o m' h...‘. ., .‘ -.~ -. < .. v...,......,..‘ v
.T - .11 ”(Jim 'W 99, '43.," ' t". vvﬂw \E‘v-x‘v manna. .Tu'v‘s‘c‘u u’uw- ~ -.,~ ~ - . , - .. . . A -‘- v" ...‘ ‘ ‘ ,. ' ’ " ‘ T T ‘w ‘
\ Hiqfél‘ﬁ‘ wrtﬁgkd‘d“; I‘ﬂ’i‘ﬁféﬁé‘é‘,‘:‘..‘1't'4'.'.'\‘.\‘-'\_'\if?1V1'.‘..‘.'.‘\‘.‘.v.3?.‘\‘.j.:\:t:T:-‘.'..Yx.~'Ttu“:\:\'L.‘u'm.' V't't" :‘.‘,‘ 2.} 'T‘t‘l'. :‘.'. ‘ 5 '_ 33:, . 3;} t. T -_ ~'\'~‘ . r I ‘ , ~ - ‘ _ g .

THE EFFECT OF ACUTE HYPOXIA 0N
GLUCOSE LACTATE, AND PYRUVATE
TRANSPORT FROM THE BRAIN

VENTRICLES OF THE DOG '

Thesis for the Degree of Ph; "D.

MICHIGAN STATE UNIVERSITY   7 7

DAVID KEITH .MmHAEL

 

\
“Mu-“am? "’Dn’um

   
      

55"" LIBRA R Y L
a Michigan State
1... University

This is to certify that the

thesis entitled

THE EFFECT OF ACUTE HYPOXIA ON GLUCOSE,
LACTATE AND PYRUVATE TRANSPORT
FROM THE BRAIN VENTRICLES OF THE DOG

presented by

David Re i th Michae 1

has been accepted towards fulﬁllment
of the requirements for

Ph.D. degree in Ehysiolggx

 

 

' Major professor /

0-7839

   

     

- \
~

i“ ‘ BINDI‘NG By
M]

  
 

HUAB 81 SUNS'
\ BHOKIB‘WQFDY ms. 51
I E I iii-fin!”

ABSTRACT
THE EFFECT OF ACUTE HYPOXIA ON GLUCOSE,

LACTATE, AND PYRUVATE TRANSPORT FROM
THE BRAIN VENTRICLES OF THE DOG

BY

David Keith Michael

The brain ventricular system of anesthetized adult
dogs was perfused from the right cerebral ventricle to the
cisterna magna with an artificial cerebrospinal fluid (CSF).
In one series of experiments, dogs were ventilated with
room-air (21% 02) during the 240 minutes of ventriculo-
cisternal perfusion. In another series, an initial 2 hour
normoxic (PaO2 = 97 i 1 mm Hg) perfusion period was followed
by a 2 hour period of ventilating the dogs with a 5-8% 02 in
N2 mixture (hypoxia; PaOZ = 39 i 2 mm Hg). The perfusion
inflow fluid contained inulin and trace quantities of radio-
actively labelled glucose or mannitol and lactate. The CSF
concentrations of glucose, pyruvate, and lactate were low-
ered, unaltered, or elevated from their normal values by
changing the concentrations in the perfusion inflow fluid.
In certain experiments, mannitol was included in the per-

fusion inflow fluid in lieu of glucose. Steady-state

measurements of inflow and outflow rates and concentrations

David Keith Michael

of the test molecules enabled the calculation of the rates
at which each was removed from or entered the perfusion
fluid.

Inulin clearance allowed estimation of CSF bulk
absorption (Ta) and formation (Of) rates. Inulin clearance
increased (p‘<0.05) with time (6:: 2 ul/min) and with
hypoxia (9:13 ul/min). vf in normoxic dogs was 50:23 ul/min.
Of decreased (p<:0.05) with time (72:2 ul/min), but decreased
more with hypoxia (15:t4 ul/min).

Transependymal net glucose flux during both normoxia
and hypoxia reflected the glucose concentration gradient
between the CSF and plasma. The glucose outflux coefficient
(KO) demonstrated saturation kinetics implying carrier-
mediated glucose efflux, whereas mannitol KO (l6zt3 ul/min)
was independent of concentration. Glucose K0 was less at
normal CSF glucose concentration (5.0 mM) than at low con—
centration (0.0 mM), but its value (78::8 ul/min) was
greater than that for mannitol, suggesting another glucose
transport system of higher capacity. Glucose Ko like that
of mannitol was unaffected by time, but in contrast to
mannitol was reduced during hypoxia, suggesting an aerobic
dependent glucose efflux.

There was a net lactate flux into CSF at CSF lactate
concentrations ranging from 0.0 to 5.0 mM. Net lactate
influx increased (p‘<0.05) with time, but increased more

(p ‘<0.05) during hypoxia. Carrier-mediated lactate

David Keith Michael

transport was indicated by the lack of a direct linear
relationship between net lactate flux and CSF lactate
concentrations. Lactate K0 was concentration dependent and
demonstrated saturation kinetics when lactate concentrations
in the perfusion inflow fluid were normal (1.6 mM). K0 was
unaffected by perfusion time, but decreased with hypoxia
when lactate concentration in the perfusion inflow was

0.0 mM.

There was always a net pyruvate flux into CSF even
at CSF pyruvate concentrations of 3.0 mM. Carrier-mediated
pyruvate tranSport was indicated by the lack of a direct
linear relationship between net pyruvate flux and pyruvate
concentrations in the CSF. Neither time nor hypoxia

affected net pyruvate influx.

THE EFFECT OF ACUTE HYPOXIA ON GLUCOSE,
LACTATE, AND PYRUVATE TRANSPORT FROM

THE BRAIN VENTRICLES OF THE DOG

BY

David Keith Michael

A THESIS
Submitted to
Michigan State University

in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Physiology

1972

To my wage, Lynn,
whose unde/wtanding and Auppott made this manubcvcép/t possibfe,
and to my chi/Edam, Scott and Euzabe/th.

ACKNOWLEDGMENTS

Many people have contributed significantly
throughout the course of this study, and the author wishes
to express his sincere gratitude to all.

The author wishes to thank his advisor, Dr. S. R.
Heisey, for his constructive criticism, and financial sup-
port throughout this program, and to the other members of
his academic advisory committee: Drs. T. Adams, W. D.
Collings, J. R. Hoffert, and W. W. Wells for their guidance
and support.

Recognition is given to Drs. C. Cress, T. A.
Helmrath, J. R. Hoffert, B. H. Selleck, and W. W. Wells
and to D. W. Bierer for their invaluable technical assis-
tance, and to D. K. Anderson for assistance with the
photography. I wish to express special thanks to Miss
Nancy Turner for her conscientious efforts in the pre-

liminary typing of this manuscript.

iii

TABLE OF CONTENTS

Page

LIST OF TABLES O O O O O O O O O O 0 O I O O O O O O O Vii

LIST OF FIGURES O O O O O O O O O O O O O O O O O O O 0 ix
I. INTRODUCTION . . . . . . . . . . . . . . . . . . 1
II. LITERATURE REVIEW . . . . . . . . . . . . . . . 5

2.1. Relationship of cerebrospinal fluid
to blood and brain . . . . . . . . . . 5
2.2. Cerebrospinal fluid (CSF) formation . . 9
2.3. Cerebrospinal fluid (CSF) absorption . . 13

2.4. Quantitative measurement of

molecular movement from the CSF . . . 15

2.5. Glucose movement among blood, brain,

and CSF . . . . . . . . . . . . . . . 18
2.6. Lactate and pyruvate movement among

blood, brain, and CSF . . . . . . . . 23
2.7. Hypoxia . . . . . . . . . . . . . . . . 28

III. STATEMENT OF PROBLEM . . . . . . . . . . . . . . 36
IV. METHODS AND MATERIALS . . . . . . . . . . . . . 37
4.1. General operative and cannulation
procedures . . . . . . . . . . . . 37
4.2. Brain ventricular and cisternal
puncture . . . . . . . . . . . . . . 40
4.3. Experimental perfusion with test
molecules . . . . . . . . . . . . . . 44
4.4. Experimental criteria . . . . . . . . . 45
4.5. Sample collection and storage . . . . . 45
4.6. Determination of normal CSF and plasma
metabolite concentrations . . . . . . 46

4.7. Measurement of blood gas tensions

and pH . . . . . . . . . . . . . . . . 46
4.8. Measurement of inflow and outflow rates

and concentrations . . . . . . . . . . 47
4.9. Experimental design . . . . . . . . . . 48

iv

VI.

VII.

J}

hub
O
.500

bnbtb

.5
o

creed rd m

0
\D
O

MUN
oo 0

mom
0 o o
NNN
o 0

“NH
0 o 0

4.9.2.4.

4.10.
4.10.1.
4.10.2.
4.10.3.
4.10.4.
4.10.5.
4.10.6.

4.11.
4.12.

RESULTS

Long duration brain ventricular
perfusions in normoxic dogs . . .
Elevated blood and low CSF inflow
concentrations . . . . . . . . . .
Normal CSF inflow concentrations . .
Elevated CSF inflow concentrations .
Brain ventricular perfusion during
normoxia and hypoxia . . . . . . .
Low CSF inflow concentrations . . .
Elevated CSF inflow concentrations .
Low CSF inflow concentrations with
mannitol . . . . . . . . . . . . .
Elevated CSF inflow concentrations
with mannitol . . . . . . . . .
Principles and calculations .
Definition of symbols .
Fluid balance . . . -.'
Bulk absorption rate, Va . .
CSF formation rate, Vf . . . .
Derivation of net flux, JX . . . . .
Derivation of transependymal outflux
coefficient, KO . . . . . . . . .
Calculated parameters . . . . . . .
Statistical methods . . . . . . . .

Arterial pH, P02, PC02, and rectal
temperature . . . . . . . . . . .
Cerebrospinal fluid bulk absorption
and formation rates . . . . . . .
Normal plasma and CSF concentrations
of glucose, pyruvate, and lactate
Glucose flux from CSF .
Lactate flux from CSF .
Mannitol flux from CSF .
Pyruvate flux from CSF .

DISCUSSION 0 O O O O O O O O O O O O O O O O

O‘C‘O‘
o o o
obWN
o o 0

SUMMARY

CSF bulk absorption and formation

rates . . . . . . . .
Glucose flux from CSF .
Lactate flux from CSF .
Pyruvate flux from CSF .

Page

51

51
52
52

53
54
54

55

55
56
56
57
57
58
58

59
60
6O

61

61
62

62
63
65
67
68

79
79
81
84
86

88

APPENDICES

A. COMPOSITION AND PREPARATION OF ARTIFICIAL

DOG CEREBROSPINAL FLUID (CSF)

B. PREPARATION OF ANESTHETIC . . .

C. DEPROTEINIZATION OF PLASMA AND CSF SAMPLES

D. SPECTROPHOTOMETRIC DETERMINATION OF D-

GLUCOSE O O O O O O O O O I O

E. FLUOROMETRIC DETERMINATION OF PYRUVATE

F. SPECTROPHOTOMETRIC DETERMINATION OF L-

I‘ACTATE O O I I O O O O O O O

G. SPECTROPHOTOMETRIC DETERMINATION OF INULIN

H. ISOLATION AND IDENTIFICATION OF
ACTIVELY LABELLED GLUCOSE . .

I. ISOLATION OF LABELLED LACTATE .
J. LIQUID SCINTILLATION COUNTING .
K. STATISTICAL FORMULAE . . . . .

LIST OF REFERENCES . . . . . . . . . .

vi

RADIO-

Page

90
92

94

96

99

105

110

113
120
127
133

139

LIST OF TABLES

Table Page

1. Arterial pH, P02, and PC02 and rectal
temperature in normoxic and hypoxic dogs
during brain ventricular perfusion . . . . . . . 69

2. Bulk absorption rate and CSF formation rate
in normoxic and hypoxic dogs during brain
ventricular perfusion . . . . . . . . . . . . . 70

3. Glucose, lactate, and pyruvate concentrations
in simultaneously obtained samples from
arterial plasma and cisternal cerebrospinal
fluid in twelve anesthetized dogs . . . . . . . 71

4. D-glucose in the inflow fluid, outflow fluid,
and arterial plasma in normoxic and hypoxic
dogs during brain ventricular perfusion . . . . 72

5. Glucose transependymal outflux coefficient
and net flux rate at low, normal, and
elevated perfusion inflow concentrations
in normoxic and hypoxic dogs during brain
ventricular perfusion . . . . . . . . . . . . . 73

6. L-lactate concentration in the inflow fluid,
outflow fluid, and arterial plasma in
normoxic and hypoxic dogs during brain
ventricular perfusion . . . . . . . . . . . . . 74

7. Lactate transependymal outflux coefficient
and net flux rate at low, normal, and
elevated perfusion inflow concentrations
in normoxic and hypoxic dogs during brain
ventricular perfusion. . . . . . . . . . . . . . 75

8. Mannitol transependymal outflux coefficient
at low and elevated perfusion inflow
concentrations in normoxic and hypoxic dogs
during brain ventricular perfusion . . . . . . . 76

vii

Table Page

9. Mean pyruvate concentrations in the inflow
fluid, outflow fluid, and arterial plasma
in normoxic and hypoxic dogs during brain
ventricular perfusion . . . . . . . . . . . . 77

10. Pyruvate net flux at low, normal, and
elevated perfusion inflow concentrations
in normoxic and hypoxic dogs during brain
ventricular perfusion . . . . . . . . . . . .. 78

H-l. Isolation and recovery of 3H-glucose in the
inflow and outflow fluids in dog Scl . . . . . 119

I-l. Isolation of 14C-lactate in ventricular
perfusion effluent . . . . . . . . . . . . . . 126

K-l. Statistical block of glucose KO data from
15 normoxic anesthetized dogs at different
inflow fluid concentrations during two
experimental brain ventricular perfusion
periods . . . . . . . . . . . . . . . . . . . 137

K-2 0 AOV table 0 C C O O O O O O O O O O O O O O O 138

viii

LIST OF FIGURES

Page
Diagram of the eXperimental animal and
equipment . . . . . . . . . . . . . . . . . . 39
Photograph of equipment used for
ventriculocisternal perfusion in the
anesthetized dog . . . . . . . . . . . . . . . 43
Diagram showing the time of experimental
manipulations and sample collections . . . . . 50
Elution profile of CSF outflow sample
containing 3H-glucose and l4C-lactate . . . . 125
133Barium external standard quench
correction curves fgﬁ differential
counting of 3H and C samples . . . . . . . . 132

ix

I . INTRODUCTION

The absence of any significant cerebral arterio-
venous (A-V) plasma concentration differences for substrates
other than glucose (Sokoloff, 1960) and a respiratory quo-
tient (R.Q.) approximating l (Gibbs et al., 1942) are con-
vincing evidence that the brain derives its energy almost
exclusively from glucose oxidation. Glucose is probably
preferentially utilized by the adult mammalian brain, since
the hypoglycemic effects on mammalian cortical electrical
activity (E.E.G.) can be counteracted by iv glucose injec—
tions, but not by injections of fructose, galactose, hexose
diphosphate, glyceric aldehyde, pyruvate, succinate,
fumerate, or glutamate (Maddox et al., 1939).

Glucose removed from blood is metabolized to
pyruvate via the Embden-Myerhoff pathway within brain tissue.
Pyruvate can be reduced to lactic acid, converted to acetyl
coenzyme A by oxidative decarboxylation, carboxylated to
form oxaloacetic or malic acids, or transaminated to form
alanine. Ultimate conversion to carbon dioxide (C02) occurs
via the tricarboxylic acid cycle and yields the high energy
nucleotide triphosphates necessary for normal tissue func-

tion. When l4C-glucose is perfused into blood, only 30-35%

of the l4C-glucose taken up by brain is directly oxidized

to 14CO2 and H20 and large amounts are converted into
lipids, proteins, and other acid soluble components (Geiger,
1958). Only 1/3 of the glucose utilized by brain is for the
direct formation of high energy molecules (i.e., adenosine
triphosphate,.ATP; phosphocreatine, PCr) and the remaining
glucose (i.e., 65-70%) is metabolized for other purposes.

Glucose and glycogen content in the mammalian brain
total 125-200 umoles/lOO g of brain and glucose consumption
is 31 pmoles/lOO 9 brain per minute (Sokoloff, 1960).
Therefore, when cerebral blood flow (CBF) is diminished or
arrested, cerebral function can only be maintained for a
short time (i.e., 4-6 minutes) through utilization of its
glucose and glycogen stores.

The large oxygen consumption (V02) of mammalian
brain tissue (approximately 20% of total body V02 at rest
(Sokoloff, 1960)) coupled with low carbohydrate reserves
accounts for the extreme sensitivity of the mammalian brain
to anoxic, hypoxic, ischemic, and hypoglycemic conditions
(Maddock et aZ., 1939; Lowry et al., 1964; Dahl and Balfour,
1964; White et aZ., 1964). Ischemia results in the reduc-
tion of brain glucose and glycogen, as well as the reduction
of high energy ATP and PCr, and an increase in lactic acid,

inorganic phosphate, adenosine diphosphate (ADP), and

creatine (Cr) (Lowry et al., 1964; Gatfield et al., 1966;

Folbergrova at al., 1970) indicating an increase in
anaerobic metabolism. Complete cessation of cerebral
circulation in man results in unconsciousness within 10
seconds (Rossen et aZ., 1943) the time interval required

to use the estimated 02 content in brain tissue (Kety, 1950).
In the absence of 02, pyruvate is reduced to lactic acid.
Lactic acid once formed cannot be used for energy production
unless there is a transformation to pyruvic acid, an event
associated with aerobic conditions (White et aZ., 1964).

It can be inferred that in the mammalian brain aerobic
glucose utilization is not only a preferential pathway but
an obligatory one.

Normal brain function depends upon the maintenance
of a relatively constant ionic composition and adequate
metabolite concentration in the interstitial fluid (ISF)
bathing the neurons (Kerr and Ghantus, 1936; Merlis, 1960).
There is substantial interest in solute exchanges between
blood and brain tissue under various physiological condi-
tions. The exchange of water and solutes can take place
between blood plasma and brain ISF by two routes. A direct
exchange may occur across blood capillaries into brain
tissue, however, brain capillaries are much less permeable
to solutes than blood capillaries in other tissues (Ferguson
and WOodbury, 1969; Brooks et aZ., 1970) and direct solute
exchange between capillary blood and brain tissue is limited

by this "blood brain" barrier. The brain is in communication

with another fluid in addition to the blood, the
cerebrospinal fluid (CSF) (Truex and Carpenter, 1969), and
an indirect solute exchange between blood and brain can
occur via the CSF. Consequently, the manner in which brain
ISF solute composition changes depends on the nature of
transfer between blood and brain ISF, CSF and brain ISF,
and blood and CSF.

Glucose moves from blood into brain and from blood
into CSF by facilitated diffusion (Fishman, 1964; Crone,
1965; LeFevre and Peters, 1966), while pyruvate and lactate
are hypothesized to move by simple diffusion (Siesjo et aZ.,
1968). Simultaneous rates of the movements of glucose,
pyruvate, and lactate across the CSFebrain and/or CSF-blood
barriers have not been measured. Knowing the flux rates and
mode of transfer for these metabolites across the barriers
separating CSF from blood and/or brain ISF during normoxia
and hYpoxia would aid in understanding brain metabolism and

brain survival in viva.

II . LITERATURE REVIEW

2.1. Relationship of cerebrospinal fluid to blood and brain

In the adult mammalian central nervous system (CNS)
the cerebrospinal fluid (CSF) occupies two main compartments,
the brain ventricles and the subarachnoid spaces. The brain
ventricles consist of four interconnecting cavities: the
two lateral ventricles, which connect by the intraventric-
ular foramina of Monroe to the midline third ventricle,
which communicates via the aqueduct of Sylvius with the
fourth ventricle. The fourth ventricle connects with the
subarachnoid spaces by the foramina of Luschka. The sub—
arachnoid spaces are formed by the inner pial membrane,
which follow the outer contours of the brain, and the
arachnoid membrane, which is in apposition with the outer-
most dural membrane.

In the subarachnoid spaces, the CSF is separated
from brain tissue by the pial membrane. As seen through the
light microsc0pe, the pia is a thin avascular membrane and
the blood (pial) vessels supplying brain tissue lie on the
surface of the membrane penetrating the pia to enter or to
return,from brain tissue (Millen and Woollam, 1961). Forbes

(1928) studied these blood vessels in vivo and failed to

observe capillary loops, although he described small
arterioles which permitted only a single column of
erythrocytes to pass through them. Similar observations
reported by Flexner (1933) and Hassin (1948) suggest that
capillaries may be present in the subarachnoid spaces and
would permit solute exchange between blood and CSF.

In four discrete regions, the pia is modified into
highly vascular tissue. During fetal development, the pia
and accompanying blood vessels invaginate and project into
the brain ventricles, and with the neural epithelium
(ependyma) lining the ventricles constitute the choroid
plexuses (Arey, 1962; Davson, 1967). In these regions,
i.e., the roofs of the third and fourth ventricles and the
walls of the lateral ventricles, the membranes separating
CSF from blood consists of two cell layers, the capillary
endothelium and the ventricular ependyma (Davson, 1967).

Within the brain ventricles, CSF is separated from
the surrounding neural tissue and brain extracellular fluid
(ECF) by a single layer of ependymal cells, which comprise
the walls of the ventricles. Some ependymal cells are close
to cerebral blood vessels (Horstman, 1954) and may possess
secretory functions (Fleischauer, 1964).

The relationship of CSF to brain ECF space and
plasma has been demonstrated by Wallace and Brody (1937;
1939) and by Olsen and Rudolph (1955). Wallace and Brody

(1937; 1939) administered salts of iodide, thiocyanate, and

bromide iv and after 24 hours determined the anion to
chloride ratios present in the plasma and in various tissues
of the dog. In all tissues examined except the brain, the
anion to chloride ratio was the same as its ratio in the
plasma, but in the brain the anion to chloride ratio was

the same as in the CSF and lower than in the plasma. More

24Na and 82Br

recently, Olsen and Rudolph (1955) studied
transfer among blood, CSF, and brain tissue. They reported
that following iv administration of these ions equilibration
between the blood and the CSF required more than 16 hours,
while equilibration between CSF and brain occurred in less
than 30 minutes. Collectively, these data suggest that
brain ECF equilibrates not with plasma but with CSF.

The first attempt to estimate the size of brain ECF
space by placing ECF markers in the CSF was reported by
Davson et al. (1961), who repeatedly replaced the CSF in
rabbits with an artificial CSF containing a known sucrose
concentration. The sucrose concentration in the subarach-
noid spaces was maintained relatively constant for 2-3 hours
by barbotage. The authors reported a spinal cord sucrose
space of 12-14%. The brain sucrose space was 8-9% and this
lower sucrose Space was presumed to be due to inadequate
mixing in CSF surrounding the brain surfaces.

Rall et a1. (1962) measured brain ECF space by a

different technique. They continuously perfused the dog's

brain ventricular system through a cannula in the lateral
ventricle and collected fluid from another cannula in the
cisterna magna, a technique first developed by Leusen (1948).
The perfusion fluid contained l4C-inulin as the ECF marker.
After 3-5 hours of perfusion the dog was killed and the
brain removed. Small coronal sections 5 mm thick were cut
from select areas of the brain at successively greater
distances from the ventricular surface and analyzed for
14C activity. Activity was greatest in the brain sections
immediately adjacent to the ventricular system and decreased
progressively with each successive section, suggesting
incomplete equilibration of the total ECF with CSF. The
authors concluded that inulin penetration into brain ECF
from CSF was slow; the inulin brain space was 12%. Recently,\
Levin et a1. (1970) perfused the cerebral subarachnoid
spaces in four different mammalian species and reported
cerebral cortical 14C-sucrose and 3H-inulin spaces of
17-20%.

The brain ECF Space estimated by the presentation
of the ECF marker via the CSF is 12-20% of the total brain
volume. Intravenous injection of ECF markers and the
resultant 2-8% brain ECF space (Davson and Spaziani, 1959;
Barlow et aZ., 1961; Reed and Woodbury, 1963) has been sug-
gested by Davson et a2. (1961) to be due to the "sink action"
of CSF, the brain ECF marker being continuously diluted by

CSF production and CSF absorption into the blood.

