cats (Fells domestigns) of either sex, weighing

2.5-4.5 kg were anesthetized either with sodium
pentobarbital (25 mg/kg iv or 36 mg/kg ip; Nembutal, Abbott
Laboratories, Chicago, IL) or with Dial-urethane (0.6 ml/kg
ip; Appendix B). .Anesthetic depth was assessed periodically
by evaluating jaw tension, palpebral, corneal and nociceptor
reflexes, and supplemental iv doses of sodium pentobarbital

were given as needed to maintain a stable anesthetic plane.

SUBQICAL_AND_EXEEBIMENIAL_RRQQEDQBE

Systemic arterial blood pressure was monitored from a
cannula (PE-60; Clay-Adams, Parsippany, NY) introduced into
the aorta via the femoral artery, which was attached to a
pressure transducer (Statham model P23Dc; Grass Instruments,
Quincy, MA) in series with a low level DC preamplifier
(model 5P1; Grass Instruments, Quincy, MA), and displayed on
a oscillograph (model 5DWC1; Grass Instruments, Quincy, MA).
Arterial blood samples were periodically collected anaerobi-
cally for pH and PC02 determinations (Appendix C) using
either Radiometer thermostatted electrodes (model PHA 927b;
Radiometer A/S Copenhagen, Denmark) or a Corning blood gas
analyzer (model 165/2; Corning Medical Instruments,

73

74
Medfield, MA). .A venous cannula (PE 60) was introduced into
the inferior vena cava via the femoral vein for injecting
supplemental doses of anesthetic and for test drug infu-
sions. Body temperature was monitored with a rectal therm-
istor probe attached to a Wheatstone bridge circuit and
meter (model 400; Yellow Springs Instrument Co., Yellow
Springs, OH) and maintained by a manually adjusted heating
pad on which the animal rested.

The trachea of each animal was intubated with an
inflatable cuffed cannula (4.0-5.0 mm ID.; American Hospital
Supply, McGraw Park, IL) which was affixed to a low resist-
ance, low deadspace respiratory valve assembly, through
which the cat breathed spontaneously. The expiration port
of the respiratory'valve was attached to a Fleisch pneumota-
chograph (model 7318; Dynasciences Medical Products, Blue
Bell, PA) and a low level DC preamplifier (model 5P1; Grass
Instruments, Quincy, MA) for monitoring respiratory
frequency and tidal volume (Appendix D). End tidal C02
(FACOZ) was monitored from a needle cannula in the tracheal
cannula which led to an infra-red C02 gas analyzer (LB-2
model; Beckman Instruments, Inc., Anaheim, CA; Appendix E).
By means of amplified signals from the pneumotachograph and
C02 analyzer, tidal volume and FACOZ, respectively. were
displayed on a calibrated strip chart recorder (model 7100
IBM; H-P Moseley, Pasadena, CA). Mean steady state tidal
volume (VT) and minute ventilation (VB) were calculated as

described in Appendix D.

75
BBAIN_yENIB1QflLAB_AND_£ISIBBHAL_EHNQIHRES

A midline skin incision was made from the skull's
interparietal bone to the spinous process of the second
cervical vertebra. Underlying neck muscles and fascia were
cleared by blunt dissection to expose the atlanto-occipital
membrane. A cisternal needle (21 ga.;:metal hubless needle;
Vita Needle Co., Needham, MA) secured in a micromanipulator
(model MM-3; Eric Sabotka Co., Inc., Farmingdale, NY) was
positioned with its point on the atlanto-occipital membrane
surface, near the dorsal edge of the axis opening, and in a
plane parallel to the stereotaxic frame; polyethylene tubing
(PE 90; nom. 25 cm in length) was attached to it as the
outflow cannula. The needle was advanced 6 mm through the
atlanto-occipital membrane and underlying dura. A
successful cisternal puncture was confirmed by the free flow
of clear cerebrospinal fluid from the outflow cannula. The
puncture site was then reinforced with collodion
(Mallinckrodt Chemical Works, St. Louis, MO), and the
outflow cannula attached to a reservoir containing
artificial cat cerebrospinal fluid (CSF; Appendix F).

A 4 to 6 cm midline incision was made in the skin of
the head from a point 4 cm caudal to the supraorbital ridge
to a point 2 cm posterior to the intra-aural line. The skin
was retracted, and the underlying muscle and connective
tissue were cleared to expose a 3 cm2 area of frontal and
parietal bone. Two 3 mm diameter holes (13.5 mm anterior to

the intra-aural line; 2.5 mm lateral to the midline; were

76
drilled through the skull using a dental burr and drill
(model 4N; Emesco Dental Co., Inc., New York, NY), exposing
the dura over each cerebral hemisphere. Two metal needles
(20 9a.; Becton, Dickinson 5 Co., Rutherford, NJ) heldin
electrode carriers (model 1204; David KOpf Instruments,
Tujunga, CA) were connected by a T-connector to a pressure
transducer (Statham model P23Ac; Grass Instruments, Quincy,
MA) for measuring intraventricular pressure (PCSF) and to
two, 20 cc syringes on a constant speed drive pump (model
975; Harvard Apparatus Co., Dover, MA) for infusion of
artificial cat CSF (Appendix F). The inflow needles were
centered over the skull openings and lowered separately
through brain tissue into the lateral ventricles; correct
needle placement was indicated by a sudden drop in inflow
pressure, and was further confirmed when changes in cister-
nal reservoir height were reflected by similar changes in

CSF pressure (PCSF).

yENIRICflLQQISIEBNAL_REBEHSIQN_EXEEBIMENIS

The infusion rate (V1) of mock cerebrospinal fluid
(Appendix F) which contained a trace amount (nom. 7.5 nCi)
of 14C-1abelled dextran (mw=70,000 D; New England Nuclear
Corp., Boston, MA) was held constant during each experiment;
Vi was 70 ill/min in some experiments and 140 nl/min in other
experiments.

The cisternal outflow line was disconnected from the

CSF reservoir and joined to an adjustable height photocell

77
drop detector attached to a fraction collector (model 568;
Instrumentation Specialties Company, Lincoln, NA) for
cisternal effluent sampling. Two drop effluent samples
(nom. 100 pl) were collected continuously during
ventriculocisternal perfusion. The height of the outflow
line could be adjusted relative to the zero level at the
animal's intra—aural line in order to regulate
intraventricular pressure (PCSF). In addition to systemic
arterial pressure and respiration, intraventricular pressure
and outflow rate were monitored and displayed continuously
using a polygraph (model 5P; Grass Instrument Co., Quincy,
MA). The arrangement of apparatus used for the

ventriculocisternal perfusions is shown in Figure l.

E l J E' H' l E E E .
Ventricular perfusion was continuous while each animal
initially breathed room air (Phase I) for 30 to 40 minutes,
and then breathed a 7% COZ-in-air gas mixture (Phase II) for
30 minutes; during these two initial phases, CSF pressure
was maintained between 0 and 5 cm of water (relative to the
level of the auditory meatus). The animal was then returned
to room air breathing for approximately 40 to 45 minutes.
During this period (Phase III), PCSF was slowly increased
and maintained between 20 and 25 cm of water by gradually
raising and keeping the cisternal outflow line above the
level of the intra-aural line. Phases I and II were then

repeated under the high CSF pressure conditions.

 

 

78

.mHmAHccm mac cooHn can we now couooHHoo maamofioowuwm mum moacemm

cooah Hmauouuc .Acsonm uocv awesomcmbu musmmoum m ou cwuowccoo mfisccmo Hmfiuwuum
HmuoEou a sm50unu cmuouflcoe ma musmmmum pooan cam Aczonm uocv poxcmah mcfiummn
counsnpm adamscme m an ucmumcoo pawn m“ can encum noumaEuwnu m cues cmuouficoa m«
on:LMuomEmu Hmuoom .wocobcoum muoumuwmmou one much 30am has mcficuoowu Lou umoscmcmLu
ousmmoum Hmfiucwuommac m on wouooccoo ma gowns accumonomu055ocm nomfioam a smsounu
concouflp ma 0>Hs> may soda mom cmuficxm one .mHsccmo Hamsomuu m 0» conomuum wahsmmmm
m>Hm> mommmcmoo 30H .mocmumfimmu 30H m caucus» Anacmmouomhnv musuxwe mum basicalmoo m
no AcacQMOOEuocv ufimIEoou donned monumwuh aawsomcmucomm anagcm one .masomaos house»
on» Low commando mum mmu ucosammm mo mmHmEmm can .uoucsoo coup m magma consumes

m« Ao¢v mumu soamuso .mcmme acumumfio on» ouch couummcfi wasccmo m an“: .umoscmcmuu
wusmmmum m an omuouficoe ma wusmmoum mmo was made m>wuclmmcwu>m m an ucmumcoo cams

we Aa>v mumu scams“ mmu .mmaowuucm> Hauhmumo Hmuouma on» one“ occasc ma Accuuxoc
Ioewv masomaos mmuma m mo aufiucmsc women m mewcwmucoo Ammo. pagan Hmcflmmoanumo
Hmaofim«uum cm dawns Acsonm Docv messy onmuomuwum 8 ca monsoon ma umo cmufiuwnummcm

as no cams one .usumummns Houcoafluwmwo cofimnmuom Hmcuoumwoousouuusob mo causewaum

.H MKDUHM

79

mwoaom—‘(mh wmnmmmcn

mmODDm2(1.—. 304...

m>4<> ><3Iw

IIIJIIJI
E< omEaxm D _ _ IJDI.

45> w4m2<m

 

 

$32.50 noun 0

cwonomzsn:
Nabwwwmn

 

.— "WEDGE

 

80

WW

These experiments were similar to those described in
protocol A, in that perfusions began while the animal
spontaneously breathed room air (Phase I), then breathed a
5% COZ-in-air gas mixture (Phase II), and next breathed room
air again (Phase III). During phase III a |'stop-flow"
perfusion was instituted by connecting the cisternal outflow
line to a closed stopcock attached to a reservoir; cisternal
effluent was stopped.: CSF pressure increased from its ini-
tial level of 0 to 5 cm of water to 20 cm of water. When
PCSF rose to 20 cm of water, the stopcock was Opened and CSF
was allowed to flow into the reservoir until PCSFldecreased
to about 12 cm of water, when stop-flow was reinstituted.
This procedure of alternately stopping and starting CSF
efflux continued for 45 minutes. Finally. PCSF was returned
to the 0 to 5 cm of water level by reattaching the outflow
line to the drOp detector. Phases I and II were then

repeated.

W

This experimental plan.for perfusion parallelslthose
described in protocols A and B, except that during Phase
III, CSF pressure was not manipulated, and 2 mg/kg of an
alpha-adrenergic blocking agent, phenoxybenzamine
hydrochloride (phenoxybenzamine hydrochloride 50 mg;
absolute alcohol 48.5% v/v; hydrochloric acid 0.5% v/v;
propylene glycol, redistilled q. 8. ad. 1.00 m1.; Smith

Kline and French labs., Philadelphia, PA)‘was slowly infused

81
intravenously. After 30 to 40 minutes, phenylephrine
hydrochloride (nom. l3ug/kg; WintrOp laboratories, New York,
NY)‘was injected iv and alpha blockade was confirmed when a
pressor response was abolished or inverted to a depressor
response. Phases I and II were then repeated under the

blocked condition.

DBIEBMINAIIQN_QE_REB£HEIQN
CQNQENIBAIIQNS_AND_BATBS

Inflow and outflow 14C-dextran concentrations were
determined using liquid scintillation spectrometry'(model
3lSOP; Beckman Instruments, Fullerton, CA; Appendix H) on
samples from the perfusion syringes and cisternal effluent,
respectively. The output from the scintillation counter was
recorded on paper tape (Epson 6100 tape punch; Shinshu Seiki
Co., Ltd., Japan), processed (Appendix H) and stored on the
PDP 11/34 computer (Digital Equipment Corp., Maynard, MA).
The data were printed on a high speed line printer (model
300; Printronix, Irvine, CA) in both tabular and graphical
formats for mathematical analysis. Inflow and outflow
rates were determined from the polygraph record as described

in Appendix G.

MAIHEflAIICAL_ANALXSI§_AND_QALCHLATIQNS
STEADX_§IAIE_ANALXSIS

The equations which have been developed for calculating
CSF formation (Vf) and absorption (Va) rates can be used

only when inflow and outflow rates and tracer concentrations

82
are stable (i.e., when the system is in a steady state;
Heisey, 1983; Fisher, 1984). In these experiments, steady
state criteria were met usually within the last 5 to 10
minutes of phases I and II; CSF formation rate was
calculated by measuring the dilution of 14C-dextran by newly

formed, tracer-free fluid:

 

where:
Vf = CSF formation rate (ul/min)
Vi = perfusion inflow rate (ul/min)
co/ci = 14C-dextran outflow concentration (co)

expressed as a ratio of perfusion inflow

concentration (ci) (dimensionless)

The clearance of dextran from the perfusion system has
previously been shown to be a pressure-dependent passive
process which is a measure of CSF bulk absorption (Heisey,

1962) and is calculated by:

 

Va = Vi _ [Vo x (co/ci)] (2)
(co/ci)
where:
Va = CSF bulk absorption rate (pl/min)

VO

perfusion outflow rate (nl/min)
During a steady state a fluid balance is defined as:

Vi+Vf = Vo+Va (3)

83

and rearranging:

 

n: = 6. - v: + va (4)
v1 [(1 ' °°/°i)] = vo - v: + va (5)
(co/ci)

The mass fluid balance described by equation 5 is a
confirmation of steady state conditions in the perfused

ventriculocisternal system.

Transient_Bssnnnss_Ana1¥ais

As shown in Figure 2, phase I steady state cisternal
outflow rate (V0) and dextran concentration are perturbed
when an animal initially breathes C02. When cerebrospinal
fluid pressure is between 0 and 5 cm of water and Vf > 0,
then Vo > Vi and co/ci < 1. With induction of hypercapnia,
Vo transiently increases (at tiv) and co/ci transiently
decreases (at tic), but both converge (at tfv and tfc,
respectively) on their room air breathing levels. These
transient responses, identified by shading in Figure 2, can
be quantified by determining their areas; the amount of
transiently displaced CSF (Vd) is defined as:

t . Vo + Vo
Vd = fV Vodt - [ €021 ra1 (tf - t-
t 2 V

 

)

1v
1v
- [Vora (l + tiv) - Vocoztiv] (6)
where:
Vd = volume of displaced CSF (pl)

84

FIGURE 2.

Representative time courses for CSF transient responses to
hypercapnia. Shown are Vo (cisternal outflow rate, ul/min),
Vi (perfusion inflow rate, ul/min), and co/ci (ratio between
inflow and outflow tracer concentrations, dimensionless) for
room-air (Phase I) and C02 (Phase II) breathing. Transient
responses of V0 (Vd; ml) and co/ci (Cd; min'l) to hyper-
capnia are defined over time interval tiv (initial time for
expression of transient Vo response and onset of C02 breath-
ing, min) to tfv (final time for expression of V0 transient
response, min), tic (initial time for co/ci transient re-
sponse, min) to tfc (final time for experession of co/ci
transient response). Cr (min) is the residual area of co/ci

over the time interval tic to tfc.

 

 

 

 

85

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

v MI
)(I
)(I
(II
V0.» ~—---/ . .
W» a a
CT E E I i m e
p 7;L 1 1
° 'iv 'fv
IROOM AIR—+i596002 ,<
l - C O/ci
M ____
i
0 l #f/
O

 

FIGURE 2 .

 

86

tiv' tfv a time for onset of C02 breathing and

V0 transient, and for the end of V0

transient, respectively (min)

vora, Voco2 s steady state perfusion effluent rate

for room air and for C02 breathing

 

(ul/min)
and,
(co/ci) + (co/ci)
Ct '-‘ ra £2 (tf " t1 ) (7)
2 C C
where:
Ct = total area of co/ci assuming no dilution

during tic to tfc (min)
ti , tf = times for initial and final co/ci
c c
transient (min)
(co/ci)ra, (co/ci)co2 8 steady state perfusion
tracer concentration

ratios for room air and

C02 breathing

 

and,
(co/ci) + (co/ci)
Cd 3 [ °°22 ra1 (tfc - tic)
(8)
tr
- , _ c
(co/ci)ra (1 + tlc) -/:ic (co/ci)dt
where:

Cd . area of co/ci dilution during transient

period (min)

87

and,
Cr = Ct - Cd (9)
where:
Cr = residual co/ci area during transient period
(min)
and,
CSth = Vd (Cd/Ct) (10)
where:
CSth = amount of unequilibrated or tracer-free
CSF in Vd (n1)
Cd/Ct = area of transient dilution of co/ci

expressed as a fraction of the total area
of co/ci if no dilution occurred during

tic to tfc (dimensionless)

Integrations in equations 5 and 7 were executed by digitiz-
ing (model GP—3; Science Accessories Corp., Southport, CT)
the graphical forms of the outflow data and processing these
results on the PDP 11/34 computer (Digital Equipment Corp",

Maynard, MA; Appendix D).

SIAIISIICAL.ANALXEES

Tabular data are presented as means and standard errors
of the mean. .A paired t-test was used to evaluate
differences between group means of CSF transient response

data. Steady state data were evaluated using a t-test

88
designed for testing differences among group means having
unequal sample sizes; a t-statistic was computed using the
difference between group means, and the best estimate of the
true population variance (a pooled variance). A 0.05 level
of significance was chosen to test the null hypothesis and

was determined prior to experimentation.

