DYNAMICS OF

ARTERIAL FLOW
ADV ANCES IN EXPERIMENTAL MEDICINE AND BIOLOGY
Editorial Board:
Nathan Back State University 01 New York at Buffalo
N. R. Di Luzio Tulane University School 01 Medicine
Bernard Halpern College de France and Institute ollmmuno·Biology
Ephraim Katchalski The Wei::mann Institute 01 Science
David Kritchevsky Wistar Institute
Abel Lajtha New York State Research Institute lor Neurochemistry and Drug Addiction
Rodolfo Paoletti University 01 Milan

Recent Volumes in this Series

Volume 105
NUTRITIONAL IMPROVEMENT OF FOOD PROTEINS
Edited by Mendel Friedman

Volume 106
GASTROINTESTINAL HORMONES AND PATHOLOGY OF THE DIGESTIVE SYSTEM
Edited by Morton Grossman, V. Speranza, N. Basso, and E. Lezoche

Volume 107
SECRETORY IMMUNITY AND INFECTION
Edited by Jerry R. McGhee, Jiri Mestecky, and James L. Babb

Volume 108
AGING AND BIOLOGICAL RHYTHMS
Edited by Harvey V. Samis, Jr. and Salvatore Capobianco

Volume 109
DRUGS, LIPID METABOLISM, AND ATHEROSCLEROSIS
Edited by David Kritchevsky, Rodolfo Paoletti, and William L. Holmes

Volume 110
HUMAN INTERFERON: Production and Clinical Use
Edited by Warren R. Stinebring and Paul J. Chapple

Volume 111
HORMONES AND ENERGY METABOLISM
Edited by David M. Klachko, Ralph R. Anderson, and Murray Heimberg

Volume 112
OV ARIAN FOLLICULAR AND CORPUS LUTEUM FUNCTION
Edited by Cornelia P. Channing, John M. Marsh, and William A. Sadler

Volume 113
PARKINSON'S DISEASE -II: Aging and Neuroendocrine Relationships
Edited by Caleb E. Finch, David E. Potter, and Alexander ·D. Kenny

Volume 114
FUNCTION AND STRUCTURE OF THE IMMUNE SYSTEM
Edited by Wolfgang Miiller·RuchhoItz and Hans Konrad Miiller.Hermelink

Volume 115
DYNAMICS OF ARTERIAL FLOW
Edited by Stewart Wolf and Nicholas T. Werthessen
DYNAMICS OF
ARTERIAL FLOW

Edited by
Stewart Wolf
St. Luke's Hospital
Bethlehem, Pennsylvania
and
Nicholas T. Werthessen
The Office of Naval Research
Boston, Massachusetts

PLENUM PRESS • NEW YORK AND LONDON

 Library of Congress Cataloging in Publication Data
Totts Gap Colloquium on Dynamics of Arterial Flow, 1976.
Dynamics of arterial flow.
(Advances in experimental medicine and biology; v. 115)
Proceedings of the Totts Gap Colloquium on Dynamics of Arterial Flow, held
in Delaware Water Gap, Pa., June 1976.
Includes index.
1. Hemodynamics - Congresses. 2. Arteries - Congresses. 3. Blood flow-
Congresses. I. Wolf, Stewart George, 1914· II. Werthessen, Nicholas The.
odore, 1911· III. Totts Gap Institute. IV. Title. V. Series.
QPI05.T67 1976 616.1'36'071 79·9170
ISBN 978-1-4684-7510-4 ISBN 978-1-4684-7508-1 (eBook)
DOr 10.1007/978-1-4684-7508-1

Proceedings of the Totts Gap Colloquium on Dynamics of Arterial Flow,

held in Delaware Water Gap, Pennsylvania, June 7-9, 1976

© 1979 Plenum Press, New York

Softcover reprint of the hardcover 1st edition 1979
A Division of Plenum Publishing Corporation
227 West 17th Street, New York, N.Y. lOOll
All righ ts reserved
No part of this book may be reproduced, stored in a retrieval system, or transmitted,
in any form or by any means, electronic, mechanical, photocopying, microfilming,
recording, or otherwise, without written permission from the Publisher
 Preface

This volume contains the edited transcript of the Second

Topical Colloquium based on leads developed at the original
conference on the artery and the process of arteriosclerosis
(the Lindau Conference of 1970). The first follow-up colloquium
on "The Smooth Muscle of the Artery" was held in Heidelberg in
1973. Planning for the present one was undertaken by the editors
with Dr. C. Forbes Dewey, Department of Mechanical Engineering,
Massachusetts Institute of Technology, Cambridge, Massachusetts.

The meeting itself was held June, 1976 at the Delaware

Water Gap, Pennsy1vani~, under the joint sponsorship of Totts
Gap Institute and the Massachusetts Institute of Technology
with financial support from the American Heart Association,
the Office of Naval Research, and the Smith, Kline and French
Company.

The objective of the series of meetings, beginning at

Lindau has been to examine from an interdisciplinary and
international point of view the fundamental physiologic and
pathophysiologic processes pertinent to the development of
arteriosclerosis. This colloquium sought to examine critically
the evidence relating hemodynamic forces to atherogenesis, to
reconcile disparate findings and interpretations in so far as
possible; and to make a synthesis of the present state of
knowledge of the dynamics of arterial flow.

Grateful acknowledgement is made for the valuable assistance

of Joan Martin and Helen Goodell in the entire editorial process.
The editors acknowledge with thanks the secretarial assistance
of Moira Martin, Colleen Nagle, Cindy Carter and Pat Ide. Special
thanks are due Joy Lowe who executed the entire final manuscript.

v
 Participants

S. BJORKERUD, Sweden A. MALL IAN I , Italy

T. CAREW, U.S.A. P. MANSFIELD, U.S.A.
C. CARO, England W. MEYER, Germany
S. CHIEN, U.S.A. R. NEREM, U.S.A.
C. COLTON, U.S.A C. SCHWARTZ, U.S.A.
R. COX, U.S.A. E. SMITH, Scotland
F. DEWEY, U.S.A. L. STONE, U.S.A.
D. FRY, U.S.A. B. TAYLOR, U.S.A.
H. GREENE, U.S.A. S. WEINBAUM, U.S.A.
D. KENYON, U.S.A. N. WERTHESSEN, U.S.A.
K. LEE, U.S.A. S. WOLF, U.S.A.

vi
 Contents

Chapter 1

ANA'I'OHICAL AND PHYSIOLOGICAL CHARACTERISTICS

OF ARTERIES 1

Chapter 2

FLUID HECHANICS OF ARTERIAL FLOW • • • • • • • • • • • •• 55

Chapter 3

CONTROL OF VASOMOTOR FUNCTION AND THE

HEHODYNAMIC CONSEQUENCES OF THE
CONTRACTILE BEHAVIOR OF ARTERIES 105

Chapter 4

HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 193

Chapter 5

HETABOLIC ACTIVITIES IN THE ARTERIAL WALL 245

Chapter 6

TRAHSPORT OF PROTEIN AND LIPID INTO THE

ARTERIAL WALL • • • • . • • • • • • 299

Chapter 7

HEMODYNAllIC CONTRIBUTION TO ATHEROSCLEROSIS 353

INDEX 467

vii
Chapter 1 ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS OF
ARTERIES

DR. SCHWARTZ: First let us review the light microscopic

structure of a variety of arteries derived from differing
anatomical sites. The common
Muscular and Elastic iliac artery, for example,
Arteries contains numerous smooth
muscle cells, and only a few
fine ramifying elastic processes. This artery is typical of
those classfied as "muscular" arteries. An anatomical
neighbor, the external iliac artery, is also predominantly
muscular with but few elastic processes. It is interesting
to note the tendency for a double type of internal elastic
lamina in this artery, the pathological or physiological
significance of which r.emains uncertain. The internal iliac
artery is an intermediate musculo-elastic artery with quite
a significant amount of elastin present relative to the
external or common iliac arteries.

DR. WERTHESSEN: You have used the name iliac continuously.

Are these from different people or from different sections
of the same artery?

DR. SCHWARTZ: Different branches of a single major

arterial trunk. The common carotid artery is classically
described as being an elastic type artery with considerable
elastin in its media. More peripherally, its major branch,
the internal carotid artery, undergoes morphological
transformation and where it enters the skull is predominantly
muscular in configuration. Also intriguing is the marked
concentration of elastin in the innermost media. The reason
for this structural appearance has yet to be explained.
Another important vessel to the brain, the vertebral artery,
is predominantly muscular in type. It is thus apparent that
some arteries are predominantly elastic, some are predominantly
muscular, and others have an intermediate structure, to
which we ascribe the category, musculo-elastic. These
structural types have been described in detail previously
(1). There is no clear cut correlation between the muscularity
or elastic composition of the artery and the propensity to

*Studies reported here done in collaboration with

Ross G. Gerrity, Ph.D. Department of Pathology, McMaster
University, Hamilton, Ontario, Canada.
1
2 CHAPTER 1

develop atherosclerosis. Some muscular arteries develop

severe disease, as do some elastic arteries. Why then are
there these dramatic differences in structure? What are
their functional correlates, if any?

Now let us turn to the development and ultrastructure

of the arterial media. Histologically, if one compares the
structure of an elastic
Arterial Development artery (such as the aorta)
from a newborn animal with the
same artery from a mature animal, several distinct differences
emerge. There is a 2-3 fold increase in wall thickness
which is not due to an increase in the number of elastic and
muscular laminae, which remains constant (2), but rather, to
a marked increase in the volume occupied by elastin, and, to
a lesser extent, collagen. For example, in the rat aorta,
the volume of aortic media occupied by elastic tissue
increases from about 12% in the newborn to 52% at three
months of age; collagen content (volume) increases from 2%
to about 20% in the same time span (2). The volume occupied
by smooth muscle cells in the same period shows an inverse
relationship to the total connective tissue volume, dropping
from about 60-70% in the newborn to about 20% in the adult
rat (2). Subsequent thickening in old age in the absence of
lesion formation occurs to some extent in the media, but to
a greater extent in the intima, which frequently exhibits a
diffuse thickening which, in some areas, can increase wall
thickness by a factor of three to five (3).

Ultrastructurally the newborn to one-week-old rat aorta

reveals very little collagen or elastin relative to a more
adult vessel (Fig.l-~.
Formation of Collagen Numerous elongated, spindle-
and Elastin shaped cells with copious
endoplasmic reticulum (ER) and
prominent Golgi are present. There are a few discernible
myofilaments in the peripheral cytoplasm of these cells, but
at this stage of development medial cells are morphologically
more like fibroblasts than smooth muscle cells. At two weeks
of age the ER and Golgi remain conspicuous, but myofilaments
are now readily discernible (Fig. l-~. Extracellularly,
there is a distinct increase in the amount of elastin, and
scattered clumps of collagen fibrils begin to make their
appearance. At four weeks of age (Fig. l-~ the lamellae are
wider, and branches of elastin extend between cells. There
is also an increased amount of collagen present in the media
at this point in time, mainly pericellular.
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 3

Fig. 1-1: Transmission electron micrograph of the aortic

media of newborn rat. Medial cells (MC) resemble fibroblasts,
with prominent Golgi (G), rough endoplasmic reticulum (ER),
and only a few myofilaments (MF) in the peripheral cytoplasm.
Only a few collagen fibrils (C) and elastic tissue bundles
(EL) are visible. X 10,000.
4 CHAPTER 1

Fig. 1-2: Transmission electron micrograph of aortic media

from a two-week-old rat. Golgi (G) and endoplasmic reticulum
(ER) are very prominent in medial cells (MC) , but myofilaments
(MF) are more numerous than in the newborn, and cells are
more spindle-shaped. Elastic tissue (EL) now fills the
laminae between cells, and collagen (C) is visible interspersed
in it. X 6, 000 .
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 5

Fig. 1-3: Transmission electron micrograph of the aortic

media of a four-week-old rat. Medial cells (MC) are more
irregular in outline than previously seen. Golgi (G) and
endoplasmic reticulum (ER) are less prominent, and myofilaments
(MF) fill much of the cytoplasm. Elastic laminae (EL) are
widened, and branches (EB) from them extend between cells.
Collagen (C) is seen in moderate quantities at the cell
periphery. X, 6000.
6 CHAPTER 1

Medial cells are decidely more irregular in shape than they

were in the earlier stages of development. The ER and Golgi
are less prominent. Myofilaments are present in large
amounts in the medial cells, which are now identifiable as
smooth muscle cells, although at 4 weeks they are still not
fully differentiated. As further differentiation occurs,
the cytoplasm fills with myofilaments, and the ER and Golgi
are less evident (Fig. 1-4). In the eight-week-old aorta
(Fig. 1-5) the trends are similar to those already described.
At eight weeks the rat is essentially a mature animal, both
sexually and from an arterial structural point of view.
Connective tissue synthesis and in particular, elastin
synthesis in the aorta, has plateaued at this point. Medial
cells are very readily identifiable as mature smooth muscle
cells, with prominent and characteristic myofilaments.

Returning to connective tissue synthesis briefly, after

a single pulse of H-proline fifteen minutes before death,
in a 4-week-old rat, a number of grains are visible within
the medial smooth muscle cells, indicating intracellular
incorporation of precursor (Fig. 1-6). Three hours after the
pulse some grains are intracellular, and most are present
within the extracellular space (Fig. 1-6). In other words,
the H-proline has been taken up and incorporated by smooth
muscle cells (SMC), into collagen and elastin, which are
exteriorized or secreted within three hours.

By six months of age, there are other changes in the

aorta. Large irregular medial SMC are packed with myofilaments
(Fig. 1-7) and considerable thinning of the elastic laminae
has occurred. Some cellular
Changes During Aging degeneration is visible.
There is distinctly more
collagen relative to elastin than at earlier ages. It
appears that in aging, as distinct from development, the
relative proportions of arterial collagen and elastin are
reversed, with more collagen and less elastin in the older
artery. At one year of age (Fig. 1-8) the appearances are
essentially similar to those described for the six-month-old
aorta. Now, however, cellular debris and dying cells are
seen scattered throughout the media together with granulo-
vesicular material. At both six months and one year the
Golgi and the ER are clearly more prominent than in the
younger mature aorta (8-12 weeks), probably reflecting
increased synthetic activity.
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 7

Fig. 1-4: Transmission electron micrograph of the aortic

media of an eight-week-old rat. Medial cells (MC) are now
readily identifiable as arterial smooth muscle cells and
their cytoplasm is filled with myofilaments (MF). The Golgi
(G) is infrequently seen, and the endoplasmic reticulum (ER)
is poorly developed. Collagen (C) fibrils now surround medial
cells, interspersed with elastic tissue branches (EB) separated
from the elastic laminae (E.). X 6,000.
8 CHAPTER 1

Fig. 1-5: Electron autoradiograph of aortic medial cells (MC),

15 minutes after intravenous injection of 3H-proline. Grains
are visible overlying Golgi (G) and endoplasmic reticulum (ER)
of the cells, but collagen (C) and elastic tissue (EL) are
not labelled. X 8,000.
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 9

Fig. 1-6: Electron autoradiograph of the aortic media 3 hours

after the injection of 3H-proline. Grains are still visible
overlying medial smooth muscle (Me), but most grains overlie
collagen (C) and elastic tissue (EL). X 15,000.
10 CHAPTER 1

Fig. 1-7: Transmission electron micrograph of the aortic

media of a six-month-old rat. Medial smooth muscle cells (MC)
are large, and irregular in outline. Their cytoplasm is
largely filled with myofilaments (MF), while endoplasmic
reticulum, Golgi and mitochondria are perinuclear (arrows).
Intracytoplasmic lipid inclusions (L) are often seen. There
is thinning of the elastic laminae (EL), collagen (C) is
prevalent, and cellular debris (DC) is visible in the extra-
cellular spaces. X 6,000.
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 11

Fig. 1-8: As in Fig. 1-7, but from a one-year-old rat.