This "sink-action,” has been circumvented by
combined presentation of the ECF marker via ventriculo-
cisternal perfusion and infusion into the general circula-
tion (Wbodward et al., 1967; Cutler et al., 1969; Baethman
et al., 1970). These reports suggest a mammalian brain ECF
space of 10-15%, a brain ECF space comparable to that

obtained by perfusion of the CSF alone (Rall et aZ., 1962)°

2.2. Cerebrospinal fluid (CSF) formation

CSF is a colorless fluid containing ions which are
also found in the blood, but unlike blood, contains no cells
and only a very small amount of protein (Davson, 1967). In
general, CSF has higher Cl— and Mg++ concentrations than in
a plasma dialysate (Kemeny at aZ., 1962; Friedman et aZ.,
1963) while CSF concentrations of K+, Ca++, and glucose are
lower (Bito and Davson, 1966; Oppelt et aZ., 1963b). Ultra-
filtration of blood is not sufficient to explain CSF forma-
tion since an energy source other than blood hydrostatic
pressure must be responsible for the observed ionic
concentrations.

Some of the first studies on CSF formation were
conducted by Dandy and Blackfan (1914) and Frazier and Peet
(1914), who produced experimental hydrocephalus in the dog
by plugging the aqueduct of Sylvius and proved beyond doubt

that CSF is formed within the brain ventricles. Dandy (1919)

showed the involvement of the choroid plexus by unilaterally

10

plexectomizing the dog while blocking both foramina of
Monroe, which resulted in the deve10pment of a unilateral
hydrocephalus in the nonplexectomized ventricle. Bering
and Sato (1963) unilaterally plexectomized dogs but blocked
the aqueduct of Sylvius or the cisterna magna with kaolin
while leaving the foramina of Monroe open, and reported the
development of an assymetrical hydrocephalus.

More direct evidence bearing on the secretory activ-
ity of the choroid plexus has been provided by studies on
the exPosed choroid plexus (Welch, 1963; Ames et al., 1965a).
Welch (1963) cannulated the large vein draining most of the
venous blood from the lateral ventricular choroid plexus in
the rabbit and reported a venous to arterial hematoCrit
ratio of 1.15, and confirmed that fluid was lost from blood
upon passage through the choroid plexus. Ames et al. (1965a)
collected under oil and chemically analyzed the freshly
secreted fluid from the exposed choroid plexus of the

lateral ventricle of the cat. The concentrations of Na+,

K+, Ca++, Mg++ and Cl- were sufficiently like those in CSF
collected from the cisterna magna, and sufficiently differ-
ent from a plasma filtrate to suggest that this fluid was a
secreted CSF and not a pathological plasma exudate.

Although such evidence favors the hypothesis that the
choroid plexuses are sites of CSF formation, the ventricular

ependyma within the aqueduct of Sylvius has also been

reported by Pollay and Curl (1967) to be a source of CSF.

11

Bering (1959) reported a CSF formation rate (Vf)
of 0.050 ml/min in the anesthetized dog by collecting CSF
outflow from a cannula placed in the cisterna magna. Vf
could be correlated with cerebral oxygen consumption (CMROZ)
and with cerebral blood flow (CBF), but neither CMRO2 nor
CBF alone could account for the total CSF produced.

The first drug unequivocally shown to affect CSF
production was the carbonic anhydrase inhibitor, acetazola-
mide (Diamox). Pollay and Davson (1963) reported approx-

imately a 50% reduction in V in the rabbit following either

f
intravenous or intracisternal administration of the drug.
Intravenous Diamox administration results in a 43% inhib-
ition of vf in both the adult (Oppelt et al., 1964) and
newborn dog (Holloway et al., 1972). Similarly, the rate
of fluid formation by the eXposed choroid plexuses is
decreased following iv Diamox administration (Ames et aZ.,
1965b). This suggests that the formation and dissociation
of carbonic acid (HZCO3) may be a necessary event in the
metabolism of the cells responsible for CSF formation, and
interference with this metabolism reduces the secretory
processes.

Oppelt et al. (1963a) studied the effects of
acidosis and alkalosis on Vf in the dog. Acidosis, produced
by either iv administration of HCl or 5-10% CO2 inhalation
had no consistent effect on V . Alkalosis induced by iv

f
bicarbonate administration caused a 23% reduction in Vf,

12

which could be further reduced to 45% by the simultaneous
administration of Diamox. Respiratory alkalosis resulted

in a 46% reduction in V and the administration of Diamox

f

during hypocapnia resulted in a 62% reduction of V Ames

f.
at al. (1965b) reported decreased fluid formation by the
exposed choroid plexus following reduction of arterial CO2
(PaCOZ) by hyperventilation. These authors also reported
an increased fluid formation when PaCO2 was increased by
10% CO2 inhalation. It was observed that the PaCO2 effects
were associated with vasomotion of the choroidal artery, so
that the rate limiting factor in CSF formation by the
choroid plexus may have been its blood supply. Since iv
Diamox results in choroidal arterial constriction in the
isolated perfused choroid plexus comparable with that pro-
duced by iv norepinephrine administration (Marci et aZ.,

1966), part of the Diamox effect on V may be mediated via

f
the vascular system.

Vf is in part an active process as shown by studies
using either dinitrophenol (DNP) or the cardiac glycoside,
ouabain. DNP blocks the oxidative phosphorylation of ADP
and in general abolishes the active transport of ions (e.g.,
the active extrusion of Na+ from the squid's giant axon
(Hodgkin and Keynes, 1955)). Davson and Pollay (1963)
perfused the brain ventricles of the rabbit and reported

Vf was reduced 45% from control values when DNP (0.05 mM)

was included in the perfusion fluid. Most secretory

13

processes involving the active transport of Na+ and K+
involve the enzyme ATPase, which is specifically inhibited
by ouabain. Some early attempts to demonstrate inhibition
of Vf following iv administration of ouabain were not very
successful (Davson and Pollay, 1963; Oppelt et al., 1964).

By contrast, Welch (1963) reported that CSF formation by

the choroid plexus could be halved by 10-5 M ouabain iv,

and 10-4 ouabain resulted in complete inhibition. Vates

et a2. (1964) have reported the in vitro Na-K ATPase
activity of the excised cat choroid plexus was reduced

about 75% when 10‘-5 M ouabain was included in the incubation
media. Vf was reduced significantly when ouabain was per-
fused through the ventricles of the cat (Vates et al., 1964),
the rabbit (Davson and Segal, 1970), and both the adult
(Cserr, 1965) and newborn dog (Holloway et aZ., 1972). The
studies of drug effects on Vf, particularly those of ouabain,
imply that CSF production probably involves the active trans-
port of Na+ and K+ across the choroidal epithelium. Such
evidence indicates that CSF is a secretion, in the sense
that active transport mechanisms are Operative during its

elaboration and metabolic energy is required.

2.3. Cerebrospinal fluid (CSF) absorption
Drainage of CSF from the ventricular system is via
the foramina of Luschka into the subarachnoid space. CSF

drainage from the subarachnoid space appears to be into

14

large endocranial venous sinuses, whose dural sheaths are
perforated in certain places by numerous fingerlike evag-
inations (villi) of the arachnoid membrane projecting into
the lumen of the sinus (Weed, 1923).

Welch and Friedman (1960) described the villi
initially found by Weed as a series of interconnecting tubes.
These investigators excised pieces of dural membrane con-
taining villi and mounted the membrane so that it separated
two fluid filled chambers. A hydrostatic pressure of about
10 mm H20 was required to initiate fluid flow from the CSF
side of the membrane to the blood side. Fluid flow was
independent of colloidal osmotic pressure on the blood side
and particulate suspensions up to 7 u diameter placed on the
CSF side passed through the villi without back filtering.
They concluded that the arachnoid villi act as one way
valves, through which both large and small molecules may
exit from CSF into blood. Since large molecules, e.g.,
inulin or dextran, exit from the subarachnoid space
(Rothman et aZ., 1961; Davson et al., 1962), smaller
molecules and probably all the dissolved substances in
CSF, whether naturally occurring or exogenously introduced

will leave CSF by this mechanism of bulk absorption.

15

2.4. Quantitative measurement of molecular movement from
the CSF

 

Molecules traverse membranes by simple diffusion
and/or by carrier-mediated transport. Diffusion is the net
movement of molecules from a region of greater concentration
to a region of lesser concentration (Stein, 1967). In a
carrier-mediated transport system the transported molecules
or ions must attach themselves to a molecule present in the
cell membrane in order to traverse the membrane. These
carriers are assumed to be limited since when the concen—
tration of the transported molecules on one side of the
membrane is large all the carriers become occupied (i.e.,
saturated) and the transport rate across the membrane cannot
be increased by further increases in the concentration of
the transported molecule unless a non-carrier transport
(e.g., diffusion) is also possible (Davson and Danielli,
1952). Among the criteria for identifying carrier-mediated
transport are demonstrations of self-saturation of the
carrier and/or competition for the carrier between struc-
turally analogous compounds. Carrier-mediated transport
along either an electrical or a chemical gradient is
described as "passive" (e.g., facilitated diffusion) and
carrier-mediated transport against an electrochemical
gradient requires the expenditure of energy and is referred

to as "active transport" (Stein, 1967).

l6

Davson et a2. (1962) have shown that when 24Na,
inulin, and p-aminohippurate (PAH) are simultaneously
injected intracisternally in the rabbit, and their concen-
trations analyzed after one hour, the rate of molecular loss
from CSF was always 24Na:>PAH:>inu1in. These results imply
that 24Na and PAH leave the CSF by route(s) other than bulk
absorption. Heisey et a1. (1962) perfused the brain ven-
tricles and estimated the rate of bulk absorption (Va) in
the unanesthetized goat. Both flow rates and steady-state
concentrations of inulin in the inflow and outflow fluids
were measured and CSF clearance calculated using a formula
analogous to that for calculating renal clearance. At
intraventricular pressures below ~15 cm H20, all the inulin
(M.W.==5000) perfused into the ventricular system was
recovered in the cisternal effluent, and inulin clearance
was zero, indicating that inulin doesn't diffuse from the
ventricles. When intraventricular pressure was raised above
-15 cm H20, inulin clearance increased linearly with pres-
sure, indicating that inulin was removed from CSF mainly by
bulk absorption distal to the fourth ventricle. Inulin
clearance was considered an accurate estimate of Va. In
addition to inulin, the CSF clearances of tritiated water
(3mg; M.W. = 20), urea (M.W. = 60), creatinine (M.W. = 113) ,
and fructose (M.W.==180) were estimated and their clearances
from CSF always exceeded that of inulin. As intraventric-

. 3
ular pressure increased the clearances of HOH, urea,

17

creatinine, and fructose increased proportionally with the
increase in inulin clearance. The clearance due to non-bulk
absorptive processes (K0) was estimated by subtracting from
the total clearance the clearance due to Va, estimated by
inulin clearance, and was found to be independent of intra-
ventricular pressure.

Pappenheimer at al. (1961) reported that Diodrast
was cleared from CSF approximately 3 times more rapidly
than creatinine in goats whose brain ventricles were per-
fused with an artificial CSF. Considering only molecular
size, creatinine (M.W.==113) should leave CSF more rapidly
than Diodrast (M.W.==401). At elevated CSF Diodrast con-
centrations, they demonstrated saturation of the Diodrast KO
indicating carrier-mediated transport of Diodrast from the
CSF. Diodrast transfer from CSF related to increasing per—
fusate concentrations was a 2 component curve: (1) a fast
rate of transfer which was non-linear at Diodrast concen-
trations below 25 ug/ml and (2) a slower, linear rise of
transfer rate, with increasing Diodrast concentrations above
25 ug/ml. More recently, Bierer (1972) reported that PAH is
transported from the perfused CSF system of the adult dog,
as indicated by competitive inhibition of transport by

another organic anion, Diodrast, and self-saturation.

18

2.5. Glucose movement among_blood, brain, and CSF

Geiger et al. (1954) perfused the vasculature of
the cat's brain with an artificial blood composed of ox
erythrocytes, serum-albumin, and Ringer's solution, and
reported that the brain was rapidly depleted of glucose
despite the maintenance of a high glucose concentration in
the perfusion fluid. Further experiments showed that glu—
cose depletion was not the result of increased metabolism,
but was primarily due to the failure of the brain to remove
glucose from the blood. Addition of liver extract to the
perfusion fluid, or the use of donor blood, abolished this
effect suggesting that the blood-brain barrier to glucose
was controlled by a substance(s) normally circulating in
the blood. When a liver extract was included in the per-
fusion fluid, brain glucose concentration paralleled vari-
ations in the glucose concentration of the perfusion fluid.
Elevations in perfusion fluid glucose concentration above
10 mM produced little or no increase in brain glucose
concentrations, suggesting the possible involvement of a
carrier-mediated transport of glucose from blood to brain.

More direct evidence of carrier-mediated transfer
from blood to brain was presented by Crone (1965), who
measured the amount of 14C-glucose extracted from blood
during a single circulation through the head. The technique
consisted of giving a steady intracarotid infusion of test

substances, whose permeability characteristics were to be

19

studied and withdrawing simultaneous blood samples from the
carotid artery and the superior sagittal sinus. Analysis

of the arterial and venous blood allowed measurement of
solute loss to the brain. Dilution due to the blood
entering brain from other sources was estimated by including
in the injections T1824 (Evan's Blue dye) which was confined
to the vascular system. Glucose extraction decreased as
plasma glucose increased. The system was saturated at
approximately 4 mM D-glucose. Fructose extraction, by
comparison, was low, did not exhibit saturation, and was

not carried by the same mechanism as glucose. This glucose
carrier system has not shown to be responsive to iv insulin
administration (Crone, 1965; Gilboe et al., 1970). More
recently, LeFevre and Peters (1966), Bidder (1968), and
Buschiazzo et a1. (1970) have presented evidence for carrier-
mediated transport of hexose sugars from blood to brain in
the rodent, as indicated by self-saturation and competition.
Mayman at al. (1964) reported a brain glucose concentration

range of 0.6-1.2:mmoles/kg H 0, while blood glucose concen-

2
tration was 5-8 mM in anesthetized mice; therefore, the
glucose concentration gradient from blood to brain favors
passive movement and suggests that glucose is transported
across the blood-brain barrier via facilitated diffusion.
Fishman (1964) was the first to report the existence

of a saturable blood to CSF glucose transport. In his study,

blood glucose was raised and held steady by iv infusion, and

20

CSF glucose concentration determined at various times.

CSF glucose concentration rose to a maximum of 11.6-13.3

mM independent of whether the plasma concentration was 16.7,
20.8, or 25.0 mM. Fructose was analyzed in a similar manner;
its movement into CSF was slight and it did not show satura-
tion when plasma concentration was raised. When 2-deoxy-
glucose was simultaneously infused with glucose, glucose
penetration into CSF was competitively inhibited. Atkinson
and Weiss (1970) repeated Fishman's work by analyzing
glucose penetration over wider perturbations of plasma
glucose concentrations. They reported a saturation of the
glucose carrier at approximately 20 mM.

Evidence of carrier-mediated glucose tranSport from
blood to CSF by the choroid plexus has been provided by
Welch and co-workers (1970). The D-glucose concentration
was measured in newly formed choroidal CSF and in arterial
plasma samples collected simultaneously from adult rabbits.
Plasma concentrations were systematically lowered by insulin
or raised by iv glucose infusion. Glucose movement into CSF
increased linearly with plasma concentrations up to 14 mM
and CSF glucose maintained a level approximately 0.6 that
of the plasma concentration. At plasma glucose concentra-
tions between 14 and 20 mM choroidal CSF glucose concentra-
tion remained essentially constant (i.e., the system

demonstrated self-saturation). When plasma glucose

21

concentration exceeded 20 mM, CSF glucose again increased
linearly with plasma glucose concentration. These results
imply the presence of at least one carrier which is opera-
tive at plasma glucose concentrations less than 14 mM and
another inoperative at plasma glucose levels below 20 mM.
Since glucose concentration in blood (6.3 mM) exceeds that
in CSF (4.2 mM) (Bito and Davson, 1966), glucose probably
moves from blood to CSF via facilitated diffusion.

The possibility of glucose transfer from CSF into
adjacent neural tissue cannot be dismissed even though the
functional activity of the isolated perfused cat spinal cord
cannot be maintained by high glucose concentrations in the
perfusion fluid (Wolff and Tschirgi, 1965). Bradbury and
Davson (1964) compared glucose, creatinine, urea, and inulin
clearance from ventriculocisternal perfusion fluid in anes-
thetized rabbits, and reported that glucose was cleared at
the fastest rate. When solute concentrations were raised in
the perfusion fluid, creatinine and urea outflux coeffi-
cients (KO) were unaffected whereas that of glucose
decreased approximately 48%, indicating saturation of
glucose carrier-mediated efflux. Csaky and Rigor (1968)
perfused the CSF ventricular system in the dog with an
artificial CSF containing 14C-glucose and reported that
removing Na+ or adding phlorizin or digitoxin in the per-

14

fusion fluid decreased the rate of C-glucose disappearance

from CSF. Bronsted (1970a and b) perfused the cerebral

22
ventricles of the cat and found that l4C-glucose efflux
from CSF was competitively inhibited by either xylose or

7 - 5 x lO—SM) was

mannose and was reduced when ouabain (10—
included in the perfusion fluid. Such results imply a
carrier—mediated transport for glucose from CSF and indi-
cate that the transport system may be involved with active
Na+ transport.

The choroid plexus is capable of accumulating
glucose and galactose in vitro (Csaky and Rigor, 1968).
Glucose and galactose accumulation was inhibited by anoxia,
ouabain, phlorizin, or DNP and by the absence of Na+ in the
incubation media. The complete inhibition of glucose and
galactose accumulation by DNP or anoxia implies dependence
an aerobic utilization of energy. Although such results
show accumulation is possible, they do not indicate whether
this represents a tendency to remove glucose from CSF or to
increase its CSF concentration. In vascular perfusions of
the isolated horse choroid plexus, galactose tranSport from
the incubation medium into the perfusate against a concen-
tration gradient has been reported (Csaky and Rigor, 1968)
and it has been inferred that the choroid plexus is sim-
ilarly capable of actively tranSporting glucose from CSF
to the blood. These in vitro results indicate the choroid
plexus may be a site of active aerobic transport of glucose
from CSF in blood, and that this transport system may be

coupled with the movement of Na+.

23

2.6. Lactate and pyruvate movement among blood, brain,
and CSF

McGinty (1929) attempted to correlate respiratory
control in the anesthetized dog with brain lactate produc-
tion as measured by arterial-jugular venous plasma lactate
concentration differences. The author reported that when
cerebral blood flow (CBF) was unimpaired, lactic acid was
extracted from blood by the brain, but if CBF was impaired
via ligation of the carotid arteries or if oxidative metab-
olism was inhibited by sodium cyanide iv, lactic acid was
released from brain tissue. He concluded that excessive
lactic acid production by brain resulted in an outward
diffusion of lactic acid into venous blood. Jugular venous
blood in the dog is not exclusively of cerebral origin and
CBF measurements were not included in this study, making
invalid McGinty's inferences regarding the production or
uptake of lactic acid by the brain. Gurdjian et al. (1944)
reported that the concentrations of lactate in the brain and
in the blood vary independently and hypothesized that the
blood-brain interchange of lactate must be slow. This
hypothesis was subsequently supported by Klein and Olsen
(1947) who increased blood lactate concentrations in the
cat by iv injections and determined the lactate concentra-
tions in arterial blood and in brain tissue at various times
following the injections. They reported that although

arterial plasma concentrations were increased 10-30 times

24

the preinjection concentration, brain lactate concentrations
at 10 and 40 minutes following iv vascular injection did not
differ significantly from brain lactate concentrations in
cats not injected with lactate.

Alexander et al. (1962) reported no rise in CSF
lactic acid concentration 30 minutes following iv lactic
acid loading in dogs. Subsequent studies (Plum and Posner,
1967; Plum et al., 1968) have shown that brain lactic acid
concentration which is increased following hyperventilation
correlates closely with CSF lactic acid concentration but
not with sagittal sinus blood concentrations. These results
suggest that CSF lactic acid concentration is not affected
by blood lactate concentration but is probably a more
accurate indicator of cerebral metabolism.

When tissue oxygenation is decreased, a metabolic
shift occurs toward the reduced state of a metabolic system
and pyruvate is reduced to lactate. However, lactate
accumulation may result from nonhypoxic causes which
increase glycolytic activity thereby increasing the con-
centrations of both pyruvate and lactate (Huckabee, 1958a).
Huckabee claimed to identify nonhypoxic causes of increased
lactate accumulation by simultaneously determining the
pyruvate and lactate concentrations and calculating an
"excess lactate": the lactate which was not due to in-

creased pyruvate concentration. Since the ratio of lactate

25

pyruvate concentrations (L/P) reflect the oxidation-
reduction state of cytoplasmic NADH/NAD+ (Huckabee, 1958a)
measurement of the L:P ratio of a tissue, and possibly the
measurement of L:P ratio in plasma in equilibrium with the
tissue could yield information on the presence of tissue
hypoxia. When tissue 02 supply is acutely decreased by
hypoxia or anemia, the blood L:P ratio increases (Huckabee,
1958b; Cain, 1965) presumably due to tissue hypoxia. Sag-
ittal sinus plasma lactate concentrations do not accurately
reflect brain lactate concentrations (Plum et aZ., 1968),

but brain ECF concentrations at least with respect to Na+,

Br-, I-, C1“, and SCN— equilibrate with the CSF (Wallace and
Brody, 1937, 1939; Olsen and4Rudolph, 1955). Consequently,
elevated L:P ratios of brain homogenates and of CSF have
been used to indicate brain anaerobiosis (Kaasik et al.,
1970).

The relationship of the L:P ratios in brain
homogenates and CSF have been studied by Granholm and Siesjo
(1967, 1969) and Granholm et al. (1968). During normoxic
normocapnia the CSF L:P ratio is less than that of homoge-
nized brain tissue and is the result of an elevated CSF
pyruvate concentration (Granholm and Siesjo, 1969). This
suggests that pyruvate may diffuse from CSF into brain
tissue down its concentration gradient to be metabolized.

However, pyruvate concentration in arterial blood plasma is

also less than that in CSF (Granholm and Siesjo, 1967)

26

implying that pyruvate movement into the CSF may involve
more than simple diffusion.

Changes in pH influence the distribution of acids
and bases between the ECF and intracellular fluid (ICF)
(Milne et al., 1958). If only the nonionized form is
diffusible and if the compound exists in equal concentra-
tions in the ECF and the ICF, the relation between the total
concentrations of the acids and their anions will be deter-
mined by the hydrogen ion activity (pH) in these reSpective
fluids. Both lactate (pK==3.8) and pyruvate (pK==2.5) are
primarily ionized at body pH and should not rapidly traverse
cell membranes.

Siesjo et al. (1968) have proposed a lactate-
pyruvate transport system between brain ICF and the ECF
(assumed in equilibrium with CSF) to explain the lower L:P
,ratio in the CSF. They hypothesized that lactate was con-
verted to pyruvate simultaneously removing a proton (H+)
from the cell into the CSF. Such a system could contin-
uously remove lactate from the cell and establish a pyruvic
acid concentration gradient from CSF into the cell. Such a
scheme is compatible with the observed brain tissue and CSF
L:P ratios in the cat (Granholm and Siesjo, 1969), if it is
assumed the conversion of lactate to pyruvate occurs faster
than the diffusion of lactic acid from brain ICF and that

of the diffusion of pyruvic acid in the Opposite direction.

27

Prockop (1968) reported that lactic acid was cleared
from CSF slower than that of a glucose analog, 3-0-methy-
glucose,and that lactic acid clearance exceeded the clear-
ance of larger molecules such as creatinine and mannitol.
Prockop suggested that since he observed no decrease in
lactic acid clearance at different CSF lactate concentra-
tions, lactic acid clearance could be explained solely by
simple diffusion and bulk absorption. CSF lactate concen-
tration in the dog is normally 1-2 mM (Prockop, 1968).
Prockop's experiments were conducted at high CSF lactate
concentrations (i.e., 5-10nwn where lactate transport may
be saturated, and limit inferences regarding the mode of
lactic acid clearance from CSF.