RESULTS

Intracranial fluid volume is defined by a dynamic recip-
rocal relationship between intracranial cerebrospinal fluid
(CSF) volume and brain blood volume (BBV). Since the skull
rigidly confines these intracranial fluid volumes, any sud-
den change in one of the volumes will necessitate a rapid
reciprocal adjustment of the other. C02 inhalation is a
potent cerebral vasodilator, and is known to increase brain
blood flow. Ventriculocisternal perfusion studies using
tracers have shown that C02 breathing also causes immediate
intracranial CSF volume and pressure changes, presumably due
to increases in BBV.

As shown in Figure 2, when the animal breathes room
air, and CSF pressure (PCSF) is less than approx 10 cmHZO,
ventriculocisternal perfusion outflow rate (V0) is larger
than inflow rate (Vi) because CSF is being formed continu-
ously within the brain's ventricles. Since the nascent CSF
is free of tracer molecules, it dilutes the tracer in the
perfusate such that co/ci, the ratio between outflow (co)
and inflow (ci) tracer concentration, is less than 1. With-
in moments of breathing C02, although perfusion inflow rate

(Vi) is constant, there is an immediate and short-term

89

90

increase in PCSF and V0, and a transient decrease in CSF
effluent tracer concentration (co/oi).

Figure 3 presents data from one animal which show these
characteristic changes in cisternal effluent tracer
concentration (co/ci), perfusate volume flow rate (V0), and
CSF pressure (PCSF) during ventriculocisternal perfusion
when the animal spontaneously breathes room air and then a
COz-in-air gas mixture. In this particular experiment when
7.2% C02 was inhaled at time - 28 minutes, Vo immediately
increased from 155 ul/min to 340 ul/min and PCSF increased
from 5 cmHZO to 7 cmHZO, and both gradually returned (approx
10 min and 3 min, respectively) to their pre-hypercapnic
levels. Further, after a brief delay (approx 2 min)
following hypercapnia induction, co/ci declined from 0.81
to 0.65 and returned to the pre-hypercapnic level within 12
min. These data are interpreted to indicate that
respiratory acidosis has no effect on steady state rates of
CSF production (Vf) and absorption (Va), but does displace a
volume of CSF (Vd) which is relatively tracer-free (CSth).
This transient expulsion of dilute CSF could be due to an
episodic increase in Vf. Alternatively, if tracer molecules
are heterogeneously dispersed in the CSF spaces, it could be
due to a shift in intracranial fluid volumes, with
consequent expulsion of tracer-free CSF into perfusion
effluent. Data from similar hypercapnic tests conducted in
experiments designed to investigate the generation of the

CSF transients are presented in the following sections.

91

.cofiuosccu mficmmouoman meadow

AcumEo b on my mummuocfi unmanned» unmwam m bosonm mmum .MHcmmouoman mo cofiuosccw

on» nouns cfia a haoumefixOLQQm bonusooo ucowmcmuu fi0\oo on» no mcwccammn one .cfie

NH away“: Hm>mH sachOLomanlwum as» ou codename can mm.o ou Hm.o soda pocaaomc «0\oo
.o> c“ mafia coppsm was» ccfihcmceoooc .Hm>ma mficmm0ummasloum on» on Ange ca xouccmv
nusuou Hannahm m an boonHou cfiE\H1 own on cfiE\Hd mmH scum o> c“ ammouocfl oumapmssfi
cm was mean» Ange mm u mefiuv mficcwouommn mo cofiuosccfi spas .NOO ww.> conusmun

assess any nuances on 0» mm Eoum «has Econ cmzumouh Ame umov Hmeficm on» oaflns cmumwxm
mumum memoum cowmsmuom m nauseae mm ou o scum .QEoc m>fiuclmmcHu>m m an A:AE\H1

mma .Eocv ucmumcoo cam: was AH>V mums scams“ acawsmumm .mous:HE ca beau ma mmmfiowhm
one .AONmEo «mmumv musmmouc mmu can «A:HE\H1 ao¢v mums ucosfimmo mmu «AH0\ooV

soaucd :oamswuom on» a“ away no :ofiuomum m we commoucxm ucmsammw mmo ca c0wumuucoocoo
cmuuxop “Asouuoh on do» Eoumv mum oumcwcuo on» so cmucmmmucom .mmanmfium> cofimsmuwc
Ammuv madam HacficmOLQwuoo co umwu owccmoumcmn a mo muomwmm oumum momwum can ucmfimcmLu

on» one noncommuc .:0nm=uuom Hmcuoumw00H50wuuco> ObnumucmmoHQOH o scum sumo

.m MKDGHm

92

on omen 0v mm mm
F

QM

NM

0
'0

mm

ONO. O

 

1.].in . . .
A656. 92....

‘

 

 

q

4‘

i

q

q
d
1
- -----.... '1

.8 .x. me.

A...

 

 

I

NI.Eu.

umul

.0

to £00.

.562...

LJJJJAALAALIIAJAJJ+JLJJJAJ

 

.m 956:

w.
00.

0?.

OO-

CNN

owN

Don

own

93
Analyses were made not only for CSF transient responses
(refer to MATERIAL AND METHODS: MATHEMATICAL ANALYSIS AND
CALCULATIONS; TRANSIENT RESPONSE ANALYSIS), but also for
steady state respiratory, cardiovascular, and CSF responses
(refer to MATERIALS AND METHODS: MATHEMATICAL ANALYSIS AND
CALCULATIONS; STEADY STATE ANALYSIS).

PROTOCOL A: Effects of High CSF Pressure on CSF Steady

State and Transient Responses to Hypercapnia

In this series of perfusion experiments, animals
spontaneously breathed room air and then a 7.2% COz-in-air
gas mixture first while CSF pressure (PCSF) was normal (nom.
0 to 5 cmHZO), and then again after their PCSF was
experimentally raised (nom. 18 cmHZO). Minute ventilation
(VB) and Paco2 increased and arterial pH decreased
significantly (P<0.05) indicating that respiratory acidosis
was established during C02 breathing both before and after
PCSF was increased (Table 1). The increase in VB was due to
a significant increase in tidal volume (VT, p<o,os) while
respiratory frequency (f) did not change (P>0.05). Blood
pressure (MABP), cerebral perfusion pressure (CPP; CPP a
MABP - PCSF), and body temperature (Tr) were unaffected
(P>0.05) by either the respiratory acidosis or the increased
PCSF.

Analyses of the steady state CSF responses to room air
and C02 breathing under normal and high CSF pressure

conditions (Table 2) revealed that increasing PCSF to

mcoduwccoo ouswmoum mmo anon Lou mo.o
u NooHa can: «nos» scum .mo.o v av neonauuse saueeoauseman .~.n a woven can: neafln> a

managed m you saw H menus but mosHm> m

 

 

o.~ « and... 86 a n8... mm. a no; Em H nmdm e H on .6 .« ..am a u no. a w a: an.» . ~82
ad « Twm 86 a 3;. me H «mm a... a «.2. m u - «a a 9% Z a 8. S a a: 8.0 . ~82
anon new:
A.

9 p... w n~.mm no... a. as... -~ .« a8... a... H and: a u. on ~.o a 9% a u on. a H m... an... . ~82
M... a 98 «a... u an... 8 u 3... a... H Tm... m H mm «a .« 9% 2 u :— 2 u a: 86 . ~82
eaaasee.

Argosy Anaemnaa-xnnv awakens-V A.e«a\nesaoenv Aces Aeneas Antony

Noon. aza us as e as see ans:

 

.sammumv ousmmmum Emu so“: car finance um c0umsmuoc an:
nuwumfiooasoauuco> mcausp Am~.h u mouHmv maccmouocmn can Amo.o u Noummv IoEuoc

ou noncommwm .A< HoooDOLm. numb uoasomm>oflcumo can muoumLficmwu mumum zcsmum .H mqm¢9

95

whammmum

mmo Hmsuo: mo mmosu scum Amo.o v my ucmummuwc maucmofiuficmfim mosam> whammmum mmo scam n

waamofiumnucmumm csonm mameficm mo umhfisc

s
0

:mm + memos mum mos~m> m

 

vao.. .u o.—m

Amvm.s .« ..=m

ASN.3 H —.mm
Ago; H v.9.

33$ H o.mm

Amvp.e .« o.e~

33.5. H Tam

“Neo.~ .« o.om

Anvee.o.« .o.o

varo.o.fl mo.o

Amvoo.o.fl mp.o

Amvwoé H N05

3:6 H m. pmp

Agofi. H n.NNP

93.0 H mim-
SZ..o H 2.35

Asvm.o.n s.am.
asvm.e.u s.am.

“5mg. H {amp
33.9.0 H 3.2:

Ames.o.« s.c.
“mvo.o.u m.e_

ASN.9 H 9...

73¢... H EN

u~.~ u wouHm
N
uo.o u oUHa

mmom :3”:

an.» n wooHn

n
ae.o u ooHa

 

 

swam a<xmoz
2.2-2.5 2:33: C ASE}: V 256:: v “ONE—.3
u> ea «exec no» «> names
.efiamumv enannend emu can:
can HmEuo: um cofimsmuwm Hanuwumwooasofluucm> ccwuso Awm.h u NOUHmV scammOmewn
can “wo.o n NoonV noELoc on nmncomnem .Aa Hooouonav when emu eueun seeeum .N mamas

96
approximately 17 cmHZO significantly decreased perfusion
outflow rate (Vo) during both normocapnia and hypercapnia
(P<0.05). Moreover, there was a tendency for steady state
CSF absorption rate (Va) to increase under the high PCSF
conditions. Steady state Vf and co/ci, however, were
unaffected (P>0.05) by either the respiratory acidosis or
the increased PCSF (Table 2).

Figure 4 presents data from one animal which are
representative of the CSF steady state and transient
responses to hypercapnia under the two pressure conditions.
The normocapnic animal was in a perfusion steady state
(i.e., co/ci and V0 were stable) after which hypercapnia was
induced (FICO2 = 7.2%; time = 28 min). When PCSF was normal
(light tracing), V0 immediately increased from 155 ul/min to
340 ul/min and PCSF from 5 cmHZO to 7 cmHZO, and both then
gradually returned (approx 10 min and 8 min, respectively)
to their prehypercapnic levels. Further, after a delay of
approximately 2 min when FICO2 was increased, co/ci declined
from 0.81 to 0.65 and returned to the prehypercapnic level
within 12 minutes. Under the high PCSF conditions, Vo
increased from 139 nl/min to 260 ul/min and PCSF increased
from 16 cmHZO to 18 cmHZO, followed by a gradual return
(approx 10 min and 3 min, respectively) to their
prehypercapnic levels. Two minutes after hypercapnic
induction, co/ci declined from 0.85 to 0.80 and returned to

the prehypercapnic level within 13 minutes.

97

.nmfin was a“ cons cmnu HmEuoc was mmom cons uwummuc

mums H0\oo ca ammouomp can o> CH mmmmLOCH ucmwmcmuu may .cwmcmaocs mums mmmusoo

65“» Leona coconua< .ucosflumm emu on» cw cowumuucwocoo cmuuch CH ammouomp ucwfimcmLu
a new .o> cg ommmuocfi unmanned» cam mumfiwwEEH cm .mmoc c“ antenna“ ucmfimcmLu

cam Hanan m maco boos©0um mficchLwcmn .sma: no HMEuoc was mmom umnuwnz .ONmEo ma

0» ONmEo N scum comamu was mmum cons unnumcoo nonwmewu w0\oo can commouooc o> mumum
wcmouw .oacmmUOEuoc mm; assess any cons .Acfie mm u mefiuv ~00 wm.h conummun.amswcm
on» cogs pounce“ was decadence»: scans bonus “awe mm 0» c oumum museum cofimsmbmc m :w
was Ame umov HmE«cm OMCQmUOEuoc one .mEsm w>wuplmmcfiu>m m wn “afiE\H: mmav ucmumcoo
camp was ~H>v muse soamcfi scamsmem .Ammmfiomhm «away mafia mo cofiuocsm m mm cmuuofic
can Asouuon ou co» Eoumv mumcwcuo on» no coucmwmucmu mum AONmEo «mmocv musmmmuc
mmo.©cm .ACHE\H1 ao>v muse ucmsammw mmo .AHO\oov scams“ scamsmumc on» Ca umnu

mo :ofiuomum m mm commwucxw ucwsammm mmo a“ :ofiumLucmocoo swuuxmo .Ammcfia xumo “Ommeo
may amen on on cwumsnpm waamucmefiuwcxm was DH :mn3 can» can Amwcfia unmflfi «ONmEo m 0»
av HMEuo: was Ammomv musmmwuc mmu mafia; umbwm cmuospcoo mums mummy oficmmoumc>m .At

Hooououmv scamsmuum Accumumwooasowuucob unannoum mmu can: ObaumucomOHQOH a mo measmom

.v mmDUHm

98

NO OO 00 On

OO. 0% 00 N
11‘! 4 1 d d ildi d 4

0n No On 0

\BOCON O

 

‘ d. ‘ ‘ ‘ d ‘ ‘

..... 1r

9. coon ov mm mm en mm on am one. o

 

all-1.1T. . . .4

“.225. NSC.

 

LII

 

.32
.9550:

 

 

wcammwca umu

 

 

 

LI in

 

 

‘ 14-: dilqlidl \eJIJIdIJ
I {III
83...... umua
«00$ «.5 A... :9 E2:

Hi...

 

3:62.:

 

LlllljllllAlliLllAjjlllLlllJ

 

.: 950E

O

O.

ON

00.

O'-

OO.

ONN

OON

OOn

OVn

a.
h.

99

Data reported in Table 3 indicate that although the
total transient co/ci area (Ct; Eq. 7), the time for the V0
transient (tfv - tiv) and the co/ci transient (tfc - tic),
and the time between the onset of Cd and the initiation of
C02 breathing (tic - tiv) were unaffected by increasing the
PCSF (P>0.05), the amount of displaced CSF during the V0
transient (Vd; Eq. 6) and the area of the transient co/ci
dilution (Cd; Eq. 8) were reduced (P<0.05) by 60% and 50%,
respectively. Under the normal PCSF conditions, an average
of 81.4 ml (CSth) of the expelled CSF was unequilibrated or
tracer-free while only 19.1 ul of Vd remained tracer free
after PCSF was increased to 18 cmHZO. This corresponds to
an 11% reduction in co/ci when PCSF was normal and a 5%

reduction when PCSF was high (Cd/Ct; Table 3).

PROTOCOL B: Effects of StOp-Flow on CSF Steady State and

Transient Responses to Hypercapnia

In this series of perfusion experiments, animals
spontaneously breathed room air and then a 4.5% COz-in-air
gas mixture before and after a 45 min period of stop-flow
(S-F) perfusion. Minute ventilation (VB) and Paco2
increased and arterial pH decreased significantly (P<0.05),
indicating that respiratory acidosis was established during
C02 breathing both before and after S-F (Table 4). The
increase in VB was due to a significant increase in tidal
volume (VT; P<0.05) while respiratory frequency (f) did not

change (P>0.05). Blood pressure (MABP), cerebral perfusion

100

anon Hesse: do each acne Amo.o v av heeneuuae saueaoamaeman eases anon gene a

mHHmoHumnucmumm csonm macawcm mo Logan: «2mm H memos mum mosam> m

 

“~36 H —.o— sync... H mo.o 32.6 H -.o

.mvoéh H 3.8 33o... H :6 3:6 H a.»

Amvmé H 0.0 Amvmé H m.~ Amvmé H ~.~p 8:6 H m.o Clo.— H m.: 3:.me H pJom amon—
:8:

ago; H o6 25.9 H m.~ “Cm; H 5.3 Amvmd H o.— 2.3.0 H v.3 33.09 H 991. anon
4:202

 

Anny Anus.

Dunn—mo .625 .10

A55 .53 25.5 253 A55 :3

> O 0 O > >
so as n as as .. Cs n8 as u e» as,

 

moo mm.» mo :ofiumamncfi on

.mammomv musmmmuc mmo ace; cam HmEuo: Lops:

noncommmm .A< Honou0umv muse sonuso emu ucmfimcmLe .m mqm<9

11()l

.HIAw

euwumme

was

.enoueh. such new wo.o

u «OOH. ewes emcee seem Amo.o v EL sentences saueeoausemfin sm.e n NooHn gene neaae> h

haamowuonucmumm czosm manages mo Cohan:

s
a

:mm + wcmwe mum mosam> m

 

 

 

.av ..~.« am.o. any ~o.e.« a.~.s Tosca—m.u emmp. .o..o...« on... 1o.vm.« an A..v~.o_« ~.om Am_es_n so .m..m.« ma um.e u mooHa
has m.m.« a.sm Ame ~e.o_« mM.. .e.vmo .« Lee Lopes.. a. ~.o~ A.FV~.« em as.vp.o.« m.om an..m.« ea Ampvm_u ma no.o . NooHe
a.» ensue
.m_v..~.« np.ms Awpvwo.o.« nm~.~ Aspeem_.u asmm. nervo.m.« no.om .e.L~.u mm Aspep.o.« F.mm As_vm.« Lop Aepcm_q mo. um.. . Noon.
83a; « e.~m 3:86 a on.» 3:: « mmo 33...... A TE 2:... H 9.... 23.6 H Nam 2:: a no 25.. a no no... ... moo:
mum mzonmm

Agree. .maem.=«-\sav Anaemxsav A.ea-\naaeoeav Loos Aeneas Agrees

.mamlmv soHMImoum nouns cam ouommh cofimsmumc as:

N“

lumumfiopooHuuco> mcfiusc Awm.¢ moonv macmmuuocmn 0cm Awo.o u ouHm. IoEuo:

ou noncommmm .Am Hooououmv sump Hmasomm>oH©umo can muoumuficmmu mumum acmoum .v mam<9

102
pressure (CPP) and body temperature (Tr) were unaffected
(P>0.05) by either the respiratory acidosis or the stop-flow
procedure.