Much cellular and granulo-vesicular debris (arrows) is present,
and remaining medial cells (Me) are highly vacuolated. Elastic
laminae (EL) are tenuous and irregular. X 6,000.
12 CHAPTER 1

As the aorta gets older, the medial cells become larger and
more irregular in shape with many more processes. This
gradual increase in cellular degeneration with normal aging
is interesting, particularly as it corresponds temporarily
with hypertrophy of remaining cells, and hyperplasia of
synthetic organelles in these cells. This pattern continues
and becomes more pronounced in two and three-year-old rats
(4).

DR. STONE: Dr. Schwartz, when you begin to see cell

death, is there any actual indication of a decrease in
protein synthetic capability at that point?

DR. SCHWARTZ: No, the evidence would point to the

contrary. However, there is no simple answer because even
at that time when one is starting to observe cell death, the
Golgi and the ER are becoming more prominent, and remaining
cells are becoming larger. There is also an increasing
amount of collagen being formed, but we have no evidence
that medial cells are dividing. It would appear that the
cells undergo hypertrophy rather than hyperplasia in order
to compensate for the death of adjoining cells. This
phenomenon appears to be a factor in biological aging of the
artery in the absence of disease.

DR. STONE: That is the question. Because the anatomical

evidence of synthesis is there, is there actual synthesis
taking place?

DR. SCHWARTZ: Yes, I believe so. But I think we

should be more specific -what is being synthesized? Obviously,
if cells are undergoing hypertrophy, we can expect that
cellular components and membranes are being produced. Also,
our data show an increase in collagen content of the media
from 24% at 12 weeks to 31% at 26 weeks on a dry weight
basis.

DR. STONE: If there is actual synthesis taking place

with thymidine labeling do you see any increased incorporation
into nuclear DNA?

DR. SCHWARTZ: Yes, but I believe we should be able to

get a comprehensive answer to that from K.T. Lee.

DR. LEE: Are you asking about thymidine incorporation

at that time?
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 13

DR. STONE: Yes, in other words, for nuclear activity.

DR. LEE: Usually when cells are dying the vessel

tries to compensate by putting
Cell Renewal new cells out to maintain the
strength of the vessel wall.
So it is likely that we have regenerating cells there and at
that particular time you will see thymidine incorporation.

DR. WERTHESSEN: But have you actually measured it?

DR. LEE; What do you mean, measured it?

DR. WERTHESSEN: By thymidine. Has anybody measured

it? You can expect it but it doesn't necessarily have to
happen.

DR. LEE: There are two ways of doing this. One way
is to get the labelling indices by counting labelled cells
and baseline cells and the other way is to measure chemically
the content of the thymidine. Probably the chemical method
may be simpler. If a large number of cells die in a vessel
and they are not replaced, the wall is weakened and results
in aneurysmal dilation or rupture. Those dying cells must be
replaced by new cells. However, we have not actually measured
thymidine incorporation in aging vessels.

DR. WOLF: I will summarize this to be sure that we

are all on the same page of the hymn book. What you are
saying is that in association with aging there is increased
cell death accompanied by increased evidence of synthesis.
Fair?

DR. SCHWARTZ: Yes, but the ultrastructural evidence of

increased synthesis is circumstantial, not affirmative. It
indicates that the mechanism for synthesis is present.

DR. WOLF: And that is why we brought these other

people into the conversation. Now are we clear on that, Dr.
Lee?

DR. LEE: I think that is what I have said.

DR. WOLF: Do you have the evidence?

14 CHAPTER 1

DR. LEE: My statement is based on our observations in

vessels stimulated by dietary means for the production of
atherosclerosis. In physiological aging, however, I am sure
the situation is different from vessels stimulated by dietary
means. But, common sense wise, if many cells in the vessel
wall die without replacement, sooner or later the vessel
will be weakened and rupture.

DR. WOLF: Fair enough, you brought it into focus.

Dr. Cox.

DR. COX: We have performed some studies in dogs and

rats on the effects of aging on secondary branches of the
aorta such as the carotid and the iliac. In general, we
have found that the total connective tissue content increases
in the aging animal. The cellular content, the smooth
muscle cell content, decreases. There are associated with
these changes decreases in contractile function as measured
by the ability of smooth muscles to generate force on a per
unit area basis. So from our results in dogs and rats, and
again these may be findings which are species dependent,
there seems to be a reduction in cellularity with aging and
in the aged. There certainly doesn't seem to be a one-for-
one replacement.

DR. BJORKERUD: The rat is a difficult example because

the rat is growing very late in life and during the growing
phase the cell has an increasing
Phases of Growth versus territory. Then the cell
Senility increases its synthetic
capacity to cope with the
increasing territory. If you study differences in synthetic
capacities, we have studied, for instance, the synthesis of
phospholipids, there is a relationship between the increase
of extracellular tissue per cell and a corresponding increase
in the rate of phospholipid synthesis per cell. I am not
certain when the rat stops growing and when the senile phase
starts. One should clearly distinguish between these two
phases. I think your discussion is pertinent for the senile
phase. That the two phases might overlap each other makes
the problem even worse.

DR. SCHWARTZ: That is a good point. I have tried to

indicate developmental changes on the one hand and aging
changes on the other, and I made the point that at eight
weeks the rat is considered to be a mature animal. There
are clearly many changes which are developmental up to the
age of eight weeks.
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 15

Just precisely what changes are part of growth and development

and what changes are degenerative or senescent, or where the
cut-off lies, is an extremely difficult question.

At the end of the spectrum, we are obviously dealing with

changes in the one to three-year-old rats that are associated
with aging, possibly including an increased turnover of
medial SMC, as well as documented increases in collagen
relative to elastin. It should also be pointed out that
only about 5% of these rats live beyond two years of age.
Beyond this age we are therefore looking at a select population,
and the changes we are observing may well be associated with
their ability to live to this advanced age.

DR. WOLF: I think Dr. Elspeth Smith has a comment.

DR. SMITH: I want to make a comment to Dr. Cox. When

you say total connective tissue increases and smooth muscle
cells decrease, is this relative to each other on a concen-
tration basis, or per unit segment of total artery?

DR. COX: It is on a basis of per unit weight.

DR. SMITH: On a per unit weight basis if one increases

the other must by definition, decrease. But do you know
what happens supposing you take a unit segment of an artery.
In that given segment, are there fewer smooth muscle cells?
Or the same number of smooth muscle cells or more smooth
muscle cells? Because I think this is one of the problems
we are up against. The concentration gives a distorted
picture to some extent and you don't really know whether one
thing is decreasing and another thing increasing relative to
each other or whether in the whole structure there is an
absolute increase or decrease.

DR. COX: This is a very difficult question to answer,

because of the problem of delineating a unit structure. I
think it is particularly
Species Differences difficult by virtue of the
fact that so many people are
studying different animals, the rat being, with Nathan
Shock's group, the principle animal that is used to study
the effects of aging on the cardiovascular system. The
differences that we have found in the rat and the dog are
considerable. I would say it is very difficult to answer
that question because of the species variability. In the
case of the dog, the changes in body size going from a young
adult, say a two year old dog, to a ten or fifteen year old
animal are relatively small.
 16 CHAPTER 1

Compare that to the rat, which as we heard continues to

increase in size from the adult stage to the senescent
stage, there is a very large difference in what one would
see in the thoracic aorta. Certainly the size of the thoracic
aorta would increase in the rat. It would not change
significantly in the dog. The wall thickness of arteries
certainly increases with age. That is one of the consider-
ations. In terms of total concept of muscle cells, I really
don't know how to answer that because I don't have the
quantitative information. We have not looked at it on any
other basis besides concentration.
DR. SCHWARTZ: Now I wish to consider changes occurring
in the arterial endothelium during growth, development and
aging. As a prelude, let me
Changes in Endothelium briefly emphasize that arterial
endothelium has a wide variety
of functions, which are summarized in Table 1. The endothelium
provides a fascinating blood-tissue interface with a regulatory
role in permeability, both in terms of influx into the
vessel and efflux from the vessel. The endothelial cell is
capable of regeneration both in vivo and in tissue culture
(5,6). It may exert some kin~of metabolic control over the
underlying media as suggested by recent studies (7). There
may be specific receptor sites on the surface, perhaps
relating to immunologic or pharmacological reactions or to a
variety of other mechanisms, including the internalization
of macromolecules. The endothelium may be contractile. It
certainly contains contractile filaments (3,8) as probably
do most other cells, but the biological significance of
endothelial contraction has yet to be determined. The
spectrum of functions of vascular endothelium is considerable
(Table 1). One needs, however, to enter a note of caution.
Many studies have, for obvious reasons, been undertaken
using capillary endothelium. This exhibits a range of structural
differences consistent with the functional specialization of
the different organs (9). As illustrated by Rhodin (10) the
capillary endothelium may be continuous, fenestrated, or
discontinuous with large gaps. These differences must be
considered in studying transport phenomena. Furthermore,
extrapolation of data derived from studies on capillaries to
the arterial endothelium may be inappropriate, or indeed
misleading.

RAT When one examines the aortic endothelium of a newborn

rat sectioned "en fac~' (Fig. 1-9), an extensive endoplasmic
reticulum and numerous mitochondria and free ribosomes are
readily seen. By 2 months of age (Fig. 1-10) the ER is
fragmented with much shorter profiles.
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 17

TABLE I
SOME PROPERTIES OF NORMAL VASCULAR ENDOTHELIUM

1. Regeneration, replication.

a) In arteries or veins
b) In tissue culture-venous or arterial endothelium

2. Control of macromolecular transport and permeability.

3. Synthesis of platelet inhibitor substances.

4. Contraction, contractile protein.

5. Plasminogen activator, fibrinolysin.

6. Tissue thromboplastin.

7. Heparin and heparatin sulphate.

8. Factor VIII synthesis.

9. Prostaglandin E synthesis and release.

10. Basement membrane formation.

11. Angiotensin conversion.

12. Serotonin uptake.

13. Histamine synthesis.

14. Phagocytosis.
18 CHAPTER 1

Fig. 1-9: Transmission electron micrograph of an "en face"

section through the aortic endothelium (E) of a newborn rat.
The endoplasmic reticulum (ER) is extensive, and mitochondria
(M) are numerous. Free ribosomes are numerous (arrows) and
cytoplasmic filaments (F) are frequently found in the peripheral
cytoplasm. X 15,000.
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 19

Fig. 1-10: As in Fig. 1-9, but through the aortic endothelium

(E) from a two-month-old rat. The endoplasmic reticulum (ER)
is fragmented and less prominent than in the newborn rat.
Weibel-Palade bodies (B) are frequently seen in the mature rat,
and a tight junction (arrow) is visible. X 12,000.
 20 CHAPTER 1

Copious plasmalemmal vesicles are apparent at all ages.

Tight intercellular junctions are present at all ages studied.
Weibel-Palade bodies are present with a variable frequency
in the mature aorta. However, they are not normally seen
with any frequency in immature animals, and also exhibit a
significant species difference with respect to frequency.
PIG Thus, they may be unreliable markers of endothelium in
tissue culture, and in our experience are rare in porcine
arterial endothelium. Some multi-vesicular bodies may be
precursors of Weibel-Palade bodies (11). Their roles,
albeit uncertain, may relate to the fibrinolytic system, or
to the development of tissue thromboplastins (12,13).

Now continuing with the mature rat aorta sectioned

"en face," the very large number of plasmalemmal vesicles is
readily apparent (Fig. 1-9,1-10), and needs emphasis,
particularly as one sees only a relatively small number of
vesicles in transverse sections. The frequency of plasmalemmal
vesicles is not age dependent (2).

Endothelial mitosis was for many years regarded as a

rare phenomenon, but is regularly
Mitosis seen up to approximately three weeks
of age in the arterial endothelium.
Endothelial cell division is definitely an age-dependent
phenomenon, the mitotic frequency declining in mature
animals,

Contractile filaments tend to have either a distinctive

perinuclear orientation in the endothelium, or alternatively,
are associated with the plasma membranes. "En face," in the
rat aorta, cytoplasmic myofilaments, sometimes with dense
bodies across them, may be seen (Fig. 1-9,1-10,1-11). They
are frequently parallel to the subendothelial anchoring
filaments which attach to the connective tissue in the sub-
endothelium (Fig. 11). These anchoring filaments are also
present in veins, arteries, capillaries and lymphatics
(14,15).

DR. DEWEY: Dr. Schwartz, with regard to these anchoring

filaments, to what extent do they penetrate
Filaments the sub-endothelial tissue?

DR. SCHWARTZ: To the best of my knowledge, and I

can't be dogmatic on this, they are quite superficial. They
emerge and attach to the connective tissue in the innermost
part of the subendothelial space.
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 21

Fig. 1-11: Transmission electron micrograph of an oblique

section through the endothelium (E) of a three-month-old rat
aorta. Cytoplasmic filaments (F) appear to merge with dense
areas (D) along the plasma membrane, and are aligned with
anchoring filaments (AF) in the subendothelial space. The
latter filaments are closely associated with (arrows), and are
similar in diameter to, the protofibrils surrounding elastic
tissue (EL). X 25,000.
22 CHAPTER 1

In young and mature animals in which the endothelium is not

widely separated from the internal elastic lamina, these
filaments merge with the protofibrils seen at the edges of
elastic tissue bundles. Don, can you comment on this?

DR. FRY: If one looks at transverse sections, one

does not see these structures. It is apparently only by
this tangential sectioning technique that they are made
visible. They must lie in the plane of the subendothelium
and be very thin in the radial direction.

DR. COX: They are extracellular?

DR. SCHWARTZ: Yes: the filaments are extracellular;

they are parallel with the intracellular filaments and fuse
into a dense body at the plasma membrane. It is impossible
to tell whether they are continuous with the cytoplasmic
filaments, or whether both types of filament attach to their
respective side of the plasma membrane.

DR. KENYON: Is there any evidence that the number of

attachment sites, if you could call them that, is a function
of the age of the endothelial layer?

DR. SCHWARTZ: I can't answer that.

DR. CHIEN: Just a comment about Dr. Cox's question.

Since the vesicles are about seven hundred angstroms in
diameter, these anchoring filaments or fibers probably are
on the order of one hundred angstroms.

DR. SCHWARTZ: That is not an unreasonable estimate.

Let us turn now to a final aspect of this part of the presen-
tation. Within the subendothelial
Subendothelial Changes space there is generally more
than one morphologic cell type
visible, even in the immature animal. Some of these cells
are readily identified as smooth muscle cells, while others
are undifferentiated and less readily identifiable. This
becomes important later when one begins to think about
atherogenesis and examines the spectrum of changes in the
endothelium and subendothelial space associated with aging.
In the 3-year-old rat aorta (Fig. 1-12), the endothelial
nuclei are dense and crenated, and irregular blebbing is
frequently seen on the surface of the endothelium. Droplets
of variable electron density, which are thought to be lipid,
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 23

are often present within the endothelial cytoplasm, sometimes

associated with lysosomes or the Golgi apparatus.
The 3-year-old rat endothelium, when viewed obliquely,
exhibits a tremendous number of cytoplasmic processes
projecting into the lumen, which may increase the surface
area considerably (Fig. 1-12). As the processes also contain
copious numbers of plasmalemmal vesicles, this change may
increase the effective number of vesicles exposed to the
lumen. At three years of age it is also of interest that
lymphocytes are often attached to the endothelial surface.
This phenomenon becomes more frequent in ~~ing animals. The
reasons for this lymphocytic attachment hav~'~et to be
explained.

The aortic subendothelial space of aged animals usually

demonstrates a diffuse intimal thickening in the absence of
any discernible lesions (2,16). When viewed "en face"
sections through the intima, the subendothelial space is
seen to be filled with amorphous granulo-vesicular material
of varying electron density (Fig. 1-13). Cellular debris
and ghost-bodies, presumably caused by the complete dis-
solution of cells, are prevalent. Other cells containing
large inclusions appear to be phagocytising this material
(Fig. 1-13).