Recently, Valenca et a1. (1971) investigated CSF
lactate clearance by intrathecal loading with either Na-
1actate or lactic acid. He reported that most of the
lactate injected into the CSF was recovered as an increase
in brain lactate concentration. Similar increases in brain
lactate concentration were found regardless of the form of
lactate injected. It can be inferred that at high intra-
thecal concentrations (e.g., 10 mM) lactate can enter brain
tissue but the mechanism of entrance cannot be deduced.
Further studies are clearly necessary to define adequately
the movements of lactate among the three extracellular fluid

compartments of the brain.

28

2.7. Hypoxia

 

Hypoxia causes an increased brain glycolytic rate
which results in increased lactate concentrations (Lowry
et al., 1964), increased cerebral blood flow (CBF) (Kety
and Schmidt, 1948; Shimojyo et aZ., 1968; Shapiro et al.,
1970; Fusjishma et al., 1971), and alterations in the
membranes of brain cells (Bakay and Lee, 1968). Membrane
alterations may be responsible for the brain edema (Rall
et aZ., 1962; VanHarreveld et al., 1965; Myers et al., 1969;
Bondareff et al., 1970) often seen during hypoxia. Changes
in CBF and/or membrane permeability could affect the amount
of a substrate (e.g., glucose) removed from the blood by the
brain, as well as the amount which enters the brain in-
directly via the CSF. Likewise, the exit of lactate from
brain either directly into the blood or into CSF could be
affected.

In the spontaneously breathing animal, hypoxia
causes hyerventilation resulting in a respiratory alkalosis
(Kety and Schmidt, 1948; Shimojyo et aZ., 1968; Shapiro et
al., 1970). Kety and Schmidt (1948) used the nitrous oxide
technique to measure CBF in man and reported that CBF
decreases during normoxic hypocapnia. This finding has been
supported by Raichle et a1. (1970), who reported decreased
CBF following sustained hyperventilation in both dogs and
man. In the dog, CBF decreased 60% within 30 minutes and

declined an additional 6% during the subsequent five hours

29

of passive hyperventilation. Restoration of normocapnia
returned the CBF to prehypocapnic control values. Similarly
in men PaCO2 was decreased to 15-20 mm Hg by voluntary
hyperventilation and within 30 minutes CBF decreased
approximately 40%, and subsequent normocapnia restored
CBF to prehypocapnic values.

Hypoxia increases CBF in man, which Kety and Schmidt
(1948) attributed to cerebral vasodilation caused by the
accumulation of metabolic products (i.e., C02, H+, lactic
acid). More recent studies in man report acute hypocapnic
(PaC02‘<30 mm Hg) hypoxia does not result in increased CBF
unless Pa02 is decreased below 40 mm Hg (Shimojyo et aZ.,
1968; Shapiro et al., 1970). Acute normocapnic hypoxia
(PaC02==37 mm Hg; Pa02==40 mm Hg) results in a 35% increase
in CBF (Shapiro et al., 1970), suggesting the vasoconstrictor
effects of hypocapnia apparently counter the vasodilator
response to hypoxia. Although total CBF increases during
normocapnic hypoxia local CBF reSponses are variable
(Fujishima et al., 1971). Fujishmia et a2. (1971) studied
hypoxic effects on local cortiCal CBF in the dog. Qualita-
tive changes in CBF were recorded by a heated thermistor,
flow probe placed either unilaterally or bilaterally in the
parietal cortex. The heated thermistor probe placed in
brain tissue was assumed to measure predominately changes

in the capillary blood flow (Fujishima et al., 1971)°

30

Local CBF decreased in 33% of the dogs subjected to 6% O2
in N2 and increased in the other 67% of the dogs. The CBF
of all dogs could be subsequently increased by normoxic
hypercapnia. A local vasoconstrictor response to hypoxia
may have occurred in areas of the cortex with a relatively
low metabolic rate which shunted blood to more rapidly
metabolizing tissue, and led to the variable CBF response
to hypoxia found by Fujishima and co-workers. Hawegawa
et a2. (1968) have shown precapillary shunts in the cerebral
cortex of the dog and direct A-V hunting in local areas
could also be responsible for the variable hypoxic response.
Brain metabolism has been studied in viva from the
A-V differences of metabolites (Gibbs et aZ., 1942; Solokoff,
1960). Brain metabolism can also be studied by decapitation,
which converts the brain to a closed system, and measuring
the rates of change in the concentrations of compounds
capable of yielding energy under these conditions (Lowry
at al., 1964; Gatfield at aZ., 1966; Folbergrova et aZ.,
1970). In the absence of oxygen (anoxic hypoxia), the major
energy reserves are ATP, PCr, and the high energy phosphate
bonds generated by the conversion of glucose and glycogen
to lactate via glycolysis. Anoxic hypoxia, produced by
ischemia results in a 4-7 fold increase in the rate of
glycolysis and causes a decrease in the concentrations of

brain glucose, glycogen, ATP, and PCr with a concomitant

31

increase in the concentration of lactate, ADP, inorganic
phosphate and Cr (Lowry et aZ., 1964).

Although these decapitation studies describe the
general patterns of changes in brain energy metabolism, they
cannot provide information concerning the metabolic events
following reoxygenation. Recently, Kaasik at al. (1970)
measured the lactate and pyruvate concentrations in blood,
CSF and brain homogenates as well as the concentrations
of brain ATP, ADP, AMP, and PCr during and after varying
periods of asphyxia induced by respiratory arrest. They
reported that during 3 minutes of asphyxia the brain tissue
L:P ratio increased and the ATP:ADP and ATP:AMP ratios
decreased. These changes were due to the rapid rise in
lactate, ADP, and AMP concentrations and the rapid fall
in ATP concentration. PCr concentration decreased but
tissue pyruvate concentration did not change during
asphyxia. In the restitution phase, ATP, ADP, AMP, and
the ATP:ADP, ATP:AMP and the L:P ratios were normalized
within two minutes. The normalization of the tissue L:P
ratio was due to a rapid decrease in tissue lactate con-
centrations and a marked increase in pyruvate concentra-
tions. These changes in the L:P ratio were interpreted
as the very fast reoxidation of cytoplasmic NADH. How-
ever, the rephosphorylation of Cr was not complete and

the tissue still showed a marked lactacidosis. In the CSF

32

there was a delayed increase in the lactate concentration
and a delayed fall toward normal values. The highest
lactate values were obtained after the resumption of normal
breathing. Lactate values as well as the L:P ratios, were
still elevated after 10 minutes. These changes were sug-
gestive of a time lag in the movement of lactate and
pyruvate from tissue into CSF, and the slow clearance of
lactate from CSF. This study suggests that elevated lactate
concentration in the CSF and the increased L:P ratio are
important in indicating a hypoxic change in brain tissue.
Anoxic hypoxia results in a decreased brain ECF
space (Rall at aZ., 1962; VanHarreveld at aZ., 1965). Rall
et al. (1962) perfused the brain ventricular system of a
dead dog and reported a brain inulin space less than that
found in brain perfusions of live dogs. VanHarreveld et a1.
(1965) reported a 6% brain ECF space measured by electron
microsc0py in rat cortical tissue frozen 8 minutes after
decapitation in contrast to a brain ECF space of 18-25%
in rapidly frozen tissue. Both groups of investigators
suggested the decreased brain ECF Space was due to brain
edema resulting from the absence of active aerobic metabolic
processes, which maintained the osmotic equilibrium between
brain ECF and ICF. Myers at al. (1969) produced hypoxia in
fetal monkeys by inducing prolonged (i.e., 4-6 hours)
maternal uterine contractions and measured hemoglobin

saturation and pH in the carotid arterial blood in the

33

fetus. When hemoglobin saturation was reduced to less than
50% and pH reduced to 6.22-7.20, the brain became edematous.
In a subsequent study, Bondareff et a2. (1970) subjected
fetal monkeys to a Similarly produced hypoxia and measured
the ECF Space of the cerebral cortex from electron micro-
graphs. They reported a 5% reduction in the ECF Space in
the hypoxic fetuses when compared to normal fetuses. In
addition, mitochondria from hypoxic fetuses were swollen
and characterized by the dissolution of the cristae. These
data suggest that brain edema concomitant with hypoxia may
have resulted in part from altered mitochondrial metabolism.
Eich and Wiemers (1950) were unable to demonstrate
the breakdown of the blood-brain barrier to trypan blue
during hypoxia. They concluded that the blood-brain barrier
was at the capillary endothelium, since this tissue would be
the last to suffer from hypoxia due to its close proximity
to blood. Supportive evidence has been provided by Bakay
and Lee (1968) and Goodale at al. (1970). Bakay and Lee
(1968) subjected cats to prolonged hypoxia (3—5 hours) and
studied the brain ultrastructure from electron micrographs.
They reported no change in the capillary basement membrane
which would indicate a change in capillary permeability.
Recently, Goodale et a1. (1970) perfused the isolated left
canine lung and filled the alveoli with albumin -l3lI in
Tyrode solution. They reported no change in the permeabil-

ity of the alveolocapillary membrane during severe hypoxia

34
(Pa02==12 mm Hg) as measured by the appearance rate of 1311
in the perfusion fluid.

Hypoxia appears to increase the permeability Of the
blood-CSF barrier to protein (Slobody at aZ., 1957; Lending
et aZ., 1961). Radio-iodinated serum albumin (RISA; iv)
appeared in the cisternal CSF Of both immature (Lending at
aZ., 1961) and adult dogs (Slobody et aZ., 1957; Lending at
aZ., 1961) more rapidly during hypoxia than during normoxia.
In puppies but not adult dogs, hypercapnic hypoxia resulted
in the faster appearance Of RISA in cisternal CSF samples
than when hypoxia alone was induced (Lending et aZ., 1961),
suggesting that age may affect the permeability Of the
blood-CSF barrier to protein during hypercapnic hypoxia.
Bakay and Lee (1968) reported increased capillary permea-
bility in the choroid plexus, which was indicated by
increase in pinocytotic vesicles and a widening Of the
capillary basement membrane. The increased RISA permea-
bility Of the blood-CSF barrier during hypoxia may be due
to increased membrane permeability in the choroid plexuses.

CSF is a secretion. Active transport mechanisms
Operate in its elaboration from blood, and metabolic energy
is required (see section 2.2). Bering (1959) reported that
Vf in anesthetized dogs was correlated with cerebral oxygen
consumption. The effects Of ouabain on Vf have been

reviewed (see section 2.2) and suggest active transport

Of Na+ and K+ contribute to CSF formation. In both the

35

newborn (Holloway at aZ., 1972) and adult dog (Michael at

aZ., 1972) V determined by the ventriculocisternal per-

f
fusion technique, was reduced during hypoxia. This suggests
that CSF formation is in part an aerobic metabolic process.
Recently, Michael et a1. (1971) reported increases
in both RISA and creatinine clearance from the perfused
brain ventricular system Of the rabbit during hypoxia and

suggested that hypoxia increased the permeability of the,

CSF-blood and/or CSF-brain barrier(s).

III . STATEMENT OF PROBLEM

The purpose of the present study is to determine
the flux rates and mode of transfer Of glucose, lactate,
and pyruvate from cerebrospinal fluid (CSF) Of the anes-
thetized dog. The simultaneous fluxes of these molecules
will be determined during both normoxia and hypoxia. Using
a ventriculocisternal perfusion technique CSF concentrations
of glucose, pyruvate, and lactate can be changed from their
normal values and net transependymal flux rate (Jx) can be
calculated. The Simultaneous calculation Of the transep-
endymal outflux coefficient (KO) for radioactively labelled
glucose and lactate will define their efflux rate from the
CSF. Both CSF formation (Vf) and bulk absorption (Va) rates
will be calculated under all experimental conditions, since

their values must be known to estimate JX and KC.

36

IV. METHODS AND MATERIALS

4.1. General Operative and cannulation procedures

Adult mongrel dogs (3-11 kg) Of either sex Obtained
from Michigan State University Center for Laboratory Animal
Resources (C.L.A.R.) were anesthetized with Dial and
urethane solution (Appendix B; 0.6 ml/kg; ip) and trache-
otomized. A diagram of the experimental equipment arrange-
ment is Shown in Figure l. The animal's ventilation was
controlled throughout the experiment by means Of a positive
pressure respirator (Model 607; Harvard Apparatus CO., Dover,
Mass.) connected to a plastic "T" tube in the trachea; lung
inflation was controlled by adjusting a clamp fitted on the
side arm which was Open to atmOSphere. The respiratory pump
was set to cycle 8—12 times/min with a stroke volume Of 160-
220 ml.

The femoral artery was cannulated (Figure l) to
collect anaerobic blood samples and to monitor arterial
pressure with a pressure transducer (Model P23D; Grass
Instruments, Quincy, Mass.) and a polygraph (Model 5P;

Grass Instruments). The arterial pressure transducer was
calibrated using a mercury manometer; the pressure response

was linear over a range Of 0 to 200 mm Hg. The cannula was

37

38

mm EU elm mo oMSmmoum SOHmSmuom o cﬁmuno Ou noumsmoo

.o

moz unmﬂon BOHMOSO one .nmoumwaom m on oouoocnoo noOSOmnoHu ounmmoum o
>n ooHOpHcoe mos onsmmoum HoHSOHHuno>ouucH .unmﬂon 30Hmuso oanoumsﬂoo
no npﬂ3 uoouooou mono o Op oouooncoo mo3 nOHnB .oHscnoO Honuoumﬂo onu
Eoum oouooaaoo Ono oHOHHuco> Honouoa unmeu onu Once oHoooc HOHSOHuuno>
on» nmsounu oodﬁsm moz Am.v oomv moasooaoe poop one mcﬂcﬂounoo mmo
HOHOHwHuum .ooowpsm Homuoo onu no oouooneo ASBOnm uocv mama nonoumcﬂ
no can ooomusm ammuno> m.ooo on» no pom mnﬂuoon o mo own unopueﬁnoucﬂ
.fn Aoom.o “v nonﬂonnﬂoa mos one .AoHBV ousuouomﬁou woon mo mcﬂnowﬂcoe
onu ooanoco Houocﬁndm Houoon onu ocowon AEO mlmv nonuomcﬂ HouoEOSHonu e
.mnoﬂuouunoocoo ouopooa ono .ono>suwm .omOOSHm oEmOHm Ono moom .Nooom
.omm maﬁnﬂauouoo now monEom UOOHn OﬁnOHooco mo coﬂuooaaoo on» no Hams
mo nmoumwaom o on oouooncoo HoOSOmcouu ousmmoum o oﬂ> oHSmmon UOOHn mo
mneuouﬂnoe oo3OHHo oascnoo Hoﬂuouuo Houoﬁom d .AOonmwn “NZ SH No mmlmv
ousuxHE NO 30H m on no AowaEHon «mo wamv HHoIEOOH Honuﬂo oomﬁdm nOHnS
Honouﬂmmou o On nouoonnoo mo3 ondu maouoonooup one .Ac3onm uonv oEoum
Oexouoououm M GA oousoom noon one an3 coenﬂmom ocoum m CH noooam moB

moo oouﬂuonumono one .ucoamﬂsvo oco Hoﬁﬁco HopcoEHHomxo one mo Eoumoﬂo

.H ousmﬂm

39

 

 

3.2.3.52. » \

0.38.30
.0.....(

. \\_....

:63: o
30:50 I
33.52:

  

 

LOUDVmCOLP

H ousmﬂm

Den—".0:

 

.0338.

t‘dl‘.‘ no.0

2‘00!
350:5

ouammoud

 

 

  

23:39

 

 
 

.9230..—

 

      

0.3002
. a. 33......)

   

dEad c073...

 

40

maintained patent by flushing with a heparin-NaCl solution
(1 I.U. sodium heparin/m1 0.9% NaCl).

The animal was placed in a prone position and the
head secured in a stereotaxic frame (Model 1504; David KOpf
Instrument CO., Tujunga, Cal.) by a snout clamp and ear-bars
inserted into the external auditory meatus. The stereotaxic
frame was tilted 45° from the horizontal plane to elevate
the animal's head and flex the neck, thereby maximizing the
area between the atlas and the base Of the Skull° The skull
and atlanto-occipital (A-O) membrane were exposed by means
of a midline Skin incision extending from the orbital
sockets caudally to the 4th cervical vertebra, and muscle
retraction with cauterization. A 6 mm diameter hole was
trephined in the right parietal bone (5.0 mm lateral and
5.0 mm caudal from the intercept of the central saggital
and coronal sutures) to eXpose the dura over the right

cerebral hemisphere.

4.2. Brain ventricular and cisternal puncture

The ventricular probe needle (20 9a.; 2" length;
Short bevel) was placed in a micromanipulator electrode
carrier (Model 1460; David KOpf Instrument CO.) and
stereotaxically placed with the tip on the dura at a point
5.0 mm lateral and 5.0 mm caudal to the intercept Of the
central saggital and coronal sutures and perpendicular to
the plane Of the stereotaxic frame. Artificial cerebro-

Spinal fluid, CSF, (Appendix A) was pumped from a 50 ml

41

disposable polyethylene syringe by an infusion pump

(Model 975; Harvard Apparatus CO.) through a length of

P.E. 50 tubing connected to a male needle adaptor on the
ventricular probe needle (Figures 1 and 2). A second outlet
on the adaptor was connected via P.E. 50 tubing to a low
pressure (0-5 mm Hg) transducer (Model P23BC; Grass Instru-
ments CO.), which enabled continuous monitoring Of perfusion
pressure. The transducer was calibrated using a water
reservoir; output was linear over a range Of 0 to 40 cm H20.
The level Of the external auditory meatus (i.e., stereotaxic
ear bars; Figure 2) was defined as a point at which zero cm
H20 perfusion pressure would be measured. The perfusion
pressure at ear-bar height was subtracted from the recorded
experimental perfusion pressure to Obtain the intraventric-
ular pressure.

The perfusion pressure with the ventricular probe
needle tip touching the dura was noted and the needle was
lowered to puncture the dura. AS the needle was lowered
into brain tissue, perfusion pressure rose, and when the
right lateral ventricle was punctured perfusion pressure

dropped immediately to 10-20 cm H O, and cardiac and res-

2
piratory pulsations were evident in the pressure recording.
The depth Of the ventricular penetration ranged 0.6-1.1 cm
below the dura.

The cisternal needle (20 9a.; 2" tube; short bevel),

held bevel up in a micromanipulator (Model MM-3; Eric

42

.AHV mnense .m.m mo mnooE mn Amv HoensOOImoeo Oeueooao

Ioeonm o oe ooeoonnoo mo3 oaooon soeeeso one .on Hoeoanmenoﬁonoee
me> onmoE onnoemeo one ne oonoeeemom Amv oaooon onmeSO one ono
ooeomneou onm oonOeeoom oeoB moeomse noon one .oHSmmon noemSenom
mnHquHnOE How AnSOnm eonv nmoumeaom o Oe ooeoonnoo Amy Hoosomnoue
ousmwoea o ono AnBOnm eonv mega o>euo omneumm o Oe ooeoonnoo mos

.on nooaon Oeeeooeo oexoeooeoem ouoonoem o en oHoeHeno> Homoeoa enmen
one ne oonOHeemom eaaooexoeooeoem .on oaooon onoum Hoasoeeeno> one
.Amv neon Moo ono Ansonm eonv mEoHo econm m an oousoom ono Adv oEoHe
Oexoeoouoem o ne ooooam mo3 moo one no ooon one .moo ooueeonemono

one ne nOHmSmuom HmnuoemeOOHsoeHeno> Hoe ooms enoEmHSwo mo nmoeooeonm

.m oenmem

43

 

Figure 2

44

Sobotka CO., Inc., Farmingdale, N.Y.), was directed
rostrally and positioned with the point on the A-O membrane
at the midline, midway between the atlas and the base Of the
skull and in a plane parallel to the stereotaxic frame. A
50 cm length Of P.E. 90 tubing was connected to the needle
as the outflow cannula (Figures 1 and 2). The cisternal
needle was lowered to puncture the A-0 and dural membranes
and movement of CSF into the outflow cannula and a further
drop in perfusion pressure indicated a connection with the
ventricular needle. The depth Of the cisternal penetration
ranged 0.3-0.6 cm below the A-O membrane. The outflow
cannula was positioned in a photocell drop-counter (Model
PTTl; Grass Instruments CO.) and the height Of the drOp
counter adjusted to Obtain a net intraventricular pressure

Of 2-4 cm H20.

4.3. Experimental perfusion with test molecules

After determining that the perfusion circuit had a
steady outflow rate (V0), and a low, stable intraventricular
pressure (4.2), a disposable syringe containing approximately
60 m1 Of artificial CSF and the dissolved test molecules
(see section 4.9) was placed in the perfusion circuit. This
interchange was accomplished by interchanging the inflow
tubing from the respective syringes at the male needle
adaptor. The introduction Of air into the perfusion circuit
was prevented by maintaining a hydrostatic head (approxi-

mately 40 cm H20) from the calibration reservoir through the

45

pressure transducer. After the inflow tubing interchange,
the hydrostatic head was removed and if necessary, intra—

ventricular pressure again set at 2-4 cm H O by adjusting

2
the outflow height.

4.4. Experimental criteria

 

Data from experiments in which intraventricular
pressure did not remain stable (i.e., varied by more than
2 cm H20) or where V0 was highly variable within the
experimental periods (4.9 ) were not used in this study.
Blood in any CSF outflow sample precluded using that sample.
Following the 240 minutes Of the experimental perfusion, an
artificial CSF (Appendix A) containing methylene blue was
perfused for 60 minutes. The animal was killed by injecting
10 ml of a saturated KCl solution iv, after which the skull
and brain were removed. Appearance Of stain (methylene
blue) was used to confirm needle placement; staining outside
the ventriculocisternal system precluded using data from

that Specific experiment.

4.5. Sample collection and storage

Following flushing Of cannula dead Space four m1
arterial blood samples were drawn in a heparinized syringe,
and immediately centrifuged (1000 x g; 0°C; 10 minutes) in
a refrigerated centrifuge (Model PR-2; International Equip-

ment CO., Needham, Mass.). Plasma was decanted into vials,

'46

capped, frozen, and stored at -20°C. CSF samples were
collected into 20 gm vials, capped, weighed, frozen, and
stored not longer than 40 hours at -20°C.

Anaerobic arterial blood samples (0.5-1.0 ml) were
drawn in a heparinized syringe immediately after each 4 ml
sample. The syringe was capped with a Hg-filled female
syringe adaptor and kept in ice water (0°C) for a maximum

Of one hour prior to PaOZ' PaCO and pH measurements (4.7).

2'

4.6. Determination Of normal CSF and plasma metabolic
concentrations

 

 

CSF and plasma samples were Obtained from 12 dogs by
puncturing the dura over the cisterna magna (4.2) and allow—
ing 1-2 ml of CSF to flow into a vial, while Simultaneously
Obtaining a femoral arterial blood sample (4.1; 4.5). These
samples were used to determine the normal concentrations of
glucose, pyruvate, and lactate in CSF and plasma (4.8). The
volume of CSF removed was replaced by 2-4 ml Of artificial

CSF (Appendix A) prior to the ventricular puncture (4.2).