Analysis of the steady state CSF responses to room air
and C02 breathing before and after S-F (Table 5) revealed
that steady state perfusion outflow rate (V0), dextran
dilution (co/0i), CSF pressure (PCSF) and rates of formation
(Vf) and absorption (Va) were maintained at the same levels
(P>0.05) for normocapnic and hypercapnic periods both before
and after S-F. '

Data in Figure 5 from one animal are representative of
the CSF steady state (i.e., V0 and 00/01 were stable) and
transient responses to hypercapnia before and after S-F.

The normocapnic animal was in a perfusion steady state after
which hypercapnia was induced (FICO2 a 4.5%; time = 0 min).
Before S-F, V0 immediately increased from 87 nl/min to 146
nl/min and then gradually'returned (approx 10 min) to the
prehypercapnic level. Approximately 2 minutes after
hypercapnia induction, co/ci declined from 0.80 to 0:72 and
returned to the prehypercapnic level within 12 minutes.
After S-F, V0 immediately increased from 86 ul/min to

140 ul/min, followed by a gradual return (approx 10 min) to
the prehypercapnic level. There was again a delay in the
dilution of the dextran marker but co/ci was reduced from a
normocapnic value of 0:78 to a minimum of 0:74 after which
it returned to its prehypercapnic level within approximately

14 minutes. Although the transient change in V0 was the

103

2mm H memos mum mosam> m

 

 

 

m; H m... fin H .2: no... H 85 En H TS is H ..:. To H Tm m um... u ~82
~2— H m.m Em H m4; mo.o H $6 ed H Ema :6 H .32. m.o H m.~ m 36 u moo:
2am zmam<
o.—.H ..o o...H u..~ ~o.o.H sp.o ..~.H o.mo m.o.H m..s m.o.H m.~ _. um.e u «ovum
m...H o.o o...H m.o~ ~o.o.H ss.o a...H m.~a m.m.H m.ae ~.o.H o.~ o. no.0 u NooHa
alwmfixfim
2.323 2.323 Guile 2:51: Emmaev
as ea Hexoo o> u> amen e
.elaumv scum noun neuue
cam ououmn :0uwsuu0m Hmcuoumu00H20uuu20> ocuusc Awm.v u Noonv mucmmouocmn
new Amo.o mouHmv loanoc ou oncocmom .Am Hooououmv sump Emu mumum acmoum .m mqm<9

 

104

.Amoum pmcuuum mnu >2 czonm may mum ououon

cmnu uouum mmou was .u0>030n .u0\oo nu oncouooc ucoumcmuu one .muspooouc mum mnu

an omuoouumcs mums u0\00 can o> mo mHm>0H mumum wbmoum onu can .momcocmou ucoumcmuu
onu mo momusoo osuu onu .o> Cu masouocfi ucoumcmLu one .u0\0o cu ommouooo ucoumcmuu
m .ncue N xoumcmv amaob umuuh m nouum .ccm o> cu 0m00u0Cu mucuceru cm comsmo

uu was new couummcH mnu ou moo wm.e acumen >h Anus o u mauuv coosccu mm: nucmmouocxm
.mnm umumm can muomoh nuon Anus o ou mlv oumum acmmum coumsmuoc 0 cu was Ahmu uwov
unaucm ouccmooeuoc one .cEsc m>uuplomCHu>m m an AcuE\H: «by ucmumcoo pawn was Auov
oumu soaucu :Oumsuumm .Ammmuomhw «quay mEuu mo couuocsu H mm oouuoac mum Aoumcucuo
Eouuon “CHE\H1 uo>v mumL ucmsamum mmu mam AmumcuCLO mou «u0\oov 30HMCH c0um=muwm

onu Cu umnu uo couuomuu m mm commoucxo ucosauuo mmv nu couuMLucwocoo cmLuxoa .Amlmv
soaulmoum uo couuoc sue me a AmocHu unabv umuum can Amocfla uncuav ououoh pwuospcoo

mums mumou oucmmou0c>m .Am H000u0um. :0umsuuom DOHHImoum obuumucomoumou m we muusmom

.m MMDOHb

105

35833:» .m "22on

ON 0. Op
Ifi fl 1

an
——---#O

aOo 8.3.1 :- see.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.
, loo.
. on.
.0:( ll
0.0.00 ..|ll "
" low.
302 no; .
" 2.5.3
0>
- O
. L
.
. J
.
. i.
.
.
. 'nll,
. .
1“ E! no
" _0\00 I

 

106
same before and after S-F, that for co/ci was less after S-F
when compared with before S-F.

Data reported in Table 6 summarize results obtained
using st0p-flow in 8 animals. They indicate that although
the volume of transiently displaced CSF (Vd), the total
transient co/ci area (Ct), the transit times for Vd and Cd
[(tfv - tiv) and (tfc:- tic), respectively], and the time
between the onset of Cd and the initiation of C02 breathing
(tic - tiv) were unaffected by the S-F procedure (P>0.05),
the dextran dilution (Cd) was reduced (P<0.05) by 50%.
Before S-F, an average of 24.1‘u1 (CSth) of the expelled
CSF was unequilibrated or tracer-free. Only 10.3 ul of Vd
remained tracer free after stop-flow. This corresponds to
an 8% reduction in co/ci before S-F and a 4% reduction after

S-F (Cd/Ct; Table 6).

PROTOCOL C: Effects of Phenoxybenzamine on CSF Steady State

and Transient Responses to Hypercapnia

In this series of perfusion experiments, animals spon-
taneously breathed room air and then a 7.2% COz-in-air gas
mixture before and after an intravenous injection of
phenoxybenzamine (P82) (2 mg/kg). Minute ventilation (VB)
and Paco2 increased and arterial pH decreased significantly
(P<0.05) indicating that respiratory acidosis was estab-
lished during C02 breathing both before and after PBz (Table
7). The increase in VB was due to a significant increase

(P<0.05) in tidal volume (VT) and respiratory frequency (f).

107

.mlm cOHOMOD: MO “$5.... EOHM Amoco V mv HCOHOMMHG ewHUCwOHMHCDfim OSHO> .mlm :HOUM‘: D

.mumeucm w you :mm H memos mum mosHm> m

 

s.m .H m.o. .o.o.H so.o o...H ..p ....H a.» m.o.H m.m s...H m.a ..o.H =.o m...H m.m s.pm.H m.mmm mum
emcee
o.o..H P.s~ mo.o.H oo.o p.o.H a.» o.o.H o.w m.o.H m.~ ..F.H a.o. m.o.H m.e s.._H a.» F.2m.H s.mmm mum
mzoamm
31v 253v 253 23.5 25.5 $3.5 25.5 :43
> O O O > >
Chemo euoxeo no so us - us H» u as see as u e» as

 

.mumlmv soawlcoum Loumm cam ououon

N8 em.e mo couueueeeu on neneoenem .Im Honouonav when aouuuao emu neeuneene .G mamas

108

Nmm cuwumflc HO OWOSH EOHN AmO.o V my ucwummmwc >HUCMOHMfir—Gflm MODHM> Nmnw cwuowwm: U

Nmm

cumum0e was

.enoumn. anon Lou mo.o

u mouHm cons muonu EOLu .mo.o v my ucououufic maucmofiuucmum mm.h n mouHm cons mos~m> n

aaamowuonucoumm czonm mHmecm uo amoeba

«2mm + memos mum mosHm> m

 

 

 

Team...« a.... .p.~a.a.« an... A.c....« ...m. Ase... .H as... Lev1.« as. use..a.« _.e. .e.. .H a. use. .a no .~.e . Na...
“use...H ¢.mm upvno.o.H o~.~ “mean .H see Imam.. .H o.mm has..H m~ Lev—.e.H o.mm .ev. .H on Lee. .H ea ue.o . ~ooH2
Nam zuam<
.m.....« as... Am.ma.a.H an... Amvsmn.u amoa~ Ance.~..u 2.... .mv~.H a_m .m...o.H ..em Ames..H o~. ance,.H an. u~.p . ~ooH.
neee.~.H ..mm .scma.a.H n... nev~e .H on. Leea.m .H a... Levm.H an .eee.a.H o.em .5.o..H an. 123°..H .m. ea.a . nova.
Nam uxomum

Annohv Ammannculxuev Ammhmnuav A.:«I\nnumonpv Aoov Annohv antoav

~82 J... as as L a. 9.8 one...

.maammv oeuEmnconwxoconm Louum cam ouomon c0umsuuom

HmcuoumuooHsOuuuco> mauusm Aw~.> u ouHmV muccmouocan cam Amo.o u Noonv IoEuo:
ou noncommom .AU HooouOLmv sump umusomm>ofi©umo cam auoumLucmou oumum acmoum .h agree

109
Both normocapnic and hypercapnic blood pressure (MABP) and
cerebral perfusion pressure (CPP) were reduced (P<0.05) by
30% after PBZ, while body temperature (Tr) was unaffected
(P>0.05) by either the respiratory acidosis or the P32
treatment.

Analysis of the steady state CSF responses to room air
and C02 breathing before and after PBz (Table 8) revealed
that after PBz the 14C-dextran perfusate concentration
(co/ci) was increased (P<0.05), and the rates of CSF
formation (Vf) and absorption (Va) were significantly
decreased (P<0.05), while PCSF and perfusion outflow rate
(V0) were unaffected (P>0.05) by either the respiratory
acidosis or the P82 treatment.

Data shown in Figure 6 are from one animal, and are
representative of the CSF steady state and transient
responses to hypercapnia before and after PBz. The
normocapnic animal was in a perfusion steady state (i.e., V0
and co/ci were stable; time -5 to 0 minutes) and hypercapnia
was induced (FICO2 = 7.2%) at time = 0 minutes. Before PBZ,
V0 increased form 160 ul/min during room air breathing to
395 ul/min following hypercapnia induction, and then
gradually returned (approx 9 min) to the prehypercapnic
level. Further, after a brief delay (approx 2 min), co/ci
declined from 0:79 to a minimum of 0.50 and returned to the
prehypercapnic level within 12 minutes. After PBZ, V0
increased from a normocapnic value of 158 ul/min to 370

ul/min immediately after the onset of hypercapnia and then

110

Nmo .ooouoo. mo onoao eoou Ame.o v as oeooouuuo sauceouuueoun mosses umo

Cumumcpp D

waamowuonucouom czosm manages uo Embasc «2mm + memos oum mo=Hm> m

 

 

 

AGO—.~.H.m.o onp.~.H m.P~ Amvmo.o.H mo.o onm.m.H p.9mu “ovm.o.H p.5m. “eva.o.H a.~ n~.s u moon
23.. H m.» E~.~ H n.- 2:86 H 85 2:; H Tr... Gem... H 92. Send H a; no... ... N82
22 22...:
3:2 H ET 3;... H 22 2:86 H 85 222 H GET 3:... H 92. 5.6 H ~.~ 2... u ~82
2%.. H Tm. 2:2 H ~..m 2:36 H ~26 3.0m H 3...; 826 H 2.9 326 H ..~ 86 u ~82
uma seamen
2:53 3253 351C 2255
on... 4:» $38 2.. 2., 282
.oammmv ocHEmuconmxocosm uoumm
cam ououon conomuom HmcuoumfiooHsofiuuco> cowusc Aw~.b u moonv owcmmouomhn
can Amo.o u ~00Ha. loanoe oo monsoonom .20 Hooouonnv when amo ououn seeoum .m mamas

111

 

.mp0uuoc owccmouomws com oficmmooEuoc nuon
mauusp Aem.o ou mh.ov Nmm uouum uoumouo >HucooHMHcmHm ouoz .uo>030n .fi0\oo mo mHo>oH
ououm mcmoum one .Aumm ououohv mooao> uncommon ucoumcmuu H0uucoo onu Eouu ucououmau

amucmowuficmum uoc mums momusoo osuu was mocsuwcmme uuosu .Nmm uouum onuaoou ou
wococcou m cm: uo\oo com o> nu mowcocmou ucoumcmLu onu smsonua< .wo\oo cu ommouoop
ucoumcmuu o .Acue N soummmv amaoc uofiun m nouns can .o> cu omoouocH oumHmoEEu

cm comomo uu can mom woufimmcfl osu ou Nou mm.v caucus an Ange o u oefiuv cocoon“

mos sucmwouommm .c0uumuumwcuecm Nmm Louum new ououoh nuon Anus o ou mlv oumum acmoum
coumsuuoc m CH mos Amu umov HoEucm OHCQMUOEuo: one .cEsm o>wuuloccfiu>m m an A2HE\H1
hmav ucmumcoo cam: mos au>v ouou soHMCH coumsuuom .Aommuomhm «away oEuu uo couuocsu
m an pouuoam mum noumcficuo Eouuon «cuE\H1 ao>v oumu ucosauuo emu cam Aoumcfiwuo

cou «u0\oov soaucu coumsuuoc onu cfl umnu uo cofiuoouu o no commoucxo ucosauuo

mmu cu couumuucoocoo couuxoa .A>u .mx\me m .Nmmv oeuemuconwxoconc mo couumuumucuecm
Amocua xumcv Louuo cam “monua uzmfiav ououon pouosccoo.ou03 mumou oflcmm0uom>m .AU

Hooououm. :0uusuuom Hmcuouuwoou=o«uuco> onwamusonhwocoam o>uumucom0umou a mo muHsmom

.@ NMDUHh

112

A 9.25 m2:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o~ o. o. o o o- 0.-
«I 4 H q q A J
_
.8 an. All ... a...
.
— J
. .400.
1
n on.
IJ
h o-
— IA
— 1 O0“
- 1
. l 00»
. 1
. 4 010
L...( . .
22.0 A 55:3 4
. . 1 can
uzmazuc» . 2., L
wz.a<~zwn>xozuxa _
_
l O.
l ..
_ .030 L

. m mmDUHh

113
gradually returned (approx 15 min) to the prehypercapnic
level. Following a 2 minute delay, co/ci declined from 0.84
to 0.60 and returned to the prehypercapnic level within 17
minutes.

CSF transient response data collected from 5 animals
that were injected with PBZ (2mg/kg: iv) are summarized in
Table 9. The data indicate that the total dextran transient
area (Ct). the time between the onset of Cd and the
induction of hypercapnia (tiC - tiv). as well as the volume
(Vd) and dilution (Cd) transients and their respective time
courses [(tfv - tiv) and (tfc - tic)] were unaffected
(P>0.05) by the PBZ treatment. Before PBZ, an average of
119.2 pl (CSth) of the expelled CSF was unequilibrated or
tracer-free, representing 15% of Vd (Cd/Ct; Table 9). After
P32, 11% of Vd, or an average of 83.9 pl, was tracer—free.
Like the Vd and Cd values, the values for CSth and Cd/Ct
tended to be reduced after P82; in all but one animal, these
values did decrease to some degree after PBz. These
differences were small, however, and the variability in CSF
transient responses among animals was large. Apparent
differences in CSF transient data from before and after PBz

were insignificant (P>0.05).

114

waamowuwnucwumm csoam mameficm mo umnEs: «2mm H mcwwe mum mosam> m

 

A.c=.o~.« o.mo Azcmo.o.« ...c ..vo.o.« ...— Aac~...« o.~F “mv~.o « m.. Amcm.._« m... “mom.o_u 2.. Anya...» c.,. Am....~._u p.mmo no;
cause

gym-.mw H N5: Asvmoé H .36 Anvm; H 06 33'.— H ~.~.. Amvmé H 0.. 2;... H 0.9 33.0 H ~2— AmvhJ H 0.: “30.29 H m6: Nam
neommm

 

Aaav Asa-V Ana-V Asa-V “cu-v Ann-v nan-V A—av

t O O O > >
guano ao\vo to so «a u «a «a u u» no «a - c» as

 

.wANmmV wcfiEmncmbmxoconm nouns can muomob

Nou w~.~ mo cofiumflmacfi 0» momcommmm .Ao Hooououmv some soauuso mmo ucmfim:Mue .m mamas

DISCUSSION

In 1954, Tschirgi et a1. administered a C02 gas mixture
to anesthetized cats and observed a rapid and transient rise
in the rate at which CSF drained from a cannula in the
cisterna magna. Subsequent ventriculocisternal perfusion
investigations using tracers have shown both an immediate
transient increase in perfusion outflow rate (60) with
hypercapnia induction, and a decrease in CSF effluent tracer
concentration (co), even though perfusion inflow rate (Ti)
and tracer concentration (ci) were stable (Martins, et a1.,
1977; Songer et a1., 1980; Fisher et a1., 1983). This
displacement of a relatively dilute CSF during early
respiratory acidosis is attributed to two phenomena. One
explanation assumes a homogeneous distribution of tracer
throughout the CSF spaces and that CSF formation rate (6f)
episodically increases during C02 breathing. Both the‘
volume and composition transients would be accounted for
were there a volume of newly formed CSF injected into the
test perfusate. The second explanation allows heterogeneous
dispersion of tracer in the CSF, an unchanging 6f, and is
based on a redistribution of craniospinal fluid volumes with
C02 breathing (Martins et a1., 1977, Songer et a1., 1983).