Crystalline bodies may be seen in the subendothelial

space from approximately 18 months of age onward (Fig. 1-14).
Some are intracellular, but most are extracellular and have
a two-dimensional periodicity of 5.67 nm (2).
They do not appear to be viruses, and their role and significance
is uncertain. Finally, one should mention the considerable
fraying of the internal elastic lamina which occurs "pari
passu" with thinning of the medial elastic laminae with
aging.

So far, we have briefly described a sequence of changes

in the media, endothelium, and subendothelial space which
occur during growth and development on the one hand, and
aging on the other.

I will review some aspects of endothelial structure and

function, with particular references to permeability later
on in chapter 6.
24 CHAPTER 1

Fig. 1-12: Transmission electron micrograph of an oblique

section through the aortic endothelium of a three-year-old
rat. The nucleus (N) is electron dense and tortuous in out-
line, and there are many cellular extensions (X) into the
lumen (L). Many lipid inclusions (I) are visible in the
cytoplasm, which appear to be associated with dense lysosomes
(arrow). The subendothelial space (SES) contains flocculent
material. X 18,500.
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 25

Fig. 1-13: Transmission electron micrograph of an en face section

through the aortic subendothelial space (SES) of a two-year-old
rat. Endothelium (E) and lumn (L) are at top left. Ghost bodies
(GB), cellular debris (DE), and intimal cells (IC) containing
large inclusions (arrows) are embedded in a matrix of granulo-
vesicular material of varying electron density. X 6,400.
26 CHAPTER 1

Fig. 1-14: Transmission electron micrograph of a crystalline

body in the aortic subendothelial space of a two-year-old rat.
Two dimensional periodicity has a spacing of 5.67 nm. X 134,400
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 27

DR. KENYON: My work concentrates on certain gross features

of blood vessel force and deformation relationships. This point
of view may be used to interpret
Effects of Shear Stress some aspects of arterial disease,
and Other Hydrodynamic but without accounting for altered
Forces on Arterial Structure vessel wall constituents. In other
words, I take the continuun view-
point for the most part. That is to view the arterial wall
as an equivalent homogeneous deformable material. I look
for some interesting behavior of this structure near bifurcations,
for example, and other illustrations that I will get into
later. I use this continuum concept not because of its
completeness but because I think it represents an area where
not enough attention has been focused up to now. Being from
the Fluid Mechanics Laboratory at M.I.T. we have emphasized
fluid stresses on the endothelial surfaces and on the vessels
themselves. I would also like to make a plea for studying
the effects of what we might call the "solid stress" in the
vessel itself.

Let me try to introduce some of the very simple notions

that are fairly standard in the engineering literature but
which deserve to be distributed to a wider audience so that
in your own studies you might better appreciate the effects
of force and deformation.

Figure 1-15 is a longitudinal section of a human descending

thoracic aorta from a standard histology textbook. I only
want to use it to illustrate the terminology of vessel wall
stresses. The force in the axial direction i.e. perpendicular
to the figure per unit area is the longitudinal stress. The
hoop stress represents the force per unit area acting perpendicular
to the wall thickness and is absorbed by the layers of
medial tissue. We have already seen in Colin Schartz's
earlier talk the detailed structure of the vessels themselves
so that this does not need reiteration at this time. What I
would point out here is that the models I will be talking
about probably are useful, if anywhere, for blood vessels
which are more or less homogeneous and I think an elastic
artery, before it becomes highly thickened, is more homogeneous
than say a typical distributing artery. So if we are talking
about macroscopic stress versus deformation relationships
they are probably going to be most meaningful before significant
arterial disease has occurred. Figure 1-16 simply represents
a similar section from the same book for a distributing artery,
a highly inhomogeneous structure. This has been stained on the
left for the muscular constituents (a) and for elastin stain on
the right (b).
28 CHAPTER 1

Subendotheliol loy,·r

longitudinolly
Infimo
~trioted loyer

Fene~troted
Media elastic
membrane

Vein

__- - Elostica e"terno

AdventItia .-.oo:;....--_
-:. ~..,.. __ ---....---~
Vein

Fig. 1-15: Longitudinal section through posterior wall of

human descending aorta. Elastic tissuents are not shown
clearly. Elastic fiber stain. X 85 (After Koolliker and
von Ebner).
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 29

odv~nl j l l o

It
•
Fig. 1-16: Sectors of two cross sections of the volar digital
artery of man. A. Stained with hematoxylin and eosin. B.
Stained with orcein to show elastic tissue. X 80. (Slightly
modified from Schaffer.)
30 CHAPTER 1

TABLE 1. Parameter values for stress-strain law

cree = A (1 -¥, ) e KAe

At ).0 K

Artery NEpi KeN NEpi KCN NEpi KCN

1 11,800 4,350 0.233 0.579 4.0 4.5

3 13,100 46,800 0.222 0.549 5.2 3.1
5 7,100 10,500 0.665 0.696 5.7 4.8

tUn1ts in dyne/cm 2

o THOR • FEM
1·8 • A800 • CAR

o 2 4 b 8 10 12 14 16 18 20
FREQUENCY Hz

Fig. 1-17: Viscoelasticity of Canine Aorta (Berge1, 1961).

ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 31

This particular artery looks as though it has a more difficult

characterization in terms of the ability of the elastin to
withstand stress because it almost looks as if it is fragmented.
It is very difficult, I think, to assign a mechanical role to
the elastin in this particular vessel. In general, I have
come to the conclusion that the small distributing arteries
are going to be very difficult to model with any meaningful
precision from the concepts that I will be talking about right
now.

When there are roughly five significant engineering

behavior modes that appear, the first one of importance is
nonlinearity in the observed versus strain law. If an artery
is excised and allowed to contract and is then pressurized the
observed diameter versus pressure relationship is highly
nonlinear, and the engineering concept of finite strain has to
be applied. This circumferential strain is based on an at-rest
dimension, radius, for example. The data shown in Table 1 are
taken from a paper by Doyle and Dobrin (17) in which they found
that they could mechanically characterize their canine carotids DOG
using an empirical factor of A (units and stress) and another
empirical factor kappa (dimensionless). As you can see, the
application of norepinephrine as a muscular stimulant greatly
increases the stress A compared to the application of KeN, a
muscular poison. Thus in the canine carotid the smooth muscle
cell, a fairly significant influence on the stress deformation
exerts relationship.

The next important property that physiologists and engineers

have examined in blood vessel mechanics is the ability of the
blood vessel to exhibit a time
Stress versus Strain dependent stress versus strain
behavior. I am not sure (for a
typical vessel) what this is really caused by in terms of a
micro-constituent. I am tempted to suggest that it is due to
the smooth muscles, and how fast they can relax. I am sure there
are people here who can comment on the relative importance of
smooth muscle in stress relaxation. Thus, the artery is a
viscoelastic material. If I stress the artery with an internal
pressure, the diameter does not immediately respond, or if I
stress it in a time dependent fashion there is a phase vessel.
In Fig. 1-17 some original data of Bergel (18) reveals that
there is an asymptotic behavior in vessels such as that at about
two Hertz and above the artery is more or less elastic in
character. That is, there is no subsequent interesting or
significant further dependence on the applied frequency.
 32 CHAPTER 1

Since this is for a dog where the normal heart rate is at

least on the order of 2 Hertz for those vessels of investigation,
it is probably a good assumption to make that at the fundamental
heart rate and all the higher harmonics the vessel properties
are governed by these asymptotic properties, such that the
dynamic stiffness modulus is about 1.7 times the static
modulus. The carotid apparently has much more frequency
DOG dependent behavior, as presumably it has more smooth muscles,
and I think these smooth muscles are allowing the material
to behave stiffer dynamically compared to its static property
than say the thoracic or even the abdominal aorta, which are
more elastic-type vessels. So it would appear as though the
smooth muscle constituent is, in part, responsible for
allowing the more peripheral arteries to behave in a much
more time dependent fashion. However, at least above two
Hertz in the dog the properties are relatively unchanged
with frequency.

Fig 1-18 provides data from a recent paper by Azuma on

more or less the same topic (19). Azuma took strips of
arteries in both longitudinal
Stress Relaxation and ring specimens. These
were five millimeter wide
strips. The samples were subjected to an applied constant
stretch and the force required to hold it at that posi-
tion diminished as a function of time. This is called
stress relaxation and stress relaxation is again another
measure of some constituent in the vessel wall allowing
itself to relax with time. The longitudinal sections,
sections along the axis of the vessel, have been stressed in
the direction that blood flows. The arch of the aorta
exhibits almost no stress relaxation for this particular
sample. In fact the only obvious amount of longitudinal
relaxation occurs in the abdominal aorta. However, the ring
sections, which are circumferential strips, exhibit somewhat
the same pattern for the more elastic arteries but as we
proceed distally, the iliac artery exhibits a very pronounced
circumferential stress relaxation and apparently this is due
again to the predominance of both tension and smooth muscle
in ring-type configurations as you proceed distally. Now the
time constants of interest here can't really be seen from
Fig. 1-18 because it is too coarse but this is a one
minute time bar, so it would appear as though the relaxations
went on within the very small part of a minute, perhaps even
on the order of a heart rate. Thus, a great part of what we
would do to attribute as gross viscoelastic character of
arteries is due to the ability of smooth muscle cells to
relax under stress. So, therefore, the proliferation of
smooth muscles with age and with disease could be very
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 33

long ring

jA';"CUIi L j:usL
h'hor~
(pro. ) G,hor°L
(pro•. ,

~Ihor~
(d,s., hthor°L
(diU'

habd·L habd0"L
......,
(prol.' (

~abd·L
AG.•bcS.