4.7. Measurement Of blood gas tensions and pH

 

PaCOZ’ PaO2 and arterial pH (4.5) were measured

using thermostated (38°C) Types E5036, E5046, and E5021
Radiometer electrodes (The London CO., Westlake, Ohio),
respectively, in conjunction with a Radiometer pH meter
2, and PCO2 scales (Model PHM27; The London

CO.). PaCO2 and Pa02 electrodes were calibrated using gas

containing pH, PO

47

mixtures verified by a Haldane-Bailey gas analyzer (Arthur
H. Thomas, Philadelphia, Pa.). The O2 mixtures were also
used to calibrate the O2 analyzer (Model C2; Beckman
Instruments, Inc., Fullerton, Cal.) which enabled the rapid
determination of the 02-N2 mixtures used tO induce hypoxia.
The pH electrode was standardized with commercial buffers
(pH==6.84; pH==7.384; Scientific Products Inc., Allen Park,
Mich.) Pa02' PaCOZ' and pH measurements were corrected to
rectal temperature with a Radiometer Blood-Gas Calculator
(The London CO.).
4.8. Measurement of inflow and outflow rates and concen-
trations

Inflow and outflow rates (Vi and V0) were determined
gravimetrically (i(Lﬁl mg; Mettler Instrument Corp.,
Highstown, N.J.) using tared vials referenced to sample time
tt0.l numb, Precision timer, Arthur H. Thomas). Vi was
determined by collecting fluid from the perfusion syringe
at the beginning and at the conclusion of the experiment.
For this study 1 mg Of fluid was assumed to occupy a volume
Of 1 pl. Inflow and outflow concentrations (ci and cc) of
D-glucose (Appendix D), L-lactate (Appendix F), and inulin
(Appendix G) were determined Spectrophotometrically.
Pyruvate concentrations were determined fluorometrically

14

(Appendix E). 3H-glucose and C-lactate were isolated

(Appendices H and I) and their radioactivity counted using

48

standard liquid scintillation procedures (Appendix J). In
some experiments 3H-mannitol was used in lieu of 3H-glucose,
and the radioactivity counted using liquid scintillation

procedures.

4.9. Experimental Design

 

Each experiment lasted 4 hours (Figure 3). Supple-
mental doseS of Dial and urethane solution (0.4 ml/kg; ip)
were administered at zero time and after 120 minutes Of
perfusion. Perfusion for 60 minutes prior to initial
sampling enabled the outflow concentrations (co) Of the
test molecules (see below) to attain steady levels. Three
19-21 minute steady-state CSF outflow samples were Obtained
between 60-120 minutes of the experimental perfusion (P1)'
Arterial blood samples were drawn midway in outflow sample
collection (i.e., at 70, 90, and 110 minutes). Another
60 minute equilibration period was begun at 120 minutes and
was followed by three 19-21 minute CSF outflow sample col-
lection periods with their corresponding blood samples (P2).
In one series of experiments (4.9.1) the dogs were main-
tained normoxic (Pa023185 mm Hg) during the 240 minutes Of
ventricular perfusion; in the other series (4.9.2) hypoxia
(PaOzjiSO mm Hg) was produced at 120 minutes by ventilating

the animal with 5-8% 02 in N2 and continued through P2.

49

.m N N

z ne 0 mmlm neHB HoEeno one mneeoaeeno> en m nmsoene oosneenoo
ono moesneE omH eo ooosooem mo3 Amm ES ommHmoomv oexommn A.m.m.wv
moeeom eoneo one nH .nOemSMHom eoHSOeeeno> mo moenneE ovm one mneuso
Amm SE mm.nmommv oexOEeon ooneoeneoﬁ oeoz mmoo one A.H.m.vv menoEHHomxo
mo moeeom ono nH .oeoeooa ono .oeo>9emm .omOOSHm mo mnOeemeenoonoo
NOom .m

IHOO 3Oameso emu nooo nmSOHne xozoen ooeooHHOO oeoB Am3oueo ooeonESnv

oEmon ono om .mm mo enoEoeSwooE one mnHHnono .ooeeom noeeooa
monEom oOOHn Hoeuoee< .Am.vv monEom mmo one Some ooneEHoeoo oeo3 moeon
BOHmeSO nOemSmeom ono mnOeeoeenoonoo SOHueSO emu .Ammv woesnee ovmtoma
noozeon ono Aamv moeanE omanom nooBeon ooeooaaoo oeos moemﬁom 30Heeso
emu oesneE om ooene .nOemSmuom mo woesneE oma ono o em ooeoemeneﬁoo

oeoB Amov oeeonemono mo momoo HoenoﬁoemmSm .moesneﬁ own How oodneenoo
ono nsmon mo3 nOemseeom Hoasoeeeno> .o oEee en .ooEeOeeom monsooooum
Movemesw ono Aoov ooueeonemono moz moo one .mnoeeooaaoo onEom ono mnOee

loaomenos Hoenoseeomxo mo Amoesnee “ommeomnov oEee one mne3onm Eoemoeo

.m oesmem

50

m oesmem

 

 

 

 

 

 

 

 

 

 

we:
3... on. 2. oo o
W a d i‘ - u u i; — an —
- - - ootod
N
n . 1 _ _. _
, - . easomlmu
_o_n_w_ rn_~.__a
O Q Q o D o_dEom oooE
o n w n a
o 0 8.0.1324.

oo

51

4.9.1 Long duration brain ventricular perfusions in
normoxic dOQS

 

 

All animals were ventilated with room air
(normoxia; Pa02==85-110 mm Hg) for the entire 240 minutes
(see Figure 3) Of brain ventricular perfusion. These
eXperimentS were conducted: (a) to examine the effect
Of 4 hours Of brain ventricular perfusion in normoxic
animals on the bulk absorption rate (Va), CSF formation
rate (Vf), the transependymal flux rates (KO) Of D—glucose
and L-lactate, and the net flux rates (JX) of D-glucose,
pyruvate, and lactate, and (b) to examine the effects Of
blood and CSF concentrations Of D-glucose, pyruvate, and
lactate on their respective movements (i.e., flux rates)

across membranes separating CSF from blood and/or brain.

4.9.1.1. Elevated blood and low CSF inflow concentrations

 

Glucose, pyruvate, and lactate flux into CSF was
induced in 6 dogs by decreasing CSF inflow concentrations
while blood concentrations were elevated for these metab-
olites. The perfusion fluid contained 1 mg/ml inulin, and
trace quantities of 3H-glucose (0.5 uc/ml; D-glucose-2-3H;
New England Nuclear, Boston, Mass.) and 14C-lactate (0.05
uc/ml; Na-L-lactate-3-14C; New England Nuclear). Inflow
fluid osmolality was 307:t2 mOsmol/l. Blood glucose,
pyruvate, and lactate concentrations were increased by
injecting a priming dose (1 ml/kg) Of 2.7 M D-glucose,

1.2 M Na-pyruvate, and 2.7 M Na-lactate into the brachial

52

vein, followed by a constant infusion (0.01 ml/min per kg)
Of a solution containing 8.9 M D-glucose, 25 mM Na-pyruvate,
and 4.4 M Na-lactate pumped by a syringe drive pump (Model

975; Harvard Apparatus CO.).

4.9.1.2. Normal CSF inflow concentrations

In 4 dogs the perfusion fluid contained normal CSF
concentrations Of D-glucose, pyruvate, and lactate; no
attempt was made to alter blood concentrations for these
metabolites. The perfusion fluid contained 1 mg/ml inulin,
4.3-4.7 mM D-glucose, 0.09-0.19 mM Na-pyruvate, 1.2-1.7 mM
Na—lactate, and trace quantities Of 3H-glucose (0.5 uc/ml;

l4

D-glucose—2-3H; New England Nuclear) and C—lactate

(0.05 uc/ml; Na-L-lactate-3-14C; New England Nuclear).
The NaCl concentration in the artificial CSF (Appendix A)
was decreased from 141.2 mEq/l to 133.0 mEq/l to adjust

perfusion fluid osmolality to 307 t2 mOsmol/l.

4.9.1.3. Elevated CSF inflow concentrations

 

In 5 dogs D-glucose, pyruvate, and lactate concen-
trations in the perfusion fluid were increased to 2—6 times
that of normal CSF concentrations. NO attempt was made to
alter blood concentrations of glucose, pyruvate, or lactate.
The perfusion fluid contained 1 mg/ml inulin, 15.7-17.9 mM
D-glucose, 0.28-0.33 mM Na-pyruvate, 4.3-5.9 mM Na-lactate,
and trace quantities Of 3H-glucose (0.5 uc/ml; D-glucose-

2-3H; New England Nuclear) and l4C-lactate (0.05 uc/ml;

53

Na-L-lactate—3-14C; New England Nuclear). NaCl concentration
in artificial CSF (Appendix A) was decreased to 114.4 mEq/l

to obtain a perfusion fluid osmolality of 309:13 mOsmol/l.

4.9.2. Brain ventricular perfusion during normoxia and

hypoxia

These experiments were designed to examine the
effects of hypoxia on: bulk absorption rate (Va), CSF
formation rate (Vf), the transependymal outflux rates (KO)
Of glucose and lactate, and the net transependymal flux
rates (Jx) Of glucose, pyruvate, and lactate at low and high
CSF concentrations for these metabolites. Dogs were venti-
lated with room air (normoxia; Pa02==85-110 mm Hg) during
P1 (see Figure 3). After 120 minutes of perfusion hypoxia
(PaozjiSO mm Hg) was induced by connecting the respirator
to a 150 liter Douglas bag (Warren E. Collins, Inc., Boston,
Mass.) filled from a high pressure gas cylinder which con-
tained 5-8% 02 in N2. Hypoxia was maintained through P2.

Since both glucose and lactate KO'S reported for
experiments described in 4.9.1 appeared to exhibit satura-
tion kinetics (indicating the presence Of carrier-mediated
transport from CSF), 3H-mannitol (considered to be a pas-
sively diffusing molecule; Prockop, 1969; Bronsted, 1970a;
Bierer, 1972), was included in the perfusion fluid in lieu
Of 3H-glucose in some experiments. Comparison of mannitol

KO'S at low and high CSF mannitol concentrations with those

54

Of lactate would provide evidence regarding a saturable

lactate carrier (Stein, 1967).

4.9.2.1. Low CSF inflow concentrations

 

In 3 dogs glucose, pyruvate, and lactate flux into
CSF was induced by decreasing CSF concentrations for these
metabolites. NO attempt was made to alter blood concentra-
tions Of glucose, pyruvate, and lactate. The perfusion
fluid contained 1 mg/ml inulin, and trace quantities Of
3H-glucose (0.5 pc/ml; D-glucose-2-3H; New England Nuclear)
and 14C-lactate (0.05 uc/ml; Na-L-lactate-3-14C; New England

Nuclear). Perfusate osmolality was 306::2 mOsmols/l.

4.9.2.2. Elevated CSF inflow concentrations

 

In 3 dogs glucose, pyruvate, and lactate concen-
trations were increased 2—5 times that of their normal CSF
concentrations; no attempt was made to alter blood metab-
olite concentrations. The perfusion fluid contained 1 mg/ml
inulin, 11.1-11.2 mM D-glucose, 0.22-0.24 mM Na-pyruvate,
3.7-4.4 mM Na-lactate, and trace quantities Of 3H-glucose

(0.5 uc/ml; D-glucose-2-3H; New England Nuclear) and 14C-

14C; New England

lactate (0.05 uc/ml; Na-L-lactate-3-
Nuclear). NaCl concentration in artificial CSF (Appendix A)
was decreased tO 114.4 mEq/l to Obtain a perfusion fluid

osmolality Of 304::2 mOsmOl/l.

55

4.9.2.3. Low CSF inflow concentrations with mannitol

 

In 3 dogs glucose, pyruvate, and lactate flux was
induced by decreasing CSF concentrations for these metab-
olites; no attempt was made to alter blood concentrations
Of these metabolites. Trace quantities Of 3H-mannitol were
included in the perfusion fluid in lieu Of 3H-glucose to
allow the calculation Of mannitol KO at low CSF inflow
concentrations. The perfusion fluid contained 1 mg/ml
inulin, and trace quantities Of 3H-mannitol (0.5 uc/ml;
D-mannitol-l-3H; New England Nuclear) and 14C-lactate
(0.05 uc/ml; Na-L-lactate-3-14C; New England Nuclear).

Perfusion fluid osmolality was 305:12 mOsmOl/l.

4.9.2.4. Elevated CSF inflow concentrations with mannitol

 

In 3 dogs pyruvate and lactate concentrations in the
CSF perfusion fluid were 2-6 times their normal CSF concen-
trations and CSF glucose concentration was 0 mM to induce
glucose influx. D-mannitol concentration in the perfusion
fluid was increased in an attempt to demonstrate saturation
Of mannitol KO. NO attempt was made to alter blood glucose,
pyruvate, or lactate concentrations. The perfusion fluid
contained 1 mg/ml inulin, 16.0-16.5 mM D-mannitol, 0.26—
0.28 mM Na-pyruvate, 4.0-5.8 mM Na-lactate, and trace
quantities of 3H mannitol (0.5 uc/ml; D-mannitol-1-3H;

14

New England Nuclear) and C-lactate (0.05 uc/ml; Na-L-

lactate-3-14C; New England Nuclear). NaCl concentration

56

in artificial CSF (Appendix A) was decreased to 114.4 mEq/l

to Obtain a perfusion fluid osmolality Of 309::3 mOsmOl/l.

4.10 Principles and calculations

 

Principles and equations used in the calculation Of
bulk absorption rate, CSF formation rate, net flux rate, and
the transependymal outflux rate have been previously
described (Pappenheimer at aZ., 1961; Heisey at aZ., 1962;

Pappenheimer at aZ., 1965) and are presented below.

4.10.1. Definition Of symbols
V = flow rate (pl/min).
i,O,p = subscripts referring to inflow, outflow, and
plasma respectively.
f,a = subscripts referring to formation and bulk

absorption Of fluid.

c = concentration (quantity/pl).

Ei= estimated mean concentration in the ventricular
system.

C. "C
J.

O .
— (quant1ty/u1)
1n ci7cO

when ci = 0, the function above is undefined
and E was estimated by: E¥=cO-+0.37 (ci-co)

(Pappenheimer at aZ., 1961).

n = steady state movement Of any substance, x, from

or into CSF perfusion.

ViciwvocO (quantity/min).

57

C = clearance of x.
= nx/c (pl/min).
c may be E’Or cO depending on whether the
substance is cleared from ventricular fluid
or from fluid reabsorbed distal to the fourth

ventricle.

4.10.2. Fluid balance

In the perfused brain ventricular system, fluid is
added to the system by the perfusion syringe (inflow fluid;
Vi) and by the amount Of CSF formed by the animal (Vf).
Fluid iS lost from the perfusion system via the outflow
cannula (outflow fluid; V0), and by the amount Of fluid
absorbed in bulk through large exit routes such as the
arachnoid villi (va). In the steady-state the total fluid

inflow equals the total fluid outflow. In summary:

Vi + Vf = VO + Va (Eq. 1)

4.10.3. Bulk absorption rate, Va

Heisey et a1. (1962) reported that diffusion Of the
polysaccharide inulin (In) from the ventricular system into
brain tissue was insignificant; the inulin lost from CSF
could be accounted for by bulk absorption distal to the
fourth ventricle. Inulin is removed from the subarachnoid

Spaces by bulk absorption at a rate which varies linearly

58

and directly with CSF hydrostatic perfusion pressure
(Heisey et aZ., 1962, Bering and Sato, 1963). Bulk

absorption can then be expressed as:

. Vici--VocO
V = = clearance of inulin, C (Eq. 2)
a cO In

 

where c's are inulin concentrations.

4.10.4. CSF formation rate, Vf

 

CSF formation rate, Vf, can be estimated by
substituting equation 2 in equation 1 to obtain:

V = V - V. + C
l

f 0 (Eq. 3)

In

4.10.5 Derivation of net flux, Jx

Ions or metabolites can enter the ventricular system
in freshly formed CSF or by exchange (active or passive)
with blood or brain through tissues lining the ventricular
system (transependymal exchange). Similarly, ions or metab-
olites can leave the ventricular system by absorption in
bulk from the subarachnoid spaces or by transependymal
exchange. Under the experimental conditions of ventri—
culocisternal perfusion, substances can also enter or leave
the ventricles with the perfusion inflow and outflow fluids.
The net transependymal flux rate (Jx' umoles/min) of any

molecule, x, during the steady-state is

Jx = Vici--VOcO-I-Vfcf--Vac0 (Eq° 4)

59

Substituting equations 2 and 4 in equation 5 and

rearranging terms:

Jx = Vi (ci-cf)-(VO-+CIn) (cO-cf) (Eq. 5)

All quantities except cf are measureable; c must be

f
estimated.

4.10.6. Derivation of transependymal outflux coefficient, KO

 

In the case where a test molecule not present in
either plasma or CSF is added to the perfusion fluid, cp and

cf are zero and equation 6 reduces to

c (Eq. 6)

J = Vici- (Vo+CIn) o

X

Jx in equation 6 is an outflux, Jxo. Jxo is determined by

the characteristics of the transependymal membranes limiting
the movement of the test molecule from the ventricular
system (outflux coefficient, K0) and the ventricular con-

centration of the test molecule, El

Jx0 = Koc (Eq. 7)

Substituting equation 7 into 6 and rearranging

_ l l o In 0

For test substances normally present in plasma or brain, Ko’
may be determined using equation 8 if an isotope of the test

substance is added to the perfusate.

60

4.11. Calculated parameters

Bulk absorption (va) was estimated by inulin
clearance (CIn; Equation 2); CSF formation rate (0f) was
calculated using Equation 3. The net flux (Jx) for glucose
(9), pyruvate (py) and lactate (l) were calculated using
equation 5 and the chemical concentrations for these metab-
olites (4.8). The calculation of Jé; the glucose concen-
tration in newly formed CSF (cf) was assumed to be 0.6 that
of the plasma concentration (cp) (Welch et aZ., 1970); for
the calculation of Jg, cf was assumed = 0. For the calcu-

lation of both J1 and pr, cf was assumed to be zero; for

1
outflux coefficient (KO) for glucose, lactate, and mannitol

J and pr' cf was assumed equal to cp. The transependymal
was calculated using equation 8 and radioactively labelled

molecules.

4.12. Statistical methods

Individual means (E) were calculated for each
parameter (4.11) from 3 values obtained during each period
of an experiment (4.9). In some instances the grand means
(i) and the standard error of the grand mean (S.E.) was
calculated (Appendix K) for all individual i during a given
perfusion period at a given inflow concentration (4.9).

Data was analyzed for significant differences using
a split-plot design analysis of variance (AOV, see Appendix K)

and critical values for Student's t distribution (P'<0.0S).

V. RESULTS

5.1. Arterial pH, P02, PC02 and rectal temperature

 

Prefatory studies indicated that body temperature,
PaCO2 and anesthesia affect CSF production and molecular
flux from CSF, possibly related to effects on cerebral blood
flow and/or cerebral metabolism. In the present study
changes in rectal temperature (Tre) were minimized by inter-
mittent external heating and those in PaCO2 by artificially
ventilating the dogs at a constant rate and depth (see sec-
tion 4.1). Anesthetic was administered at the same time
during experiments to minimize its variability as an operant
factor.

Ventriculocisternal perfusions were performed on
27 dogs. Fifteen were ventilated with room-air (normoxia);
12 were normoxic only for the initial 2 hours and were then
ventilated with 5-8% 02 in NZ-(hypoxia) for the remaining
2' PaOZ'

and Tre for each dog were calculated from 3 measurements

2 hours (see section 4.9). Mean values of pH PaCO

a!
between 60-120 minutes (P1) and between 180-240 minutes (P2)
of the perfusion (Figure 3; section 4.12). Data reporting

mean pHa, PaCOZ' Paoz, and Tre for all normoxic and hypoxic

dogs are summarized in Table l. Tre (approximately 38°C)

61

62

and PaCO2 (approximately 38 mm Hg) were considered to be
normal in both normoxic and hypoxic dogs. There was no
difference (p >0.05) in PaO2 among normoxic perfusion
periods (A-Pl and P2; B-Pl) and PaOZ was decreased (p‘<0.05)
during hypoxia (B-Pz) producing normocapnic hypoxia.
Arterial pH was not different (p:>0.05) among normoxic
perfusion periods. During hypoxia, pHa decreased (p<:0.05),
suggesting the release of metabolic acids from hypoxic

tissue into the blood.

5.2. Cerebrospinal fluid bulk absorption and formation rates
Data for CSF bulk absorption rate (Ha) and formation

rate (0f) in normoxic and hypoxic dogs during brain ventric-

ular perfusions are summarized in Table 2. 6a during the

initial 2 hours of perfusion (Pl; Conditions A and B) was

the same. 6a increased similarly (p:<0.05) during P2 in

both normoxic and hypoxic dogs. 0f decreased (p‘<0.05)

during P2 in both normoxic (Condition A) and hypoxic

(Condition B) dogs.

5.3. Normal plasma and CSF concentrations of glucose,
pyruvate, and lactate

 

 

Table 3 contains data for D-glucose, pyruvate, and
L-lactate concentrations in arterial plasma and cisternal
CSF samples obtained simultaneously from normoxic dogs prior
to ventriculocisternal perfusion. Plasma glucose concentra-

tion was greater (p<:0.05) than CSF glucose concentration.

63

The plasma and CSF concentrations of either lactate or

pyruvate were not different (p >0.05).

5.4. Glucose flux from CSF

 

The effect of different D-glucose concentrations
in the perfusion inflow fluid (Ci) and plasma (cp) on out-
flow concentrations (co) is shown by data summarized in
Table 4. When ci was less than cp (i.e., ci==0.0 and 4.5 mM),
co exceeded ci, indicating that glucose moved into the per-
fusion fluid. At elevated ci (i.e., 16.9 and 11.1 mM), ci
exceeded co showing that glucose moved out of the perfusion
fluid. In one group of dogs (Condition A; ci==0.0 mM) both
00 and cp were less (p‘<0.05) during P2 compared with P1'
suggesting that alterations in cp affected Co°

In 5 dogs (Condition A), elevation of glucose ci
resulted in a glucose extraction from the perfusion fluid
of 35% (percent extraction==100 (ci-cO)/ci) during both
P1 and P2. In 3 dogs with elevated ci (Condition B),
glucose extraction was increased during P2 to twice the Pl
value. These data suggest an increase in transependymal
outflux during hypoxia possibly due to increased glycolysis
in brain tissue.

Glucose Ko's when glucose ci's were low (cl), normal
(c2) and elevated (c3, c4) are summarized in Table 5. K0

was independent of time at all ci's under normoxic condi-

tions (A-Pl and P2). KO at G1 was greater (p‘<0.05) than

64

at either c2, c3, or c4, and Ko's at c2, c3, and c4 were
not different (p=:0.05) from each other during normoxia
(A-Pl and P2;
low and elevated ci's, suggesting that hypoxia may affect

B-Pl). Hypoxia-reduced (p<<0.05) KO at both

metabolic processes responsible for glucose movement from
CSF.

Net glucose transependymal flux rates, summarized
in Table 5, were calculated assuming that cf==0.0 mM (Jg)

and that c =0.6 c (Jé; Welch et aZ., 1970). The differ-

f p

ence between Jg and J; at comparable ci's represents glucose
influx which could be attributed to that entering in CSF
produced by the animal and is a function of 6f (which was
measured) and cf (which could not be measured). Since J9
and J; represent glucose movement only by diffusion or
carrier-mediated transport, a glucose influx will be larger
when cf is assumed equal to 0.0 mM (Jg) than when cf==0.6 cp.
Both J9 and J; were highly variable among dogs, but were
constant during normoxia in individual animals. When
perfusion inflow concentrations were low (C1) or normal (c2)
glucose net flux rates were negative indicating a net flux
into CSF. At elevated ci (c3, c4), both J9 and J; were
positive, indicating a net glucose outflux. These data
corroborate the changes in glucose concentration. There

was no change (p:>0.05) in J; with time during normoxia
(i.e., A-P1

cannot be explained. In condition B—P

and P2). The change (p‘<0.05) in Jg at c1

at c and c4 glucose

1 l

65

flux (J9 and Jé) was less than that in A-P (c1 and c3,
respectively) and probably reflect different ci's and cp's
in the two conditions (see Table 4). During hypoxia (B—Pz)
there was an increase (p‘<0.05) in net glucose influx at C1’
Glucose (J9 and Jé) at c4 did not increase (p:>0.05) during
hypoxia even though glucose extraction from the perfusion
fluid doubled (Table 4). Changes in Jg were similar to
those of J; except that flux rates did not change (p >0.05)

during hypoxia at CI.