The increase in brain blood flow and volume during

115

116
hypercapnia (Reivich, 1964; Rich et a1., 1953; Grubb et a1.,
1974; Adams et a1., 1980) empties CSF into the perfusate
from spaces that were inadequately mixed by convection and
diffusion. Both phenomena, of course, may contribute to the
short-term displacement of a dilute CSF during hypercapnia.
A major goal for the present eXperiments was to establish a
data base to select between these explanations.

The mass balance equations developed for
ventriculocisternal perfusion tests (See METHODS; Equations
l-S) accurately assess CSF dynamics only when the rates and
tracer concentrations of perfusion inflow and outflow are
stable and when craniospinal fluid volumes are unchanging.
CSF transient responses (see Figure 2) to stimuli which
perturb the craniospinal fluid balance must be analyzed
using alternative methods. In the studies reported here,
equations were developed to assess transient CSF responses
to hypercapnia during ventriculocisternal perfusion. These
equations, which were derived from the relationship between
and the time courses of CSF outflow rate (To) and tracer
concentration (co/ci), permit the quantification of the
volume of perfusate (Vd) and tracer-free CSF (CSth)
transiently extruded from the cisternal cannula during early
respiratory acidosis, and define the amount of transient
dilution (Cd) associated with Vd.

Ventriculocisternal perfusion, which is widely accepted
for measuring the rates of CSF formation (6f) and absorption

(Va), was used in these experiments. This technique

117
involves placing cannulas into the lateral ventricles for
artificial CSF inflow, and positioning a large-diameter
needle in the cisterna magna for measuring outflow (Figure
78). Because the cisternal cannula provides an alternative
flow route of low resistance in parallel with the
subarachnoid spaces, normal CSF flow distribution (Figure
7A) is disrupted. During ventriculocisternal perfusion at
normal pressure, both newly formed CSF and perfusion flow
are shunted preferentially along the highest conductance
course, that through the needle in the cisterna magna. Flow
rate through the subarachnoid spaces increases, however,
when CSF pressure rises, since high CSF pressure increases
CSF bulk flow through the arachnoid villi (Welch and
Friedman, 1960; Pollay and Welch, 1962). Because mixing of
tracer in the total CSF volume during ventriculocisternal
perfusion studies depends primarily on convection, enhanced
subarachnoid CSF flow will increase the rate at which
tracers are distributed by mass flow in these spaces.

CSF transient responses to hypercapnia were measured
before and after a period of facilitated flow through the
subarachnoid space. Subarachnoid perfusate flow was
augmented either by impeding cisternal efflux of perfusate
by raising the level of the cisternal outflow cannula (i.e.,
high CSF pressure; Protocol A), or by preventing CSF outflow
by clamping the cisternal outflow cannula (i.e., stop-flow

perfusion; Protocol B).

118

The question of whether incomplete mixing of tracer
throughout the CSF spaces and/or an increased 6f accounts
for the CSF transient responses to hypercapnia also was
approached in another way. CSF transients were assessed
before and after the intravenous administration of a
vasodilating agent, phenoxybenzamine, which is a competitive
alpha-antagonist that binds irreversibly to both alphal and
alphaz adrenoceptors (Goodman et a1., 1980). Chemical
denervation was chosen over surgical denervation as a method
of achieving cerebral vasodilation since it has been
demonstrated that not all adrenergic fibers supplying brain
parenchymal vessels and the choroid plexuses degenerate
after denervation of sympathetic ganglia (Lindvall et a1.,
1978a; Edvinsson et a1., 1975), and because there is a
source of adrenergic fibers originating in the locus
coeruleus (Hartman et a1., 1972; Raichle et a1., 1975;
Rennels and Nelson, 1975). Cerebral vasodilation was
produced by phenoxybenzamine prior to a hypercapnic test in
order to evaluate the contribution of sympatho-adrenergic
activity to the craniospinal redistribution and mixing pro-
cess to the CSF transient response thereby unmasking any
transient change in 6f associated with early respiratory
acidosis.

Results from stop-flow (S-F) experiments (Table 6)
indicate that at least 50% of the transient dilution of
cisternal outflow associated with C02 breathing during

ventriculocisternal perfusion is due to the craniospinal

119
shunting of CSF unequilibrated with tracer molecules. The
transient dilution of dextran (Cd) decreased by 50% after
S-F, even though the total transient co/ci area (Ct) and the
volume of transiently displaced CSF (Vd) remained constant.
Further, Cd decreased during high CSF pressure perfusion
(Table 3). This supports the S-F finding that mixing of
tracer in the CSF spaces is incomplete during
ventriculocisternal perfusion at normal CSF pressure.

The time course for Vd and Cd [(tfv - tiv) and (tfc -
tic)' respectively], and the delay between the beginning of
Cd and the induction of hypercapnia [(tic - tiv); Figure 2)
were unaffected by both S-F (Table 6) and high CSF pressure
perfusions (Table 3). These data indicate that while the
s-F and high CSF pressure treatments enhanced convective
mixing of tracer throughout the CSF spaces, they did not
change either the reciprocal relationship between CSF and
brain blood volume or the time course of craniospinal fluid
redistribution during acute hypercapnia.

The decrease in Cd during high CSF pressure perfusion,
however, in addition to being due to convective mixing may
also be a function of a decrease in Vd (Table 3). This may
be due to a high CSF pressure-induced decrease in transmural
pressure of the pial vessels, which go through the
subarachnoid space. A decreased transmural pressure in the
highly compliant pial veins, might decrease venous
distensibility so that the C02 stimulus effects a smaller

increase in total brain blood volume which is smaller than

120
when CSF pressure and venous transmural pressure are normal.
A smaller increase in brain blood volume would induce a
smaller compensatory shunting of CSF from the cranium (i.e.,
Vd would decrease).

Although this explanation may be valid when
intracranial pressure is markedly increased, cerebrovascular
responses to inhaled C02 (FICO2 = 5%) are undisturbed when
intracranial pressures are less than about 45 Torr (Avezaat
et a1., 1980). Moreover, neither sequential bouts of C02
inhalation (Avezaat et a1., 1980; Lofgren, 1973) nor
physiological shifts of resting intracranial pressure
(Marmarou et a1., 1973) independently alter the craniospinal
fluid redistribution process.

An alternative explanation for the decrease in Vd
during high CSF pressure perfusion is that the elevated CSF
pressure, produced by raising the height of the outflow
line, increased CSF absorption (Ta) and increased cisternal
outflow pressure so that CSF outflow rate (To) decreased.
Although data in Table 2 show that To decreased during high
CSF pressure, they do not show a significant increase in Ta.
This may be due to the small number of experiments (n a 2).
Other studies demonstrate that increasing CSF pressure
causes an increase in 6a (Heisey et a1., 1962; Hochwald and
Sahar, 1971; Mann et a1., 1979; Davson et a1., 1970;
Marmarou et a1., 1978; Welch and Friedman, 1960). If there
were an increase in 6a during high CSF pressure perfusion,

the total volume of CSF leaving the cranium during acute

121
hypercapnia may be unchanged, but there may be more cranial
CSF flowing through the arachnoid villi, and less flowing
through the cisternal cannula than when the hypercapnic test
was conducted at normal CSF pressure.

These data indicate that overestimates of steady state
6f would result if measurements are made prematurely after
perturbing the craniospinal fluid balance (Haywood and Vogh,
1978). This does not invalidate, however, the use of mass-
balance equations for calculating steady state 6f and ya
(Tables 2, 5, and 8), because they are appropriate for
stable conditions of dynamic nonhomogeneity (Heisey et a1.,
1962; Heisey et a1., 1983) as they are for steady states
when perfusate mixing is complete and tracer distribution is
uniform.

Nonuniform tracer distribution occurs during
ventriculocisternal perfusion because of the technique and
the anatomy of the CSF spaces. The craniospinal CSF system
is a multi-chambered labyrinth in which CSF circulates along
a tortuous path from its sites of production in the
ventricles to where it is absorbed in the craniospinal
subarachnoid spaces. The normal CSF flow pattern (Figure
7A) is disrupted in ventriculocisternal perfusion because
the perfusate mainstream is directed from the cerebral
ventricles to the cisterna magna (Figure 7B). The perfusate
flow and newly formed CSF are directed away from the sub-

arachnoid spaces which reduces mixing in these spaces.

 

122

.ham>auoommmu .mmocmm mmo cfinuws can ou Hanuouxo mcofiuooufic
30am ucomoumou mzouum cwmoHo can some .moommm cwocsomHmnsm a“ mcfixfia can cmsousu

30am nuon mommmuocfi o AI: o> umsu om Hmccwso ucwsammm cofimsmumm on» msamEmHo no

.wommm cwocnocumnsm

on» a“ mcfixwe o>fiuoo>coo cocoon cfisos o> cmumfl xamumcofiuquOummww < .Ao>v mummsmuwm
ucosammo nomHHoo ou mamas maumumfio oucfi condoms“ canoe: oocmumfimwu 30H m cmsousu

omam pan .Amm> can om>v mousou cofiumuomnmou assoc: >9 moommm mmo w>mmH ou maoccmno
Hmficcuomuucfi can unasowuuco> smsouzu mmmm won» umnuooou can .mm> can om> no pan» 0»
upon much 30am wasao> muH .moaowuucw> Hmuwuca oucfl Aw>v mums accumcoo m an comssm ma

moasowaoe accouu mcficfimucoo mmo HmonMfluum .consmucm HmcuoumfiooHsofiuucm> mewusa um
.Amu>v mmufim HmcfiOuosomuuxm um can nom>v momsxmam QHOuono
an coauou m“ a“ gowns um mmonu cu Hmsvm mound um mwommm owocnocumnsm Amm>v Hmcfimm

can Aom>v anaembo scum conuomnmwu >Hm>fimmmm ma mmu waamEuoz "d .mCuouumm scam mmo

oh MfiDUHh

.m. magi

         

        

 

s .>
0
Snow: — ouonu\ @
32230 205.9052.» 0
f 3ch A R; :3...
AH - .ofiuzuotxo
[I
u u
. 0:32.. (A333...
30:5 3965

cocoa
N/I/ u_occoo.oa=n\\

3:38 R\R\

124

When CSF pressure is increasing during ventriculocis-
ternal perfusion either by raising the height of the cister-
nal outflow line (Protocol A), or by occluding the cisternal
cannula (Protocol B),icisternal efflux (60) diminishes
(Table 2) or stops, respectively, and circulation of
perfusate through the subarachnoid spaces increases. Con-
vective mixing is increased and the transient dilution of
cisternal effluent during early respiratory acidosis (Cd) is
reduced (Tables 3 and 6).

Data from the high CSF pressure and S-F perfusions do
not explain why there was residual Cd after convective
mixing was increased. It is unlikely that mixing was
inadequate, since the time for the initiation of dilution
[(tic - tiv); Figure 2] was the same before and after
periods of facilitated flow distal to the cisternal cannula
(Tables 3 and 6). Transit time for unequilibrated CSF would
be expected to vary as a function of its distance from the
cisternal outflow needle (Figure 7B). In addition, data
from an auxiliary study (n = l) in which an animal underwent
70 minutes of S-F, and then a hypercapnic test (FICO2 =
7&2%) during ventriculocisternal perfusion (Figure 8), indi-
cate that a transient dilution of dextran was still detect-
able even after nearly doubling the duration of the S-F
treatment. The percentage of Cd remaining could not be
determined since there was no hypercapnic test preceeding

the S-F period.

 

125

.wo\oo c“ owmcuocc ucomeMuu m .Acfie

N xoummmv umaow «camp a sound .ccm o> aw wmmouocfi ucofim:Muu can mumficmeew cm comsmo
mane .mwm ocufimmcfi may on Nov m~.h weapon an Acfie o u oefiuv pounce“ mm3 cacmmouwamn
con3 Ache o o» mlv mumum acmwum c0wmsmuom w ca mc3 Ava“ umov Hmeficm UHCQMUOEuoc

on» .mlm moum< .mEsm w>fiuolomau>m w an ACHE\H1 chv ucmumcoo mac: mm3 AH>V

much 30amca scamsmumm .Ammmwomnm «:fiEv 08“» mo cofiuocsm m mm kuuoam mum Aouwcacuo
Eouuon «cfiE\H1 «o>v much ucosammm mmo can AmuMchuo mo» «fio\oov BOHMCH cofimsuuwm

on» ca can» no :ofiuomum 0 mm pommwumxm unusammw mmu cfi cofiuMuucoocoo swuuxmn

.cofimsmumm Hanumumfioofisofiuucm> m mcfiusc omuosocoo ummu ofismmouomas mnu wommomum

Amlmv soHMImoum mo cofiumm cfiE on c .sofimsuumm toHulmoum womGOHOHQ a scum mama

om NMDUHh

126

35va2:

 

 

 

 

 

6. on on ma o... m. o. n o n-
I 1 d I d I q d - 1
_
_
.
.
.
«CO INK]. :: Eco.
.
.
.
.
.
_
1 E525
.
. O9
.
_
.
.
.
.
.
-
.
.
-
hlLlIlllul.‘ r. .
.
llLllL

.oxoo

J

.m mmDUHh

cap

on.

OON

 

 

127

One way to define mixing time would be to complete a
series of experiments similar to Protocol B, but use stOp-
flow periods with different durations. The transient CSF
responses induced by hypercapnia both before and after the
assigned S-F period would be analyzed for the volumes of
displaced CSF (Vd) and the associated dilutions of dextran
in the cisternal effluent (Cd). For those animals in which
Vd remained a stable and independent variable, a difference
between the before and after Cd values (dCd) would be deter-
mined for each experiment and plotted as a function of S-F
time. It is anticipated that for relatively short stop-flow
periods, dCd will be small and vary directly as a function
of S—F time. JFurther, it seems likely that as stop-flow
perfusion time is extended, the dCd will increase to some
maximum value and become independent of time. ‘At a certain
S-F interval, complete mixing of tracer in the CSF spaces
will be achieved and dCd will be at a maximum; further
prolongation of stop-flow*will no longer have an effect on
dCd. When dCd is maximized, the associated duration of S-F
would be equivalent to the time it takes to establish a
state of perfusion homogeneity. Then any transient dilution
of tracer (Cd) occurring with hypercapnia during ventriculo-
cisternal perfusion, and after a period of S-F, would neces-
sarily be attributed to sources other than the contamination
of the perfusion mainstream with poorly mixed CSF.

CSF transient responses to hypercapnia may not be due

entirely to craniospinal fluid redistribution. Increases in

128

CSF formation rate (9f) contribute to both Vd and Cd and
account for the residual dilution after mixing (Tables 3
and 6). It has not been demonstrated, however, that 6f
transiently increases with C02 breathing, partly because
equations for CSF hydrodynamics require steady states
(Heisey et a1., 1962; Heisey et a1., 1983). Also, the
transient 6f response to C02 has not been investigated
because steady state responses are assumed to reflect
accurately transient responses.

Portnoy et al. (1980) studied the interrelationships
among systemic arterial, CSF, and superior sagittal sinus
pressures, and the pulse waves of these pressures during
normo- and hypercapnia. They discovered that for the
”closed skull" preparation (idh, no ventriculocisternal
perfusion), transient and steady state cerebrovascular
responses to hypercapnia are importantly different. During
normocapnic and hypercapnic steady states, the CSF pressure
pulse wave, which reflects the cerebral venous pulse wave,
is transmitted to the superior sagittal sinus unchanged.
This suggests that impedance to cerebral venous outflow
provided by the intradural venous channels (lateral lacunae)
is minimal. Impedance to blood flow from the cerebral veins
to the superior sagittal sinus depends on the degree of
compression of the lateral lacunaeu Conversely, during
hypercapnia induction, the shape of the sagittal sinus
pressure pulse wave is different from that of the CSF.

This suggests an increased impedance to cerebral venous

129
outflow and an increased compression of the lateral lacunae.
Portnoy and co-workers ascribe the transient compression of
the lateral lacunae during hypercapnia induction to the
transient increase in CSF pressure associated with hyper-
capnia induction. CSF pressure increases because vascular
expansion is more rapid than CSF absorption (Ga), and the
increase in brain blood volume exceeds the capacity of the
spinal CSF reservoir to accommodate shunted cranial CSF
(Rich et a1., 1953; Marmarou et a1., 1978). {As hypercapnia
continues and 6a progressively diminishes the total CSF
volume (Martins et a1., 1972) to compensate for the expanded
cranial blood volume, CSF pressure begins to fall and
lateral lacunae compression is relieved.

Compression of the lateral lacunae during hypercapnia
induction retards drainage from its tributary intracerebral
veins (Portnoy et a1., 1980) causing venous congestion.
Venous congestion, in turn, is associated with an increased
extravasation of Evans blue from venules located upstream to
the congestion (Auer et a1., 1980). This suggests that
hypercapnia induction may be associated with a transient
' disturbance of the blood-brain barrier as well as an in-
creased capillary filtration rate. The increased rate of
plasma filtration may effect a transient and significant
increase in brain interstitial fluid flow into the brain
ventricles (Rosenberg et a1., 1982) and this may be an
extrachoroidal source of CSF (Rosenberg et a1., 1980). CSF

production at extrachoroidal sites may be regulated

130
primarily by the rate of capillary filtration while
choroidal CSF production is primarily an active process
since acetazolamide effects nearly a 100% cessation of CSF
production in the isolated perfused choroid plexus and only
a 60% decrease in 6f in 111g_(Pollay, 1975).