(dIS\ )
p.\.'~

[:'-1 ~'-L
~

[ ~--.usL
~
~~~

Fig. 1-18: Stress relaxation in canine arterial segments.

(Azuma, 1972).
34 CHAPTER 1

important in controlling their elastic characteristics,

apart from the obvious changes in collagen to elastin which
also may playa lesser role in viscoelasticity.
The third mechanical property obtained from the existing
literature says that there is a pronounced anisotropy of the
vessels. Now anisotropy, in engineering usage, means that
the stretch versus deformation relation of an element depends
on the orientation of the element before stressing. If I
do not get the same value for force versus deformation, I
call the material anisotropic. There are some data for
canine aortas from Patel (Fig. 1-19) which should look
familiar to most of you (20-27). The hoop modulus E's is
about (7-12) X 10 9M/Cm at heart rate frequencies. The
axial modulus E's for this particular test was about twice
as large. We would conclude from this particular test that
within the incremental conditions of this test, the arteries
should be about twice as stiff axially as circumferentially.
Now this does not agree with some measurements of wave
speeds in the canine carotid obtained by Moritz and Anliker
(21) Their wave speed data was only consistent with a hoop
stress modulus which was roughly twice as big as the axial
modulus. These results are completely at odds for these two
vessels. Now it may well be that there are functional needs
for these two vessels which vary, and therefore, we would
not expect a consistent pattern of relative stiffness in the
axial direction compared to the circumferential. On the
other hand, I would also point out that these values are
highly dependent on the state of pre-stress in the vessel
and that the carotid, for example has to undergo quite
severe longitudinal stretch as a function of head position.
It seems quite realistic to assume that not only is anisotropy
very likely to significantly vary between vessels but within
anyone vessel it is very likely to be a function of what
one might call the physiologic state.

The prime notation of Fig. 1-20 is a measure of whether

the deformation occurs in phase with applied force or out of
phase and the values of the out of phase components for
either the hoop stress, the axial stress or the radial
stresses appear to be very small compared to the real part.
So in a very crude sense, these vessels are anisotropic
and they are roughly within 10% of being perfectly incrementally
elastic.

DR. DEWEY: What artery was the Anliker data on?

DR. KENYON: Carotid.

DR. DEWEY: And this is the aorta?

ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 35

E'
• II E', , .. YIVO 0.1. 11000011
•• OIlOUlI IU. ~ LOW 1tIa •
.. OIQUP. 1111011 "'OW ",- •
• • OIlOU' 111 HOW •• "'''' •• '
., OIlOU.]I 1111'" .... N'''' ".'

i
IN Y"IO OAI. '6 0001'
•••
~---.•----....-----.----...... ~
I

i; • ..··-····..--.·····1
~';;
ti .. 1
« !
=~.
0 -
~
:;:
I

. s. . ·. . ~ I.
t'"
-, _ _ _ _......._ _ _ _ L -_ _ _ _ _ _......._ L -_ _'''''__ _ _ __

1 • • I I • •

fREQUENCY-HI
Fig. 1-19: Anisotropy of canine aorta (Patel, 1973).
 W
0-

,
~
er ••
' .0 r fI!..
IN VIVO DATA I I() DOGS I
I o ill
._0-' I " .. )., , 0 . A.'
.A. OIOtlU' I 1.10_ "" 10_ ),,'
•• 010111".
• • 0 1011...
• 'OW
t "'0111 " .
1II. .. ,Co" "', •
... eM ", J
I"
IN VI 110 DATA (6 DOGS :
...2 05 . e---·. ---. --e·- e-----e-----e---- -0 - - - -..
- '"" .... : --- -
:...... - =*==-0___
. ~~ -~
-•..
I
-~ .
~----.-- -----e,~_-; :-'1-
____ - - . .... .
a"" . ..... -+--. . -. -+-- - - ....- --
---
--

...~ oI . I

c(. -'
.o 0.5 <. ~.

.....-------
o ~ •• ,., ~ .-
Wri-W---- «
.
e F •• ...

-1.1 .~-
2 s 5 o 2 5 2 2
• , • o • •
'UQUINCY -HI

Fig. 1-20: Anistropy of canine aorta (Patel, 1973).

o
I
}>
-0
-i
m
:lJ
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 37

DR. KENYON: Yes, apparently it seems not to be

consistent between those two. If you want to characterize,
and I think it is a valid thing for one to do, the anisotropic
properties of the vessels, you have to be very careful which
one you are talking about and how you perform your tests.
For example, do you leave the vessel in the animal when you
make your longitudinal displacement measurement, or do you
excise it? I would believe that there is a significant
amount of longitudinal strengthening of at least some vessels
in situ. This may account for some of the discrepancy but
again the comparison between Patel, and Moritz and Anliker
should be quite close because, after all, they were both in
situ measurements.

DR. CAREW: Just a short comment. There are some

older data that tell an interesting story with regard to
wave speed and that is the observation of the amount of
longitudinal displacement during the cardiac cycle. These
are data back in the early 60's. In the thoracic aorta,
Hall et al showed a lengthening with passage of the pressure
faults underneath that particular site which was being
measured. But in the abdominal aorta just the points of
attachment, the initial stretch on the vessels as well as
perhaps the inherent properties can vary enormously from
site to site.

DR. KENYON: I cannot really justify the differences I

see reported in the literature, variations in values may be
due to experimental technique. None the less, I think
there are some real fundamental differences in each vessel.
Fig. 1-20 is of more recent vintage but is a generalization
of an earlier study of corresponding static properties.

DR. FRY: Yes, there was a series of papers about that

time from our laboratory by Patel and co-workers in which
they studied systematically the geometry; longitudinal
vascular tethering; elastic symmetry; and the anisotropic,
static and dynamic rheologic properties of the aorta, (Fig.
22-26) (Fig. 1-5). Later, Vaishnav, Young, and Patel
studied the nonlinear viscoelastic properties of the aorta
(27). Many of these studies are summarized by Patel and
Vaishnav in review articles (28,29).

DR. KENYON: That takes us through three fairly conventional

engineering descriptions of blood vessel rheology. I think
the next important point to
Retraction of Cut Vessel make about blood vessels is
that they do appear to be
under a state of pre-stress and this is a fairly common
38 CHAPTER 1

observation of the physician, or the physiologist, that when

he snips a vessel it contracts. It seems to withdraw. I
have always wondered why it should be. I have read through
Bergel's article very carefully and he said that it might be
due to a polymerization process within the vessel during its
maturation and that there are certain internal stresses set
up in the vessel leading to longitudinal stress. However,
there is another concept that may be involved. I imagine
the systemic circulation as being some kind of very peculiar
shaped glove in which you have an aorta, as "wrist," say,
and a peripheral circulation of small distributing arteries
and capillary networks as fingers with many small perforations.
One concept which could explain the longitudinal pre-stress
in the vessel is to visualize how one could maintain equilibrium
in the system assuming there is no restraint on this glove
due to the tissue around it. The wall tension that would
exist in the vessel under normal hydrostatic pressures,
systemic circulation, is about the right order of magnitude
to explain the pre-stress in the vessel. Now I am not
saying that this is what happens, in fact, because I don't
know what the time mean value of the tethering in the longitudinal
direction is from systemic circulation, but I do think that
it offers a mechanical reason for "pre-stress" longitudinally.
Whatever vessel we are looking at, it would appear to be
necessary to maintain equilibrium with a longitudinal tensile
force in the wall.

DR. CARO: Could a very simple minded biologist comment?

I can't think of any tissue in th~ body that isn't pre-
stressed, much less those of the blood vessels.

DR. WOLF: There are other alternative explanations

that take into account the innervation of the arterial wall
The trauma of snipping may stimulate the nerve network to
bring about the retraction you observe.

DR. KENYON: I quite accept your comments as pertinent.

This survey completes the picture for this particular view
of the blood vessel and its dynamic and anisotropic properties.
One comment I would like to make on this view of arteries in
general is that you have to make some assumptions about the
mechanical behavior of the vessel under pulsatile strain,
and the assumptions that were made by Patel and others were
these. First that the vessel is incompressible under rapid
changes of pressure (heart rate frequency). The second
assumption is that the artery is in a state of stress such
that excursions are defined by a free energy function for
the vessel.
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 39

While I can accept the concept of incompressibility for

rapid load changes, I find it much more difficult to appreciate
the pertinence or mechanical correctness of the assumption
that these changes are dominated or controlled by energy
function. That does not follow from thermodynamics because
this is a highly memory-sensitive system.

I would also like to mention the fifth and final property

of blood vessels which has been described, the permeability
function. Permeability, we think, is probably control for
the large particles determined by the vesicular transport
mechanisms. On the other hand, we know that the arteries
are capable of supporting a water flux through the wall.
This water flux is extremely small and is therefore very
difficult to measure ••• a typical number might be 10 to the
minus 6 centimeters per second or about l/lOOth a micron per
second. It is a very, very small fraction of the endothelial
cell thickness per second. An interesting calculation you
can do is to use the Palade data and calculate a flux
of water using their vesicular transport rates and you come
up with about the same number. That is, it would appear
that you could account for all the water flux in a typical
endothelial cell just by assuming that every time a vesicle
came up to the surface that it gulped in mostly water. It
may be low molecular weight, it may be L.n.L. or whatever,
but some way or other arteries are able to transmit a small
and yet measurable water flux through their surface. Now
the quantity that controls this rate of flux is the hydraulic
permeability and I will denote it as K~. For arteries it
is about ten to the minus thirteen. The units are square
centimeters per second. This represents the water flux on a
bulk area basis produced by a unit pressure drop across the
unit thickness of tissue, and the relative smallness of this
number has important implications for the assumption of
incompressibility. If you make a model of the arterial wall
of two phase material capable of exuding water, then it
exudes it at a rate which, in part, is determined by this
number. The other quantity which enters into this exudation
rate is the compliance of the tissue itself. The relative
smallness of the permeability is in part responsible, I
should think, for the success with which the artery may be
viewed as incompressible for short term loading. For long-
time loading, I am much less secure about the assumption of
constant volume. You don't have to squeeze an artery between
two porous plates for example, to exclude water. I believe
it is possible to change its volume if you wait long enough,
by producing a variety of load patterns which is not a
concept which I think many people have previously considered.
40 CHAPTER 1

excess pressure 6P

.,
A A'

Fig. 1-21: Model for aortic arch.

_ l ___ ~E~t---

Fig. 1-22: T-junction geometry.

ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 41

I first of all want to get into some very simple-minded

concepts of stress at places in the arterial system which
are points of interest which are likely to develop plaques.
Fig. 1-21 is a model of the aortic arch and I assume that
the aorta is isotropic and linearly elastic. I can use
known solutions to find the hoop stresses in the arch under
a small excursion of pressure. There is an axial hoop
stress, i.e., a stress along the flow direction. There is
also stress in the circumferential or hoop direction,
Ne . Now it turns out that the value of this hoop stress
is a function of position around the hoop and it is greater
on the inside of the bend and least on the outside of the
bend. The difference between the maximum and the minimum
hoop stress divided by their mean value is equal to twice
the ratio of vessel radius to the radius of the bend, R/Ro.
If as typical data for the aortic arch R = 1.5 cm. Ro = 5
cm., then you should get about a 60% increase in hoop
stress on the inside of the turn in the aorta compared to as
opposed to the outside. I am neglecting bifurcations and
all the other obvious complications which would make it very
difficult to solve this precisely. But this indicates, for
a crude approximation, how we could view the difference in
uptake, for example, between inside and outside of a turn.
I do not intend that this be considered complete but it does
indicate an interesting trend.

The next model I want to talk about is a model of a T-

branch (Fig. 1-22) as representative of something we can
actually calculate.
Model for Stress The increase in vessel stress
Concentration is at the branch point
itself. The Y-branch is also
of interest but is very difficult to calculate. It should be
done experimentally first, I think. Fig. 1-22 represents a
geometry solved in an approximate way and verified with a
photoelastic model. We may view this, for example, as the
aorta with the intercostal branchings. This has been solved
for a number of ratios of parent vessel radius to daughter
radius, and as a function of parent vessel thickness to
daughter thickness. In this particular solution it was
assumed that there was no strengthening of the material
locally at the bifurcation, so naturally we would expect
some elevated stresses. For example the hoop stress N*,
which represents the membrane stress at the point at which
these two cylinders intersect is quite high. And in fact
one can picture this mechanically by considering the effect
of a "hole" in the parent tube. Slice it along the bottom
surface longitudinally, open it up, and then stretch it.
42 CHAPTER 1

We have an equivalent to what the engineering literature

calls stress concentration due to a hole in a plate. The
stresses at the hole are a function of position around the
hole. The membrane stresses are most severe at point B.
The theoretical solution for a plate predicts an increment
of stress at point B of three times the hoop stress far away
from the junction. Again, however, I must emphasize that
this assumes there is no extra material thickening at the
junction. Table 2 gives the stress concentration factor K.
(30). K here represents the ratio of predicted stress at
point B from that which would occur if there were no junction
present. The numbers are, I think, interesting. These are
certainly factors which seem to be more important than
taking account for example, thick vessel effects or vessel
inhomogeneity which have been suggested as important to
atherosclerosis. These results suggest to me that it would
be worthwhile to do some geometric mappings of bifurcations
and make models of these materials to find out what is the
stress distribution at these sites because it looks to me as
though they are also points where there is augmentation of
shear stress and these two mechanical factors may "complement"
one another at bifurcations. The theory shown here relates
only to unstiffened junctions for various radius ratios and
thickness ratios. Results for what might be called structural
reinforcement at the junction have been worked out.

DR. CARO: I haven't studied this in detail but I have

the impression, indeed strong impression, from casts that
the intercostal junction which you mentioned is not really a
T junction. It is a Y junction, it seems as if the daughter
vessel is coming off at right angles to the parent, but in
fact, its initial course is tangential to it, and then it
rapidly turns through 90 degrees.

DR. KENYON: The geometry (Fig. 1-22) is highly idealized

for use in physiological junctions. But I think the numbers
would still be something quite interesting in the more
physiological geometrics.

DR. CARO: I am sure you are right, but would that

modify your conclusions?

DR. KENYON: Yes, very much. The angle of bifurcation

is a very strong factor in the stress concentrations. A Y
configuration would reduce the stress on the inside of the
bend. Thank you for bringing that out.
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 43

TABLE 2

MalO-vessel
Diameter Stress thlC'kness Deviation
K I (exp.) K-K1'"'o
ratio ratio ratio K (theory) Model
.B/ a'" .B T / a"', 2a.\t/T [t. 2] K
I

0.50 0.98 13.4 3.74 3.50 7 E-4

0.50 0.99 13.2 3.73 3.50 7 E-4B
0.50 0.98 13..1 3.73 3.74 0 E-4E
0.80 1.00 13.0 4.32 4.10 5 C-7A
1.00 1.00 13.0 4.40 3.90 13 C-8A
0.28 1.23 26.9 3.54 4.00 -12 C
0.63 0.92 19.0 4.52 5.00 -10 R
1.00 1.00 19.0 5.06 5.43 -7 S
0.34 1.02 98.0 6.53 5.70 15
0.29 0.56 13.1 2.66 3.32 -20 E-l
0.29 0.59 12.9 2.67 2.90 -8 E-2
0.29 0.58 13.0 2.67 3.28 -18 E-J
0.29 0.56 13.3 2.67 3.36 -20 E-7
0.57 0.41 13.1 2.56 2.70 -5 C-5H
0.30 0.48 26.9 3.24 3.00 8 B
0.50 0.512 30.0 9.02 8.33 8 C-l
44 CHAPTER 1

Arteries do have their permeability, and they also have

their ability to withstand normal force, and when you put
these two together you come
Contribution of Arterial out with a model that was
Permeability originally developed in soil
mechanics, incidentally, called
Consolidation Theory. This means that if an elastic material
with a pore structure squeezed it, it doesn't empty spontaneously.
It takes time for the liquid to squeeze through the pores
and the time it takes is a function of how fine the pores
are. If you model the artery as a two-phase material with a
compliance and permeability of correct order of magnitude,
one estimates that it takes a time on the order of several
minutes at least, to release water from an artery when
suddently squashed and this is due to its relatively fine
pore structure. The consequence of this is that the volume
change is really very small during normal rates of application
of pulsatile blood pressure. But that volume change, if it
occurs at all, is going to be confined to those areas of the
vessel that have, what you might call, free access to fluid
the lumen end at the capillary network between the media and
adventitia. This suggests that in pulsatile studies there
may be a time dependent volume change locally of a very
small amount, such that if you are interested in mass transfers,
this could be significant. Now why is it significant?
Because it is a small distance through which these volume
changes are propagated. Incidentally, it is governed by a
diffusion equation. The volume change within an artery
satisfies the diffusion equation. And yet the diffusion
constant itself is on the order of ten to the minus 6 square
centimeters per second. So during one heart beat the diffusion
distance is the square root of this diffusion coefficient
divided by frequency. The volume change during each heart
beat will propagate about ten microns.

This is a theoretical calculation. It reveals that the