5.5. Lactate flux from CSF

 

Data reporting the effect of different L-lactate
ci's and cp's on co are summarized in Table 6. At low

(0.0 mM) and normal (1.5 mM) c.'s, c

always exceeded c.
l o 1

indicating lactate movement into CSF from blood or brain.
At low and normal ci's, cO increased (p<<0.05) with time
although cp did not change (p >0.05), suggesting that the
lactate came from brain. During hypoxia, at low ci's, cO
also increased (p‘<0.05) but was accompanied by an increase
(p < 0.05) in cp.

At elevated ci's (5.2 and 4.5 mM), cO was the same

as ci during normoxia (A-P and P2) and always exceeded cp.

1
These data indicate that lactate movement from CSF is slow.
During hypoxia (B-Pz), both c0 and cp increased (p<:0.05),

but the increase in co could result from lactate diffusion

from plasma.

66

Lactate Ko (l4C-lactate) at low (cl), normal (c2),

and elevated (c3, c4) ci's are summarized in Table 7.

During normoxia KO values did not change (p:>0.05) with

time (A-P1 and P2). Ko at G1 was greater (p<<0.05) than at
either c2 or c3 and the Ko's at c2 and c3 were not different
(p:>0.05). During hypoxia (B—Pz) KO decreased (p‘<0.05) at
cl, but did not change (p:>0.05) at c4.

The net lactate transependymal flux rates, reflected
by data reported in Table 7, were calculated: assuming
cf==0.0 mM (J1) and cf==cp (J1). The difference between J1
and J1 at comparable ci's (range==80-230 nmoles/min) rep-
resents the maximum effect which changes in plasma lactate
concentration and/or Hf could have on net lactate flux. The
first assumption may be more nearly correct since data in
Table 6 suggest that CSF lactate concentration is indepen-
dent of that in plasma. J1 and Ji were negative at all ci's
indicating a net transependymal influx of lactate. During
Pl (Condition A) lactate influx at C1 was greater (p;<0.05)
than at either c2 or c3, while there was no difference

(p:>0.05) between net influx at c and c3. During normoxia,

2
net lactate influx increased (p‘<0.05) with time at all
ci's. Since cp was unchanged (Table 6), this suggests
lactate movement from brain during P2. At c1, lactate
influx was greater (p<:0.05) during A-P1 than during B-Pl.
During A—Pl, plasma glucose, pyruvate, and lactate were

increased (section 4.9.1.1) and may have increased brain

67

lactate production. At c3 and c4 blood metabolite
concentrations were not elevated and lactate influx in these
dogs were not different (p:>0.05). Lactate diffusion from
plasma cannot explain net lactate influx, since at c3 and
c4, during P1, ci exceeds cp (Table 6). During hypoxia
(B-Pz), net lactate influx was greater than during B-Pl
(p‘<0.05) at both c1 and c4 and the increase at c4 was
greater (p‘<0.05) than the increase during normoxia (A-Pz;

c3).

5.6. Mannitol flux from CSF

Both glucose and lactate Ko's appear to demonstrate
saturation kinetics (Tables 5 and 7) suggesting carrier-
mediated transport for these materials from CSF. 3H-
mannitol, considered to be a passively diffusing molecule
(Prockop, 1968; Bronsted, 1970a; Bierer, 1972), was in-
cluded in the perfusion fluid in lieu of 3H-glucose (sec-
tions 4.9.2.3 and 4.9.2.4). D-mannitol Ko's at low (cl)
and elevated (c2) ci's are summarized in Table 8. During
normoxia (P1) and hypoxia (P2), K0 was independent of con-
centration, supporting the hypothesis that D-mannitol leaves
the CSF only by bulk absorption and simple diffusion
(Prockop, 1968; Bronsted, 1970a; Bierer, 1972). Ko's for

mannitol, unlike those for glucose and lactate (Tables 5

and 7), were not different (p:>0.05) during hypoxia.

68

5.7. Pyruvate flux from CSF

 

The effects of pyruvate ci and cp on C0 are shown
by data reported in Table 9. Neither time (Condition A)
nor hypoxia (Condition B) affected Co' When cp was ele-
vated (Condition A; ci==0.00 mM), cO reflected cp and was
higher than when cp was normal (Condition B; ci== 0.00 mM),

indicating pyruvate may diffuse from plasma into CSF. When

c. was normal (0.12 mM) or elevated (0.24 and 0.29 mM), c

1 O

was the same as CJ.- and always greater (p<<0.05) than cp.
These data indicate that pyruvate moves into more readily
than it moves out of the CSF.

The net pyruvate transependymal flux rates, sum-
marized through data reported in Table 10, were calculated:
assuming cf==0.00 mM (pr) and that cf==c (J' ). The

P PY
difference between pr and J' at comparable ci's (range =

PY
2-12 nmoles/min) represents the maximum effect which changes
in cf and/or Vf could have on net pyruvate flux. Net
ruvate flux J and J' was ne ative at all c.'s
py ( py py) g 1
indicating a net transependymal influx. Net influx was

not different (p:>0.05) during P at any Ci' and indicates

2

that neitherAtime (A-P and P2) nor hypoxia (B-PZ) affected

1
pyruvate influx. During A—Pl, net influx at low ci (c1) was
greater (p‘<0.05) than at either normal (c2) or elevated
(c3) ci's. There was no difference (p:>0.05) in net influx
at c2 and c3. The greater (p<=0.05) net influx of pyruvate
at c1 in A-P1 when compared with B—Pl could be due to the

elevated cp in the former (see Table 9).

69

.mmx am “3:39.

.mommnucmumm cw moon mo Hon-52

 

 

 

 

 

 

 

 

 

o.on.~.o HHHI «Hﬂmm «Ho.OHmo.OI mmHXQ
m.OHH.mm Hﬂhm «Ham mo.0HhN.h Hmﬁm
mm
a: m.oe.o.mm a: 3mm 8: :8 a: No.0...mmK mmhm
Hm
maxomamlmaxoauoz .m
H.0HH.O HHO HHHI Ho.OHHo.OI mmeM
m.OHm.mm Hﬁmm NHOOH No.0Hmm.b mmﬁm
mm
3.3 m.0Hv.mm 3.3 HHmm Ami NHOOH 3.3 No.0meK mmHM
Hm
‘ aﬂxoshoz coaumusa msoq .¢
mu m m m
Auov a Amm say «00 m Amm say No m an coauwonou
scamsmnom unasownusm> sauna unwuso moat
oﬁxomhn can owxoﬁuoc cw “muev musumuwmaou Havoc“ can moum can .mom .mm Hmwumuu< .H manna

70

Table 2. Bulk absorption rate (Va) and CSF formation rate (Vf) in
normoxic and hypoxic dogs during brain ventriculation

perfusion

 

 

Condition 6a (ul/min)

A. Long Duration

Normoxia
P1—
XiSE 24:2 (15)
P2—
XtSE 30i3
Iii-(:53 6i2*
B. Normoxia-Hypoxia
P1—
XiSE 23:3 (12)
P2
X+SE 32i5
A 1513 9i3*

Vf (pl/min)

52 i 4 (15)

45:3
-7:2*

45 i 3 (12)

30 i 4*
-15 i 4”

 

Number of dogs in parentheses.
* < . ; .
p 0 05 Plale

+p< 0.05; A75B.

71

Table 3. Glucose, lactate, and pyruvate concentrations in simultaneously
obtained samples from arterial plasma and cisternal cerebro-
spinal fluid (CSF) in twelve anesthetized dogs

 

 

 

Glucose Lactate Pyruvate
Sample (mM) (mM) (mM)
Plasma 7.0i0.8 1.6i0.3 0.12:0.02
CSF 5.0i0.2* 2.0:0.4 0.13:0.01

 

Data expressed as mean tSE.

*p < 0.05, plasma #CSF.

72

m

.AH.H.m.v :oHuomm mmmv scamsmcw >ﬂ wn vmum>mam 0+

.me m an a36v m .

.mmmmnucmumm cw mmoc mo Hmnﬁsz

 

 

 

®.OHN.OI N.HH@.HI H.HHM.O Q.OHB.O mmuﬂm
v.0Hh.m N.HHO.® N.._..H.H.m m.OHm.N mmﬂm
Nnm
N.0Hm.m V.OH®.m H.0HH.HH m.OHm.m N.OHO.N 0.0 mmumm
Amy Amy Amy Amy Amy Amy Hm
maxomhm
unﬂxosuoz .m
¢.OHM.O m.OHH.OI m.OHH.O m.OHN.O «$.0HN.N *N.OH®.OI mmuux<
m.OHv.h m.HHO.HH h.NH©cOH m.HHm.0 +m.QHN.m 0.0Hm.N mmnﬂm
Nam
m.OHN.h m.HHH.HH OoHHmomﬂ F.NHm.OH N.HHM.® N.Oﬂm.¢ +O.HHv.HH 0.0Hv.m 0.0 MMHWW
Amy Amv Amv Avv Aﬁv Avv Amy Amv Amv Hm
mwxoauoz
coaumusa mcoq .¢
mo 00 an mo 00 Ho mo 00 an sowuﬁvaoo

 

 

0 Q

 

“may MEmMHm Haﬁumuum can .Aoov vwaam 3onuao .Aaov Uﬂaam scams“ on» cw coﬂumuucoosou mmoosamua

SE :oﬂumuucoosoo mmoousIo

\
5.

 

gownsmumm HMHﬁOﬁuusm> Gamma Uganda mmoc aﬂxomhn can owxoshoc ca

.e manta

73

v

.UHO

.m

o .No x HD um

m m
.n . n “ox amoovm...

.>Hw>ﬂuoommou
.28 H.0HH.HH can .o.HHm.©H .N.0Hm.v .o.o ou mcﬁuuwmmu aoﬂumuucmocoo mmooaamln 30Hmcﬂn“

.Nmk Hm Nmo.o Vans

an .m .~ .a

O O U

.mmmonusmumm ca mmoc mo Honssz

 

 

 

 

 

OMNHOHG «OBHOHNI Ommﬂomm OmHOMH «MHmH unﬂﬂmm mmuﬂm<
+O¢NHONm OOHHOva +OmNHO®¢ OVHHOMOI +OHHmm ¢HMHH mmufm
N
m
+OmHOHH OmHONNI +ONHOF OMHOOmI +mHmm NHHmVH mmuNm
E as E as E is He
aﬁxomam
Imaxosnoz .m
¢o Ho eo Ho v0 Ho
oaﬂﬁom OOHOVI OQHOO ONHHOG OmHON «OmHOhH GHMI NHO mHm mmuTMC
+OQHHOmm OBNHOmMI OMHHOHVI +O®HHOhm ommﬂomwl OVHomOI +hHva +Mva NHHHVH mmuﬂm
Nm
+OHHHOmh ONNHOﬂMI ONNHOBVI +OOHHOHm OHMHOmwl omHHONmI +0HHhm +Hva VHHmMH mmuNM
5 E as E E as E E so am
. owxoeuoz
coaumnso anon .¢
no No Ho no No Ho no «0 Ho casuaocoo
a Aeas\mmaosec me Aeaexaac ox

AcH8\meoEsv .o

[I]
(III‘

 

 

meanso mwop canon»: can u
Haﬁuos .AHUV 30H an A

h

w

 

HXOEOG Cd

 

 

 

 

 

 

macamSMHom unasowuucm> sauna

msoﬂumuuswocoo 30HmcH scamsmumm Ava «may wmum>mam can “Nov
u «may mung xdam um: can AOMV ucmﬂowmuooo xsamuso Hosapcommmsmuu mmoosao

.m wanna

74

m

.AH.H.m.v coauomm mmmv GOHmsmcH >w an umum>wam 0+

.me am 306 Va.

mononucoumm Ga mmOp mo Honasz

 

 

 

«m.OHm.¢ «N.OH©.H «H.HHv.m «H.0HN.H mmuNMQ
O.HHo.b ¢.0Hm.m H.HHm.m N.OHb.N mmuﬂm
Nm
m.OHN.N m.OHN.v m.OHm.v m.0Hm.m N.0Hm.a 0.0 mmuﬂm
on Amy Amy Amy Amy Ami Hm
mwxomwm
Inﬂxosuoz .m
N.OH0.0 m.OH®.O N.0Hm.o sN.OHo.H 0.0HM.HI«H.OHG.O mmuﬂm<
N.OHB.H N.0Hm.m H.0HV.N N.OHo.m +m.OH@.N m.OHm.N mmuﬂm
N
m
m.OHh.H m.OHm.v v.0HN.m n.0HN.N H.0HO.N N.0Hm.H +0.0Hm.m N.OHH.N 0.0 mmuﬂm
Amv Amy Amy Avv Adv Avv Amy Amy Amy Hm
waxoauoz
:oaumusn mcoq .¢
mo 00 go mo 00 ac mo 00 ﬂo coauﬂpsoo

 

 

 

ZS cowumuusmocou mumuomqun

 

 

 

cowmamuom umaaowuusm> sauna mcwusv mmov canon»: can owxoeuo: :« Amov

mammam Hmﬂuouum can .Aoov pﬂsam 30Hmuso .AH

.ov Uﬂsam 30Hmsw on» :w :oﬂumuusoocoo mumuomalq .m wanna

75

oOHO

.m

o .Noxao an.

.H

O
. n a x amodvm...

.>Hw>auommmmu
.zs n.0Hm.¢ can .v.oam.m .m.on.H .o.o 0» mcahummmu coaumnucmoeoo mumuomHuA soamca

40 .mo .mo .H

.max J 2.6.0 Va...

0

mononusmnmm ca moon mo umnﬁsz

 

 

 

 

 

«OQHOBHI «OMHOONI «OVHOmNI «omﬂomml OHHmH «mHmHl mmuﬂm<
+omHOMNI OVHONﬁI +OmHOHVI omﬂomOI mHHm hﬂmm Wmuﬂm
N
m
+Omﬂoml OMHONNI +OmHO®HI OmHOBMI +VHNM MHmh mmufm
so as so as as so Ha
manomwm
Imaxosuoz .m
v0 Ho v0 Ho go HD
«OQHOMHI «OGHOGHI «OMHOMHI ¥OFHONHI «ONHOQHI «ONHOmI VHGI ﬁHH m“? MWMTMQ
+O¢HOHNI +OMHONMI Ohﬁomﬂl +OBHOHMI +ONHOOQI OhHOhml +OHHMV +MHm¢ hﬂwh mmufm
Nm
+ONHOml +OHHOMHI OQHONMI +OMH0mHI +OHHONNI OGHONmI +mﬁmﬂ +®Hmv mHVB Mmuﬂm
E E a: E E as E 3 so He
.mwxoauoz
cowumuso mcoq .¢
no No Ho MO ND Ho no No do coﬁuﬂccoo
155385. we “558355 a... 2233 ex

 

 

 

 

 

Haﬁuo: .Aaov 30H no Amp “any mums stm um: can AOMV ucmﬁowmmmoo asamuso Hmsacsmmomcmuu «panama

cowmsmumm Hmaaowupcm> sauna
msﬁuso moon owxomas can owxoauo: cw msoaumnusmosoo 30Hmsw coﬁmSMHmm Ava «mov woum>mao can .Amov

.h wanna

76

Table 8. Mannitol transependymal outflux coefficient (KO) at low (cl)
and elevated (c2) perfusion inflow concentrations in normoxic
and hypoxic dogs during brain ventricular perfusion

 

 

KO (pl/min)

 

Condition c1 c2

 

B. Normoxia-Hypoxia

l
XiSE 14:2 (3) l8i5 (3)
P2—
XiSE 14:1 1614
Axisn 0:2 1:2

 

Number of dogs in parentheses.

= inflow mannitol concentration referring to 0.0 and 16.8:t0.2 mM,

c , c
1 2 respectively.

77

m

.AH.H.m.¢ coﬂuowm mmmv sawmamcﬁ >ﬁ an vmum>0ao 0+

.momwnuswumm ca mmoo mo umnssz

 

111!

 

 

H0.0HH0.0 H0.0HO0.0 H0.0HH0.0 H0.0HH0.0 mmuNMG
HO.OHOH4O H0.0HMN.O H0.0H¢H.O H0.0Hm0.0 Mmuﬂm
Nam
H0.0Hm0.0 H0.0HNN.O H0.0H¢N.O H0.0HMH.O H0.0Hm0.0 0.0 mmuﬂm
on Rwy va Amy Amv Amy Hm
Maxommm
Inﬂxosuoz .m
H0.0HH0.0 O0.0HO0.0 H0.0HO0.0 H0.0HO0.0 HO.OHNO.OI H0.0HH0.0 mmuﬂm<
N0.0HNH.O H0.0HON.O H0.0Hm0.0 m0.0HmH.O +H0.0HNN.O N0.0HNN.O Wmuﬂm
N
m
N0.0HHH.O H0.0HmN.O H0.0HmN.O H0.0H®0.0 v0.0HmH.O N0.0HNH.O +N0.0H¢N.O N0.0HHN.O 0.0 mmuﬂm
Amy Amy Amy Aev Avv Avv Amy Amv Amy Hm
maxosuoz
coﬁumusa mcoa .m
m0 00 “0 m0 00 do m0 00 «0 soﬂuﬂvcoo

 

 

Amuv mammam Haﬁnmuum can .Aouv candy onmuao .Aauv cﬂaam scamsw 0:0 Cw mcoﬂumuucuosou muo>9uhm s00:

 

SE soaumuucwusou oum>5u>m

 

 

 

sowmnmuom Hoanuwuuco> sauna mcﬂuaw mmoc 0wxom>£ can oﬂxoauoc ca

.m manna

78

w

.OHO

am

0 .muxao 00

mm >m
L... a h. .mo.o vm...

.>H0>ﬂu00mmwu .25

 

 

 

 

 

Ho.onam.o .Ho.OHam.o .mo.oama.o .o.o op maauumumh coaumuuemueoo mum>nuam onmca u we .mo .mo .Ho
.mmmwnucwumm cw mmoc mo Honﬁaz
who vwm man «.3: mm H M<
+HHNI mHmHI +mHv: «Hum: mmnﬂm
mm
3: +37 8V 3:: g +73: Sc 38: mm Hm
mwxomxmlmaxoauoz .m
m0 H0 MU HO
. H3 .30 Tam: NHH «Hun NHHI mm “no.3
+HH>I +vﬁml wﬂdvl +NHNHI +mHVHI MHOmI mm HIM
N
m
Amy +me| va +mﬁm| on mammn Amy +NHmHI Avv +onHI “mo «Howl mmnﬂm
-- Hm
mwxoauoz
sowumusa anon .d
m0 N0 H0 m0 . «0 H0 coauﬁvcoo
Acas\mmaoasv amp Aswﬁ\moaosav man

 

 

 

 

sowmsmuom Hudsowuuc0> Gamma msﬁusv mmoc oﬂxommn can oﬂxoshoc Ga maowumuucmosoo scams“

cowmzwuom avo .m0v noum>0H0 can .~N0V awake: .AHUV 30H no a

a?

.%m

. by xaam 00¢ oum>suhm

.OH GHQMB

VI . DISCUSSION

6.1. CSF bulk absorption and formation rates

Perfusion of the brain ventricles with artificial
cerebrospinal fluid (CSF) containing test molecules permits
the study of exchanges between this fluid and the blood or
the brain. Molecular clearance is expressed as the volume
of CSF from which a substance is removed per unit time. It
is a function of the rate at which the test molecule enters
and leaves the ventricles through a perfusion inflow and
effluent, respectively, as well as its concentration in the
ventricles.

Inulin, a large molecule, does not penetrate the
ventricular ependyma, and its clearance reflects CSF bulk
absorption rate (Pa) (Rall et aZ., 1962; Heisey et aZ.,
1962). Inulin clearance which is a direct, linear function
of intraventricular pressure, was used to predict 6a in the
present study (Heisey et aZ., 1962; Bering and Sato, 1963;
Bierer, 1972): Intraventricular pressure was maintained at
a steady level (section 4.2) so that 6a was constant.

Data reported in Table 2 indicate that 6a was un-

affected by hypoxia. Increased Va in experiments testing

the effects of hypoxia is attributable to a factor of time

79

80
since 6a increased similarly in both normoxic and hypoxic
animals. A similar judgment may be valid for analogous data
reported earlier (Michael et aZ., 1971) in which radio-
iodinated human serum albumin (RIHSA) was used as the test
molecule for calculating Ha.

The increase in 6a with time (Table 2) may be attrib-
uted to preparation deterioration (Cserr, 1965). Data
reported in Table 8 indicate that mannitol Ko did not change
during hypoxia, implying that neither time nor hypoxia
affect mannitol transependymal permeability. Since inulin
is a larger molecule than mannitol, it is presumed that its
ventricular permeability would be similarly unaffected, and
inulin clearance would remain an accurate estimation of Pa
in these experiments.

Changes in 6a (Table 2) may be related to changes in
saggital sinus venous pressure (SSVP; unmeasured in this
study), since both intraventricular pressure (IVP) and an
IVP-SSVP pressure difference linearly and directly vary Ga
(Heisey et aZ., 1962; Bering and Sato, 1963). Since IVP was
maintained constant, SSVP appears an effective independent
variable in these studies, and changes in 6a may have
occurred distal to the fourth ventricle at the arachnoid
villi.

The mean CSF formation rate (Hf) of 50 ul/min
(Table 2; A-Pl; B-Pl) is similar to that previously reported

for the anesthetized dog (Bering and Sato, 1963; Oppelt et

81

aZ., 1964; Atkinson and Weiss, 1969). Similarly, the range
of 9f reported here (30-85 ul/min) is the same as that
reported earlier (Oppelt et aZ., 1964; Cserr, 1965).

0f is proposed to be an active transport system
(Vates et aZ., 1964; Cserr, 1965; Holloway et aZ., 1972)
and would be expected to vary with cerebral oxygen con-

sumption and cerebral blood flow (CBF; Bering, 1959). Vf
decreased significantly during hypoxia (Table 2; B-P2) as
reported earlier (Holloway et aZ., 1972; Michael et aZ.,
1972), possibly through an indirect effect on CBF (Kety and
Schmidt, 1948; Fujishima et aZ., 1971); it also decreased
with time. Although total brain perfusion increases with
hypoxia (Kety and Schmidt, 1948; Shapiro et aZ., 1970),
decreases in regional CBF may affect 9f.

CSF formation may be reduced not only due to the
direct and indirect effects of hypoxia but also due to the
effects of the anesthetic. Barbiturates affect 9f possibly
by inhibiting active transport of ions (Lorenzo et aZ., 1968;
Bering, 1959) and may be implicated in the decrease in 9f

with time reported here, related to the anesthetization

procedure (section 4.9).

6.2. Glucose flux from CSF
Carrier-mediated glucose transport has been reported
for the membranes which separate the blood, brain and CSF

compartments, and which moves glucose from blood into brain

82

tissue (Crone, 1965; Gilboe et aZ., 1970), from blood into

CSF (Fishman, 1964; Atkinson and Weiss, 1969; Welch et aZ.,
1970) and from the CSF into either blood or brain (Bradbury
and Davson, 1964; Bronsted, 1970a and b). In vitro studies
using the choroid plexus, further suggest active transport

of glucose from the CSF (Csaky and Rigor, 1968). In brain

ventricular perfusion studies, glucose can go from CSF into
the cerebral capillary blood, the brain parenchyma (neurons
and glial cells), or the epithelium of the choroid plexus,

however the specific course is unidentifiable.

As glucose concentration in the perfusion inflow
increased, its efflux (KO; Table 5) decreased, implying
saturation of a carrier-mediated transport system, as
reported earlier (Bradbury and Davson, 1964; Bronsted,
1970a and b). This occurred when the ventricular glucose
concentration was nearly equal to that in normal CSF (5.0
mM; Table 3).

Since mannitol (which moves only by diffusion) and
glucose show similar diffusion characteristics, the inhibi-
tion of glucose active transport should result in similar
Ko's for both molecules. Even with glucose transport
saturated, its Ko (Table 5) remained 4 times greater than
that for mannitol (Table 8), suggesting additional transport

systems with a higher saturation point.