Whether or not the lateral lacunae are significantly
compressed in the I'open skull' preparation (i.e., during
ventriculocisternal perfusion) cannot be answered from data
presented here, nor has it been addressed in other studies.
Future investigations using systems analysis of CSF and
sagittal sinus pressure pulse waves during ventriculocister-
nal perfusion may provide valuable data regarding the
mechanisms underlying transient responses of the CSF system
to hypercapnia.

The CSF transient responses to hypercapnia, derived
from the craniospinal redistribution of CSF and possibly
from an episodic burst in 9f, do not depend on the
activation of alpha-adrenoceptors. Data in Table 9
indicate that all aspects of the CSF transient phenomena
were the same before and after phenoxybenzamine (PBZ)
administration. These data indicate that cerebral alpha-
adrenoceptors do not influence either the transient behavior
of 6f or the craniospinal fluid redistribution process
associated with the onset of hypercapnia. Since these
phenomena are intimately related to cerebral hemodynamics,
it is not surprising that investigations of the

cerebrovascular C02 response have yielded similar results.

131
Both surgical and pharmacological treatments to eliminate
brain perivascular sympathetic nerves and circulating
catecholamines (Dahlgren and Siesjo, 1981; IFAlecy et a1.,
1979; Hardebo et a1., 1982; Davis and Sundt, 1980; Waltz et
a1., 1971; Lopez de Pablo et al., 1982a; Kawamura et a1.,
1974) have left the COz-induced immediate and sustained
elevation of cerebral blood flow unaffected. Moreover,
other studies report a decreased response of cerebral
vessels to adrenergic stimuli (Page and Olmstead, 1975;
Lopez de Pablo et al., 1982b; Auer et a1., 1982) as well as
to some nonadrenergic agents (Altura and Altura, 1977)
during hypercapnia. Apparently hypercapnia can overwhelm
neural mechanisms and the influences of some non—neural
vasoactive substances such that an elevated Paco2 becomes
preeminent in determining cerebral hemodynamics.

The cerebrovascular C02 response is not dependent upon
an intact cerebral sympathetic innervation. The volume of
CSF transiently displaced from the cranium to effect spatial
compensation for an expanded intracranial blood volume would
therefore not depend upon alpha-adrenoceptor activation.
Since both Vd and Cd did not change after PBZ (Table 9) any
episodic increase in .Vf is not due to an alpha-
adrenergically mediated process.

Conversely, the steady state rates of both CSF
formation (Vf) and absorption (Va) were influenced by alpha—
adrenoceptor activity; Vf and Va declined after intravenous

PBZ (Table 8). The decrease in Va may reflect either a

132
decrease in the pressure gradient between the CSF
subarachnoid space and the dural venous sinus and/or an
increase in resistance to CSF outflow at the arachnoid
villi. Since CSF pressure remained constant after PBZ
(Table 8), venous dural sinus pressure may have increased
causing a decreased hydrostatic gradient favoring CSF
absorption. Moreover, an increased cerebral venous pressure
without a concomitant increase in CSF pressure could cause
Va to decrease via another mechanism. Since cerebral venous
pressure is that which surrounds the villi, and CSF pressure
is that within these valvular structures (Welch and
Friedman, 1960), a decrease in the trans-villus hydrostatic
pressure gradient may have occurred to cause a narrowing of
the diameter of the villi, much like the passive change in
blood vessel diameter effected by a decreased transmural
pressure. The decrease in villi diameter in turn, would
effect an increase in resistance to absorption and thereby
cause a decrease in Va.

A mechanism underlying such an increase in sagittal
sinus pressure after PBZ has not been proposed. It is
known, however, that venous sinus obstruction and conditions
such as subarachnoid hemorrhage or expanding cerebral tumors
are often accompanied by impaired absorption of CSF
(Marmarou et a1., 1975).

Unlike Vf and Va, CSF pressure (PCSF) and v.0 remained
stable after PBZ (Table 8). Perhaps the decrease in Va,

which would tend to increase PCSF and V0, was offset by the

133

concomitant decrease in Vf, which would tend to decrease CSF
pressure and V0.

The decline in Vf after PBZ may be explained on the
basis of at least three different phenomena. The first
explanation supposes that the PBZ-induced decrease in MABP
(Table 7) caused a significant decrease in blood flow to CSF
secretory sites which, in turn, caused a decrease in steady
state Vf (Table 8). This hypothesis stems from a study
conducted by Davis and Sundt (1980) in which they examined
the relationship of cerebral blood flow to cardiac output,
mean arterial blood pressure, and blood volume. ‘Using

Xenon133

clearance as an indicator of cerebral blood flow,
they found that the administration of PBZ (1.5 mg/kg) to
normotensive cats caused a 24% and a 22% decrease in MABP
and cerebral blood flow, respectively: the majority of their
data points fell above a MABP level of 90 Torr, and cerebral
blood flow remained stable to changes in MABP induced by
intravenous angiotensin. They explained the fall in
cerebral blood flow with administration of PBZ by
speculating that, "Within the range of autoregulation, cere-
bral blood flow is relatively insensitive to changes in
MABP; However, increasing or decreasing MABP may still have
some effect on cerebral blood flow.“ Eklof and his
colleagues (1971) found similar results, reporting a 29%

decrease in cerebral blood flow after bilateral chronic

cervical ganglionectomy while autoregulation mechanisms

134
remained intact. They failed to speculate on the mechanism
responsible for the decline in cerebral blood flow.

In the present study, the administration of PBZ may
have decreased cerebral blood flow, even though MABP, which
was lowered by 30% after PBZ (Table 7), remained within a
pressure range over which cerebral blood flow normally'
autoregulates (Harper, 1966; MacKenzie et a1., 1976; Page et
a1., 1980). Consequently, the proposed decrease in cerebral
and perhaps choroidal blood flow (Page et a1., 1980)
critically limited the passive transudation of plasma
ultrafiltrate across choroidal and/or brain parenchymal
capillaries. The delivery of solute and water to metabolic
transport mechanisms of CSF secretory sites is crucial for
sustaining Vf (Pollay, 1975; Sklar et a1., 1980). The
present data (Table 8) may show the effects of an inadequate
capillary transudation process on Vf and represent a situa-
tion in which cerebral blood flow limits CSF formation rate.

The second explanation for the lowered Vf after PBZ is
also based on the cerebrovascular response to the alpha
blockade but ascribes the decreased Vf to a decreased brain
capillary permeability; It is‘well known that a relatively
rich network of noradrenergic fibers from peripheral ganglia
accompany the extracerebral vessels (i.e., pial vessels) and
intracerebral vessels including arterioles and cerebral
veins (Edvinsson et a1., 1971/72; mean et a1., 1974); the
main receptor involved appears to be alpha-adrenergic,

mediating vasoconstriction (Edvinsson and mean, 1974).

135
More recent evidence indicates that central adrenergic
neurons, the cell bodies of which are densely assembled in
the locus coeruleus, may directly innervate the
intraparenchymal microvessels (Hartman et a1., 1972;
McCulloch et a1., 1982; Swanson et a1., 1977) and regulate
microcirculatory hemodynamics and transcapillary transport
mechanisms: cerebral hemispheric blood flow decreases when
central noradrenergic cell bodies in the locus coeruleus are
stimulated by carbachol (Raichle et a1., 1977), angiotensin
II (Grubb and Raichle, 1981), and electrical pulses (Raichle
et a1., 1976), and brain vascular permeability concomitantly
increases (Raichle et a1., 1977; Preskorn et a1., 1980).
These effects can be blocked by the intraventricular (but
not systemic) administration of phentolamine (Raichle et
a1., 1977) or by bilateral locus coeruleus lesions (Bates et
a1., 1977).

Since PBZ blocks all alpha-adrenoceptors regardless of
their location and origin, it can effect a centrally-
mediated decrease in capillary permeability. The lowered
capillary'permeability'may decrease capillary filtration
rate, thereby imposing a limitation on the availability of
plasma ultrafiltrate to choroidal and/or extrachoroidal CSF
secretory sites. Vf declines (Table 8) according to this
theory because of an inadequate delivery of fluid and
electrolytes to secretory sites.

The final explanation for the Vf values measured after

PBZ (Table 8) rests primarily on the direct effects of PBZ

136

on the active component of CSF synthesis, the metabolic
apparatus. Electron microscopic studies reveal that
autonomic axon terminals in the choroid plexus have a close
association not only with the smooth muscle wall of small
arterioles, but also with the secretory epithelial cells
(Edvinsson et a1., 1975; Edvinsson et a1., 1977). Further,
investigations of the local adrenoceptors mediating the
sympathetic nerve-induced effect on the CSF secretory sites
reveal that although(alpha-receptor-mediated local
vasoconstriction may assist in reducing CSF production by
decreasing plexus blood flow (Cserr, 1971; Page et a1.,
1981), beta-receptor stimulation can decrease Vf presumably
by direct inhibitory actions on the plexus epithelium
(Lindvall et al., 1978a; Lindvall et a1., 1978b).

Phenoxybenzamine exerts its effects primarily through
inhibitory actions on both pre- and post-junctional alpha-
adrenoceptors (i.e., alphal and alphaz). Recent evidence,
however, disputes the validity of ascribing the net effects
of PBZ administration solely to the non-specific blockade of
alpha-adrenoceptors. Rather, the non-specificity of the
blockade imposed by such antagonists as PBZ and phentolamine
may actually cause the expression of sympathomimetic actions
(Patel et a1., 1981; Saeed et a1., 1982). More than a
decade ago it was shown that the normal vasoconstrictor
responses to sympathetic nerve stimulation in the splanchnic
bed were inverted to a beta-receptor-mediated vasodilation

after PBZ, yet the normal vasoconstrictor response was not

137
changed by prOpranolol, a non-specific beta-adrenoceptor
antagonist (Greenway and Stark, 1970). It was suggested
that the beta-receptors were not innervated by the
sympathetic nerves but, after PBZ, were stimulated by an
increased overflow of noradrenaline; PBZ inhibits the
re-uptake of the transmitter. Further, it is now known that
blockade of prejunctional alpha-adrenoceptors increases the
release of noradrenaline from sympathetic nerve terminals by
dis-inhibiting transmitter release (Langer, 1977). Saeed et
a1. (1982) administered phentolamine systemically to
conscious dogs at rest and found an excess of plasma
catecholamines, as well as hypotension, vasodilation,
tachycardia, and an increase in cardiac output. After
subsequent beta-adrenoceptor blockade, however, the
vasodilation was not observed. It was concluded that when
both alpha—adrenoceptor subtypes are blocked, as they are
after PBZ, catecholamine release is enhanced and re-uptake
is inhibited, such that the amount of plasma catecholamines
increases. The larger amounts of circulating catecholamines
allow diffusion in sufficient concentration to stimulate
adjacent beta-adrenoceptors (Patel et a1., 1981).

In the present studies, PBZ presumably blocked both
alpha-receptor subtypes and in the process directly deranged
MABP (Table 8), and perhaps changed cerebral blood flow
(Davis and Sundt, 1980; Eklof et a1., 1971) and/or the
capillary transudation process (Raichle et a1., 1977;

Preskorn et a1., 1980). Even more importantly, PBZ may have

138
indirectly caused the activation of beta-adrenoceptors by
increasing the level of plasma catecholamines. Since beta-
adrenoceptors are associated with both cerebral vessels
(Edvinsson and mean, 1974) and choroidal secretory
epithelial cells (Lindvall et al., 1978b: Nathanson, 1979),
and may be accessible to circulating catecholamines which
traverse leaky choroidal capillaries (Wright, 1979), these
receptors may have been activated. In turn, and perhaps in
conjunction with alpha-mediated cardiovascular responses,
beta adrenoceptor stimulation caused a significant decrease
in vf (Table 8) by direct inhibition of the metabolic
machinery of the secretory epithelium (Lindvall.et a1.,
1978a; Lindvall et al., 1978b; Nathanson, 1979).

Steady state CSF formation rate (Vf), however, remained
insensitive to increases in CSF pressure (Table 2) and/or in
Paco2 (Tables 2, 5, and 8). That CSF pressure is without
effect on Vf has been demonstrated by Heisey et a1. (1962)
in goats, Hochwald and‘Wallenstein (1967) in cats, Eisenberg
and co-workers (1974) in hydrocephalic cats, and Sklar et
a1. (1980) in dogs. Likewise, data from a number of
ventriculocisternal perfusion studies show that increases in
Paco2 are without effect on Vf (Oppelt et a1., 1963; Oppelt
et a1., 1964; Hochwald et a1., 1973; Tschirgi et a1., 1954;
Davson and Segal, 1970).

It has been hypothesized that only when cerebral blood
flow autoregulatory mechanisms are disrupted does Vf become

susceptible to increases in CSF pressure and Paco2 (Sklar et

139
a1., 1980; Weiss and Wertman, 1978; Heisey et a1., 1983).
Cerebral autoregulation can be disturbed by the inhalation
of a C02 gas-in-air mixture of 9% or more (Raichle and
Stone, 1972), by extensive cerebral surgical intervention
(Ames et a1., 1965), and by hypotensive episodes in which
blood pressure draps below 60 Torr (Harper, 1966; MacKenzie
et a1., 1976); the latter condition has been shown to be
associated with a depressed Vf (Carey and Vela, 1974; Sklar
et a1., 1980).

During all phases of every experiment MABP levels
remained within a pressure range (Tables 1, 4, and 7) over
which cerebral blood flow normally'autoregulates (Harper,
1966; MacKenzie et a1., 1976; Page et a1., 1980). Further,
hypercapnic tests were conducted using C02 gas-in-air
mixtures that were less than 9%. Vf was not sensitive to
either increases in PCSF (Table 2) or Paco (Tables 2, 5,

2 and 8), suggesting that cerebral blood flow autoregulatory

capacity was not impaired during these studies.

CONCLUSIONS

Mixing of tracer is incomplete in CSF spaces during

ventriculocisternal perfusion.

Craniospinal fluid redistribution significantly
contributes to the CSF transient responses to

hypercapnia.

Overestimates of CSF formation rate will result if
measurements are made prematurely after perturbing the

craniospinal fluid balance.

The craniospinal fluid redistribution process and CSF
transient responses are independent of alpha-

adrenoceptor activity.

CSF rates of formation and absorption are decreased
after phenoxybenzamine treatment due to alpha-mediated
cerebrovascular responses and/or beta adrenoceptor
inhibition of the metabolism of CSF secretory

epithelium.

Steady state CSF formation is independent of increases

in CSF pressure or PaC02°

140

APPENDICES

BBV

Cd

Cd/Ct

co/ci

CPP

Cr

CSF
CSth

Ct

APPENDIX A

DEFINITION OF TERMS

brain blood volume

area of co/ci dilution during transient period
(min)

area of transient dilution of co/ci expressed as
a fraction of the total transient area of co/ci
if no dilution occurred during the period tic to
tfc (dimensionless)

perfusion effluent concentration (co) of 14C-
dextran expressed as a ratio to perfusion inflow
concentration (ci; dimensionless)

cerebral perfusion pressure (CPP a MABP - PCSF;
Torr)

residual co/ci area during transient period

(Cr = Ct - Cd: min)

cerebrospinal fluid

amount of unequilibrated or tracer-free CSF in
V6 (:11)

total area of co/ci assuming no dilution during

t.

1c to tfc (min)

respiratory frequency (breaths/min)
fraction of C02 in alveolar (end-tidal) gas
(% (:02)

141

FIC02
MABP

PBZ

PCSF

Tr
Va

Vd

142
fraction of inspired C02 (% C02)
mean arterial blood pressure (Torr)
phenoxybenzamine hydrochloride
partial pressure of C02 in arterial blood (Torr)
cerebrospinal fluid pressure (cmHZO)
(logarithmic) hydrogen ion concentration
stop-flow perfusion method
time of onset of co/ci transient, and of the end
of the co/ci transient, respectively (min)
time of onset of C02 breathing and V0 transient,
and of the end of the V0 transient, respectively
(min)
rectal temperature (°C)
CSF bulk absorption rate (pl/min)
volume of displaced CSF during Vo transient
period (ul)
minute ventilation (ml/min; BTPS)
CSF formation rate (ul/min)
perfusion inflow rate (ul/min)
perfusion outflow rate (ul/min)

tidal volume (ml; BTPS)

APPENDIX B

PREPARATION OF DIAL-URETHANE ANESTHETIC SOLUTION

REAGENTS:

l. Diallyl barbituric acid (crystalline; K a K

Laboratories, Inc., Plainview, NY)

2. Disodium calcium ethylene diamine tetra acetate

trihydrate (Pfaltz & Bauer, Inc., Flushing, NY)

3. Monoethyl urea (Pfaltz & Bauer, Inc., Flushing, NY)

4. Urethane (Aldrich Chemical Co., Milwaukee, WI)

METHOD:

Add 10 g reagent #1, 50 mg reagent #2, 40 g reagent #3,
and 40 g reagent #4 to 10 cc of deionized water. Heat the
mixture in a water bath to dissolve the chemicals. Cool to
room temperature and store in a stoppered dark glass bottle

at room temperature.