region undergoing volume change is such a small fraction of
the vessel thickness that the incompressibility assumption
would be a good one. Yet having the ability to distribute
forces that vary significantly over a ten micron distance,
means that I will have a potential for a large transfer via
the water flux at those points and only at those points. If
you have ever done any heat transfer problems, here is an
analogy. In heat transfer we might think of how much heat
flux occurs if sinusoidal temperature variation of high
frequency exits at the surface.
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 45

If the material has a fairly low thermal diffusion constant,

it means that all the temperature gradients are established
over a very thin surface layer and that means that the heat
flux which accomapnies that applied temperature difference
is going to be extremely large. The same concept, at least
in theory, applies to the artery. By establishing significant
pressure differences over small thicknesses we have in
principle a mechanism for very high augmentation of water
flux during the heart beat.

All of the approximations which have been made here

are important because the artery is not homogeneous. If
there is any surface resistance to the water flow this would
be extremely important to account for. The model would then
be an approximation for a vessel denuded of endothelium.

The system of hoop stress and liquid stress which is

produced by this model is shown in Fig. 1-23 The thickness
to radius ratio is denoted by ( 3 = h/R) With a two-phase
view of an artery we have to somehow distribute this stress
system to the fluid and solid constituents. If you assume
that the solid is both isotropic and elastic, then there is
an incremental hoop tension of amount. There is a radial
compression ofAP/23 amount. On the other hand there is a
system of fluid stresses and, of course, the stress in fluid
is the same in all directions on a bulk basis. It is also
of amount6P/23 and it is a tension. It is extremely important
to realize that if I apply internal pressurization AP
against the lumen that I actually reduce the fluid inside
the vessel, (where there is no significant amount of fluid
flow). The pressure drop occurs over thickness, which
is on the order of ten microns. I think you can see what is
happening. We have the potential for a highly augmented
plasma flux within this superficial intimal layer.

These results are summarized in Fig. 1-24. We have at

the lumen of the blood vessel an applied pressure increment
~ P, and I have normalized the associated incremental
liquid pressures,4P, within the artery to the magnitude of
applied pressure, 6 P. P rapidly drops to a value which is
about -511 P. I t stays constant throughout the thickness of
the vessel because for those deep regions of the vessel no
significant volume change occurs in the time between heart
beats. There is not enough time for significant fluid flux
to occur. During diastole, on the other hand, applied
pressure is reversed inside and according to this model,
plasma flows back to the lumen. Mechanically, this is due
to a kind of "wringing out" of the surface layer during
diastole.
46 CHAPTER 1

e
e
e
+

B B
incremental fluid incremental matrix
stresses stresses

B :: t::.P/2r.
Fig. 1-23: Incremental stresses (two phases).

.
p

1I
I
"0 _________
I
diastole ,
\
\
\
o.-..-..-..-..-..-..-------~--
-1 distance

systole

f inti.l
lutnen t vasa
VlSO.,.

Fig. 1-24: Pulsatile plasma pressure distribution.

ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 47

So there is with this model a sponge-like property of

the intima which could be quite relevant to plasma wall
flux. I would like to leave time for questions because I
think apart from its highly speculative outlook, I would
like to convey to you that even though we think that the
permeability is controlled by events on a molecular scale,
there are circumstances where this motion could be affected
by plasma squeezing. Thus, continuum mechanics at this
stage can offer some interesting if not formative descriptions
of what we might call vessel rheology. This is only a start.
The problems yet to be solved depend very intimately on
detailed architecture of the wall at the point under investigation
and the magnitude of the forces in the longitudinal direction,
which I do not yet know how to estimate.

DR. STONE: Peterson (31) showed quite a few years ago

that there was about a one percent change during the cardiac
cycle in the radius of the carotid artery in the in situ-
artery. Conversely, Arndt (32) showed to the contrary that
there was an average of 15 percent change in that same
location, but the artery had not been handled before measurement.
So we are now not only talking about in vitro, in vivo, in
situ but intact as well. There seemed to be a large discrepancy
in these radius change figures from the intact to in vitro
experiments.

DR. KENYON: I am aware of the discrepancy and I

failed to point out that in fact the X-ray and ultrasonic
determinations of compliance for a cat, at least, seemed to
indicate that it is more flexible in the body than out. But
on the other hand, I think Strandness has done some measurements
which indicate that there is not that mucb difference. I
still think that there is a possibility that the Arndt data (32)
on cats are not necessarily invalid because there is yet to
be an exact determination of in vivo compliance data. My
intuition suggests that, if anything, the vessel should be
stiffer circumferentially in situ and in vivo than in vitro.
I don't know ••• There is an interaction between the radial
and the longitudinal behavior and you have to snip a significant
amount of tissue away from the artery to make your in situ
determination, and that in itself may affect the stretch
during pressure excursions.

DR. STONE: You will get a tremendous vasoconstrictor

response when you manipulate most blood vessels. It lasts
for extended periods of time
Vasomotor Effects so, to me, anyway, it seems
like my reaction would be just
the reverse of yours.
48 CHAPTER 1

That the vessel would be much more compliant normally than

it is once it has been manipulated.

DR. KENYON: I think it is a question of how much more.

I think there are compliance data where it was about twelve
to one, I don't think it was quite that bad •••

DR. STONE: Arndt's data showed some 15 percent change

in carotid radius with the cardiac cycle as opposed to 1
percent for Peterson.

DR. KENYON: It is believed, but I am sure more people

here can testify than I can, that to make a diameter determination
in a vessel undergoing pressurization, the vessels don't
remain round and the action of your gauge is sometimes
uncertain •••

DR. NEREM: Would you comment on the difference between

the coronary system and the aorta from the viewpoint that
you are bringing us?

DR. KENYON: I am excited about the potential importance

of strain on coronaries because it seems as though the
distributing arteries and coronary vessel remain relatively
disease free when they dive into the myocardium. This
therefore, suggests that if the artery is free to distend
under pressure, it might undergo larger changes in longitudinal
or axial strain, than when it is imbedded in a highly dense
tissue. But I am only speculating. In terms of this permeability
process, it is going to be very difficult to establish the
pressures within the arterial wall by any experimental means
I can think of.

DR. COX: I had a couple of comments, with regard to

your questions about deformation, the magnitude of deformation,
and comparing Arndt's and Peterson's work. (32,31) There
are many factors that can contribute to the determination of
values of circumferential strain during the cardiac cycle.
One of the most important of these is the value of mean
pressure that exists at the time of measurement. Peterson's
measurements were in an anesthetized dog. As soon as you
start tampering with an unanesthetized animal, you get
moderat1y severe hypertension, 130 to 150 mm Hg arterial
pressure are not unusual. This would result in an increase
in stiffness due to the nonlinear nature of vascular elasticity
and you would certainly expect a smaller value of wall
strain.
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 49

With regard to Arndt's work, he used two types of measurements,

one using an ultrasonic transducer on the carotid artery
which he subsequently used on the aorta, and for some other
measurements he used radiographs. He measured wall diameter
from radiographs at different pressures and used these data
to define values of strain. To some extent there are methodolo-
gical reasons for the differences in wall strain which he
reported to an extent in comparing Arndt's data with other
data in the literature especially data on the thoracic
aorta, the effects of thoracotomy, of surgery in general, of
trauma, all contribute to a determination of the value of
radial wall strain both directly and indirectly. You also
have to consider that the value of the pulse pressure which
is dependent on the physiological state of the animal, also
is affected by thoracotomy, by fluid loss, the state of
hydration, by the nature of blood loss in an animal. These
factors also contribute to wall strain. So just saying wall
strain is 5% or 10% is really meaningless because of these
large numbers of factors. It must be related to both mean
and pulsatile pressure under the conditions of measurement.

The other comment I had was concerned with the continuum

models you were talking about. One of the things I think
you have to remember is that the blood vessel is an amazingly
complex structure that has an autoregulatory capacity.

By autoregulation I mean a structural autoregulatory

capacity. The smooth muscle cells apparently respond to the
state of wall strain, or
Autoregulation rate at which connective tissue
elements are synthesized and
extruded into the extracellular space. One of the determinents
of that rate of synthesis is growth. During growth and
development when there are large increases, for example,
in bone size, I am sure you can expect increases in the
axial strain of blood vessels in the axial direction through
increased connective tissue synthesis. Also in some of your
models with complicated geometry and stress concentrations
at certain points, one would expect that smooth muscle cells
in those areas that sense the elevated wall stress or elevated
deformation would increase locally the synthesis of connective
tissue proteins, to minimize differences in wall stress.
This would be expected to be true at complicated geometry
such as branching and curvature. Also, the movement of the
heart and the associated deformation of coronary arteries
with cardiac contraction would produce time varying changes
in deformation of coronary arteries. These may also be
signals to coronary artery smooth muscle which would control
connective tissue synthesis.
50 CHAPTER 1

So, for example, in aging as we get dilatation of the

cardiac chamber there may be different signals impinging on
coronary artery smooth muscle cells that may regulate their
synthesis of connective tissue proteins and this, instead of
having a negative feedback effect may produce a positive
feedback effect at least in the direction of coronary artery
pathology.
DR. KENYON: I think I am a convert to your point of
view in the sense that it would appear that we would really
like to know something more about the actual stress distribution
within the arteries as they exist. I am perfectly sympathetic
to your qualifications that must be put on these models. I
would reiterate, they were only presented, not in completeness
but because I hope they will stimulate further serious
investigation of what I consider physiologically important
applications of continuum mechanics to rheology of blood
vessels. The finite strain models are interesting in a
sense but they are not, of themselves, necessarily going to
shed light on how the day-by-day physiological state of
arteries is affected by stress. That is just a personal
point of view. In other words, the natural state of the
artery is not something, that to my mind can be achieved by
letting it come to rest in a beaker.

DR. CARO: I believe I found a slight discrepancy. I

think you said that you account for the observed water
transport by vesicular movement. But my unpublished data on
sodium acetate transport, indicate that it is transported at
a far higher rate than protein and because of its small
size and low lipid solubility I expect it to be transported
mainly via the endothelial inter cell clefts rather than the
vesicles. Thus, I suspect that vesicle transport is by
comparison quite slow.

DR. KENYON: Perhaps I was misunderstood. A long time

ago I made the calculation but I assumed that what was being
transported was some volume of material of a water-like
nature. When I used the average vesicular density in the
apparent front movement (as pointed out in the transmission
electron micrographs of Pallade) that it would appear to
account for most of the water flux, not necessarily the
protein flux. It was a very crude calculation and it was
only interesting in comparison to measured water fluxes.
That is really what I was estimating, not the protein.
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 51

DR. WEINBAUM: We have been interested, as you might

guess, in the viscoelastic behavior of the endothelial
cells. At least three different
Viscoelastic Behavior relaxation times are involved.
of Endothelial Cells There is going to be a viscous
motion of intracellular fluid
contents and an associated relaxation time. There is also a
membrane flow, relaxation time, and finally a longer time
volume change effect of the nature that you are talking
about. What I find really interesting is that the relaxation
time you find in your macroscopic models is of the order of
half a second or a second. This time is of the same order
that Shu Chien and Dick Skalak observe in their studies of
the viscoelastic behavior of a red cell. In the red cell
this time is associated with membrane flow. My intuition
would be that the behavior you are observing that corresponds
to that is viscoelastic membrane behavior integrated now
over many cells. I would like to hear your own thoughts on
this.

DR. KENYON: Could be. But I just thought it could be

due to internal mechanisms in smooth muscle itself. Whatever
its source, it appears to be associated with the muscular
constituent because it appeared as though it had the most
dramatic and long-lasting relaxation pattern compared to,
for example, the more pure collagen type tissue or the more
pure elastintype tissue, which apparently have a shorter and
less dramatic relaxation problem. I don't know what the
relative role of the three major constituents is. Let us
not forget mucopolysaccharide. I think to some extent it
might contribute to viscoelastic phenomena on perhaps a
longer time scale.

DR. DEWEY: One of the interesting things to come out

of Dr. Kenyon's calculations is that there is, by virtue of
the structure of the artery, a tremendous potential to drive
the fluid into and out of the vessel. I emphasize the word
potential because in the presence of an intact endothelial
barrier it appears that the transport mechanisms are reduced
by as much as one to two orders of magnitude. In the absence
of an intact endothelial barrier, however, the pressure
gradients which are produced by the stretch in the artery,
and the fact that the fluid has to enter the artery to
accommodate this incompressible condition, movement of water
would be relatively great, perhaps one to two orders of
magnitude greater with every heart beat but over a distance
of only a few tenths of a micron at the most.
52 CHAPTER 1

DR. KENYON: I would suggest that if endothelial cells

were ever susceptible to the ballooning mechanism that I
just illustrated you might well get a little more water flux
there than you would account for if you just used the mean
driving force for fluid flow. Apparently you only need to
move them apart a little bit to obtain fluid flow but when
they are on the order of 40 to 50 angstroms at their junctions,
it is very difficult to get through them at high rates. In
fact, here again you can estimate what the typical mean
pressure drop is across the endothelial cell for the observed
flux and it turns out to be in the neighborhood of 10 millimeters
of mercury, a significant fraction of the total plasma
pressure drop which may well occur across the endothelial
junction.

BIBLIOGRAPHY

1. Mitchell, J.P.A. and Schwartz, C.J.: In "Arterial Disease"

Blackwell Scientific Publications, Oxford, England,
1965.

2. Gerrity, R.G. and Cliff, W.J.: The aortic tunica media of

the developing rat. I. Quantitative sterologic and
biochemical analysis. Lab. Invest. 32: 585-600, 1975.

3. Gerrity, R.G. and Cliff, W.J.: The aortic tunica intima in

young and aging rats. Exp. Molec. Pathol. 16: 382-402,
1973.

4. Cliff, W.J.: The aortic tunica media in aging rats. ~

Molec. Pathol. 13: 172-189, 1970.

5. Schwartz, S.M. and Benditt, E.P.: Cell replication in the

aortic endothelium: A new method for study of the
problem. Lab. Invest. 28: 699-707, 1973.

6. Jaffee, E.A., Nachman, R.L., Becker, C.G., and Minick, C.R.:

Culture of human endothelial cells derived from
umbilical veins. J. Clin. Invest. 52: 2745-2756, 1973.

7. Morrison, A.D., Berwick, L., Orci, L., and Winegrad, A.L.:

Morphology and metabolism of an aortic intima-media
preparation in which an intact endothelium is
preserved. J. Clin. Invest. 57: 650-667, 1976.
ANATOMICAL AND PHYSIOLOGICAL CHARACTERISTICS 53

8. Gabbiani, G., Badonnel, M.C., and Rona, G.: Cytoplasmic

contractile apparatus in aortic endothelial cells of
hypertensive rats. Lab. Invest. 32: 227-234, 1975.

9. Majno, G.: Ultrastructure of the vascular membrane. In

"Circulation." Handbook of Physiology. Amer. Physiol.
Soc., Washington D.C., Vol. 3, 1965.

10. Rhodin, J.A.G.: In "Histology. A Text and Atlas."

Oxford University Press, New York, pp. 331-370, 1974.

11. Sengel, A. and Stoebner, P.: Golgi origin of tubular

inclusions of endothelial cells. J. Cell Biol.
44: 223-226, 1970.

12. Astrup, T. and Buluk, K.: Thromboplastic and fibrinolytic

activities in vessels of animals. Circ. Res.
13: 252-260, 1963.

13. Warren, B.A.: Fibrinolytic properties of vascular

endothelium. Brit. J. Exp. Pathol. 44: 365-372,
1963.

14. Leak, L. V. and Burke, J.F.: Ultrastructural studies on

the lymphatic anchoring filaments. J. Cell Biol.
36: 129-149, 1968.

15. Ts'Ao, C. and Glagov, S.: Basal endothelial attachment.

Tenacity at cytoplasmic dense zones in the rabbit
aorta. Lab. Invest. 23: 510-516, 1970.

16. Moss, N.S. and Benditt, E.P.: The ultrastructure of

spontaneous and experimentally induced arterial
lesions. Lab. Invest. 23: 231-245, 1970.

17. Doyle, J.M., and Dobrin, P.B.: "Finite deformation

analysis of the relaxed and contracted dog carotid
artery." Microvascular Research 3: 400-415, 1971.

18. Bergel, D.H.: "The dynamic elastic properties of the

arterial wall." J. Physiol. 156: 445-457, 1961.

19. Azuma, T., et al: Viscoelastic properties of large arteries

Proc. 5th. International Congress of Rheology, Vol. 2
University of Tokyo, 1969.

20. Patel, D.J., et al: Static anisotropic elastic properties