83

The decrease in glucose Ko during hypoxia (Table 5)
suggests the inhibition of an aerobic process. Glucose
accumulation by the choroid plexus (in vitro) is inhibited
by DNP and anoxia suggesting an active transport mechanism
(Csaky and Rigor, 1968). Intraventricular administration
of ouabain, a Na+ -K+ ATPase inhibitor, reduces both glucose
efflux and 9f (Bronsted, 1970a), suggesting further that
both are dependent on active transport. In the present
study, hypoxia also may have inhibited the high capacity
glucose transport system.

Consistent with the observation that glucose KO
decreased during hypoxia (Table 5) are data indicating net
glucose influx during hypoxia increased when glucose concen-
tration in the perfusion inflow was less than that in plasma.
When perfusion inflow concentration was elevated above that
of plasma, however, there was no change in net glucose out-
flux, even though glucose KO decreased. However if the
active transport component of glucose efflux makes a small
contribution to its net flux, net glucose flux would reflect
concentration differences among compartments. In the pres-
ent study, net glucose flux during normoxia and hypoxia
reflected glucose concentration differences between the CSF
and plasma. This supports the concept of a low capacity
active transport system for glucose in the choroid plexus

(Csaky and Rigor, 1968).

84

6.3. Lactate flux from CSF

 

Lactate concentrations in the blood are independent
of those in either brain or CSF (Klein and Olsen, 1947;
Alexander et aZ., 1962), whereas CSF lactate concentrations
correlate with those in brain (Plum and Posner, 1967; Plum
et aZ., 1968; Kaasik et aZ., 1970). Hydrogen ion activity
(pH) influences the distribution of lactate among these
fluids (Gurdjianet aZ., 1944; Milne e1: aZ., 1958). Lactate
exists in brain, CSF and blood as an anion (pK==3.8) but
diffuses across compartment membranes only when non-ionized.

Net lactate flux rates were negative (Table 7) even
at elevated lactate concentrations in the perfusion inflow,
indicating a nEt lactate influx. Blood pH exceeds that of
CSF during normocapnic normoxia (Davison, 1967), and would
favor lactate diffusion from CSF. When lactate concentra-
tions in the perfusion inflow exceeded those in plasma, net
lactate influx must be either from brain or by an active
process from blood. When perfusion inflow concentration
was 0.0 mM net lactate influx (cl; Table 7) was more when
plasma glucose and pyruvate concentrations were elevated and
further suggests that the increased lactate influx was due
to increased lactate production by brain. The lack of a
direct linear relationship between perfusion inflow con-
centrations of lactate and its net influx (net influx was

same at normal and elevated perfusion inflow concentrations)

85

are consistent with the hypothesis of carrier-mediated
lactate transport (Stein, 1967).

Net lactate influx increased with time, but
increased more during hypoxia (Table 7). The increased
lactate influx with time may be attributable to mitochon-
crial inhibition by a barbiturate anesthetic (Aldridge,
1962). Hypoxia results in an increased glycolytic rate and
increased brain lactate concentration (Lowry et aZ., 1964;
Kaasik et aZ., 1970), which would increase net lactate
influx.

As lactate concentration in the perfusion inflow
increased, its outflux coefficient (KO; Table 7) decreased,
implying saturation of carrier-mediated transport. This
conflicts with the hypothesis that lactate clearance from
CSF is only by non-carrier mediated diffusion and bulk
absorption (Prockop, 1968). Lactate transport was satu-
rated when CSF lactate concentrations were normal (1.6 mM;
Table 3) and explains the failure to identify a saturable
carrier at CSF concentrations of 5-10 mM (Prockop, 1968).
When the lactate carrier was saturated lactate K0 was
approximately twice that of mannitol as previously reported
(Prockop, 1968). Further, the inclusion of mannitol in lieu
of glucose in the perfusion inflow (section 4.9.2) allows
the inference that lactate efflux is independent of glucose

flux (Prockop, 1968).

86

Lactate KO (Table 7) was unaffected by time but
at c1 KO decreased from normoxic values during hypoxia.
The KO decrease could be due either to inhibition of an
active aerobic transport or to competition for the lactate
carrier due to increased lactate influx from brain. Lactate
Ko at elevated perfusion inflow concentrations of lactate
did not change significantly with hypoxia, indicating no

change in transependymal permeability to lactate.

6.4. Pyruvate flux from CSF

Klein and Olsen (1947) suggested that pyruvate
moves rapidly from plasma into brain tissue and is rapidly
converted to lactate. Data in Table 9 show that elevated
plasma pyruvate concentrations may be reflected by those
in CSF. Rapid equilibration of pyruvate (pK==2.5) among
the blood, CSF and brain would suggest some process other
than simple diffusion.

There was always a net pyruvate flux into CSF
(Table 10). Simple diffusion of pyruvate into CSF from
plasma cannot explain the net influx at elevated perfusion
inflow concentrations (c3, c4), since neither the pyruvate
concentration gradient nor the pH of these fluids favor it.
When plasma pyruvate concentrations were the same (Table 9),
there was no direct linear relationship between net pyruvate

flux and pyruvate concentrations at normal and elevated

87

concentration in the perfusion inflow fluid (Table 10)
further implying a carrier-mediated pyruvate transport.
Data in Table 9 suggest pyruvate movement from CSF
is not rapid and the failure of net pyruvate influx to
change with either time or hypoxia (Table 10) would suggest

that movement from CSF into brain may be slow.

VII. SUMMARY

1. Inulin clearance a measure of the bulk
absorption rate of cerebrospinal fluid (CSF) increased
with perfusion time.

2. Mean CSF formation rate (Vf) in normoxic dogs
was 50 ul/min. Vf decreased both with perfusion time and
hypoxia, and the latter decrease was greater.

3. Mannitol efflux from CSF was unaffected by time,
hypoxia, or concentration, implying movement by non-carrier
mediated diffusion.

4. Carrier-mediated glucose efflux from CSF may be
by two types of carriers. The first, a low capacity-carrier,
was saturated at normal CSF glucose concentrations. The
second, a high capacity-carrier, accounted for an outflux
coefficient which was significantly greater than that of
mannitol.

5. The glucose outflux coefficient was unaffected
by time but was significantly reduced during hypoxia, sug-
gesting inhibition of an-aerobic dependent glucose transport.

6. Net glucose flux during normoxia and hypoxia
reflected glucose concentration differences between the CSF

and plasma.

88

89

7. There was a net lactate flux into CSF at CSF
lactate concentrations ranging from 0.0 to 5.0 mM. Net
lactate influx increased with time, but increased more
during hypoxia. Carrier-mediated lactate transport was
indicated by the lack of a direct linear relationship of
net lactate flux with perfusion inflow concentrations of
lactate.

8. Carrier-mediated lactate flux from CSF was
indicated by concentration dependent outflux coefficient
(K0). K0 was unaffected by perfusion time, but was
decreased with hypoxia when lactate concentration in the
perfusion inflow was 0.0 mM.

9. There was a net pyruvate flux into CSF at CSF
pyruvate concentrations ranging from 0.00 to 0.30 mM.
Carrier-mediated pyruvate transport was indicated by the
lack of a direct linear relationship between net pyruvate
flux and pyruvate concentrations in the CSF. Neither time
nor hypoxia affected net pyruvate influx.

10. Net fluxes of glucose, lactate, and pyruvate
were calculated assuming two different values for their
concentration in newly formed CSF (cf). The direction of
their net flux was independent of cf's indicating minimal

contribution of Vf to exchange of metabolites.

APPENDICES

APPENDIX A

COMPOSITION AND PREPARATION OF ARTIFICIAL

DOG CEREBROSPINAL FLUID (CSF)

Dog CSF contains: Na+, 150 mEq/l; K+, 3.0 mEq/l;
+

3 I

Ca++, 2.3 mEq/l; Mg+

-3
4

, 0.8 mEq/l; 01', 133.2 mEq/l; HCO

25.0 mEq/l; PO , 0.5 mEq/l (Cserr, 1965).

Reagents

l. NaZHPO4-H20, analytical reagent (A.R.)

2. KCl, A.R.

3. NaHCO A.R.

3'
4. NaCl, A.R.

5. CaClZ, A.R.
6. MgC12'6H20, A.R.
Solutions

 

A. 34 mg of reagent 1; 224 mg of reagent 2; 2.1 g of
reagent 3; 7.25 g of reagent 4 are dissolved in
distilled H20; q.s. to one liter.

B. 12.76 g of reagent 5 is dissolved in distilled H20;
q.s. to 100 ml.

C. 8.13 g of reagent 6 is dissolved in distilled H20;

q.s. to 100 ml.

90

91

CSF for ventricular perfusion was made by adding
0.1 ml of both solutions B and C to 100 m1 of solution A.
The resultant mixture is adjusted to approximately pH 7.3

by bubbling with 3-9% CO2 at room temperature for one-half

hour prior to use.

APPENDIX B

PREPARATION OF ANESTHETIC

The anesthetic, Dial and Urethane solution, was
prepared from a commercial formula (CIBA Pharm. Co.,
Summit, N.J.). Each m1 of anesthetic solution contained:
100 mg diallylbarbituric acid, 400 mg urethane, and 400 mg

monoethyl urea.

Reagents
l. Diallylbarbituric acid (Dial; K & K Lab. Inc.,

Plainview, N.Y.).

2. Monoethyl urea (Pfaltz and Bauer, Inc., Flushing,
N.Y.).

3. Disodium calcium ethylene diamine tetraacetate
trihydrate (Pfaltz and Bauer, Inc.).

4. Ethyl carbamate (Urethane; Aldrich Chemical Co.,

Inc., Milwaukee, Wis.).

Procedure

 

To make 100 ml of anesthetic, place 10 g of reagent
1; 40 g of reagent 2; and 40 g of reagent 4 in a 250 ml
beaker. Dissolve 50 mg of reagent 3 in one ml of distilled

H20, and add solution to powders. Place beaker in water

92

93

bath (100°C) and stir occasionally until solution is
complete. Cool to room temperature; q.s. to 100 ml with
distilled H20. Place solution in dark bottle, cap, and

store at room temperature.

APPENDIX C

DEPROTEINIZATION OF PLASMA AND CSF SAMPLES

Reagents

l.

2.

HClO 70-72%, w/v; Sp. Gr. 1.6; A.R.

4'

K CO

2 3, A.R.

Solutions

 

A.

HClO4, approx. 6%, w/v
Dilute 15.6 ml of reagent 1 to 300 ml with

distilled H20.

K2CO3,

Dissolve 69 g of reagent 2 in approximately 80 ml

5M

of distilled H20; q.s. to 100 ml.

Deproteinization procedure

(0°C);

Frozen plasma and CSF samples were thawed in ice

0.5 m1 aliquots were added to 4.0 ml of ice-cold

solution A, mixed, and centrifuged (1000 x g at 0°C) for

10 minutes to precipitate and remove protein from the sample.

Three-tenths ml of solution B was added dropwise, with mix-

ing, precipitating KClO3 and neutralizing the solution to a

pH of 6.8-7.2 (addition of solution B is done slowly to

94

95

prevent the loss of sample by rapid CO evolution). The

2
samples were centrifuged (1000 x g; 0°C; 10 min.) to remove
KClO3. The supernate was decanted and used for subsequent

analyses. Total dilution of deproteinated sample was 9.6x.

APPENDIX D

SPECTROPHOTOMETRIC DETERMINATION OF D-GLUCOSEl

Principle

 

The glucose oxidase method is a coupled enzyme
system for the quantitative, colorimetric determination of
D-glucose (Hugget and Nixon, 1957) based upon the following
reactions:

glucose oxidase
l. D-glucose-toz-tHZO :H202-+D-g1uconic acid
peroxidase
2. H202-+reduced chromogen t 1: oxidized chromogen

 

 

Reaction 1 is specific for D-glucose. The amount of
oxidized chromogen present in the assay mixture is deter-

mined by its absorbance at 400 mu.

Reagents

l. Glucostat reagent kit (Worthington Biochemical Corp.,
Freehold, N.J.)
a. buffered glucose oxidase vial
b. chromogen vial

2. HCl, concentrated, A.R.

 

1From H. U. Bergmeyer and E. Bernt (1965).

96

97

3. Glucose stock solution, 10 mg glucose/ml (55.6 mM)
containing 0.25% benzoic acid (Fisher Scientific Co.,

Fairlawn, N.J.).

Solutions

 

A. Glucostat reagent (for approximately 80 determina—
tions)
Dissolve contents of two vials of reagents 1a and 1b
in distilled H20; q.s. to 160 ml.

B. HCl, 4N
172 ml of reagent 2 diluted to 500 ml with

distilled H20.

Glucose standard solutions

 

Glucose standards (2.78, 5.56, 8.34, and 11.1 mM)
are made within 24 hours of the assay by diluting 0.5, 1.0,

1.5, and 2.0 ml of reagent 3 to 10 ml with distilled H O.

2

Procedure
To 0.5 ml duplicate aliquots of deproteinated

samples, standards, and distilled H 0 (see Appendix C),

2
add 2.0 ml of solution A, mix, and incubate at room temper-
ature. After exactly 10 minutes incubation, add two drops
of solution B, mix, and let stand for at least 5 minutes.
Read absorbancy (O.D.) of standards and samples in spectro-
photometer (Model DB, Beckman Instruments Inc., Fullerton,

Cal.) against the distilled H 0 blank at 400 mu.

2

98

Calculations

 

Plot absorbancy (O.D.400) as a function of the
concentration of the glucose standards. The plot should
be linear through 11.1 mM. The concentration of glucose
in the unknown samples is calculated using the following
formula:

2(C /OD )
C = S S x ODu

11 n
S

 

where:
CD = absorbancy
C = concentration
5 = standard
u = unknown

n = number of samples.

APPENDIX E

FLUOROMETRIC DETERMINATION OF PYRUVATEl

Principle

 

Lactic dehydrogenase (LDH) catalyzes the reduction
of pyruvate by reduced nicotine-adenine dinucleotide (NADH)

to form lactate and nicotine—adenine dinucleotide (NAD).

+ LDH +
Pyruvate-FNADH-tH —————+ Lactate-tNAD

The equilibrium of the reaction favors lactate formation at
pH 7.5. With excess NADH in the mixture, the reaction pro-
ceeds to completion and pyruvate is quantitatively converted

to lactate. When NADH is oxidized to NAD+, absorbancy and/or

fluorescence decreases, and this decrease is measured.

Reagents
1. NaOH, A.R.

2. (NH 504, A.R.

4)2

3. NaHCO A.R.

3!
4. Triethanolamine-HCl, A.R. (Sigma Chemical Co.,

St. Louis, Mo.)

 

1From J. R. Williamson and B. E. Cory (1969) and
T. Buchner et al. (1965).

99

100

Ethylene—diamine-tetraacetic acid, EDTA-disodium
salt, EDTA-NazHZ-ZHZO.

Reduced nicotine-adenine dinucleotide, NADH

(Sigma Chemical Co.). Stable for at least a year
when stored dessicated in dark at 0-4°C.

Lactic dehydrogenase suspension (LDH; Sigma Chemical
Co.). Crystalline preparation from rabbit skeletal
muscle which is suspended in 2.1 M ammonium sulfate;
stable for periods up to a year if kept between
0-4°C. The concentration of LDH is 5 mg protein/ml
with a specific activity of at least 600 units/ml.

Sodium pyruvate, A.R. (Calbiochemical Corp., Los

Angeles, Cal.). Stored dessicated in dark at 0-4°C.

 

Solutions
A. NaOH, 18N
Dissolve 360 g of reagent 1 in distilled H20;

q.s. to 500 m1.

(NH4)280 2.1 M

4!
Dissolve 27.8 g of reagent 2 in distilled H20;
q.s. to 100 ml.

NaHCO 1%, w/v

3’

Dissolve l g of reagent 3 in distilled H O;

2
q.s. to 100 ml.

Triethanolamine buffer (TEA), pH 7.5
Suspend 37.2 g of reagent 4 and 7.4 g of reagent 5

in 400 ml of distilled H20. Add solution A until

101

pH 7.5 is obtained; q.s. to 500 ml with distilled
H20.
E. NADH

1. To 30 mg of reagent 6 add solution C; q.s. to
100 ml. This solution is for the spectrOpho-
tometric determination of the pyruvate standard
concentration.

2. Dilute 0.1 ml of solution E-l to 10 ml with
solution C for use in fluorometric assay.

F. LDH (for approximately 50 determinations)

Dilute 0.1 ml of reagent 7 to 1.0 ml with solution B.

Pyruvate standard (approximately 0.1 mM)

 

Dissolve approximately 11 mg of reagent 8 in
distilled H20; q.s. to one liter. Prepare fresh solution
just prior to assay.

Spectrophotometric determination of pyruvate standard
concentration (External standard)

 

 

To three test tubes containing 4.0 ml of solution D,
2.0 ml of pyruvate standard, and 0.025 ml of solution E-l,
add 0.020 ml of solution F and mix. To another test tube
(reference blank) containing the above constituents add
0.020 ml of distilled H20 in lieu of the LDH. Read in spec-
trophotometer (Model DB, Beckman Instruments Inc.) at a wave-
length of 340 mu. Since the absorbance of the sample (with
LDH) is less than that of the blank (without LDH) the refer-

ence blank is inserted in the site normally used for samples.

102

Calculations

 

Calculate the concentration of the pyruvate standard

from the following formula.

V1/6.22 x L/V2 x OD = C8

where:
Vl = total volume in test solution (6.045 ml)
V = volume of standard added (2.0 m1)
L = length of light path (1 cm)
6.22 = molar extinction coefficient of NADH at 340 mu

(cmz/ 11M)

OD = mean absorbancy of three pyruvate standard
samples

C = pyruvate concentration of standard solution (mM).

Fluorometric assay_for pyruvate

 

To obtain assay range prepare a test tube containing:
4.0 ml of solution D, 0.2 m1 of deproteinated standard
pyruvate solution, 0.010 ml of solution E-2, and 0.02 ml
of distilled H20. Mix and read sample fluorescence against
that of a distilled H20 blank in Model 110 Fluorometer
(G.K. Turner Assoc., Palo Alto, Cal.) with:
a. UV light source, 4 watt, Cat. #110-850 (G.K. Turner
Assoc.)

b. Primary filter (360 mu), Cat. #110-811 (G.K. Turner

Assoc.)

103

c. Secondary filter, 2A (415 mu), Cat. #110-816

(G.K. Turner Assoc.)

d. Neutral density filter (A.H. Thomas, Philadelphia,

Pa.).

Use the appropriate neutral density filter such that the
fluorescence reading for the assay range is between 80-100
scale units. Record this as Rr‘ This reading is the
maximum fluorescence due to NADH.

For both deproteinated samples, and the pyruvate
standard (the concentration of which has been determined
spectrophotometrically) add to test tubes, in order, 4.0 ml
of solution D, to one test tube 0.2 ml and to another 0.1 ml
of sample (0.2 ml and 0.1 ml of sample will suffice as two
readings per sample), and 0.020 ml of solution F, mix and
read fluorescence. Record as R1. (This is the sample
background fluorescence). Add 0.010 ml of solution E-2,
mix, and incubate at room temperature, and measure fluo-
rescence after exactly 15 minutes. Record fluorescence
(R2). R2 of tubes containing 0.2 ml of sample should be
one-half that containing 0.1 ml of sample. If 0.2 ml sample
reads zero or R1, more NADH must be added or sample size
decreased. Check reaction completeness by adding 0.010 ml
of solution E-2 and fluorescence should increase by a value

of R .
r

104

Calculations
The pyruvate concentration in the samples are

determined from the following equation.

Rr - (RLu—Rl,u)
Rr- (R2,s -Rl,s

) X C8 = Cu
where:
R = fluorometer reading
1 = background (i.e., no NADH)
2 = assay reading (i.e., after NADH added)
r = assay range (i.e., no LDH and full NADH deflection)
C = concentration

u,s = unknown and standard samples, respectively.

Special precautions

 

NADH fluorescent properties are temperature depen-
dent, thus all assay mixtures must be read at the same
temperature. All tubes used in the direct reading of
fluorescence must be soaked in soapy water overnight and
then rinsed a minimum of 6 times in hot tap water, followed
by 6 rinses in distilled H20, and finally 2 rinses in
deionized distilled H O to remove all trace of fluorescent

2

compounds in the soap.

APPENDIX F

SPECTROPHOTOMETRIC DETERMINATION OF L-LACTATEl

Principle

 

Lactate dehydrogenase (LDH) catalyses the oxidation
of L-lactate with nicotine-adenine dinucleotide (NAD) to
form pyruvate and reduced nicotine-adenine dinucleotide

(NADH).

LDH
L-lactate-tNAD t pyruvate-tNADH -+H

 

Since the equilibrium of the reaction lies far to the left,
the reaction products must be removed from the mixture to
obtain quantitative oxidation of L-lactate. Protons are
bound by use of an alkaline reaction medium (pH==9.5) and
pyruvate is trapped as the hydrazone. The basic equation
describing the reaction used in the spectrophotometric assay

of L-lactate is:

. L +
L-lactate-+NAD-+Hydra21ne —2§» pyruvate hydrazone-+NADH+-+H

0

The amount of NADH+ formed is determined by its absorbancy

at 340 mu.

 

1From J. P. Ellis et a1. (1963).

105

106

Reagents

l. NaOH, A.R.

2. (NH4)ZSO4, A.R.

3. Hydrazine sulfate, A.R. (Sigma Chemical Co.)

4. Glycine, A.R. (Sigma Chemical Co.)

5. Ethylene-diamine—tetra-acetic acid, EDTA disodium
salt, EDTA-Nasz-ZHZO

6. Nicotine-adenine dinucleotide (NAD; Sigma Chemical
Co.)
NAD is stable for periods up to a year when stored
frozen and desicated.

7. Lactic dehydrogenase suspension (LDH; Sigma
Chemical Co.)
Crystalline preparation from rabbit skeletal muscle
which is suspended in 2.1 M ammonium sulfate and
stable for periods up to a year if kept between
0-4°C. The concentration of LDH is 5 mg protein/ml
with a specific activity of at least 600 units/ml.

8. L-lactic acid standard solution (4.4 mM; Sigma

Chemical CO.).

Preparation of solutions

 

IA.

NaOH, 18N
Dissolve 360 g of reagent 1 in distilled H20;

q.s. to 500 ml.

107

B. (NH 2.1 M

4’2504'

Dissolve 27.8 g of reagent 2 in distilled H O;

2
q.s. to 100 ml.

C. Hydrazine-glycine buffer (0.4 M Hydrazine;
1 M glycine; pH 9.5). Suspend 7.5 g of reagent
4, 5.2 g of reagent 3, and 0.2 g of reagent 5 in
approximately 30 m1 of distilled H20. Add solution
A until pH 9.5 is obtained and dilute mixture to

100 ml with distilled H O. Dilute 50 m1 of this

2

solution with equal volume of distilled H O for

2
assay. Undiluted hydrazine-glycine buffer solution
is stable for at least two weeks when stored at
0-4°C.

D. LDH
Dilute 0.4 ml of reagent 7 to 2 ml with solution B.

E. NAD
Dissolve 160 mg of reagent 6 in 100 ml of diluted

solution C just prior to use.

Lactate standard solutions

 

Lactate standards (0.6, 1.1, 2.2, 3.3 mM) were
prepared by diluting 1.25, 2.5, 5.0, and 7.5 ml of reagent 8
to 10 ml with distilled water. L-lactate standards are

stable indefinitely, when stored at 0-4°C.

108

Procedure

 

To 0.3 ml duplicate aliquots of deproteinated
samples, standards, and distilled H20 (see Appendix B) add
2.0 m1 of solution E and mix. Add 0.04 ml solution D and
mix. Incubate at room temperature for exactly 35 minutes.
Read absorbancy (O.D.) of standard and sample solutions in
spectrophotometer (Model DB, Beckman Instruments, Inc.) at

340 mu against the distilled H 0 blank. Read O.D. again

2
after 45 minutes, and at 10 minute intervals thereafter
until readings stabilize. A change in O.D. from the initial
reading indicates the reaction was not complete. If the
reaction does not reach a constant end point within 60
minutes, more enzyme or a fresh preparation of enzyme is
needed because the activity of the LDH (solution D) was too

low. All reactants must be at room temperature prior to

adding solution D.