143

APPENDIX C
ARTERIAL pH AND P002 NEASURENENTS

Reference: Radiometer Instruction Manual, Model PHA 927b,

Radiometer A/S, Copenhagen, Denmark
OPERATING PRINCIPLE:

pH Electrode:

The electrode (model E5021; London Co., Cleveland, OH)
is a H+-sensitive glass capillary surrounded by a glass tube
containing a reference fluid. .A test solution is drawn into
the glass capillary and contact between the glass electrode,
sample and calomel electrode is made through a saturated KCl
bridge» The glass electrode develops a potential which
varies directly with the H+ concentration in the test
sample; the calomel electrode potential is constant. The
potential difference between the glass and calomel
electrodes is amplified and displayed on a pH meter (model

PHM27; The London Co., Cleveland, OH) in mv or pH units.

PCO2 Electrode:

The PCO2 electrode (model E5036; The London Co.,
Cleveland, OH), a specialized H+-sensitive glass electrode
with a built-in Ag-AgCl electrode, measures the pH of a
NaHCO3 + NaCl solution in which it is immersed. 'The

144

145
solution is separated from a test sample by a Teflon
membrane which is permeable to C02. The C02 in the test
solution or calibration gas diffuses into the buffer
solution until a equilibrium is established; the amount of
C02 diffusing from the test solution to the buffer will
depend on the Pco2 of the sample. The (H+) of the buffer
solution changes and the H+ sensitive glass electrode
deve10ps a potential proportional to this change which is
referenced to the potential of the silver chloride
electrode. The potential difference is amplified and
displayed on an indicator scale calibrated in mmHg PC02

(model PHM27; The London Co., Cleveland, OH).

CALIBRATION:

The C02 electrode and the pH electrode are calibrated
using two samples of known C02 concentrations (5% and 12%
C02; The London Co., Cleveland, OH) and commercially
prepared buffer solutions (pH = 6.840 and pH = 7.383;
Scientific Products Inc., Allen Park, MI) at 38 1 0.2°C,

respectively.

SAMPLE HANDLING:

Arterial blood samples (nom. lcc) were collected
anaerobically in syringes which had their dead space filled
with heparinized (1000 U.S.P. units/ml; Heparin Sodium, The
Upjohn Company, Kalamazoo, MI) saline “Ll cc heparin per m1

of 0.9% saline solution). The pH and PCO2 were measured*

146
and accepted if duplicate readings were 1 0.01 pH units or 1,
1.0 Torr, respectively. Measurements were corrected for the
difference between the electrode and animal's body
temperature ('TCORR'; HP67 programmable calculator
algorithm) which was measured at the time the blood sample

was taken.

 

* Radiometer electrodes were used for some pH and PCO2 meas-
urements and for the Operating principles. Other pH and
PCO2 measurements were made using a Corning blood gas
analyzer (model 165/2; Corning Medical Instruments,

Medfield, MA).

APPENDIX D
RESPIRATORY NEASURENENTS AND CALCULATIONS

PRINCIPLES OF THE PNEUMOTACHOGRAPH:

The pneumotachograph (Fleisch flowtransducer; Dyna-
sciences, Blue Bell, PA) is a cylindrical flow transducer
containing a small metal grid which is heated to prevent
water condensation, and which offers constant resistance to
gas flow through it. A pressure difference develops across
the grid (constant resistance) which is proportional to the
rate of gas flow through it. A differential manometer
(model PT 5A; Grass Instruments, Quincy, MA) is connected by
tubing to a port on either side of the grid and transduces
the pressure drOp to a prOportional voltage output which is
directed through a low level DC preamplifier (model 5P1;
Grass Instruments, Quincy, MA) to a strip chart recorder
(model 7100 BM; H-P Moseley, Pasadena, CA) calibrated in
l/minm Calibration is accomplished by directing known gas
flows through the pneumotachograph from a calibrated gas
flow meter (tube #604-1A; Matheson Gas Products, East
Rutherford, NJ). During experiments expiratory flow rate
was monitored and displayed on the chart recorder by attach-
ing the pneumotachograph downstream from the respiratory

valve assembly.

147

148
The following measurements and calculations were made

from the flow rate tracings:

Respiratory Frequency
Every 10 minutes respiratory frequency (f; breaths/min)
was determined by counting the number of expirations occur-

ring during the 30 sec period and multiplying it by 2.

Tidal Volume

At 10 minute intervals during respiratory steady
states, the speed of the strip chart (recording expiratory
flow rate) was increased from 1.25 cm per min to 2.5 cm per
sec and 4 to 5 expiratory flow rate tracings were recorded.
These flow vs. time tracings were analyzed using 'TIDVOL'*,
a program written for use on the digitizer (model GP-3;
Science Accessories Corp», Southport, CT) interfaced with
the PDP 11/34 computer (Digital Equipment Corp., Maynard,
MA). “TIDVOL' determined the area under the respiratory
flow curve (VT; m1; ATPS (ambient temperature and pressure,
saturated with water vapor)) by approximating a definite

integral using the trapezoidal rule (Eulem's method):

where,
VT a (0.5 Y0 + Y1 + Y2 + . . . + Yn_1 + 0.5 Yn) dx
(D-l)

 

* The 'TIDVOL routine" is also appropriate for approximating

the definite integrals in equations 6 and 8 in the

Materials and Methods section.

149
and,
VT = area under one respiratory flow curve
(m1)

successive expiratory flow rates

Y0, Y1, coo Yn
(ml/min)
dx = time for one expiration (min); derived

from recorder chart speed

The mean of 4 - 5 tracings, the standard deviation, and the
standard error of the mean (SE) were calculated. If the SE

was i 1.0 ml the data were accepted.

Minute Ventilation

Another PDP 11/34 algorithm (“TIDBITS") converted VT in
ATPS units to BTPS (body temperature, ambient pressure, and
saturated with water vapor) and to STPD (standard
temperature and pressure, dry), and calculated minute

ventilation (VB; ml/min in BTPS and STPD), where:

VB 3 VT x f (D‘Z)

APPENDIX E

ANALYSIS OF CARBON DIOXIDE IN ALVEOLAR
AND NIX. EXPIRED GAS

 

SAMPLE gELL

 

—ub——

 

 

wmoow
SAMPLE GAS
f? ETECTOR CELL
INFRARED A ,z I
sounce '

 

 

co, MOLECULES
/

 

 

 

 

I R BEAM

I a BEAM 0

 

METAL
DIAPHRAGM

 

INFRARED1
SOURCE

 

DETECTOR CELL
(REFERENCE
HALF)

 

 

 

T t - ZERO CONTROL

 

L__ . -_—————-——nerensncs CELL

FIGURE E-l .

Reference: Beckman Instruction Manual, LB-2 Medical Gas

Analyzer, Beckman Instruments, Inc., Anaheim, CA

150

151
OPERATING PRINCIPLES:

Carbon dioxide, unlike nitrogen or oxygen, absorbs
energy from a narrow band of the IR spectrum; the amount of
energy absorbed will vary directly with the C02 concentra-
tion. The gas analyzer is equipped with two matched IR
sources and 3 gas cells: a sample cell (containing the gas
to be analyzed); a reference cell (containing no C02); and a
detector cell (containing C02). The detector cell is
divided by a deformable metal membrane and is located behind
the other 2 cells (see Figure E-l). IR radiation passes
from the 2 matched sources through the sample and reference
cells and into the two halves of the detector cell. When
C02 molecules within the closed chambers of the detector
cell absorb IR radiation, their kinetic energy and the
pressure within the cell proportionally increase. If there
is no C02 in the sample cell, the amount of radiation
reaching the sample side of the detector cell will equal the
amount reaching the reference side, the pressure in each
half of the detector cell will be equal, and the membrane
will not deform. If the test gas contains C02, the amount
of radiation reaching the sample side of the detector cell
will cause the dividing membrane to bow toward the sample
side changing the capacitance between the metal diaphragm
and a fixed plate within the detector cell. The resulting
change in voltage is directly related to the percentage of
C02 in the test gas sample and is displayed on the console

indicator in C02 percentage units.

152
CALIBRATION PROCEDURES:

The instrument contains a pump which continuously draws
gas through the sample cell at an adjustable rate; the
recommended air flow through the cell is 500 ml/min. there
are two reasons for this suggested flow rate: first, a
particular pressure in the sample cell is associated with a
given inflow rate and resistance (inflow resistance is a
function of the length and diameter of the inlet tubing);
second, the absorption of energy from the sample IR beam is
not a linear function of the C02 concentration in the
sample. Instead, this relationship is inherently nonlinear
and requires a complimentary linearizing circuit. This
circuit is preset to correct for nonlinearity only when
pressure conditions associated with an inflow rate of 500
ml/min (standard conditions) exist in the sample cell. The
device can be operated under nonstandard conditions (flow
rate < 500 ml/min) and linearity retained if the pressure in
the sample cell under standard conditions is duplicated:
sample cell pressure will be unchanged if inflow rate and
resistance are concurrently and proportionately decreased
and increased, respectively; In these experiments it was
necessary to operate the instrument under nonstandard
conditions (inflow rate = nom. 200 ml/min) since an
anesthetized cat breathing air may have a minute ventilation
less than 500 ml/min. A.two-stage calibration procedure was
devised to maintain linearity under measuring (nonstandard)

conditions.

153

Standard Conditions (flow rate = 500 ml/min)

1. Depress POWER and OPERATE buttons and allow the
instrument to equilibrate for at least 30 minutes

prior to calibration.

2. Adjust ZERO control such that the panel meter reads
+0.03% C02 when ambient air is drawn through the

heated inlet tubing and the sample cell.

3. Sample from a Haldane analyzed C02 gas mixture (> 8%
C02; < 10% C02). Adjust the GAIN control so that
the panel meter reads the calibration gas concentra-

tion.

4. Sample from another analyzed C02 mixture (between 0%
and 8% 002). The panel meter will read this concen-

tration if the response is linear.

Nonstandard Conditions (flow rate < 500 ml/min)

1. Attach a 23 ga needle onto the heated inlet tubing

to increase inflow resistance.

2. Sample from same calibration gas as in A-3 and
adjust FLOW (decrease inflow rate) until the panel

meter records the same value as in A-3.

3. Repeat A-4.

APPENDIX P

COMPOSITION AND PREPARATION OF ARTIFICIAL
CAT CEREBROSPINAL FLUID (CS?)

ELECTROLYTE COMPOSITION OF CSF (from Vogh and Maren, Am. J.
Physiol. 228: 673-683, 1975):

Substance mull
Na+ 151.0
x+ 2.5
Ca++ 2.5
Mg++ 1.0
Cl' 138.0
Hco3' 20.0
apo4' 1.0
304' 1.0

REAGENTS USED IN PREPARING ARTIFICIAL CAT CSF:
1. NaCl, Analytic reagent (A.R.)

2. NaHCO3, (A.R.)

3. NaH2P04'H20, (A.R.)

4. NaZHPO4'H20, (A.R.)

5. KCl, (A.R.)

6. MgSO4'7H20, (A.R.)

7. CaC12, (A.R.)

154

155

SOLUTIONS:

A. Dissolve 7.6 g of reagent #1, 1.68 g of reagent #2,
169.0 mg of reagent #3, 134.0 mg of reagent #4, and
186.4 mg. of reagent #5 in deionized water; q.s. to one
liter.

B. Dissolve 24.65 g of reagent #6 in deionized water; q.s.
to 100 ml.

C. Dissolve 27.75 g of reagent #7 in deionized water; q.s.
to 100 m1.

METHOD:

Fifty m1 of solution A is bubbled at room temperature
(nom. 15 minutes) with 5% COz-95% 02 gas mixture to adjust
the pH to approximately 7.40. Then 50 ul each of solutions
B and C are added to 50 ml of solution A to establish the

normal ionic composition of cat CSF.

APPENDIX C
DETERNINATION OP INPLOW AND OUTPLOW PERPUSION RATES

The inflow and outflow drop volumes were determined
gravimetrically'by collecting a known number of drOps of
fluid in tared vials from PE tubing connected to the inflow
syringe (inflow) or to the cisternal cannula (outflow).
Assuming a specific gravity of 1.0, the volume per drOp (V;

pl) could be determined by:

V = (G-l)

5 IS

where:
W = net weight of fluid (mg)

n = number of drops

During an experiment the cisternal outflow cannula was
directed through a photocell drop detector attached to a
fraction collector controller (model 568; Instrumentation
Specialties Co., Lincoln, NA). The controller was
programmed to advance the collector after a pre-set number
of drOps were detected; normally 2 drop samples (nom. 100
pl) were collected. Further, the controller circuit was
modified to provide a voltage input to the polygraph (model

5P; Grass Instrument Co., Quincy, MA) triggering an event

156

157
mark at the end of each sample. The polygraph record was
analyzed using a digitizer (model GP-3; Science Accessories
Corp”, Maynard, MA) for measuring distance between
successive outflow sample events. A computer program

('OUTFLOW') was used to calculate outflow rate by:

{, ._. (33.3331 (0-2)
D
where,
V = flow rate (ul/min)
V = volume per drop (pl; see equation G-l)
S a polygraph chart speed (mm/min)
D a distance between event marks on polygraph record

(mm)

At the completion of each experiment, the outflow cannula
was disconnected from the drOp detector and replaced by the

inflow cannulae for determining inflow rate using equations

1 and 2.

APPENDIX E
BETA RADIATION COUNTING

Reference: Beckman Liquid Scintillation Counter Manual,
Model 3150P, Beckman Instruments Inc.,

Fullerton, CA.

OPERATING PRINCIPLES:

Liquid scintillation (LS) counting is a method for
detecting beta particle emission from decaying nuclei of
unstable isotopes. When a radioisotope such as Carbon-l4
emits beta particles and is suspended in a photoluminescent
medium (iJL, LS cocktail; a mixture of organic solvent and
fluor molecules), the medium absorbs energy from the beta
particles and releases it in the form of light. The LS
counter houses special vacuum tubes (photomultiplier tubes;
PMT) which are light-sensitive and which produce voltage
pulses in response to photons that strike their surfaces;
the number of voltage pulses produced is proportional to the
number of fluorescent events detected. These voltage pulses
are summed, amplified, directed to a sealer, and recorded as
“counts.“ Counting rate (counts per min; CPM) is based on
the total number of counts accumulated in a one-minute time

period.

158

159

The liquid scintillation counting technique relys on
critical reactions between the radioactive sample and the
light-producing chemical to indicate the presence of a
radioisotoPe. Anything which promotes the interaction of
the emitted particle with the fluorescing chemical will
enhance the efficiency of detection; anything which
interferes with it will reduce the photon output of the
scintillation fluor mixture and the detection efficiency.
This interference process is known as I'quenching" and can be
caused by any of the following agents or conditions: color
in the sample; the presence of salts in the medium which do
not fluoresce; incomplete dispersion of the radioactive
sample in the fluorescing mixture; or inadequate
concentration of fluor or solvent in the LS cocktail.
Quenching reduces the final count (“counting efficiency;'
EFF) of a radioactive sample by shifting the isotOpes light
emission spectrum toward lower pulse heights.

The external standard channels ratio (ESCR) method,
based on the effects of gamma radiation on a 14C-quenched
standard set (Amersham/Searle Corp», Boston, MA), was used
to detect quenching in the samples and to provide a means of
correcting for it. The LS counter is equipped with a gamma
radiation source (137-cesium) which is brought in proximity
to a series of sealed standards containing equal amounts of
radioactive matter (carbon-l4) but with varying amounts of
chemical quencher. The gamma rays interact with matter and

discharge compton electrons which interact with solvent

160
molecules in the LS cocktail in the same way as beta
particles to produce light scintillations. .Additionallyy
compton and beta emission spectra exhibit similar quench-
induced shifts. Each standard is analyzed in two channels
which count over two different portions of the energy
spectrum both before and after exposure to gamma radiation
(to determine sample counts due to beta emission, and total
counts due to beta compton emission, respectively). The
counts from each channel due to gamma radiation (the
difference between total and sample counts) are determined
and expressed as a ratio (ESCR) which is indicative of the
amount of shift in the compton spectrum and inversely
proportional to the amount of quenching in the sample.
Counting efficiency (EFF) of the standards was determined
as:

(CPM - background CPM) x 100
DPM

 

EFF = (H-l)
where,
CPM a counting rate due to beta emission (counts/min)
DPM = known activity of 14C-standards (disintegrations/

min)

A quench correction curve is constructed by plotting EFF for
the standards as a function of their corresponding ESCR

values. The curve can be described as:

Y = mX + b (H-2)

161
where,
Y = counting efficiency (EFF) for the standards

x = ESCR value for the standards

EFF (standard #1) - EFF (standard #2)
ESCR (standard #1) - ESCR (standard #2)

 

b = counting efficiency for the standards when ESCR = 0

Each sample undergoes the same counting procedure as
described for the 14C-standards, and their CPM and ESCR
values determined. From the quench correction curve or from
the mathematical expression describing it (equation H-2),
sample counting efficiencies (EFF) are determined. Finally,
the corrected activity of each sample (DPM;

disintegrations/min) is calculated as:

sample CPM

DPM ==
sample EFF

(H-3)

 

CALCULATIONS or 14c-DEXTRAN CONCENTRATIONS:

The disintegration rates (DPM) and concentrations
(DPM/ul) for 14C-dextran in samples from the inflow syringe
and cisternal effluent were calculated using the PDP 11/34
computer (Digital Equipment Corp., Maynard, MA). A paper
tape punch (Epson 611; Shinshu Seiki Co., Ltd., Japan)
attached to the LS Counter (model 3150P; Beckman Instruments
Inc., Fullerton, CA) provided an output which contained ESCR
and counts per min (CPM) for quenched 14C-standards,
background, perfusion inflow (Ci) and cisternal effluent

(co) samples. These data were entered into the computer

162
through a program called 'SDPMl' using a paper tape reader,
as were inflow and outflow sample volumes (ul) which were
manually entered. A second program ('SDPMZ') using the
output data from 'SDPMl,‘ calculated the following
information for each standard and sample: the corrected
activity (DPM) (see equations H-l to H-3); the 14C-dextran
inflow and outflow concentrations (Ci and co, respectively;
DPM/pl); and a ratio of outflow to inflow 14C-dextran
concentration (co/Ci; dimensionless). .A final program,
'SDPM3,‘ a versatile plotting routine, accepted the stored
output from 'SDPM2,‘ combined these data with the manuallyb
entered collection time and outflow rate (V0; Appendix G)
data for each sample, and plotted CO/Ci and V0 as a function

of time.