of the aorta in living dogs. Cire. Res. 25: 765-779,
1969.
54 CHAPTER 1

21. Moritz, W.E., and Anliker, M.A.: Wave transmission

characteristics and anisotrophy of canine carotid
arteries. J. Biomechanics 7: 151-154, 1974.

22. Patel, D.J., Mallos, A.J., and Fry, D.L.: Aortic mechanics
in the living dog. J. Appl. Physiol. 16: 293-299,
1961.

23. Patel, D.J., and Fry, D.L.: Longitudinal tethering dogs.

Circ. Res. 19: 1011-1021, 1966.

24. Patel, D.J., and Fry, D.L.: The elastic symmetry of arterial
segments in dogs. Circ. Res. 24: 1-8, 1969.

25. Patel, D.J., Janicki, J.S., and Carew, T.E.: Static

anisotropic elastic properties of the aorta in living
dogs. Circ. Res. 25: 765-780, 1969.

26. Patel, D.J., Janicki, J.S., Vaishnav, R.N., and Young, J.T.:
Dynamic anisotropic viscoelastic properties of the
aorta in living dogs. Circ. Res. 32: 93-107, 1973.

27. Vaishnav, R.N., Young, J.T., Patel, D.J.: Nonlinear

viscoelastic theory for large blood vessels.
Proceedings of the International Congress on
System Dynamics, Valley Forge, Pa., 1975.

28. Patel, D.J., and Vaishnav, R.N.: Rheology of large blood

vessels. In Bergel, D.H. (Ed.): Cardiovascular
Fluid Dynamics. Academic Press, Inc., London, 1-64,
1972.

29. Patel, D.J., and Vaishnav, R.N.: Mechanical properties of

arteries. Proceedings of the NATO-Advance Study
Institute in Cardiovascular Fluid Dynamics,

30. Lind. N.C.: An elastic analysis of the stress concentration

of a pressurized T-branch pipe connection, in First
Intll Conf. Pressure Vessel Technology, Part 1, AS ME ,
269-275, 1969. Approximate stress-concentration
analysis for pressurized branch pipe connections, in
Pressure Vessels and Piping: Design and Analysis
Part 2, ASME, 952-958.

31. Peterson, L.H., Jensen, R.E., and Parnell, J., Mechanical

properties of arteries in vivo. Circ. Res. 8: 622, 1960.

32. Leitz. K.H., and Arndt. J.O.: Die Durchmesser-Druck-

Beziehung des intakten Gefassebietes der A. carotis
communis von Katzen. Pf1ugers Archiv. 301-50. 1968.
Chapter 2 FLUID MECHANICS OF ARTERIAL FLOW

DR. DEWEY: The primary objective of my talk* today is

to summarize those aspects of fluid mechanics that relate to
arterial disease. Some simple quantitative estimates of
fluid shear stress will be made and the effects of flow
pulsatility and vessel geometry will be noted. I hope
specific topics that warrant additional research will be
suggested.

In viewing the great body of literature that exists on

arterial hemodynamics, one comes face to face with the fact
that a large fraction of the existing literature is specifically
limited to either a particular artery or a particular experiment
in vitro for which the fluid dynamicist has some hope of
making theoretical interpretation. I emphasize this because,
if we are really serious about understanding arteriosclerosis
and possible exacerbating effects arising from hemodynamics,
we have to talk about complex arterial geometries for which
accurate theoretical calculations are nearly impossible. We
have to be, I think, more sensitive than we have collectively
been in the past to the full complexity of the problem that
faces us.

For this reason, I have chosen as Figure 2-1 an x-ray

angiogram of the coronary circulation. Understanding this
complicated flow network is
The Coronary Circulation the real problem we are trying
to solve. The coronary circulation
is a very tenuous delivery system. Compromise of one or
more of the small arteries which feed the myocardium can
lead to death. We are talking about a very detailed tree, with
many bends, bifurcations, and collateral branches; if we are
going to make any sense out of this problem, this is the
situation we have to address.

Most quantitative hemodynamic information that is

available refers to the larger isolated vessels such as the
aorta and the femoral arteries. I would like to take the
point of view that such information is useful insofar as it
helps us to understand more complicated problems such as the
coronaries.

*The research reported in this talk was sponsored by

the National Heart, Lung, and Blood Institute, Grants HL14209
and HL2l859.
55
56 CHAPTER 2

Fig. 2-1: X-Ray angiographic picture of the human coronary

circulation. The angiogram was taken using an excised human
heart (from Fulton (52) p. 164).
FLUID MECHANICS OF ARTERIAL FLOW 57

In spite of these pejorative statements, of necessity

one must refer to experiments in the large isolated arteries
simply because the amount of data available on the coronary
circulation is woefully inadequate to the tasks that one
would like to confront as a fluid dynamicist.

One encouraging fact is that Drs. Caro, Schwartz, Nerem

and others have been sensitive to this problem and have been
looking at coronary circulation with more than just a passing
glance. One of the things we will talk about later is the
problem of modeling; ways in which information gleaned from
one artery may be used to estimate the flow parameters in a
different artery. In particular, I would like you to be
conscious of the differences
In Vitro Modeling between animal models and the
human circulation. These
differences are often larger than the variations that exist
between animal species or between the human circulation and
a well-designed in vitro model. We are really interested in
the human condition and have to be realistic with regard to
whether or not the animal experiments we are doing are in
elucidation of the human condition or simply of the particular
test animal.

Figure 2-2 was borrowed from McDonald's book (1); it

gives you a simplified overview of the variety of velocities
and pressure waveforms that
Velocities and Pressure one finds in the systemic arteries.
Wave Forms in Systemic Figure 2-3 comes from Mills and
his co-workers (2); it gives a
a sequence of waveforms in vivo in a single human male at
various locations in the systemic circulation. Similar results
have been obtained by other workers, notably Nerem et a1., (3)
and Clark and Schultz (4). The amount of information that
could potentially be generated in experiments of cardiac
output and vessel tone is enormous. When one is trying to
measure a particular physical quantity (for example the wall
shear stress or the relation between velocity waveforms and
local pressure waveforms), one is forced to concentrate on a
small fraction of the total available information and then
work very hard to acheive a consistent physical interpretation
of the data obtained. Examples of such careful work are two
papers by Ling, Atabek and co-workers (5,6) reporting blood
velocity measurements at the aorto-i1iac trifurcation in
dogs. Their results are applicable to the particular state
of cardiac output and muscle tone existing in an anesthetized
animal.
58 CHAPTER 2

--
~
E

-
u
>-
"ou
Qj
>

30

Fig. 2-2: Representative waveforms in the major human

arteries (from McDonald (1) p. 356).
FLUID MECHANICS OF ARTERIAL FLOW 59

!tIGHT SUllClAVIAN .RT.

-0-!70 ]:~nY
,.IIE$$UIIE ~IIESSUItE

c""-.
,......._ - - 0

oUCENOING AORTA
I'IIE$SUIIE

7' ~ ·
1I~1\ ~
VElocn~t
.""_.
o

N!iHl IIfNAl All'.

DESe , _II. I,ll)

v=!l~
---4:1!:...-.
10 ]10 VE~ITY

lllOHT ro.«lN ILIAC AR'.

---- 0 "",.."

MOO, _II. II

=:[~
~IIES$URE

~I=
Fig. 2-3: In Vivo pressure and velocity waveforms in a
human male (from Mills et al. (2».
60 CHAPTER 2

How do we generalize these results to the aorto-iliac

bifurcation of man, or to different states of cardiac output
and flow division between the distal vessels? The velocity
waveforms in these latter instances will have different
harmonic contents and different ratios between maximum and
minimum flow; there may even be retrograde flow in some
circumstances and not in others. In the ensuing portions of
this presentation, I will attempt to illustrate some simple
methods of fluid dynamic scaling and assess their accuracy.

Many investigators, including Dr. Fry and others attending

this Conference, have suggested that fluid shear stresses
acting on the artery
Estimates of Wall wall may participate in the
Shear Stress physical processes which lead to
atherosclerosis. This particular
fluid dynamic parameter has been implicated by in vivo measure-
ments, (7) where very large wall shearing stresses were
chronically or acutely induced and degradation of the arterial
wall structure ensued. The wall shear stress, 'w , is
simply the force per unit area exerted on the artery surface
by the motion of blood adjacent to the surface. This force
on the endothelium acts in the direction of the flow.

Physiological values of Tw (dynes/em 2) are determ!ned

primarily by the time-varying volumetric mean velocity, U (t)
(em/sec), the tube radius R (em),
In Vivo and the geometrical and temporal
history which the flow experiences
both locally as well as proximally to the site of observation.
To a high degree of approximation, blood flowing in the larger
arteries (D > I em) may be regarded as having a Newtonian
viscosity which is denoted as ~ The kinematic viscosity,
~/p is v(cm 2 /sec).

In vivo situations are characterized by secondary flow

and large temporal variations, so that a single set of
parameters, such as U(t), the Reynolds number Re = UD/V,
unsteady, or Womersley, parameter a = R ~ are not
sufficient to characterize the local flow dynamics without a
detailed geometrical description of the arterial lumen and a
specification of the flow variation during the cardiac
cycle. Aortic flow is dominated by a primary systolic flow
pulse containing some 60-90% of the total stroke volume;
coronary flow, on the other hand peaks during diastole and
is strongly influenced by myocardial contractility during
systole.
FLUID MECHANICS OF ARTERIAL FLOW 61

Available evidence (8) (9) (10) suggests that the harmonic

content of coronary flow is at least as high as that of the
major systemic arteries, with significant content in the
first 5 or 6 harmonics of the heart period.

In spite of these regional differences there are many

useful conclusions regarding flow in human arteries that may
be drawn from animal and
Turbulent Flow laboratory experiments. We
may examine possible velocity
profiles and scaling laws for wall shear stress induced by
flow. We may estimate conditions under which disturbed flow,
including turbulent flow, can be produced. And we can suggest
the effects which flow separation may have on fluid properties
at the endothelial surface.

In arriving at predictions of wall shear in various

human arteries, I would like to begin with two very simple
estimates. The first is steady flow in a long straight tube
(Poiseuille flow) for which

Poiseuille

Here U is the volumetric mean velocity (i.e., half the

centerline velocity in Poiseuille flow). This result under-
estimates the shear in regions of reattachment of a separated
flow, on a surface where the adjacent velocity is high
because of a skewing of the velocity profile (i.e., strong
secondary flow), in pulsatile flow, and near the entrance of
a vessel.

A better estimate for oscillating flow (again, a long

tube is assumed) is the Womersley result for flow sinusoidally
modulated at a frequency f = w/2TI The parameter ex ,
which is the ratio of the unsteady acceleration of the fluid
to the viscous forces acting on the fluid, is ex = R~
If one chooses w to be equal to 2TI times the basic
heart frequency fH, typical values of ex for the ma)or
human arteries range from 3 to 15. A more appropr1ate value
of ex for the strongly-peaked waveforms such as those seen
in the aorta would be based on (approximately) the third
harmonic of the heart frequency, and we shall denote this
value of ex'~ which is equal to ~ cx.
62 CHAPTER 2

U a =R..jWlV

,.- - ....

C':~[,:exp-{~a(~)}]
/ For a» 1

( aU)
ay y=o a U
=../2 Rand

Tw=(A) (~) ,URlU

COMPARE P01SEU1LLE:

Tw =4(,)

Fig. 2-4: Diagram illustrating the velocity profile near

the arterial wall in pulsatile flow.
FLUID MECHANICS OF ARTERIAL FLOW 63

For large a , the flow near the wall at peak systole

is described by the equation shown in Figure 2-4, and the
wall shear stress is given by

Womersley T
W
a
= /2 (~ ~ ) .
The appropriate value of U to be used in this relation is
the volumetric mean velocity. For a» 1, the Womersley
relation predicts a larger shear stress than Poiseuille
flow; the numerical factor is (0./12) rather than 4. In the
aorta, where values of 0.* are less than 20, the shear
stress in pulsatile flow would exceed the instantaneous
value given by Poiseuille flow by a factor less than 5. For
the arteries of interest, the coronaries and the carotids,
the effects of pulsatility are much less dramatic because,
in those arteries, an appropriate value of 0. is 3 to 10 and
Poiseuille flow provides a useful estimate of TW The
effects of a distensible arterial wall on the estimated
velocity U (t) and Tw, at least at peak flow velocity, can be
shown to be on the order of the arterial distensibility,
6. R/R, during the pulse.
Except for the pulmonary artery, this effect is unimportant.

Serious errors would result from applying the Womersley

and Poiseuille flow estimates to entrance regions of arteries
and to bifurcations. Distal to rapid changes in cross-
sectional area or changes in flow direction. strong secondary
flows and regions of flow reversal may be produced. yielding
wall shears which exceed the previous estimates by factors
exceeding 5. Further consideration will be given to these
effects shortly.

On the basis of the previous discussion, it is useful

to estimate the limits of Tw which might be encountered in
human and animal arteries, and to compare these values with
the critical shear stress that was found, by Fry (7) and
Carew (11), to produce significant changes in the endothelial
surface of dog aortas. This shear stress, approximately 400
dynes/cm 2, has been referred to as a "critical" stress,
although subtle endothelial changes were observable at much
lower values of shear. Don Fry has commented upon the
effects of long-term "sub-critical" shear in previous publications,
and his observations are most valuable (11,12).
64 CHAPTER 2

Typical peak systolic velocities in the aorta of a

resting man or dog is about U 100 cm/sec, and in the
carotids and femorals it is about U 60 cm/sec.* The
radii of these vessels vary substantially, and typicall values
are as follows:

RADIUS (cm)

Aorta Femoral Carotid

Man 1. 25 0.5 0.4

Dog 0.50 0.3 0.2

For the larger arteries, such as the human aorta, Womersley's

estimate yields (assuming ~ = 0.04 poise, fH = 1 Hz, U
100 em/sec, and a = a*):

T
w
~
,...., 61 dynes/cm 2 .

For the small arteries, such as the femoral, the flow at

peak systole more nearly resembles a Poiseuille velocity
profile as you can see in Figure 2-5. Hence the Poiseuille
estimate is appropriate:

19 dynes/cm 2

*The estimate of 60 cm/sec for the human carotids and

femorals is based on experiments by several authors and
summarized by McDonald (1974). This velocity is larger than
some ultrasound measurements would suggest. (See, for example,
Fronek et al. 13)) as contrasted to the results of Hisland
et al. (14) and Fronek et al. (15). The large discrepancy
between various ultrasound measurements and previous data (1)
remains unresolved.
FLUID MECHANICSOF ARTERIAL FLOW 65

o
I" "
U/UCL.
-0- 1·0
o
o
o

o
R/R W
Fig. 2-5: Predicted and measured velocity profiles in the
femoral artery of a man (from Schultz (53». The solid line
is calculated from Womersley's theory.
 66 CHAPTER 2

Both of these estimates of Tw are less than the so-

called critical shear stress by a factor of more than 6.
The estimates are for resting subjects, and an increase of a
factor of 2-4 is not unrealistic at peak cardiac output.
(But note that carotid flow remains essentially constant,
irrespective of cardiac output or flow distribution to the
abdominal organs.) And the estimates themselves are most
certainly not accurate to better than a factor of 2. But
taking all of these considerations into account, one is
forced to conclude that the only fluid-mechanical mechanisms
available to increase local shear rates to the "critical"
value suggested by Fry are very strong secondary flows, as
in the vicinity of bifurcations, or in flow profiles modified
by atherosclerotic narrowing or separation phenomena.

By way of comparing these crude estimates to actual

DOG in vivo data, I'd like to refer to the experiments of S. C.
Ling (5) in which shear stress measurements were made in the
descending aorta of a dog. Although the details of the
experiment are lacking. we estimate R = 0.5 cm. U = 150
cm/sec,u =0.04 poise and fH = 3 Hz. Again, using a* in
place of a for the aorta, the Womersley model gives

2
= 160 dynes/em

which is in fortuitous agreement with the measured shear

stress of 160 dynes/cm 2 •

Clearly, the higher linear velocity and heart rate of

these dog experiments by Ling et al. (5) yield shear stresses
which are larger than one
DOG Shear Rates would expect in humans. These
measurements made with a hot
film probe, bear on an important point and deserve to be
repeated. Perroneau et al. (16) have used Doppler ultrasound
to measure maximum velocity gradients in the aortic arches
of anesthetized and implanted conscious dogs. Peak shear
rates did not exceed values corresponding to TW = 10 dynes/cm 2
Atherosclerosis is observed in the abdominal aorta where
strong secondary flows and flow separation would not be
expected. And yet in some regions where the shear is known
to be high, such as the carina (or flow divider) side of a
bifurcation, the presence of atherosclerosis is not common
(12).
FLUID MECHANICS OF ARTERIAL FLOW 67

The original in vivo measurements of "critical" stress

by Fry, and the subsequent experimental results of Carew
(11) on excised dogs' aortas, DOG
"Critical" Stress were obtained under conditions
producing laminar flow. No
analogous experiments have been performed to determine if an
equivalent "critical" stress exists for human endothelium,
or if the "critical" value is influenced by the presence of
substantial turbulence or other disturbed character of the
flow.

To conclude this discussion of shear stress estimates,

let's return to the major coronary vessels and estimate the
shear stresses there. Figure
Shear Stress in 2-6, taken from Gregg's useful
Coronary Vessels book (8), illustrates the type
of flow rate variation one
may expect in the coronaries during the cardiac cycle. Taking
R = 2.5 mm, a peak velocity of 50 cm/sec, and an effective
value of a of 5 (10) we conclude from either the Poiseuille
or Womersley formulas that the wall shear stress is on the
order of 30 dynes/cm 3. This is some 10 times smaller than
the "critical" shear. Even during extreme exercise, critical
values would not be approached. However, in Figure 2-1, you
will recognize that the major coronary arteries twist, turn,
and branch profusely in their course over the myocardium. Clearly,
the peak shear stress will be strongly increased by secondary
flows and may exceed these simple estimates by a substantial
amount.

I want to proceed now to a discussion of flow separation

and secondary flow. Frequently the two effects occur at the
same location, as has been
Significance of observed at bifurcations. Because
Bifurcations of the importance which has been
attached to atherosclerosis
occurring near bifurcations, many ill vivo and in vitro
experiments have been performed on bifurcation geometries.
Some of the particular experiments which come to mind are