Calculations

 

Plot absorbancy (O.D.340) as a function of lactate
concentration in the lactate standard solutions. The plot
should be linear through a concentration of 4.4 mM. Lactate
concentration in the unknown samples can be calculated using
the following equation:

2(C /OD )
c = S S x on
0. ns u

 

where:

OD

109

absorbancy
concentration
standard
unknown

number of samples.

APPENDIX G

SPECTROPHOTOMETRIC DETERMINATION OF INULIN

(Direct Resorcinol Method Without Alkali Treatment)1

Principle

Inulin, a polysaccharide, is hydrolyzed into its
fructose moieties; fructose reacts stoichiometrically with
resorcinol forming a colored complex with peak absorbancy

at 490 mu.

Reagents

1. Resorcinol, A.R.
2. Ethanol, 95%
3. HCl, concentrated, A.R.

4. Inulin, A.R. (Pfanstiehl Lab. Inc., Waukegan, Ill.)

Solutions
A. Resorcinol (100 mg%)
Dissolve 100 mg of reagent 1 in reagent 2; q.s. to
100 ml. Prepare fresh just prior to assay.
B. HCl, 10 N

Add 224 ml of distilled H20 to 776 ml of reagent 3.

 

1From H. w. Smith (1956).

110

lll

Inulin standard solutions

Dissolve 200 mg of reagent 4 in distilled H O;

2
q.s. to 100 m1. Pipette 7.5, 5.0, 4.0, 3.0, 2.0, and 1.0 ml
of this solution (2 mg/ml) and dilute each to 10 ml with
distilled H20 obtaining 1.5, 1.0, 0.8, 0.6, 0.4, and 0.2

mg/ml standards, respectively. Store inulin standards at

0-4°C.

Procedure

 

To duplicate 0.05 ml CSF samples (nondeproteinated),
inulin standards, and distilled H20, add 1.0 ml of solution
A and mix. Add 2.5 ml of solution B (in fume hood) and mix.
Place marbles on top of test tubes and heat at 80°C for
25 minutes. Cool tubes to room temperature in cold water
and read O.D. of samples and standards at 490 mu in spec-

trophotometer (Model DB, Beckman Instruments, Inc.) against

distilled H20 blank within one hour.

Calculations

 

Plot absorbancy (O.D.490) as a function of the
concentration of the inulin standards. The plot should be
linear through a concentration of 2 mg/ml. The concentra-
tion of the unknown samples can be determined from the

following formula:

£(C /OD )
C = S S x ODu

u n
S

 

where:

OD

112

absorbancy
concentration
standard
unknown

number of samples.

APPENDIX H

ISOLATION AND IDENTIFICATION OF

RADIOACTIVELY LABELLED GLUCOSE1

Principle

 

The sample containing various solutes is passed
through an anion exchange resin removing the anions in the
sample. The glucose, a neutral molecule, in the effluent
is reacted with adenosine triphosphate (ATP) in the pres-
ence of hexokinase (HK) and Mg++ ion at a pH 8.0 forming

g1ucose-6-phosphate (G6P) and adenosine diphosphate (ADP).

HK ++
Glucose-tATP ’ Mg > G6P-+ADP

 

This reaction mixture is passed through another anion
exchange resin and G6P is retained on the resin. This G6P
is eluted with 3N HCl. Any radioactivity in the eluted
sample will be due to radioactively labelled glucose in

the initial sample.

Reagents

1. MgCl -6H 0, A.R.

2
2. (NH4)2SO

2

4, A.R.

 

1From H. J. Horhost (1965) and S. Englard and K. R.
Hanson (1969).

113

114

 

3. HCl, concentrated, A.R.

4. KOH, A.R.

5. Tris-hydroxymethyl-aminomethane (Trizma base;
Sigma Chemical Co.)

6. Adenosine Triphosphate (ATP; Sigma Chemical Co.)
disodium salt, ATP-NazHZ-BHZO. When stored des-
sicated and frozen it is stable for at least one
year.

7. Hexokinase (HK; Sigma Chemical Co.)

Crystalline preparation from yeast which is sus-
pended in 3.2 M (NH4)ZSO4 and stable for at least
one year when stored at 0-4°C. HK concentration is
at least 410 mg protein/ml with a Specific activity
of 2000 units/m1.

8. Dowex l-8x, Anion Exchange Resin (Cl-; Sigma
Chemical Co.)

100-200 mesh; 8% crosslinked.
Solutions

A. MgClz, 0.1 M
Dissolve 2.0 g of reagent 1 in distilled H20;
q.s. to 100 ml.

B. (NH4)ZSO4, 2.1 M

Dissolve 27.8 g of reagent 2 in distilled H20;

q.s. to 100 ml.

115

HCl, 3 N

Add 258 ml of reagent 3 to approximately 500 ml of
distilled H20; q.s. to one liter.

Tris buffer, approximately 0.1 M, pH 8.0

To 1.21 g of reagent 5 add approximately 30 ml of

distilled H 0. Adjust to pH 8.0 with solution C;

2
q.s. to 100 ml with distilled H O. Tris buffer

2
solution is stable for at least 6 months at 0-4°C.
ATP, approximately 0.01 M
Dissolve 30 mg of reagent 6 and dilute to 5 ml with
solution D. ATP dissolved in Tris buffer is not
stable and must be prepared just prior to use.
Hexokinase, HK
Dilute 0.02 ml of reagent 7 to 1.0 ml with solution
B. Diluted HK is stable for at least 6 months at
0-4°C.
KOH, 5%, w/v.
Dissolve 50 g of reagent 4 in distilled H20; q.s.
to one liter.
Dowex-18X (Cl-)

Reagent 8 is washed with distilled H O, removing

2
excess acid, and adjusted to pH 6.8 with solution G.
The resin is stored at 0—4°C as a slurry and must be

rewashed every month.

116

Procedure

 

a.

1.0 m1 of deproteinated sample (dilution factor 9.6,
see Appendix C) is passed through a Dowex-l resin
(solution G) column (0.51:4.0 cm) followed by 1.0 m1
of distilled H20. The total effluent is adjusted to
2 ml with distilled water.

To convert glucose to glucose-6-phosphate in the

2.0 ml of effluent from part a, add 0.04 ml of
solution A, 0.2 ml of solution E, and 0.02 ml of
solution F (total volume==2.26 ml; dilution factor==
21.7). Mix and incubate 10 minutes at room temper-
ature.

Add 2.0 m1 of incubation mixture (part b) to another
Dowex-1 resin (Cl_) column (see part a). Collect
effluent (total volume 2.0 ml; dilution factor==21.7).
Elute glucose-6-phosphate by adding 1.0 m1 of solu-
tion C and collect 1.0 ml of eluate (dilution factor
10.8).

Repeat steps a-d, substituting distilled H20 for
Hexokinase (solution F) in step b. The radioactiv-
ity in this eluate is due to any anions passed
through the resin column in step a and must be
subtracted from the radioactivity measured in the

g1ucose-6-phoSphate eluate (procedure d).

117

Recovery profile of 3H-glucose isolation

 

The recovery of 3H-activity is shown in Table H for
two CSF inflow samples (R1, R2) and for six CSF outflow
samples (2, 3, 4, 6, 7, and 8) in dog Sc-l. Duplicate
0.1 ml aliquots of each fluid sample were counted using
standard liquid scintillation procedures (Appendix J) on:
(1) nondeproteinized sample; (2) following deproteinization;
(3) following conversion to glucose-6-phosphate; (4) efflu-
ent containing cations and neutral compounds passed through
the second Dowex column; and (5) the fluid eluted from the
second column with HCl. The deproteinization procedure
resulted in a 2-6% loss of 3H activity; an additional 8-9%
of 3H activity was retained on the first resin column
(procedure b). Less than 1% of the 3H activity was due
to conversion to cations or nonglucose neutral molecules
(procedure c). Only 1-2% of 3H activity was lost due to
either glucose conversion to anions which passed through
column 1 (procedure a) or retention by column 2 (procedure
e). In dog Sc-l only 75% of the original inflow fluid 3H
activity was due to 3H-glucose; of this 75% there averaged
a 92% recovery in the outflow fluid. The relative lack of
interconversion of isotopically labelled glucose to other
molecules during the perfusion of the brain ventricular

spaces has been reported previously (Bronsted, 1970b) and

is supported by Table H-l. Similar recoveries were obtained

118

in all dogs; the amount of 3H-glucose activity in the
perfusion inflow fluid ranged from 79-84% of the original

3H activity.

. mamEmm quﬂmﬂno c.0coo

n Numboomu w

 

 

119

coaeauouomm coausHHpeamamEmm c.0cou 9

.H8 H.0\Em©m n sowumuucwusoum
m.mo o.mm m.oe m.mo o.me m.mo m.oe e.me sum>oomu a
mmmv anv omnv momv swam Hang omooa mmmm Amldv coaumuusmosou
How mom men mma mom mom nma mma Amy coaumuuco0soo
m.oa Am muspwooumv
mmov voom mmom mmmv mmvm moom namoH omom Adv coaumuucmocou
m.oa A0 musomoonmv

mumsam m QESHOO
m.o 0.0 0.0 H.o o.o m.o e.o 0.5 sumaoumu a
mm o m m m ma mv mo cowumuucoocou
>.Hm A0 ousomooumv

usmsammm N meadow
o.nm m.ow v.mm m.mm m.mm n.0m v.em m.om >u0>000u w
mmmm ommm oaom omOM oHOM ommm oanm @vmm coaumuucmocou
n.am An waspmooumv

ucmsammm H cadaou
m.mm n.vm w.mm N.hm m.nm m.mm m.om m.vm muo>oomu w
swan vmmn How» come Nome mmmn vvmva ovnma coaumuucmocoo
o.m A0 xatcmmmmv

coaumuﬂcﬂmuoummo
Hmmmm ovmmh mmomn vmmme vmawn ommmn monava ovomma H coaumuucmucoo

mHmEMm Hmcﬂmwuo

 

m e o v m m «m m uouomm mmum

 

cowusaﬂa
Hmnadz 0HmEmm

 

 

Aae\ea\aao Hum moo ca

mcﬂsam Am .... .mv onmuso new Amm «Hmv BoncH 050 CH 0m00samlmm mo >H0>000u 0cm cOHUMHOmH .Hum magma

APPENDIX I

ISOLATION OF LABELLED LACTATE1

Principle

 

Ion-exchange chromatography utilizes the
differential affinity of charged molecules in solution for
inert, immobile oppositely charged substances on the resin
bed. L-lactate, an anion, is bound to an anion (C1_)
exchange resin. Elution of lactate from the resin bed is
accomplished by using a progressively more acid (HCl) eluent.
The C1- of the acid displaces the weaker anions (such as

lactate) bound to the resin bed.

Reagents
1. HCl, concentrated, A.R.

2. Dowex 1-8X Anion Exchange Resin (Cl-; Sigma
Chemical Co.)
100-200 mesh; 8% crosslinked.

3. KOH, A.R.

 

1From R. w. VonKorf (1969).

120

121

Solutions

 

A. KOH, 5%, w/v.
Dissolve 50 g of reagent 3 in distilled H20;
q.s. to one liter.

B. _HC1, 0.01 N

Add 0.8 m1 of reagent 1 to distilled H20;
q.s. to one liter.

C. HCl, 0.05 N
Add 4.2 ml of reagent 1 to distilled H20;
q.s. to one liter.

D. HCl, 0.1 N
Add 8.4 m1 of reagent 1 to distilled H20;

q.s. to one liter.

Preparation of resin column

 

Reagent 2 is washed with distilled H20 to remove
impurities and acid. After washings reach approximately
pH 4.5, the resin suspension in water is adjusted to pH 6.8
with solution A and stored as a slurry at 0-4°C.

A resin column (1 X 17 cm) is formed by pouring
water-suspended resin into a 50 ml burette. Packing of
the column is aided if the tip of the burette is attached
to an aspirator. The packed resin column should have about

2 mm of water remaining above the resin at all times.

122

Procedure

 

a.

A 2 m1 aliquot of the deproteinated sample (pH 6.8-
7.2; Appendix C) is added to the resin column and
allowed to drain into the resin bed. The burette
stOpcock is closed until sample is to be eluted.
Begin elution with distilled H20 via gravity flow
from an eluant flask into a reservoir flask (con-
tinuously stirred). Fluid from reservoir flask
passes via siphon to the resin column. Elute and
collect ten 5 ml fractions using an automatic
fraction collector (Model 1205 DE, Warner-Chilcott
Lab., Richmond, Cal.).

After 10 fractions have been collected, empty eluant
flask, add 0.01 N HCl (solution B) and continue
elution. Collect ten 5 ml fractions.

Again empty eluant flask, refill with solution C
and collect twenty 5 m1 fractions.

Repeat step d refilling with solution D and
collecting twenty 5 m1 fractions.

Radioactivity of labelled material in each fraction
is counted using 0.1 ml duplicate aliquots from
each 5 m1 sample tube (see Appendix J). The eluted
fractions containing L-lactate (lactate peak) are

identified by chemical methods (Appendix F).

123

L-lactate elution profile

 

The results from eluting a single CSF sample (vial
8; dog Sc-l) are shown in Figure I-l. There is a single
peak corresponding to 3H activity (fractions 2-6) and a

14

single peak corresponding to C activity (fractions 29-33).

3H-glucose, a neutral molecule, is eluted first since it is
not bound by the anion exchange resin. The total 14C activ-
ity (dpm) in fractions 29-33 (Table I-l) corresponds to 95%
of the total 14C activity in the sample. The specific
activity (dpm/mg) (Table I-l) is the same in fractions
30-32 as in the initial CSF sample (collected after 4 hours
of ventricular perfusion) suggesting that 14C activity
remains on L-lactate during perfusion. Similar recoveries

and Specific activities were obtained from single samples

for seven dogs.

124

 

1.4.. l.‘ I,II lull. Ill

.samaauomemmn .Hom z H.e the .z mo.o
.z Ho.o on 0mm omaaﬂpmﬂp Eoum Ummcmno mmz ucmsam cobz 00m0ﬂpcﬂ Aunmﬂn
on puma Eouwv m3ouu< .pﬂsam Damsam mo HE OOH omcﬂmucoo nonsmno ocﬂxﬂz
mumDUMHIA dime I
eat o a

va
EQU m 0

m
.AHE my nomads 0390
coﬂu0muw mnu ou mpcommouuoo mmmﬂomnm 0:9 .AHE\mEV mumuomauq mo soap
Imuus00co0 may on Osman on“ so myocapuo 0:0 “Amoa x Empv muﬂ>ﬂuomoﬂpmu

ou mucommmuuoo umoa on» so mumcﬁpuo 0:9 .mpmbomHIUva cam whoosHmnmm

mcﬂcﬂmucoo AHIOm moo am HmH>v 0Hm80m 30Hmuso mmo mo 0HHMOHQ coﬂusam

.HIH mudmﬂm

HIH musmﬁh
a one.

- e... 9. on
88

 

 

s. +

‘-‘

 

125

_E\mE

 

 

 

 

% o—

126

 

 

 

 

 

0 000.0 om 00H N o.m mm
ooo.aov o~o.o omo.m ooa.oe mom o.m mm
mov.HHv ~eo.o oem.oa omm.em emm.a o.m Hm
emo.odv mmo.o oem.m ome.mv emm o.m om
0 000.0 oe omm a o.m mm
* :oHuomum
Hmm.moe mam.o omm.Hm oom.~ma mmH.m o.m m Hma>
aa2\smtv He\me He\eme ent Hence as H.o\sme lass mamsmm
>0ﬁ>wuom oﬂma0mmm mEsHo>
>0H>Huom 0
ea
“He\eH\HH
«HI0m mom amw H0ﬂ>v usmsammm coﬁmsmumm Hoasoﬂuucm> CH mumuowalo mo GOADMHOmH .HIH manna

Va

APPENDIX J

LIQUID SCINTILLATION COUNTING1

Principle

 

Liquid scintillation counting is a method of
detecting radioactivity. The radioactive substance is
dissolved in or completely wetted by the scintillation
solution. The scintillation solution converts the energy
of the primary particle emitted by the radioactive sample
to light energy. The light quanta entering the multiplier
phototube are amplified and counted by a scaling circuit.

The widest application of liquid scintillation
counting has been in the counting of low energy beta
emitters such as 3H and 14C. In any radiation detection
method, the counting efficiency (ratio of observed counting
rate to the actual rate of radioactive disintegrations in
the sample) is greatest when the maximum number of emitted
particles reaches and interacts with the multiplier photo-
tube. Absorption along their paths within the sample and

between the sample and the detector is most severe for low

energy beta emissions; such losses are reduced and counting

 

1From Instruction Manual, Mark I liquid scintilla-
tion computer, Model 6860, Nuclear-Chicago Corp., Des
Plaines, I11.

127

128

efficiency is increased by dissolving the radioactive sample
directly in the scintillation solution.

Scintillation counting is a proportional counting
method, i.e., the magnitude of the output signal from the
detector is proportional to the energy given up to the
detector by the primary particle. This signal is amplified
and fed into a pulse height analyzer which compares the
signal to reference voltages. A discriminator circuit
passes the signal if it falls between two selected voltage
levels. Counting between two finite discriminator levels
is referred to as differential counting.

The energy spectra of 3H and 14C overlap, and when
both beta emitters are present in the same sample they are
counted simultaneously on two separate analyzer channels.
Channel A amplifiers were adjusted to give high efficiency
of counting 3H with minimal interference of 14C; the
amplifiers of channel C were set so that counting efficiency

14

of 3H was insignificant while efficiency for C was maximal.

Channel B amplifiers were set to maximize energy pulses from

an external standard of known disintegration rate (133Ba).

Quench"correction curve

Any nonfluorescent solute or solvent will absorb or
quench energy emitted from the primary particle and reduce
the efficiency of counting the radioactivity. A set of

3H and 14C standards (Nuclear-Chicago Corp.) containing

129

known amounts of isotope (492,000 dpm and 255,000 dpm,
respectively) and varying degrees of quenching for each
isotope over the range used in these experiments were used

to determine:

1. 3H counting efficiency in channel A
2. 14C counting efficiency in channel A
3. 14C counting efficiency in channel C.

133B

The external standard, a, is used to determine the

amount of quenching present in each sample. The channels

133

ratio relates the net Ba count rate in the channel (B)

133 133

(i.e., preset to maximum Ba efficiency) to the net Ba

count rate in channel A. Since the sample will quench
energy emitted by the gamma source, 133Ba, the channels
ratio (B/A==(cpm) channel B/(Cpm) channel A) increases as
the amount of quenching decreases.

The quench correction curve (Figure J-l) is a plot

3H and 14C standards as a function

of the efficiencies of
of the standards' channels ratio (B/A). Counting efficiency
of each isotope in an unknown sample can be read directly

from the quench correction curve corresponding to the

unknown sample's channels ratio (B/A).

Calculation of disintegration rates of 3H and 14C

 

The disintegration rates of unknown samples can be
calculated from the net count rate of 3H and 14C and the
efficiencies of counting each isotope in channels A and C

from the following equations:

where:

130

_ N1C2

-N

201

 

l l

disintegration rate
disintegration rate
count rate (Cpm) in
count rate (cpm) in
counting efficiency
counting efficiency

counting efficiency

(dpm) of
(dpm) of
channel A
channel C

(%) for 3

(%) for 14

(%) for

14

3H
14

C

H in channel A

C in channel A

C in channel C.

Figure J-l.

131

133Barium external standard quench correction

curves for differential counting of 3H and 14C
samples. Efficiencies are calculated as net
Cpm on scaler A or C divided by dpm of 3H or
14C standards, respectively; multiplied by 100.
14C quench correction curve for 14C efficiency
in channel C (squares); 3H quench correction
curve for 3H efficiency in channel A (circles);

14 14

C quench correction curve for C efficiency

in channel A (triangles) are plotted as a

 

function of channels ratio (B/A) of 1338a.
Discriminator settings:
Channel Window Attenuation
A 0.0-2.5 B524
B 0.0-9.9 ' F670

C l.5-9.9 E738

ﬁ—ﬂ—E“ ———.—__

 

‘—¥__ﬁ___—_—_—_—__,-.__*_

132

50 x EFFICIENCY

 

APPENDIX K

STATISTICAL FORMULAE

Grand Mean: XI: %§
where:
Zx'= sum of individual means

n = number of individual means.

Standard error of grand mean (SE).

 

 

_ 2372 — (YEW/n
SE " / n (n-I)

Split plot design (Steel and Torrie, 1960)

 

The split plot analysis of variance (AOV) was
employed since the experimental design incorporated the
inclusion of two variable factors: perfusion time and
metabolite concentration. The statistical formulae employed
in the calculation of the sums of squares (SS) and the eval-
uation of the F statistics (Table K-2) are given below and
illustrated from data in Table K-l.

Xijk is the kth observation on the ith concentration
in the jth period, and X... is the sum of all observations

at all concentrations in all periods. Then nij==number of

observations at the ith concentration; jth period.

133

(I)

(II)

(III)

(IV)

(V)

(VI)

134

j=2 i=3
2 ni. n = Z ni. n = Z ni.
j=l 3 3 i=1 3 ij 3
Total sum of squares (SS):=ZXijk"C=:lOOZ

+ooo+902-C

 

 

 

 

=43,019.37
2
Where c==—§—3LL = 31972/30==340,693.63
Xi" 21524- +3522
Concentration 85:: - C = ‘°’
ni 12
+...-C
= 21,598.08
Dogs within Concentration SS = 2x: k-—C-—(III)
2+ 2+ 2
= 1673 629 855 ._C._(III)
= 20,790.79
£xg" 15962 16012
Period 88 = ———l—-- c = —————u+-————-c
n. 15 15
J
= 0.87
2x?..
Interaction SS = Tril—l- C - (III) - (V)
ij
5 2
= 825 2848 +...-c- (III)- (V)

73.24

135

(VII) Error b SS (1) - (II) - (IV) - (V) - (VI)

1135.10

Analysis of mean differences employed Student's t distribu-

tion and was calculated by the following formulae.

Test statistic t = 5L
s s—
d
where:
3': mean difference
53-: standard error of the difference.

A. If there was no significant interaction between
concentration and test periods.
1. To compare the difference between individual

period means (j).

2(x.1.-X.2-) = (loo-115)'t...-+(175-177):=_
nij . 6

0|
u

7

 

 

 

, 5
_ - .
S /2 (Error b)/n 12 13.8

 

Q.-

2. To compare the difference between grand means of any

two concentrations within a period.

136

X...-X!.. = 825-335 = 490
13 13

n.. +113.
Error a —1-1——.il = 20,790.79 ”“3 = 65.8
d nijnij 96

0|
n

 

 

 

m
I
II

If there was a significant interaction between
concentration and the test period. (This situation was
not evident in the present example, nor was it evident
for any of the other molecules analyzed in this manner.)
To analyze differences between two j means at the same

concentration (i), or between i means in the same

 

period.
ni.-+ni.
SE): Bb-l) Error b-tError a] -—J———11-
n..n..
1] 13
where

number of different concentrations.

0'
ll

137

Table K-l. Statistic block of glucose KO data from 15 normoxic
anesthetized dogs at different inflow concentrations (A, B,
and C) during two experimental brain ventricular perfusion

 

 

 

 

 

 

 

 

periods
Period (j)
Concentration Concentration
(1) xilk xi2k Total ni
100 115 215
177 161 338
126 139 265
xljk (A) 145 157 302 12
102 99 201
175 177 352
Period total 825 848 1673
80 83 163
83 78 161
x2jk (B) 86 91 177 8
86 82 168
Period total 335 334 669
146 145 291
64 69 133
x3.k (C) 81 76 157 10
3 57 39 96
88 90 178
Period total 436 419 855
Grand total 1596 1601 3197
n 15 15 n=30

 

138

Table K-2. AOV table

 

 

 

Source df SS MS F
Concentration (III) 2 21,598.08 10,799.04 6.23*
Error a (IV) 12 20,790.79 1,732.56

Period (V) l 0.87 0.87 <1”0
Interaction (VI) 2 73.24 36.62 <1J0
Error b (VII) 12 1,135.10 94.59

 

*p < 0.05.

df

MS

Test statistic F (AOV)
To test variances between concentrations:

F = MS /MS

III

degrees of freedom

mean square = SS/df

To test variances between periods:

F = MSV/MSV

To test interaction between concentrations and periods:

= M
F SVI/MS

 

LI ST OF REFERENCES

L IST OF REFERENCES

Alexander, S. C., R. D. Workman, and J. C. Lambertsen. 1962.
Hyperthermia, lactic acid infusion, and composition of
arterial blood and cerebrospinal fluid. Amer. J.
Physiol. 202:1049-1054.