BIBLIOGRAPHY

BIBLIOGRAPHY

Adams, T., S. R. Heisey, M. C. Smith, M. A. Steinmetz, J. C.
Hartman, and H. K. Fry. Thermodynamic technique for
the quantification of regional blood flow. Am. J.
Physiol. 238: 682-696, 1980.

Altura, B. M. and B. T. Altura. Vascular smooth muscle and
neurohypophyseal hormones. Fed. Proc. 36: 1853-1860,
1977.

Ames, A., M. Sakanoue, and S. Endo. Na, K, Ca, Mg, and Cl
concentrations in Choroid plexus fluid and cisternal
fluid compared with plasm ultrafiltrate. {L Neuro—
physiol. 27: 672-681, 1964.

Ames, A. III, R. Higashi, and F. Nesbitt. Effects of PC02,
acetazolamide and ouabain on volume and composition of
Choroid-plexus fluid. Am. J. Physiol. 229: 415-419,
1965.

Auer, L. M. and B. B. Johansson. Pial venous constriction
during cervical sympathetic stimulation in the cat.
Acta. Physiol. Scand. 110: 203-205, 1980.

Auer, L., B. Johansson, and E. T. MacKenzie. Cerebral
venous pressure during actively induced hypertension
and hypercapnia in cats. Stroke 11: 180-183, 1980.

Auer, L. M., W. Kuschinsky, B. B. Johansson, and L.
Edvinsson. Sympatho-adrenergic influence on pial veins
and arteries in the cat. In: Cerebral Blood Flow:
Effects of Nerves and Neurotransmitters. Heistad, D.
IL, and M. L. Marcus (edsJ. Elsevier North Holland,
Inc. pp. 292-300, 1982.

Avezaat, C. J. J., D. H. M. Eijndhoven, and D. J. Wyper.
Effects of hypercapnia and arterial hypotension and
hypertension on cerebrospinal fluid pulse pressure and
intracranial volume-pressure relationships. CL Neurol.
Neurosurg. and Psychiat. 43: 222-234, 1980.

163

164

Bates, D., R. M. Weinshiboum, R. J. Campbell, and T. M.
Sundt Jr. The effect of lesions in the locus coeruleus
on the physiological responses of the crebral blood
vessels in cats. Brain Res. 136: 431-443, 1977.

Bering, E.1L Cerebrospinal fluid production and its rela-
tionship to cerebral metabolism and cerebral blood
flow. Am. J. Physiol. 197: 825-828, 1959.

Bering, E. A. and O. Sato. Hydrocephalus: changes in
formation and absorption of cerebrospinal fluid within
the cerebral ventricles. J. Neurosurg. 20: 1050-1063,
1963.

Carey, M. E. and A. R. Vela. Effect of systemic arterial
hypotension on the rate of CSF formation in dogs. J.
Neurosurg. 41: 350-355, 1974.

Cervos-Navarro, J. and F. Matakas. Electron microsc0pic
evidence for innervation of intracerebral arterioles in
the cat. Neurology 23: 282-286, 1973.

Cserr, H. Physiology of the Choroid plexus. Physiol. Rev.
51: 273-311, 1971.

Curl, F. D. and M. Pollay. Transport of water and electro-
lytes between brain and ventricular fluid in the rab-
bit. Exp. Neurol. 20: 558-574, 1968.

D'Alecy, L. G., C. J. Rose, and S. A. Sellers. Sympathetic
modulation of hypercapnic cerebral vasodilation in
dogs. Circ. Res. 45: 771-775, 1979.

Dahlgren, N. and B. K. Siesjo. Cerebral blood flow and
oxygen consumption in normocapnia and hypercapnia:
modulation influence of paravertebral sympathetic
blockade at the low thoracic level. Acta. Anesth.
Scand. 45: 331-336, 1981.

Davis, D. H. and T. M. Sundt, Jr. Relationship of cerebral
blood flow to cardiac output, mean arterial pressure,
blood volume, and alpha and beta blockade in cats. J.
Neurosurg. 52: 745-754, 1980.

Davson, H. Physiology of the Ocular and Cerebrospinal
Fluids. Little and Brown, Boston, 1956.

165

Davson, H. and C. P. Luck. The effect of acetazolamide on
the chemical composition of the aqueous humor and
cerebrospinal fluid of some mammalian species and on
the rate of turnover of 24Na in these fluids. J.
Physiol. 137: 279-293, 1957.

Davson, H. and E. Spaziani. Effect of hypothermia on
certain aspects of the cerebrospinal fluid. Exp.
Neurol. 6: 118-128, 1962.

Davson, H. Physiology of the Cerebrospinal Fluid. J. and
A. Churchill Ltd., London, 1967.

Davson, H., G. Hollington, and M. B. Segal. The mechanism
of drainage of the cerebrospinal fluid. Brain 93: 665-
678, 1970.

Davson, H. and M. B. Segal. The effects of some inhibitors
and accelerators of sodium transport on the turnover of
2Na in the cerebrospinal fluid and the brain. J.
Physiol. 209: 131-153, 1970.

Deane, R. and M. Segal. The effect of vascular perfusion of
the Choroid plexus on the secretion of cerebrospinal
fluid. Proc. Physiol. Soc. 285: 18-19, 1978.

Diamond, J. M. and W. H. Bossert. Standing-gradient osmotic
flow. A mechanism for coupling of water and solute
transport in epithelia. J. Gen. Physiol. 50: 2061-
2083, 1967.

DiMattio, J}, G. M. Hochwald, and C. Malhan. The effects Of
the hydrocephalic process on cerebral blood flow in the
cat. In: Cerebral Circulation and Metabolism.
Langfitt, T. W., L. C. McHenry, Jr., M. Reivich, and H.
Wollman, (eds.). Springer-Verlag, New York. pp. 223-
227,1975.

Duckles, S. P. and J. A. Bevan. Pharmacological characteri-
zation of adrenergic receptors of rabbit cerebral
artery in vitro. J. Pharmacol. Exp. Ther. 197: 371-
378, 1976.

Edvinsson, L., R. Hakanson, Ch. mean, and K. A. West.
Sympathetic influence on carbonic anhydrase activity in
choroid plexus. Exp. Cell Res. 67: 245, 1971.

166

Edvinsson, L., Ch. mean, E. Rosengren, and K. A. West.
Concentration of noradrenaline in pial vessels, choroid
plexus, and iris during 2 weeks after sympathetic
ganglionectomy or decentralization. Acta Physiol.
Scand. 85: 201-206, 1972.

Edvinsson, L., K. C. Nielsen, Ch. mean, and K. A. West.
Sympathetic adrenergic influence on brain vessels as
studied by change in cerebral blood volume of mice.
Europ. Neurol. 6: 193-202, 1971-72.

Edvinsson, L. and Ch. mean. Pharmacological characteriza-
tion of adrenergic alpha and beta receptors mediating
the vasomotor reSponses of cerebral arteries in vitro.
Circ. Res. 35: 835-849, 1974.

Edvinsson, L., K. C. Nielsen, Ch. mean, and K. A. West.
Adrenergic innervation of the mammalian choroid plexus.
Amer. J. Anat. 139: 299-308, 1974.

Edvinsson, L. Neurogenic mechanism in the cerebrovascular
bed. Autonomic nerves, amine receptors and their
effects on cerebral blood flow. .Acta Physiol. Scand.
427: 1-36, 1975.

Edvinsson, L., P. Aubineau, Ch. mean, R. Sercombe, and J.
Seylaz. Sympathetic innervation of cerebral arteries:
Prejunctional supersensitivity to noradrenaline after
sympathectomy or cocaine treatment. Stroke 6: 525-
530, 1975a.

Edvinsson. In, R. Hakanson, M. Lindvall, Ch. mean, and K.
(L Svenson. Ultrastructural and biochemical evidence
for a sympathetic neural influence on the choroid
plexus. Exp. Neurol. 48: 241-251, 1975b.

Edvinsson, L. and E. T. MacKenzie. Amine mechanism in the
cerebral circulation. Pharmacol. Rev. 28: 275-347,
1977.

Edvinsson, L. and M. Lindvall. Autonomic vascular
innervation and vasomotor reactivity in the choroid
plexus. Exp. Neurol. 62: 394-404, 1978.

Eklof, B., D. H. Ingvar, E. Kagstron, and T. Olin. Persist-
ence'of'cerebral blood flow autoregulation following
chronic bilateral cervical sympathectomy in the monkey.
Acta Physiol. Scand. 82: 172-176, 1971.

167

Eisenberg, H. M., J. E. McLennan, and J. Welch. Ventricular
perfusion in cats with kaolin-induced hydrocephalus.
J. Neurosurg. 41: 20-28, 1974.

Epstein, M. H., A. M. Feldman, and S. W. Bradbury. Science
196: 1012-1977.

Exton, J. Molecular mechanisms involved in alpha-adrenergic
responses. Trends Pharm. Sciences 3: 111-115, 1982.

Fenstermacher, J., C. Li, and V. Levin. Extracellular space
of the cerebral cortex of normothermic and hypothermic
cats. Exp. Neurol. 27: 101-114, 1969.

Fisher, M. J., S. R. Heisey, and T. Adams. Cerebrospinal
fluid transients induced by hypercapnia. Am. J.
Physiol. 245: 701-705, 1983.

Goldstein, D. A., M. Romoff, E. Bogin, and S. G. Massry.
Relationship between the concentrations of calcium and
phosphorous in blood and cerebrospinal fluid. J. Clin.
Endocr. Metab. 49: 58-62, 1979.

Goodman, L. S. and A. Gilman. The Pharmacological Basis of
Therapeutics. Macmillan Publishing Co., Inc., New
York, 1975.

Greenway, C. V. and R. D. Starke. The vascular responses of
the spleen to intravenous infusion of cats, angio-
tensin, and vasopressin in the anesthetized cat. Br.
J. Pharmacol. 38: 583-592, 1970.

Grubb, R. L. and M. E. Raichle, Intravenous Angiotensin II
increases brain vascular permeability. Brain Res.
210: 426-430, 1981.

Grubb, R. L., M. E. Raichle, and J. O. Eichling. The ef-
fects of changes in PaCOZ on cerebral blood volume,
blood flow, and vascular mean transit time. .Stroke 5:
630-639, 1974.

Hamer, A. and A. Sahar. Absence of formation of cerebro-
spinal fluid in the spinal subarachnoid space of the
cat. Israel J. Med. Sci. 14: 741-744, 1978.

Hammerstad, J. P., A. V. Lorenzo, and R. W. P. Cutler.
Iodide transport from the spinal subarachnoid fluid in
the cat. Am. J. Physiol. 216: 353-358, 1969.

168

Hardebo, J. E., L. Edvinsson, and Ch. mean. Influence of
the cerebrovascular sympathetic innervation on blood-
brain barrier function. Acta. Physiol. Scand. 452:
65-68, 1977a.

Hardebo, J. E., L. Edvinsson, P. C. Emson, and Ch. mean.
Isolated brain microvessels: enzymes related to
adrenergic and cholinergic functions. In: Neurogenic
Control of the Brain Circulation. mean, C. and L.
Edvinsson, (eds.). Pergamon Press, oxford. pp. 105-
113, 1977b.

Hardebo, J. E., P. C. Emson, B. Falk, Ch. mean, and E.
Rosengren. Enzymes related to monoamine metabolism in
brain microvessels. J. Neurochem. 35: 1388-1393,
1980.

Hardebo, J. E. and Ch. mean. Barrier mechanisms for
neurotransmitter monamines at the blood brain
interface. Ann. Neurol. 9: 1-6, 1980.

Hardebo, J. E., O. Lindvall, and B. Nilsson. On the
possible influence of adrenergic and cholinergic
mechanisms in normo- and hypercapnia. In: Cerebral
Blood Flow: Effects Of Nerves and Neurotransmitters.
Heistad, D. D. and M. L. Marcus (eds.). pp. 378-383.

Harper, A. M. Autoregulation of cerebral blood flow:
influence of the arterial blood pressure on the blood
flow through the cerebral cortex. J. Neurol. Neuro-
surg. Psychiat. 29: 398-403, 1966.

Hartman, B. K., D. Zide, and S. Udenfriend. The use of
dopamine beta-hydroxylase as a marker for the central
noradrenergic nervous system in rat brain. Proc. Nat.
Acad. Sci. 69: 2722-2726, 1972.

Haywood, J. R. and B. P. Vogh. Some measurements of
autonomic nervous system influences on production of
cerebrospinal fluid in the cat. J. Pharmacol. Exp.
Ther. 208: 341-346, 1978.

Heisey, S. R., D. Held, and J. R. Pappenheimer. Bulk flow
and diffusion in the cerebrospinal fluid system of the
goat. Am. J. Physiol. 291: 1564-1567, 1962.

169

Heisey, S. R., T. Adams, M. J. Fischer, and W. Dang. Effect
of hypercapnia and cerebral perfusion pressure on
cerebrospinal fluid production in the cat. Am. J.
Physiol. 244: 224-227, 1983.

Hernandez, M. J., M. E. Raichle, and H. L. Stone. The role
of the sympathetic nervous system in cerebral blood
flow autoregulation. Europ. Neurol. 6: 175-179, 1971-
72.

Hernandez-Perez, M. J., M. E. Raichle, and H. L. Stone. The
role of the peripheral sympathetic nervous system in
cerebral blood flow autoregulation. Stroke 6: 284-
292, 1975.

Hochwald, G. M. and Wallenstein. Exchange of albumin be-
tween blood, cerebrospinal fluid and brain in the cat.
Am. J. Physiol. 212: 1199-1204, 1967.

Hochwald, G. M. and A. Sahar. Effect of spinal fluid
pressure on cerebrospinal fluid formation. Exp.
Neurol. 32: 30-40, 1971.

Hochwald, G. M., C. Malhan, and J. Brown. Effect of
hypercapnia on CSF turnover and blood-CSF barrier to
protein. Arch. Neurol. 28: 150-155, 1973.

Katzman, R. and H. M. Pappius. Brain Electrolytes and Fluid
Metabolism. Williams and Wilkins Co., Baltimore, 1973.

Kawamura, J., M. Rennels, and E. Nelson. The innervation of
intracerebral arteries of the human brain: an electron
microscopic study. AMNA 15-17, 1972.

KnOpp, L. M., J. R. Atkinson, and A. A. Ward, Jr. Effect of
Diamox on cerebrospinal fluid pressure of cat and
monkey. Neurol. 7: 119-123, 1957.

Langer, S. 2. Presynaptic receptors and their role in the
regulation of transmitter release. Br. J. Pharmacol.
60: 481-497, 1977.

Lindvall, M., L. Edvinsson, and Ch. mean. Histochemical,
ultrastructural, and functional evidence for a
neurogenic control of cerebrospinal fluid production.
Acta Physiol. Scand. Suppl. 452: 77-86, 1977.

170

Lindvall. 1L, L. Edvinsson, and Ch. mean. Reduced cerebro-
spinal fluid formation through cholinergic mechanisms.
Neuroscience Letters 10: 311-316, 1978a.

Lindvall, M., L. Edvinsson, and Ch. mean. Sympathetic
nervous control of cerebrospinal fluid production from
the choroid plexus. Science 201: 1767-1768, 1978b.

Lindvall. IL, L. Edvinsson, and Ch. mean. Effect of sym-
pathomimetic drugs and corresponding receptor antago-
nists on the rate of cerebrospinal fluid production.
Exp. Neurol. 64: 132-145, 1979.

Lindvall, M. and Ch. mean. Evidence for the presence of
two types of monoamine oxidase in rabbit choroid plexus
and their role in breakdown of amines influencing cere-
brospinal fluid formation. J. Neurochem. 34: 518-522,
1980.

Lindvall, M. and Ch. mean. Automonic nerves in the
mammalian choroid plexus and their influences on the
formation of cerebrospinal fluid. CL Cereb. Blood Flow
and Metab., 1981.