those of Gutstein and Schneck (17), Ferguson and Roach (18),
Attinger (19), Rodkiewicz and Roussell (20), Fuerestein et
(21), Freedman (22), and Ling et al. (5). There are other
data and calculations which will be mentioned later during
the talk. But to fix some of our quantitative ideas, let's
look at Figure 2-7 which contains data from Brech and Bellhouse
(23). The model was a symmetric Y-branch with a 90°
included angle between the two branches and an area increase
from parent vessel to two daughter vessels of 1:1.12.
68 CHAPTER 2

10
..,., ..
,
'
20

,,' FLOW

0 10

O~~--~------~~

LEFT CORONARY ARTERY RIGHT CORONARY ARTERY

Fig. 2-6: Flow rates in the coronary arteries (from Gregg (8)).
Reconstruction and comparison of typical fl ow curves obtained with
the orifice meter from the anterior descending branch of the left
coronary artery in a small dog and from the right coronary artery
in a large dog. AP, aortic pressure eF, coronary inflow. Ordi-
nates, upper, mm. Hg; lower, flow in cc. per min. Vertical inter-
cepts demarcate systole and diastole. Dotted lines, predicted
intramural velocity curves.
FLUID MECHANICS OF ARTERIAL FLOW 69

10
9
8
7

.
'\
,,: \ ,,...
6
T
TO 5
4 "', 0

3 ..........:::::::----.-O-INSIDE (CARINA)
2 ~.- TOP/BOTTOM
I ~'------O-OUTSIDE
0
0 I 2 3 4
DAUGHTER TUBE DIAMETERS
DISTAL TO POINT OF CARINA

Fig. 2-7: Wall shear stress distribution in a 90 0 y- tube

with equal branches (from Brech and Be11house (22)). Cross-
sectional area increases (1:1.12) at bifurcation.
70 CHAPTER 2

The abscissa is distance along the daughter vessel, measured in

daughter diameters, and the ordinate is local wall shear stress
T normalized to the upstream (parent) shear stress To.

The shear stress on the flow divider or carina side of this

bifurcation is substantially elevated -- by factors in the
vicinity of 10. On the other hand, on the outside surface of
this region you see a dramatic drop in the wall shear stress,
and studies that have been done with appropriate instrumentation
have actually demonstrated a reverse flow and the shear stress
in the opposite direction from the main flow.

In these bifurcations, we find substantial secondary flows,

portions of the wall that experience very high shear rates
(factors anywhere from three to ten higher than you would
estimate on the basis of a simple undisturbed axial-flow
model), and the potential for separation and substantial
oscillation of the local shear as a function of the cardiac
cycle.

Let me take a moment to discuss the difficulty of

obtaining results similar to those shown in Figure 2-7 from
a theoretical calculation. In a branching vessel with real
physiological contours and pulsatile flow, you have all of
the complications which a theorist would like to avoid.
Even recent calculations such as those of Kandarpa and
Davids (24) and O'Brien et al. (25) have, of necessity, made
dramatic simplifications of the actual geometry and flow
characteristics to allow numerical solutions of the Navier-
Stokes equations with a reasonable amount of computer time.
To obtain accurate values for velocity profiles and wall
shear. it is necessary to take very closely-spaced grid
points in the numerical calculations, both in space and
time, and also to extend the physical domain covered by the
calculation to include substantial distances both proximal
and distal to the site of interest. These calculation are
very expensive to program and debug, and are also expensive
to run on a computer. An
Reynold's Number additional limitation of substantial
Calculations importance is that the calculations
become geometrically more difficult
as the Reynolds number R = 2UR/v increases. Perhaps
for the coronary arteries where the Reynolds numbers are on
the order of 100-300, such calculations might be constructive.
But for these calculations to be meaningful, the complicated
geometries which we saw in Figure 2-6 would necessitate very
small calculation steps in time.
FLUID MECHANICS OF ARTERIAL FLOW 71

As an experimentalist, I find in vitro model studies

much more exciting and efficient, and an absolute necessity
for high Reynolds numbers and complicated geometries. In a
sense, you can view in vitro models as analog computers
which have substantial flexibility and the important advantage
of being able to reproduce turbulence when and if it occurs.
The experimentalist must defend his in vitro models against
a number of criticisms if he purports his data to be
representative of in vivo conditions. Figure 2-6 indicates
that coronary artery flow is not in phase with aortic pressure,
and a substantial contribution to this behavior is the
change in coronary artery cross-section caused by contraction
of the myocardium. A laboratory model with rigid walls
will not model this situation.

But available evidence indicates that the carotid and aorto-

iliac bifurcations do not change shape appreciably during
the cardiac cycle, and rigid models should be reasonably
accurate representations of in vivo conditions. Theoretical
calculations (and I include here dimensional analysis) are
invaluable in estimating the quantitative values of the
parameters to be measured, in assessing potential systematic
experimental errors (including the effects of unsteady flow
and wall distension), and in scaling experimental results to
conditions that differ from those used to obtain the experimental
data.

Let's return to Figure 2-7 to examine the details of

the flow in bifurcations and the complicated patterns which
arise in realistic geometries. For purposes of our discussion
I'm going to limit my presentation to in vitro data obtained
in our laboratory by Scott Smith (26), even though a number
of the phenomena he observed have been described by previous
authors (17, 18, 20, 21, 23, 27, 28, 29, 30).

First, let me describe the experimental techniques

which were used. Figures 2-8 and 2-9 show, respectively,
the overall experimental arrangement and cross-section of
the glass model bifurcation which was used. The experimental
protocol is as follows: first, the flow is clamped off and
a fluorescent dye is injected into the model and mixed with
the working fluid. A laser beam is spread out into a thin
sheet of light by a cylindrical lens and passed through the
center of the model so that the laser-induced dye fluorescence,
as seen from above, gives a clear cross-section of the model
(31) •
72 CHAPTER 2

inflow

to the
video
tape
recorder

laser

cylindrical
focuelns len.e
lenee

o
Greduatlns cylinder
and Itop witch for
flow meaBurementl

Fig. 2-8: Facility for laser fluorescence visualization of

flow in a model carotid bifurcation (Smith (25».
FLUID MECHANICS OF ARTERIAL FLOW 73

second /
boundary

Fig. 2-9: Outline drawing of the glass model bifurcation,

illustrating the observed secondary flow patterns (Smith (25)).
74 CHAPTER 2

Now we start the flow; as fresh fluid enters the model,

the fluid containing fluorescent dye is washed out, first
in the regions where the velocity
Steady Flow is highest and later in regions
where the velocity is low. So
in the first instant, what one sees is essentially equivalent
to a cross-section of the model artery just as one would
obtain with an x-ray angiogram (32) and as the flow continues
streaks of dye will be formed which will move with the fluid
and describe the complicated motions across the entire
vessel section. The flow pictures in the next 5 figures
were taken with a camera photographing the plane of fluorescence
from above, although we have also used a TV camera to record
the flow with excellent results.

Figures 2-l0a, 2-l0b, 2-l0c, and lOd, in chronological

order, show how the flow starts from rest and washes out the
dye. The first is with no flow. We now start the fluid
motion; the Reynolds number is 1000, characteristic
of peak systolic flow in the carotids, and the flow branching
ratio at the bifurcation is (Qexternal/Qinternal = 1.3,
which is the physiological ratio for an undiseased artery.
By the time the fluid has moved 10 or so diameters along the
artery, the flow field has settled down to a quasi-steady
state and subsequent motion of the dyed fluid presents a
very useful picture of the dynamics of the fluid motion.

In Figure 2-11, we have another picture at Re = 1000

and QE/QI= 1.3, selected to most clearly illustrate the
complicated patterns we see.
Separation of Flow There are several features
that I want you to focus on.
First, separation occurs on the outise of both branches, not
just the internal carotid branch. The second remarkable
feature is that the separation regions extend a long way
down the daughter vessels, which was indeed a surprise to us
when we first observed this. A third feature, which was
anticipated from the detailed velocity profiles obtained by
Brech and Bellhouse (23), is the very high-velocity flow on
the carina (bifurcation) side of each branch. The dye is
very rapidly washed away from this region, and the resulting
shear stresses are extremely high as we observed in the
previous Figure 2-7.

Our conclusion from these preliminary studies (and I

want to emphasize that much work remains to be done) is that
the detailed geometry, branching ratio, and Reynolds number
are very important to the local shear stresses along the
external sides of the bifurcation in both branches.
 FLUID MECHANICS OF ARTERIAL FLOW 75

Fig. 2-10: A sequence of laser fluorescence pictures showing

washout of dyed fluid in the carotid bifurcation model
(Smith (25)).
76 CHAPTER 2

I emphasize particularly the geometry because the complicated

vorticity pattern you see in the carotid sinus (in Figure 2-
11) would be entirely different. both qualitatively and
quantitatively. if we had a straight tube without a sinus.

Now all of the fluorescence pictures I've shown you so

far have been for steady flow. Scott Smith (26) a preliminary
study of the vorticity patterns
Pulsatile Flow in the sinus with a pulsatile
flow. The technique was to use
a peristaltic pump instead of a constant-head gravity feed
as the flow source. The temporal waveform was far from
physiological for the carotids (it looked. in fact. more
like the coronary waveforms we saw earlier in Figure 2-6,
but the qualitative results are interesting nonetheless).

Without going into a lot of detail. Scott's conclusion

was that the flow near peak systole was quasi-steady; by
that. I mean that the vorticity patterns in the carotid
sinus at peak systole are. for all intents and purposes, the
same as those one sees in a steady flow at the same Reynolds
number and branching ratio.

DR: WEINBAUM·: Let me understand your use of the term

"unsteady" here. The whole experiment is based on a transient
effect, and you have used the term quasi-steady in two
different contexts.

DR. DEWEY: I apologize if this has been confusing.

When the flow starts from rest and instantaneously rises to
a steady value. this produces
Steady vs. Unsteady Flow in the first instant, a starting
vortex phenomenon which rapidly
washes downstream. It is analogous to the old aerodynamics
problem where an airplane accelerating off the runway
leaves behind a trailing vortex but very rapidly this vortex
becomes unimportant to the steady lift on the plane. The
flow field in the vicinity of the plane is then considered
steady. By analogy, after the starting vortex has passed
through the model artery, the remaining dye is set into
motion in a manner which, for steady inlet flow, persists in
time. Eventually, the dyed fluid washes out and is replaced
by undyed fluid. but the detailed fluid motion persists and
is unchanging in time. Even though the experimental technique
is based on a transient phenomenon, the flow patterns are
essentially steady, hence the use of the term quasi-steady.
FLUID MECHANICS OF ARTERIAL FLOW 77

Fig. 2-11: Flow in the model carotid bifurcation at a proximal

Reynolds number of 1000 and a branching ratio (Qexternal/Qinternal)
of 1.3 (Smith (25)).
78 CHAPTER 2

With cyclic unsteady inlet flow, the fluid motions

within the model artery go through cyclic variations. The
detailed flow field at any instant depends on the instantaneous
value of volumetric flow as well as the previous time history
of the fluid oscillation. The term quasi-steady in this
instance refers to the fact that the instantaneous flow
patterns quantitatively resemble a steady flow with the same
volumetric flow existing at that instant. We observe this
quasi-steady behavior near peak systole in our models of
bifurcations, as well as our models of stenosed arteries:

DR. COLTON: Dr. Dewey, isn't the tracer dye subject

to dispersion with time, so that the dye doesn't really
follow the fluid streamlines?

DR. DEWEY: That's correct. The saving grace is that

the dye molecular weight is large, on the order of 500, and
diffuses across streamlines only very slowly.

In a steady separated region, for example, some small diffusion

takes place from the flow within the cavity to the external
fluid, and the dye concentration decreases with time. But
this process is quite slow; and, furthermore, we can follow
the dye patterns over several decades of concentration.

DR. CARO: You haven't told us what the waveform was

for the unsteady tests.

DR. DEWEY: The unsteady flow was created by a peristaltic

pump which produced flow rate variations ranging between 0.5
and 1.5 of the mean during each cycle. The detailed waveform
contained significant harmonics up to about the third. But
the waveform was not modeled after the human circulation and
so these tests were only suggestive of the unsteady contribution
to flow dynamics.

DR. CARO: Why I raised the question was because two

possibly relevant studies have been undertaken: One is by
Howard Lyne (33) and the other is by Peter Blennerhasset
(34) which may still be in thesis form. Howard Lyne looked
at the secondary flow due to sinusoidal oscillations in a
curved pipe and found essentially that this is directed
inwards rather than outwards as in the work of Dean (35)
with steady flow. Blennerhasset studied the secondary flows
due to a combination of steady and non-steady pressure
gradients.

DR. COLTON: Our experiments show very clearly separation

in the internal carotid as you go around the bend.
FLUID MECHANICS OF ARTERIAL FLOW 79

DR. DEWEY: In all of the steady-flow models of bifurcations

which we have studied, separation occurs on the outside of
the branch just distal to the bifurcation. All of the high
Reynolds number numerical calculations I have seen support
this view.

DR. CARO: In fact a very crude experiment was done.

Lyne injected a dye stream and under appropriate conditions
found the secondary flow to be directed inwards, i.e.,
towards the inner wall of curvature of the tube.

DR. DEWEY: I think it's clear that we need more

definitive data to resolve some of these questions. And it
is especially imperative that the geometry of the bifurcation
or branch or aortic arch be modeled accurately.

But let me return to a review of the phenomena we have

observed in our model studies, because there is still one
important point I want to make.

Let's get back to Figure 2-11 which illustrates the

flow patterns we see when the ratio of flows in the external
and internal carotids is 1.3. We're now going to decrease
the branching ratio to 0.73 and then again to 0.40, simulating
what would happen with substantial obstruction of the external
carotid.

Figure 2-12 is for branching ratio of 0.73, and you can

see that the flow patterns have changed dramatically. With
the largest amount of flow
Flow Patterns with going through the external
Simulated Obstruction carotid, we saw very little
detailed vorticity in the
external branch. The flow hugged the carina side of the tube
but the streamlines emanating from the point of separation at
the junction were smooth and laminar. Now we see substantial
vortex shedding and vortex formation at the mouth of the
vessel. We even notice that some of these disturbances
propagate upstream and lift dye filaments off the proximal
tube wall.

Figure 2-13 indicates where the external flow has been

further reduced, the branching ratio being 0.4. The unsteady
disturbances produced near the mouth of the external branch
are so strong that they seriously affect the flow patterns
in the internal branch.
80 CHAPTER 2

Fig. 2-12: Carotid flow as in Fig. 2-11, but with a

branching ratio of 0.73.
FLUID MECHANICS OF ARTERIAL FLOW 81

Fig. 2-13: Carotid flow as in Fig. 2-11, but with a branching

ratio of 0.4.
82 CHAPTER 2

In this figure, we even see what appears to be a small

separation bubble on the carina side of the internal branch,
with a high-velocity jet-like flow near the carina side of
the lumen and adjacent to this region of separation. A
large separation region persists in the sinus of the internal
branch, and the overall sinus flow pattern resembles that
found at equivalent internal branch velocities but with
physiological branching ratios.

These florescence pictures dramatize what I believe to

be a most important parameter in characterizing flows through
physiologically-dimensional
Scale of Vorticity vessels: that is the scale
of vorticity. In several of the
figures, you saw vorticities whose dimensions were some
substantial fraction of the vessel diameter. This is what I
would call large-scale vorticity. Many authors such as Dr.
Nerem, have referred to these patterns as disturbed flow,
and I think that their terminology is also very appropriate.
It is not turbulence in the ordinary sense because the
small-scale vorticity and accompanying high-frequency fluctuations
in wall pressure and wall shear are absent.

The presence of large-scale vorticity in the carotid

sinus is very evident in all of the measurements which O'Brien
Erlich and Friedman (25) have made. They observed that as
the flow velocity was increased and decreased, either with a
slow variation or with the unsteady waveforms produced by
the peristaltic pump, the large-scale vortex pattern in the
sinus expanded and contracted. So if I were an endothelial
cell sitting on the surface of the carotid sinus, I would
see first flow in one direction and then, as the vorticities
within the sinus expanded and contracted, flow in the other
direction, scrubbing first in one direction and then in the
other during the cardiac cycle.

We have no idea what this oscillatory shear stress does to

the endothelium.

In this connection, it is interesting to note that we

have examined some 200 x-ray angiograms of atherosclerotic
carotid bifurcations in
Atherosclerotic Carotid connection with our work on
phanoangiography (36) and in a
substantial majority of these cases we find the focal point
of lesions to be in the sinus area. The plaques build up from
the external sinus wall toward the interior lumen.
FLUID MECHANICS OF ARTERIAL FLOW 83

In many advanced cases, the only open area is a small orifice

on the carina side located where we observe high-velocity
flow in models of the patent artery. I think it is provocative
that the carina side. the region of very high shear stress,
is spared whereas the sinus area, which contains large-scale
vorticities which scrub back and forth during each heartbeat,
exhibits a propensity for lesion formation.

Let me digress for a moment to discuss a method we have

developed for modeling unsteady flows. Figure 2-14a is a
photograph of our pulsatile