Aldridge, W. N. 1962. Enzynes and Drug Action. CIBA Found.
Sympos., London, 1961. ed. by A.V.S. DeReuck. Boston:
Little, Brown & Co., pp. 155-169.

Ames, A., K. Higaski, and F. B. Nesbett. 1965a. Relation
of potassium concentration in choroid plexus fluid to
that in plasma. J. Physiol. 181:506-515.

Ames, A., K. Higashi, and F. B. Nesbett. 1965b. Effects of
PC02, acetazolamide and ouabain on volume and composition
of choroid plexus fluid. J. Physiol. 181:516-524.

Arey, L. B. 1962. Developmental Anatomy. 6th ed.
Philadelphia: W. B. Saunders Co., pp. 472-506.

Atkinson, A. J., Jr., and M. F. Weiss. 1970. Kinetics of
blood-cerebrospinal fluid glucose transfer in the normal
dog. Amer. J. Physiol. 216:1120-1126.

Baethmann, A., U. Steude, S. Horsch, and W. Brendel. 1970.
The thiosulphate (35$) space in the CNS of rats after
ventriculo-cisternal perfusion. Pflugers Arch 316:51-63.

Baky, L., and J. C. Lee. 1968. The effect of acute hypoxia
and hypercapnia on the ultrastructure of the central
nervous system. Brain 91:697-706.

Barlow, C. F., M. S. Domek, M. A. Goldberg, and L. J. Roth.
1961. Extracellular brain space measured by 35$-
sulphate. Arch. Neurol. 5:102-110.

Bergmeyer, H. U., and E. Bernt. 1965. Methods of Enzymatic

Analysis. ed- by H. U. Bergmeyer. New York: Academic
Press, pp. 123-130.

139

140

Bering, E. A. 1959. CerebrOSpinal fluid production and its
relationship to cerebral metabolism and cerebral blood
flow. Amer. J. Physiol. 197:825-828.

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

Bidder, T. G. 1968. Hexose translocation across the blood
brain interface: configurational aspects. J. Neurochem.
15:867-874.

Bierer, D. W. 1972. Active transport of organic anions
from the brain ventricles of the dog. (Master's thesis),
Michigan State University. East Lansing, Michigan.

Bito, L. Z., and H. Davson. 1966. Local variations in
cerebrospinal fluid composition and its relationship
to the composition of the extracellular fluid of the
cortex. Exp. Neurol. 14:264-280.

Bondareff, W., R. E. Myers, and A. W. Brann. 1970. Brain
extracellular space in monkey fetuses subjected to
prolonged partial asphyxia. Exp. Neurol. 28:167-178.

Bradbury, M. W. B., and H. Davson. 1964. The transport of
urea, creatinine, and certain monosaccharides between
blood and fluid perfusing the cerebral ventricular
system of rabbits. J. Physiol. 170:195-211.

Bronsted, H. E. 1970a. Ouabain-sensitive carrier-mediated
transport of glucose from the cerebral ventricles and
surrounding tissues in the cat. J. Physiol. 208:187-201.

Bronsted, H. E. 1970b. Transport of glucose, sodium,
chloride, and potassium between the cerebral ventricle
and surrounding tissues in cats. Acta Physiol. Scand.
79:523-532.

Brooks, G. F., A. Koch, and J. Wong. 1970. Kinetics of
sodium transfer from blood to brain of the dog. Amer.
J. Physiol. 218:693-702.

Bucher, T., R. Csok, W. Lamprecht, and E. Latzo. 1965.
Methods of Enzymatic Analysis. ed. by H. U. Bergmeyer.
New York: Academic Press, pp. 253-259.

Buschiazzo, P. M., E. B. Terrell, and D. M. Regen. 1970.
Sugar transport across the blood-brain barrier. Amer.
J. Physiol. 219:1505-1513.

141

Cain, S. 1965. Appearance of excess lactate in anesthetized
dogs during anemic and hypoxic hypoxia. Amer. J.
Physiol. 209:604-610.

Crone, C. 1965. Facilitated transport of glucose from
blood into brain tissue. J. Physiol. 181:103-113.

Csaky, T. Z., and B. M. Rigor. 1968. The choroid plexus
as a glucose barrier. Prog. in Brain Res. 29:147-154.

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

Cutler, R. W. P., R. J. Robinson, and A. V. Lorenzo. 1968.
Cerebrospinal fluid transport of sulfate in the cat.
Amer. J. Physiol. 214:448-454.

Dahl, N. A., and W. M. Balfour. 1964. Prolonged anoxic
survival due to anoxia pre-exposure: brain ATP,
lactate, and pyruvate. Amer. J. Physiol. 207:452-456.

Dandy, W. E. 1919. EXperimental hydrocephalus. Ann. Surg.
70:129-142.

Dandy, W. E., and K. D. Blackfan. 1914. Internal hydro-
cephalus. An experimental clinical and pathological
study. Amer. J. Dis. Child. 8:406-482.

Davson, H. 1967. Physiology of Cerebrospinal Fluid.
lst ed. London: J. & A. Churchill Ltd.

Davson, H., and J. F. Danielli. 1952. The Permeability of
Natural Membranes. 2nd ed. Cambridge, England:
Cambridge University Press.

Davson, H., C. R. Kleeman, and E. Levin. 1961. Blood-brain
barrier and extracellular space. J. Physiol. 159:67-86.

Davson, H., C. R. Kleeman, and E. Levin. 1962. Quantitative
studies of the passage of different substances out of the
cerebrospinal fluid. J. Physiol. 161:126-142.

Davson, H., and M. Pollay. 1963. Influence of various
drugs on the transport of 1311 and PAH across the cere-
brospinal fluid-blood barrier. J. Physiol. 167:239-246.

Davson, H., and M. B. Segal. 1970. The effect of some
inhibitors and accelerators of sodium transport on the
turnover of 22Na in the cerebrospinal fluid and the
brain. J. Physiol. 209:131-153.

142

Davson, H., and E. Spaziani. 1959. The blood-brain barrier
and the extracellular space of brain. J. Physiol. 149:
135-143.

Dunn, R. B., and F. Lioy. 1971. Lactate and pyruvate
levels in brain and skeletal muscle during hyperthermia
in dogs. Can. J. Physiol. Pharmacol. 49:520-524.

Ellis, J. P., S. M. Cain, and E. W. Williams. 1963. Rapid
accurate analysis of blood lactate. Tech. Doc. Report
No. SAM-TDR-63-49. USAF School of Aerospace Medicine,
Aerospace Medical Div. (AFSC) Brooks Air Force Base,
Texas.

Englard, S., and K. R. Hanson. Methods in Enzymology. XIII.
ed. by J. W. Lowenstein. New York: Academic Press,

Eich, J., and K. Wiemers. 1959. Cited by Davson, 1967.

Ferguson, R. K., and D. M. Woodbury. 1969. Penetration of
C-inulin and 14C-sucrose into brain, cerebrospinal
fluid, and skeletal muscle of developing rats. Exp.
Brain Res. 7:181-194.

Fishman, R. A. 1964. Carrier transport of glucose between
blood and cerebrospinal fluid. Amer. J. Physiol. 206:
836-844.

Fleischhauer, K. 1957. Untersuchungen am Ependym des
Zwischen-und Mittel-hirns der Landschildkrote. Z.
Zellforsch. 46:729-769.

Flexner, L. B. 1933. A note on the rate of circulation of
cerebrospinal fluid. Amer. J. Physiol. 106:201-203.

Folbergrova, J., O. H. Lowry, and J. Passoneau. 1970.
Changes in metabolites of the energy reserves in
individual layers of mouse cerebral cortex and
subjacent white matter during ischaemia and anaesthesia.
J. Neurochem. 17:1155-1162.

Forbes, H. S. 1928. Observation and measurement of pial
vessels. Arch. Neurol. Psychiat. 19:751-761.

Frazier, C. H., and M. M. Peet. 1914. Factors of influence
in the origin and circulation of the cerebrospinal fluid.
Amer. J. Physiol. 35:268-282.

143

Friedman, S. B., W. G. Austin, R. E. Rieselbach, J. B. Block,
and D. P. Rall. 1963. Effect of hypochloremia on
cerebrospinal fluid chloride concentration in a patient
with anorexia nervosa and in dogs. Proc. Soc. Exp.

Biol. N. Y. 114:801-805.

Fujishima, M., P. Scheinberg, and R. Busto. 1971.
Cerebral cortical blood flow. Variable effects of
hypoxia in the dog. Arch. Neurol. 25:160-167.

Gatfield, D. P., O. H. Lowry, D. W. Schulz, and J.
Passonneau. 1966. Regional energy reserves in mouse
brain and changes with ischaemia and anesthesia. J.
Neurochem. 13:185-195.

Geiger, A. 1958. Correlation of brain metabolism and
function by use of a brain perfusion method in situ.
Physiol. Reviews 38:1-20.

Geiger, A., J. Magnes, R. M. Taylor, and M. Veralli. 1954.
Effect of blood constituents on uptake of glucose and
metabolic rate in brain perfusion experiments. Amer.
J. Physiol. 177:138-149.

Gibbs, E. L., W. G. Lennox, L. F. Nims, and F. A. Gibbs.
1942. Arterial and cerebral venous blood. Arterial-
venous differences in man. J. Biol. Chem. 144:325-332.

Gilboe, D. D., R. L. Andrews, and G. Dardenne. 1970.
Factors affecting glucose uptake by the isolated dog
brain. Amer. J. Physiol. 219:767-773.

Goodale, R. L., B. Goetzman, and M. B. Visscher. 1970.
Hypoxia and iodoacetic acid and alveolocapillary

barrier permeability to albumin. Amer. J. Physiol.
219:1226-1230.

Granholm, L., A. E. Kassik, L. Nilsson, and B. K. Siesjo.
1968. The lactate/pyruvate ratios of cerebrospinal
fluid of rats and cats related to the lactate/pyruvate,
the ATP/ADP, and the phosphocreatine/creatine ratios of
brain tissue. Acta Physiol. Scand. 74:398-409.

Granholm, L., and B. K. Siesjo. 1967. Lactate and pyruvate
concentrations in blood, cerebrospinal fluid, and brain
tissue of the cat. Acta Physiol. Scand. 70:255-256.

Granholm, L., and B. K. Siesjo. 1969. The effects of
hypercapnia and hypocapnia upon the cerebrospinal fluid
lactate and pyruvate concentrations and upon the lactate,
pyruvate, ATP, ADP, phosphocreatine, and creatine con-
centrations of cat brain tissue. Acta Physiol Scand.
75:257-266.

144

Gurdjian, E. S., W. E. Stone, and J. E. Webster. 1944.
Cerebral metabolism in hypoxia. Arch. Neurol. Psychiat.
51:472-477.

Hasegawa, T., J. R. Ravens, and J. F. Toole. 1967. Pre-
capillary arteriovenous anastomoses "thoroughfare
channels" in the brain. Arch. Neurol. 16:217-224.

Hassin, G. B. 1948. The morphology of the pial blood
vessels and its bearing on the formation and absorption
of the cerebrospinal fluid. J. Neuropath. 7:432-438.

Heisey, S. R., D. Held, and J. R. Pappenheimer. 1962.
Bulk flow and diffusion in the cerebrospinal fluid
system of the goat. Amer. J. Physiol. 203:775-781.

Hodgkin, A. L., and R. D. Keynes. 1955. Active transport
of cations in giant axons from Sepia and Loligo. J.
Physiol. 128:28-60.

Holloway, L. S., S. Cassin, and I. Weinstein. 1972. The
effect of hypoxia, acetazolamide and ouabain on CSF
production rate in the newborn dog. Fed. Proc. 31:223.

Horhost, H. J. 1965. Methods of Enzymatic Analysis.
ed by H. U. Bergmeyer. New York: Academic Press,
pp. 117-123.

Horstman, E. 1954. Die Faserglia des Selachiergehins.
Z. Zellforsch. 39:588-617.

Huckabee, W. E. 1958a. Relations of pyruvate and lactate
during anaerobic metabolism. 1. Effects of infusion of
pyruvate or glucose and of hyperventilation. J. Clin.
Invest. 37:244-254.

Huckabee, W. E. 1958b. Relationships of pyruvate and
lactate during anaerobic metabolism. III. Effect of
breathing low oxygen gases. J. Clin. Invest, 37:264-271.

Huggett, A. G., and D. A. Nixon. 1957. Use of glucose
oxidase, peroxidase, and o-dianisidine in determination
of blood and urinary glucose. Lancet 273:368-370.

Kaasik, A. E., L. Nilsson, and B. K. Siesjo. 1970. The
effect of asphyxia upon lactate, pyruvate, and bicar-
bonate concentrations of brain tissue and cisternal CSF,
and upon the tissue concentrations of phosphocreatine
and adenine nucleotides in anesthetized rats. Acta
Physiol. Scand. 78:433-477.

145

Kemeny, A., H. Boldizsar, and G. Pethes. 1961. The
distributions of cations in plasma and cerebrospinal
fluid following infusion of solutions of salts of sodium,
potassium, magnesium, and calcium. J. Neurochem. 7:218-
227.

Kerr, S. E., and M. Ghantus. 1936. The carbohydrate
metabolism of brain. II. The effect of varying the
carbohydrate and insulin supply on the glycogen, free

sugar, and lactic acid in mammalian brain. J. Biol.
Chem. 116:9-20.

Kety, S. S. 1950. Circulation and metabolism of the human
brain in health and disease. Amer. J. Med. 8:205-217.

Kety, S. S., and C. F. Schmidt. 1948. The effects of
altered arterial tensions of carbon dioxide and oxygen
on cerebral blood flow and cerebral oxygen consumption
of normal young men. J. Clin. Invest. 27:484-492.

Klein, J. R., and N. G. Olsen. 1947. Distribution of
intravenously injected glutamate, lactate, pyruvate,
and succinate between blood and brain. J. Biol. Chem.
167:1-5.

LeFevre, P. G., and A. A. Peters. 1966. Evidence of
mediated transfer of monosaccharides from blood to
brain in rodents. J. Neurochem. 13:35-46.

Lending, M., L. B. Slobody, and J. Mestern. 1961. Effect
of hyperoxia, hypercapnia, and hypoxia on blood-

cerebrospinal fluid barrier. Amer. J. Physiol.
200:959-962.

Leusen, I. 1948. Influences des modifications Simultanées
de la concentration en ions K et Ca du liquide
cephalorachidien sur 1e systéme vasomoteur. Experientia
4:154-155.

Levin, V. A., J. D. Fenstermacher, and C. S. Patlak. 1970.
Sucrose and inulin Space measurements of cerebral cortex
in four mammalian Species. Amer. J. Physiol. 219:1528-
1533.

Lorenzo, A. V., J. P. Hammerstad, and R. W. P. Cutler. 1968.
The effect of anaesthetic agents on the cerebrospinal
fluid clearance of 35S-sulfate and 125I-iodide. Biochem.
Pharmacol. 17:1279-1283.

146

Lowry, O. H., J. V. Passonneau, F. X. Hasselberger, and
D. W. Schultz. 1964. Effect of ischemia on known
substrates and cofactors of the glycolytic pathway
in brain. J. Biol. Chem. 239:18-30.

Macri, F., A. Politoff, R. Rubin, R. Dixon, and D. Rall.
1966. Preferential vasoconstriction properties of
acetazolamide on the arteries of the choroid plexus.
Int. J. Neuropharmacol. 5:109-115.

Maddock, S., J. E. Hawkins, and E. Holmes. 1939. The
adequacy of substances of the glucose cycle for the
maintenance of normal cortical potentials during
hypoglycemia produced by hepatectomy with abdominal
evisceration. Amer. J. Physiol. 125:551-565.

Mayman, C. I., P. D. Gatfield, and B. McL. Breckenridge.
1964. The glucose content of the brain in anesthesia.
J. Neurochem. 11:483-487.

Merlis, J. K. 1940. The effect of changes in the calcium
content of the cerebrospinal fluid on spinal reflex
activity in the dog. Amer. J. Physiol. 131:67-72.

Michael, D. K., D. W. Bierer, and S. R. Heisey. 1971.
Hypoxia-induced changes in clearance from cerebrospinal
fluid. Fed. Proc. 30:215.

Michael, D. K., D. W. Bierer, and S. R. Heisey. 1972.
Effects of anesthesia and hypoxia on cerebrospinal fluid
production in the adult dog. Fed. Proc. 31:395.

Millen, J. W., and D. H. M. Woollman. 1961. On the nature
of the pia mater. Brain 84:514-520.

Milne, M. D., B. H. Scribner, and M. A. Crawford. 1958.
Non-ionic diffusion and the excretion of weak acids
and bases. Amer. J. Med. 24:709-725.

Myers, R. E., R. Beard, and K. Adamsons. 1969. Brain
swelling in the newborn rhesus monkey following pro-
longed partial asphyxia. Neurol. 19:1012-1018.

McGinty, D. A. 1929. The regulation of respiration. XXV.
Variations in the lactic acid metabolism in the intact
brain. Amer. J. Physiol. 88:312-325.

Olsen, N. S., and G. G. Rudolph. 1955. Transfer of sodium
and bromide ions between blood, cerebrospinal fluid, and
brain tissue. Amer. J. Physiol. 183:427-432.

147

Oppelt, W. W., T. H. Maren, E. S. Owens, and D. P. Rall.
1963a. Effects of acid-base alterations on cerebro-
spinal fluid production. Proc. Soc. Exp. Biol. N. Y.
114:86-89.

Oppelt, W. W., E. S. Owens, and D. P. Rall. 1963b.
Calcium exchange between blood and cerebrospinal fluid.
Life Sci. 8:599-605.

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

aP‘

Pappenheimer, J. R., V. Fencl, S. R. Heisey, and D. Held.
1965. Role of cerebral fluids in control of respiration
as studied in unanesthetized goats. Amer. J. Physiol.
208:436-450.

 

Pappenheimer, J. R., S. R. Heisey, and E. F. Jordan. 1961.
Active transport of diodrast and phenosulfonphthalein
from cerebrospinal fluid to blood. Amer. J. Physiol.
200:1-10.

Plum, F., and J. B. Posner. 1967. Blood and cerebrospinal
fluid lactate during hyperventilation. Amer. J. Physiol.
212:864-870.

Plum, F., J. B. Posner, and W. W. Smith. 1968. Effect of
hyperbaric-hyperoxic hyperventilation on blood, brain,
and CSF lactate. Amer. J. Physiol. 215:1240-1244.

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

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

Prock0p, L. D. 1968. Cerebrospinal fluid lactic acid:
Clearance and effect on facilitated diffusion of a
glucose analogue. Neurol. 18:189-196.

Raichle, M. E., J. B. Posner, and F. Plum. 1970. Cerebral
blood flow during and after hyperventilation. Arch.
Neurol. 23:394-403.

Rall, D. P., W. W. Oppelt, and C. S. Patlak. 1962.
Extracellular Space of brain as determined by diffusion

of inulin from the ventricular system. Life Sci. 2:43-
48.

148

Reed, D. J., and D. M. Woodbury. 1963. Kinetics of
movement of iodine, sucrose, inulin, and radio-
iodinated serum albumin in the central nervous system
and cerebrospinal fluid in the rat. J. Physiol. 169:
816-850.

Rossen, R., H. Kabat, and J. P. Anderson. 1943. Acute
arrest of cerebral circulation in man. A. M. A. Arch.
Neurol. & Psychiat. 50:510-528.

Rothman, A. R., E. J. Freireich, J. R. Gaskins, C. S. Patlak,
and D. P. Rall. 1961. Exchange of inulin and dextran
between blood and cerebrospinal fluid. Amer. J.

Physiol. 201:1145-1148.

Shapiro, W., A. J. Wasserman, J. P. Baker, and J. L.
Patterson, Jr. 1970. Cerebrovascular response to
acute hypocapnic and eucapnic hypoxia in normal man.
J. Clin. Invest. 49:2362-2368.

Shimojyo, S., P. Scheinberg, K. Kogure, and O. M. Reinmuth.
1968. The effects of graded hypoxia upon transient
cerebral blood flow and oxygen consumption. Neurol.
18:127-133.

Siesjo, B. K., L. Granholm, and A. Kjallquist. 1968.
Regulation of lactate and pyruvate levels in the
cerebrospinal fluid. Scand. J. Clin. Lab. Invest.
22:Suppl. 102:1:F.

Slobody, L. B., D. C. Yang, M. Lending, F. J. Borrelli,
and M. Tyree. 1957. Effect of severe hypoxia on blood-
cerebrospinal fluid barrier. Amer. J. Physiol. 190:365-
370.

Smith, H. W. 1956. Principles of Renal Physiology. New
York: Oxford University Press, p. 206.

Sokoloff, L. Handbook of Physiology, III, ed. by J. A.
Field. Chap. 77. Washington, D.C.: American
Physiological Society, pp. 1841-1864.

Steel, R. G. D., and J. H. Torrie. 1960. Principles and
Procedures of Statistics. New York: McGraw-Hill.

Stein, W. D. 1967. The Movement of Molecules Across Cell
Membranes. VI. New York: Academic Press.

Truex, R. C., and M. B. Carpenter. 1969. Human Neuroanatomy.
6th ed. Baltimore: Williams and Wilkens Co.

149

Valenca, L. M., D. C. Shannon, and H. Kazemi. 1971.
Clearance of lactate from the cerebrospinal fluid.
Neurol. 21:615-620.

VanHarreveld, A., J. Cromwell, and S. K. Malhotra. 1965.
A study of extracellular space in central nervous
tissue by freeze substitution. J. Cell. Biol. 25:
117-137.

Vates, T. S., S. L. Bonting, and W. W. Oppelt. 1964.
Na-K activated adenosine triphosphatase and formation

of cerebrospinal fluid in the cat. Amer. J. Physiol.
206:1165-1172.

VonKorff, R. W. 1969. Methods in Enzymology. XIII.
ed. by J. W. Lowenstein. New York: Academic Press,
pp. 425-430.

Wallace, G. B., and B. B. Brodie. 1937. The distribution
of iodide, thiocyanate, bromide, and chloride in the
central nervous system and spinal fluid. J. Pharmacol.
& EXper. Therap. 61:397-411.

Wallace, G. B., and B. B. Brodie. 1939. The distribution
of administered bromide in comparison with chloride and
its relations to body fluid. J. Pharmacol & Exper.
Therap. 65:214-219.

Weed, L. H. 1923. The absorption of cerebrospinal fluid
into the venous system. Amer. J. Anat. 31:191-221.

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

Welch, K., and V. Friedman. 1960. The cerebrospinal fluid
valves. Brain 83:454-468.

Welch, K., K. Sadler, and R. Hendee. 1970. Cooperative
phenomena in the permeation of sugars through the lining
epithelium of choroid plexus. Brain Res. 19:465-482.

White, A., P. Handler, and E. L. Smith. 1964. Principles
of Biochemistry. 3rd ed. New York: McGraw-Hill.

Williamson, J. R., and B. E. Corey. 1969. Methods of
Enzymology. XIII. ed. by J. W. Lowenstein. New York:
Academic Press, pp. 435-444.

150

Wolff, P. H., and R. D. Tschirgi. 1956. Inability of
cerebrospinal fluid to nourish the spinal cord.
Amer. J. Physiol. 184:220-222.

Woodward, D. L., D. J. Reed, and D. M. Woodbury. 1967.
Extracellular space of rate cerebral cortex. Amer.
J. Physiol. 212:367-370.