Lindvall, M., Ch. mean, and B. Winbladh. Effect of
sympathetic denervation on various aspects of transport
function in the mammalia Choroid plexus. In: Cerebral
Blood Flow: Effects of Nerves and Neurotransmitters.
Heistad, D. D., and M. L. Marcus, (eds.). Elsevier
North Holland, Inc. pp. 393-401, 1982.

Lofgren, J. Effects of variations in arterial pressure and
PaCO2 tensions on the cerebrospinal fluid pressure-
volume relationships. .Acta Neurol. Scandinav. 49:
586-598, 1973.

LOpez de Pablo, A. L., M. C. Gonzalez, G. Dieguez, B. Gomez,
and S. Lluch. Reduction of cerebrovascular reactivity
during hypercapnia. Am. J. Physiol. 242: 441-446,
1982a.

Lopez de Pablo, A. L., M. C. Gonzalez, G. Dieguez, B. Gomez,
and S. Lluch. Cerebrovascular responses to C02 after
inhibition of sympathetic activity. J. Appl. Physiol.:
Respirat. Environ. Exercise Physiol. 53: 873-878,
1982b.

171

Lux, W. and J. Fenstermacher. Cerebrospinal fluid formation
in ventricles and spinal subarachnoid space of the
rhesus monkey. J. Neurosurg. 42: 674-678, 1975.

MacKenzie, E. T., S. Strandgaard, D. I. Graham, J. V. Jones,
A. M. Hargper, and J. K. Farrar. Effects of acutely
induced hypertension in cats on pial arteriolar cali-
ber, local cerebral blood flow, and the blood-brain
barrier. Circ. Res. 39: 33-41, 1976.

Mann, D. D., A. B. Butler, R. N. Johnson, and N. H. Bass.
Clearance of macromolecular and particulate substances
from the cerebrospinal fluid system of the rat. J.
Neurosurg. 50: 343-348, 1979.

Maren, T. Bicarbonate formation in cerebrospinal fluid:
role in sodium transport and pH regulation. Am. J.
Physiol. 222: 885-899, 1972.

Marin, J. and F. Rivilla. Nerve endings and pharmacological
receptors in cerebral vessels. Gen. Pharmacol. 13:
361-368, 1982.

Marmarou, A., K. Shulman, and J. LaMogese. Compartmental
analysis of compliance and outflow resistnace of the
cerebrospinal fluid system. J. Neurosurg. 48: 332-
344, 1978.

Martins, A. N., J. K. Wiley, and P. W. Myers. Dynamics of
the cerebrospinal fluid and the spinal dura mater. J.
Neurol. Neurosurg. and Psychiat. 35: 468-473, 1972.

Martins, A. N., T. Doyle, and N. Newby. PC02 and rate of
formation of cerebrospinal fluid in monkey. Am. J.
Physiol. 231: 127-131, 1976.

Martins, A. N., N. Newby, and T. F. Doyle. Sources of error
in measuring cerebrospinal fluid formation by
ventriculocisternal perfusion. J. Neruol. Neurosurg.
and Psychiat. 40: 645-650, 1977.

Masuoka, D. and E. Hansson. Autoradiographic distribution
studies of adrenergic blocking agents. I. 14C-
phenoxybenzamine (Bensylyt NFN), an alpha-receptor-type
blocking agent. Acta Pharmacol. Toxicol. 25: 113-122,
1967.

172

Maxwell, Pu S. and D. C. Pease. The electron microscopy of
the choroid plexus. J. Biophys. Biochem. Cytol. 2:
467-474, 1956.

McCulloch, J., H. E. Savake, and W. Angerson. Regional
water permeability in CNS of conscious rats: effects of
hypercapnia and locus coeruleus lesions. In: Cerebral
Blood Flow: Effects of Nerves and Neurotransmitters.
Heistad, D. D. and M. L. Marcus (eds.). Elsevier North
Holland, Inc. pp. 509-516, 1982.

Milhorat, T. H., M. K. Hammock, J. D. Fenstermacher, D. P.
Rall, and V: A. Levin. Cerebrospinal fluid production
by the choroid plexus and brain. Science 173: 330-
332, 1971.

Millen, J. W. and G. E. Rogers. An electron microsCOpic
study of the choroid plexus in the rabbit. J. Biophys.
Biochem. Cytol. 2: 467-474, 1956.

Nakamura, S. and T. H. Milhorat. Nerve endings in the
choroid plexus of the fourth ventricle of the rat:
electronic microsc0pic study. Brain Res. 153: 285-
293, 1978.

Nakamura, S. and G. M. Hochwald. Effects of arterial PC02
and cerebrospinal fluid volume flow rate changes on
choroid plexus and cerebral blood flow in normal and
experimental hydrocephalic cats. {L Cerebral Blood
Flow and Metab. 3: 369-375, 1983.

Nathanson, J.IL B-Adrenergic-sensitive adenylate cyclase
in secretory cells of the choroid polexus. Science
204: 843-844, 1979.

Nelson, E. and M. Rennels. Innervation of intracranial
arteries. Brain 93: 475-490, 1970.

Nielsen, K. C., Ch. mean, and B. Sporrong. Ultrastructure
of the autonomic innervation apparatus in the main pial
arteries of rats and cats. Brain Research 27: 25-32,
1971.

Oldendorf, W. The Nervous System. In: The Basic Neuro-
sciences. 1. Raven Press, New York. pp. 279-289,
1975.

173

Oppelt, W. W., T. Maren, E. S. Owens, and D. P. Rall.
Effects of acid-base alterations on cerebrospinal
fluid formation. Proc. Soc. Exp. Biol. Med. 114: 86-
89, 1963.

Oppelt, W. W., C. S. Patlak, and D. P. Rall. Effect of
certain drugs on cerebrospinal fluid production in the
dog. Am. J. Physiol. 206: 247-250, 1964.

mean, C., L. Edvinsson, and K. C. Nielsen. Autonomic
neuroceptor mechanisms in brain vessels. Blood Vessels
11: 2-31, 1974.

mean, C. and L. Edvinsson. Histochemical and pharmacologi-
cal approach to the investigation of neurotransmitters,
with particular_regard to the cerebrovascular bed. In:
Neurogenic Control of Brain Circulation mean, C. and
L. Edvinsson (edsJ. Pergamon Press, Oxford. pp. 15-
38, 1977.

mean, C., L. Edvinsson, and J. E. Hardebo. Pharmacological
in 113:9 analysis of amine-mediated vasomotor functions
in the intracranial and extracranial vascular beds.
Blood Vessels. 15: 128-147, 1978.

mean, C. and J. E. Hardebo. Functional aspects of the
blood brain barrier with particular regard to effects
of Circulating vasoactive:neurotransmitters. In: Cere-
bral Blood Flow: Effects of Nerves and Neurotransmit-
ters. Heistad, D. D. and M. L. Marcus (eds.). pp.
119-128, 1982.

Page, I. H. and F. Olmsted. Influence of respiratory gas
mixtures on arterial pressure and vascular reactivity
in normal and hypertensive dogs. Circulation 3: 801-
819, 1975.

Page, R. H., D. J. Funsch, R. W. Brennan, and M. J.
Hernandez. Choroid plexus blood flow in the sheep.
Brain Res. 197: 532-537, 1980.

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

174

Patel, P., D. Bose, and C. Grenway. Effects of prazosin and
phenoxybenzamine andd- andp-receptor-mediated re-
sponses in intestinal resistance and capacitance ves-
sels. J. Cardiovasc. Pharmacol. 3: 1050-1059, 1981.

Pollay, M. and K. Welch. The function and structure of
canine arachnoid villi. J. Surg. Res. 2: 307-311,
1962.

Pollay, M. and H. Davson. The passage of certain substances
out of the cerebrospinal fluid. Brain 86: 137-150,
1963.

Pollay, M. and F. Curl. Secretion of cerebrospinal fluid by
the ventricular ependyma of the rabbit. Am. J.
Physiol. 213: 1031-1038, 1967.

Pollay, M., A. Stevens, E. Estrada, and R. Kaplan. Extra-
corporeal perfusion of choroid plexus. J. Appl.
Physiol. 32: 612-617, 1972.

Pollay, M. Formation of cerebrospinal fluid. J. Neurosurg.
42: 665-673, 1975.

Portnoy, H. D., M. ChoPP: C. Branch, and M. B. Shannon.
Cerebrospinal fuid pulse wave form as an indicator of
cerebral autoregulation. J. Neurosurg. 56: 666-678,
1982.

Preskorn, S. H., B. K. Hartman, M. E. Raichle, L. W.
Swanson, and H. B. Clark. Central adrenergic regula-
tion of cerebral microvascular permeability and blood
flow: pharmacological evidence. Adv. Exp. Biol. Med.
131: 127-138, 1978.

Preskorn, S. H., G. H. Irwin, S. Simpson, D. Firesen, J.
Rinne, and G. Jerovich. Medical therapies for mood
disorders alter the blood-brain barrier. Science 213:
469-471, 1981.

Preskorn, S. H., B. K. Hartman, G. H. Irwin, and C. W.
Highes. Role of the central adrenergic system in medi-
ating Amitriptyline-induced alteration in the mammalian
blood-brain barrier in vivo. J. Pharm. Exp. Ther.

223: 388-395, 1982.

175

Raichle, M. E. and H. L. Stone. Cerebral blood flow
autoregulation and graded hypercapnia. In: Cerebral
Blood Flow and Intracranial Pressure. I Europ. Neurol.
6: 1-5, 1971-72.

Raichle, M. E., B. K. Hartman, J. O. Eichling, and L. G.
Sharpe. Central noradrenergic regulation of cerebral
blood flow and vascular permeability. Proc. Nat. Acad.
Sci. 72: 3726-3730, 1975.

Raichle, M. E., J. O. Eichling, M. G. Straatman, M. J.
Welch, K. B. Larson, and M. M. TerPogossian. Blood-
brain barrier permeability of 11C-labeled alcoholics
and ISO-labeled water. Am. J. Physiol. 230: 543-552,
1976.

Raichle, M. E., R. L. Grubb, and J. O. Eichling. Neural and
hormonal regulation of brain water permeability. In:
Neurogenic Control of Brain Circulation. mean, C. and
L. Edvinsson (edsJ. Pergamon Press, Oxford. pp. 465-
470, 1977.

Reed, D. J. and M. H. Yen. The role of the cat choroid
plexus in regulating cerebrospinal fluid magnesium. J.
Physiol. 281: 477-485, 1978.

Reivich, M. Arterial PC02 and cerebral hemodynamics. Am.
J. Physiol. 206: 25-35, 1964.

Rennels, M. L. and E. Nelson. Capillary innervation in the
mammalian central nervous system: an electron micro-
scopic demonstration. Amer. J. Anat. 144: 233-241,
1975.

Rich, M., P. Scheinberg, and M. S. Belle. Relationship
between cerebrospinal fluid pressure and cerebral blood
pressure. Circ. Res. 1: 389-395, 1953.

Rosenberg, G., L. Wolfson, and R. Katzman. Pressure-
dependent bulk flow of cerebrospinal fluid into brain.
Exp. Neurol. 60: 267-276, 1978.

Rosenberg, G. A., W. T. Kyner, and E. Estrada. Bulk flow of
brain interstitial fluid under normal and hyperosmolar
conditions. Amer. J. Physiol. 238: 42-49, 1980.

176

Rosenberg, G., W. Kyner, and E. Estrada. The effect of
increased CSF pressure on interstitial fluid flow
during ventriculocisternal perfusion in the cat. Brain
Res. 232: 141-150, 1982.

de Rougemont, J., A. Ames III, F. B. Nesbitt, and H. F.
Hofmann. Fluid formed by choroid plexus. A technique
for its collection and a comparison of its electrolyte
composition with serum and cisternal fluids. J. Neuro-
physiol. 23: 485-495, 1960.

Ryder, H. W., F. F. Epsey, F. D. Kimbell, E. J. Penka, A.
Rosenauer, B. Podolsky, and J. P. Evans. The mechanism
of the change in cerebrospinal fluid pressure following
an induced change in the volume of the fluid space. J.
Lab. Clin. Med. 41: 428-435, 1953.

Sadoshima, S., M. Thames, and D. Heistad. Cerebral blood
flow during elevation of intracranial pressure: role
of sympathetic nerves. Am. J. Physiol. 241: 78-84,
1981.

Sadoshima, S. and D. Heistad. Sympathetic nerves protect
the blood-brain barrier in stroke-prone spontaneously
hypertensive rats. Hypertension 4: 904-907, 1982.

Sadoshima, S., D. W. Busija, and D. D. Heistad. Mechanism
of protection against stroke in stroke-prone spon-
taneously hypertensive rats. Am. J. Physiol. 244:
406-412, 1983.

Saeed, M., O. Sommer, J. Haltz, and E. Bassenge. -Adreno-
ceptor blockade by phentolamine causesfi -adrenergic
vasodilation by increased catecholamine release due to
presynapticcfl-blockade. J. Cardiovasc. Pharmacol. 4:
44-52, 1982.

Sata, O. and E. A. Bering. Extra-ventricular formation of
cerebrospinal fluid. Brain Nerve (Tokyo) 19: 883-885,
1967.

Sercombe, R., P. Aubineau, L. Edvinsson, H. Mamo, C. mean,
E. Pinard, and J. Seylaz. Neurogenic influence on
local cerebral blood flow: Effect of catecholamines or
sympathetic stimulation as correlated with the sympa-
thetic innervation. Neurol. 25: 954-963, 1975.

177

Sklar, F.1L and.I. Elashvili. The pressure-volume function
of brain elasticity. J. Neurosurg. 47: 670-679, 1977.

Sklar, F. H., J. Reisch, I. Elashvil i, T. Smith, and D. M.
Long. Effects of pressure on cerebrospinal fluid for-
mation: nonsteady-state measurements in dogs. Am. J.
Physiol. 239: 277-284, 1980.

Snodgrass, S. R. and A. V. Lorenzo. Temperature and
cerebrospinal fluid production rate. Am. J. Physiol.
222: 1524-1527, 1972.

Songer, M. J., W. Dang, S. R. Heisey, T. Adams, and H. K.
Fry. .Redistribution of intracranial fluid volumes
during hypercapnia (Abstract). Physiologist 23: 98,
1980.

Starke, K. :Regulation of noradrenaline release by presynap-
tic receptor system. Rev. Physiol. Biochem. Pharmacol.
77: 1-124, 1977.

Stone, H. L., M. E. Raichle, and M. Hernandez. The effect

of sympathetic denervation on cerebral C02 sensitivity.
Stroke 5: 13-18, 1974.

Swanson, L. W., M. A. Connel ly, and B. K. Hartman. Ultra-
structural evidence for central monoaminergic innerva-
tion of blood vessels in the paraventricular nucleus Of
the hypothalamus. Brain Res. 136: 166-173, 1977.

Tschirgi, R. D., R. W. Frost, and J. L. Taylor. Inhibition
of cerebrospinal fluid formation by a carbonic anhy-
drase inhibitor, 2-acetylamino-l,3,4-thiadiazole-5-
sulfonamide (Diamox). Proc. Soc. Exp. Biol. and Med.
87: 373-376, 1954.

Vates, T. 8., S. L. Bonting, and W. W. Oppelt. Na-K acti-
vated adenosine triphosphatase formation of cerebro-
spinal fluid in the cat. Am. J. Physiol. 206: 1165-
1172, 1964.

Vogh, B. 'The relation of choroid plexus carbonic anhydrase
activity to cerebrospinal fluid formation: study of
three inhibitors in cat with extrapolation to man. J.
Pharmacol. Exp. Ther. 213: 321-331, 1980.

178

Wald, A., G. M. Hochwald, and C. Malhan. The effects of
ventricular fluid osmolality'on bulk flow'of fluid into
cerebral ventricles of cats. Exp. Brain Res. 25: 157-
167, 1976.

Waltz, A. G., T. Yamaguchi, and F. Regli. Regulatory re-
sponses of cerebral vasculature after sympathetic
denervation. Am. J. Physiol. 221: 298-302, 1971.

Weed, L. H., L. B. Flexner, and J. H. Clark. Effect of
dislocation of cerebrospinal fluid upon its pressure.
Am. J. Physiol. 100: 246-261, 1932.

Weed, L. H. and L. B. Flexner. The relations of the intra-
cranial pressures. Am. J. Physiol. 105: 266-272,
1933.

Weiss, M. H. and N.‘Wertman. Modulation of CSF production
by alterations in cerebral perfusion pressure. Arch.
Neurol. 35: 527-529, 1978.

Welch, L. and V. Friedman. The cerebrospinal fluid valves.
Brain 83: 454-469, 1960.

Welch. K. Secretion of cerebrospinal fluid by the choroid
plexus of the rabbit. Am. J. Physiol. 205: 617-624,
1963.

Welch, K. The principles of physiology of the cerebrospinal
fluid in relation to hydrocephalus. In: Advances in
Neurology. I W. J. Friedlander (ed.). Raven Press,
New York, 1975.

Wright, E. and J. Prather. The permeability of the frog
choroid plexus to non-electrolytes. J. Membrane Biol.

Wright, E., G.‘Wiedner, and G. Rumrich. Fluid secretion by
the frog choroid plexus. Exp. Eye Res. 25: 149-155,
1977.

Wright, E. M. Transport process in the formation of the
cerebrospinal fluid. Rev. Physiol. Biochem. Pharmacol.
83: 1'34, 1978.

Wright, E. Relations between the Choroid plexuses and the
cerebrospinal fluid. TINS 1: 13-15, 1979.