Method for Modeling flow source which we have used
Unsteady Flow for several investigations (37).
Our objective was to develop a
system which could be programmed to deliver flow pulses containing
up to 8 harmonics of the fundamental period and could be
easily reset to any desired Reynolds number and waveform.
The system contains a 16-segment diode function generator to
provide the programmed waveform. The resulting electrical
signal drives a servo-controlled piston which delivers a
pulsatile flow component; this is superimposed on a steady
flow to achieve the desired waveform. Figure 2-14b, you see
that the results are quite good; we can simulate all types
of waveforms, all the way from aortic flows to coronary
flows.

There are two final topics which I wish to cover briefly

before concluding the formal portion of my remarks. The
first has to do with flow instabilities which may lead to
disturbed flow and turbulence; and the second deals with the
effects of arterial stretch.

I mentioned previously that we must be careful to

distinguish between highly-disturbed flows, with discrete
large-scale vorticity, and
Highly Disturbed Flows turbulence, which exhibits a
wide frequency spectrum. One
question many of you may ask is, "what causes the instabilities
which lead to these disturbed conditions?"

One very common source of large-scale vorticity in the

cardiovascular system is curvature of the axis of the vessels
as, for example, in the aortic
Curvature of Axis of arch. A centrifugal force is
Vessels. A Source of generated when the flow goes
Vorticity around a bend and one expects the
velocity to be highest on the
inside of the bend, with secondary flow around the circumference
from the outside to the inside.
84 CHAPTER 2

a. b.
Fig. 2-l4a & l4b: Pulsatile flow system (Pitts and Dewey (54)).
a. Photograph of system
b. Oscilloscope record of command signal (upper trace)
and instantaneous test section flow rate (lower trace).
FLUID MECHANICS OF ARTERIAL FLOW 85

Dye tracer studies, such as those of Stehbens (27) and

Brech and Bellhouse (23) show that at the bifurcations, this
pattern is more complicated; helical motion is set up which
transports fluid from the carina side (internal side) of the
bifurcation to the external side along the circumference, with
a return flow from outside to inside along the mid-plane.
Thus, a double helix pattern of secondary flow is established.
This secondary flow can become unstable with appropriate
geometric and temporal perturbations, and can break down
into highly-disturbed and even turbulent motion.

And Lowell Stone has some data which suggest that vessels in
the heart and brain can change dramatically in caliber on the
basis of both nervous and
Endothelial Response to pharmacological stimulus. So
Fluid Stress Depends on again we have a very important
State of Arterial Stretch coupling between the arterial
mechanics and the fluid flow:
endothelial response to fluid stresses may depend strongly
upon the state of arterial stretch.

Dr. Cox will present data on arterial properties, the

numbers you see on Figure 2-15 are simply a small representation
of the data which are available in the literature. This table,
compiled from citations in McDonald's book, (1) lists the
changes in radius which one would expect in several major DOG
arteries during the cardiac cycle for humans and for dogs and
cats. The column labeled A R/R (calc.) is the percentage CAT
change in radius one would calculate from elastic moduli,
arterial radii, and systolic-diastolic pressure changes which
have been reported in the literature. The predications are
in reasonable agreement with those measurements which have
been made.

From these data, there are two conclusions one can draw,
and I invite your comments on them. The first is that there
are very substantial differences between the distensibility of
human arteries and those of common test animals upon which we
lavish so much experimental attention.

DR. WOLF: All animals are not alike!

DR. DEWEY: Precisely. If you take the human carotid and

the dog carotid artery, for example, the value of L\R/R for the DOG
human artery is 17% whereas it is only 4.5% for the dog.
86 CHAPTER 2

STATIC RESTING
VESSEL Ep (X 10- 6 ) boP (-~t) X 100
(DYNES/CM 2 ) (MM HG) CALC EXPT'L

HUMAN CAROTID 0.4 50 17 14.3

DOG " 2.0 67 4.5

HUMAN Asc. AORTA 0.47 45 11.3 12.0

DOG " " 0.8 40 7.5 5.0
CAT " " 18.5

DOG FEMORAL 3 90 4

HUMAN PULMONARY 20 64

NOTE: TW SHOULD BE SIMILAR IN THE PULMONARY ARTERY

AND THE AORTA.
Fig. 2-15: A table of arterial distensibilities compiled from
data in McDonald (1).
FLUID MECHANICS OF ARTERIAL FLOW 87

What I would conclude from this

Be very cautious in is that, if arterial stretch is
Inferences Regarding really important in determining
Human Atherosclerosis trans-endothelial flux into the
from Results of Animal artery, then there is a factor of
Tests about 3 or 4 differences between
the experimental animal you are
testing and the human situation you are really interested in.
If a facotr of 3 or 4 is important, then you better be very
cautious about what you infer about human atherosclerosis from
the results of animal tests.

There appear to be less dramatic differences between the DOG

dog and human descending aorta. But the only piece of data
I have been able to find on the pulmonary artery, quoted in
McDonald's book, is quite dramatic. This experiment, based
on serial x-ray angiograms, concludes that t. R/R is 64% for this
vessel. This occurs in spite of the fact that the internal
pressure change between systole and diastole is half that in
the systematic circulation. If one takes the flow rates and
vessel diameters of the aorta and pulmonary artery, you
estimate that the wall shear stresses in the two vessels are
quite comparable, give or take a factor of 2.

Yet the pulmonary artery does not exhibit atherosclerosis.

DR. MANSFIELD: We do find arteriosclerosis in the

pulmonary arteries of patients with pulmonary vascular
hypertension.

DR. DEWEY: That implies that it is not only the delta

P but also the absolute value of the internal pressure which
is important.

DR. CARO: There could of course be quite different

explanations. For example it might be that thrombosis has a
more important role in the development of atherosclerosis in
the pulmonary system than in the systemic arterial system.

DR. MANSFIELD: Yes, but you see arteriosclerotic

disease in the major branches of the pulmonary vessel under
circumstances of high pulmonary flow (such as left to right
intra cardiac shunts) with only modest elevations in pulmonary
artery pressures.
88 CHAPTER 2

DR. CARO: Well another difference is, of course, that

in the presence of pulmonary hypertension the pulmonary
artery is stretched, and if the cardiac output is unchanged
the mean wall shear stress will be reduced. Some of us have
suggested that that would favor the development of atheroma.

DR. DEWEY: One of the things I want to suggest here

is that wall stretch and shear stress are both important to
the properties of the endothelial
Approximately same Shear layer and the uptake of labeled
Pulmondary Arteries as materials. In the pulmonary
in Aorta arteries under normal circumstances
there is approximately the same
shear as in the aorta; there is also an enormous amount of
stretch, but no disease.

I have a number of additional comments concerning

uptake and transport, as well as further remarks to make
concerning the character of vascular turbulence and the
importance of proper scaling in experiments. I think I will
save those for Dr. Nerem when he presents some of his information
on Wednesday morning.

To conclude, let me identify several specific examples

of arterial fluid dynamics which are of particular interest
for future research. The first
Examples of Arterial is suggested by an experiment
Fluid Dynamics for performed by Flaherty et al. (38)
Future Research in which they produced arterio-
venous shunts both in the carotids
and the iliac arteries of the dogs. They found that substantial
disease developed in the iliacs but not in the carotids. Yet
the shear stresses produced in the two regions were very
similar and, in both instances, were factors of 8-10 below
the so-called critical shear stress. I have reread this
paper several times, and each time I find the results to
be most provocative. An adequate explanation of this experiment,
and perhaps additional confirmatory data, would be a very
impor~ant contribution to our understanding of endothelial
integrity.

Several times in this meeting we have touched upon the

topic of arterial distensibility and its potential importance
to the properties of the endothelium. Let me reiterate a
statement by Dr. Kenyon this morning: when the coronary
arteries pass underneath bridges of myocardial tissue they
are free from disease; but both proximal and distal to the
bridges, one can find substantial atherosclerotic deposits.
FLUID MECHANICS OF ARTERIAL FLOW 89

The minute the artery dives underneath the myocardium and is

supported on both sides, there is no disease. I find this
to be another very provocative result. One may speculate
about wall vibrations and the lack of damping, or possibly a
dramatic change in distensibility between the tethered and
untethered artery. If Professor Stehbens were here, I'm
sure he would have addressed this subject. I don't have the
answer to this problem; but inasmuch as the coronaries and
carotids are our preeminent targets, these facts demand
attention.

There are, of course, many more questions which we

would like to have answered. We need additional data on the
scale and the power and
Unsteady and Turbulent spectral density of unsteady and
Vascular Flows and Effects turbulent vascular flows. In
Unsteady Shear Stress particular, we need to determine
on Endothelial Integrity the effects of unsteady shear
stress on endothelial integrity.
We are beginning a series of experiments in this direction in
the Fluid Mechanics Laboratory at MIT. At present, there
are only two important experiments--one by Fry (7) and the
other by Carew (11) upon which all of our concepts of "critical
shear" are based. All of our speculations about what happens
to the endothelial cell under stress are based on these two
sets of data. The implications of this concept are sufficiently
powerful to motivate much additional research in this area.

DR. MALLIANI: Do you think that important differences

may exist between your experimental preparation and a normal
in vivo situation where the elastic vessels undergo relevant
movements?

DR. DEWEY: I would like to answer that in two different

ways. Basically it means are you going to sit on an erythrocyte
and move through the fluid and
Importance of Tethering try to figure out things that are
going on in the fluid side, that
is one answer. And the second one is if you are sitting on an
endothelial cell in the wall, you want to understand what is
going to happen to you. From the point of view of understanding
fluid mechanics it is not, in my opinion, essential that the
wall move. If you go through any reasonable estimate of
secondary flows and so forth, the maximum extent of the
secondary flow is roughly on the order of magnitude of
delta Rover R, in fact it is usually substantially less.
 90 CHAPTER 2

What this means is that if you are talking about shear

stress on the wall, for example, trying to estimate what
shear stress exists in the carotid bifurcation, you can do
an experiment in vitro with a solid wall and expect to get a
number which is clinically significant in terms of predicting
what it would be in vivo. On the other hand, the response
of the endothelial cell can vary dramatically if the wall is
tethered or untethered. For example, in the external carotid
artery and probably the coronary which are free on the
outside, tethering is all important.

DR. SMITH: I would just like to present one figure

that seems to be relevant both to Dr. Dewey's last figure
and to Dr. Schwartz's first
Developing Arteries group of figures. This is some
Before and After Birth work from the Chicago group (39)
who looked at the developing
RABBIT arteries in young rabbits and examined the ascending aorta and
adjacent-pulmonary artery segment before birth and during
the period after birth. In the neonatal period these are,
of course, joined at the ductus arteriosus and they are both
at the same pressure and in the rabbit they are similar in
size and structure. The ductus closes over the first two
months after birth, and the pressure in the ascending aorta
rises from about 30 mm up to about 90 mm of mercury whereas
the pressure in the pulmonary trunk falls from about 30 to
15 mm. There was, of course, rapid growth; DNA anc the
number of cells in the segments increased greatly and at the
same rate in both vessels, but the ratios of collagen to DNA
and of elastin to DNA, showed marked changes as in Figure 2-16.
In the ascending aorta, the elastin to DNA ratio increased
ten fold, whereas in the pulmonary artery it remained almost
constant and again for collagen there was a ten fold increase
in aorta and for the pulmonary artery it remained almost
constant. Either the pressure or the stretch is having an
enormous influence on connective tissue production and
obviously on the whole rheological properties of the vessels.

DR. NEREM: I believe your flow visualization studies

are particularly useful to this group because too often when
physiologists hear fluid
Flow Visualization Studies mechanics people talk, they
actually think that we know what
is going on. When you look at those flow visualization studies,
you realize that we have a long way to go in our understanding.
FL.UID MECHANICS OF ARTERIAL FLOW 91

160

« 120
z
0
......
Z
~
(f)
«
....J
w 80
ex:
0

z
w
(!)
«
....J
....J
0 40
u

2 :3 4
AGE (MONTHS)

Fig. 2-16: (Smith) The ratio of collagen/DNA and elastin/DNA

in ascending aorta (AA) and pulmonary artery (PA). Collagen
------) and elastin (------) contents of rabbit arteries during
early growth. (From Leung et al., ref. Reproduced by permission
of Academic Press.)
92 CHAPTER 2

As you know, we have been doing some studies on the coronary

system and we don't have quite as vivid pictures to show,
but we feel the complexities there are similar. One of the
things that complicates the coronary system, forgetting
about your first figure, if one wants to take a simple
minded view of the extramural vessels, is the fact that the
major coronary bifurcations are essentially in a plane which
is in itself curving as the vessels move distally over the
heart. You have the curve of the bifurcation together with
the curve of the entire plane turning as they move around
the heart, thus the secondary flows have to be extremely
complex and important to what is going on. I did want to
come back to the question of shear stress in the arterial
system. You mentioned Peronneau's data which he reported at
our Ohio State meeting (16). In fact. from his numbers for
shear rate. one does compute a shear stress of 10 dynes per
square centimeter. However. I find it hard to place much
faith in a shear stress which is deduced from a velocity
profile. As a fluid mechanics person, I have had little
success in the past trying to make that kind of an estimate.

I would rather turn to the type of analysis of Ling and his

co-workers (40,41). They have developed a mathematical
model into which experimental waveform and pressure information
is entered; the details of the flow are then calculated.
Pedley (42) has carried out a similar type of calculation
for the aorta and both of those give values which are more
in line with the measurement, albeit crude, that was reported
in the 1968 paper of Ling et al. (43). In other words, without
other mitigating factors, one might still have a stress of
100 dynes per square centimeter and then, with the addition
of some of the phenomena you have discussed, it may be even
higher. So I am not sure that 400 dynes per square centimeter
is out of the question in some locations. Furthermore, and
I am sure Dr. Fry will agree, these critical values that are
often quoted could be considerably altered by various biochemical
influences that might affect the endothelial cells.

DR. DEWEY: To respond to the last statement first. I

emphasize first of all I could not agree with you more that
the "critical" shear stress is
Critical Shear Stress simply a suggestion of the level
at which substantial rapid changes
take place. They are quantifiable in some systematic statistical
way.
FLUID MECHANICS OF ARTERIAL FLOW 93

There are potentially other things that happen at lower

shear levels which. over a long period of time. may be
equally as important to the survival of the vasculature as
we know it in a young person. But in addition to that I find
it interesting that the critical shear stress is available
only for dog aortic endothelium and not for man. or for DOG
other arteries. In view of the other differences between
the animal models and human models I would suggest that this
is a potential contribution that somebody can make in giving
us equivalent data. hopefully on a human endothelium. I
think you made the point about the coronary arteries and so
I have nothing to add.

DR. NEREM: You mentioned turbulence and either I misheard

you or else I would like to discuss this further. You talked
about these flow disturbances essentially being initiated. I
thought you said during the diastolic portion of the beat.
Did you mean systolic. or did I hear you wrong?

DR. DEWEY: What I meant to say was in the diastolic

side of the systolic beat. On the backside of the systole.

DR. STONE: Yes. conceptually. Dr. Dewey, in your

models of coronary vessels what would be the changes if the
coronaries were highly tethered?

DR. DEWEY: I can only suggest one thing. I am very

slow when it comes to this kind of speculation. There is
only one thing that has occurred
Tethering to me and that is that an
untethered artery. or a potentially
untethered artery is subject to a substantial wave propagation
phenomenon and potential resonances which are not present in
a tethered artery.

This has been true in many experiments where people have

exposed. for example. femoral and carotid arteries and we
have done it too in dogs and got absolutely erroneous results
with respect to the paraspectral density of the unsteady
flow. They have been absolutely and totally wrong and
people still do that with unsupported tubes of latex to get
these absolutely wrong results with respect to tethered
arteries in the circulation. There is a long standing
controversy about this in which I have a very strong opinion.
Arteries like femoral and carotids in vivo are well tethered
and they don't resonate.
 94 CHAPTER 2

However. we have seen resonant spectra obtained from the

external carotid artery. I would expect that the same kind
of resonance phenomena could occur in the coronary arteries
on the upper cardiac surface.

DR. SCHWARTZ: I have two questions. One relates

specifically to the concept
Critical Shear Assoicated of critical shear. My
with Significant Endothelial understanding is that it is
Injury associated with significant
endothelial injury. Is this
correct?

DR. FRY: Correct. The definition of the acute critical

yield stress (t c ) implies detection of identifiable structural
changes in certain members of the endothelial cell population
(44,45). This is not to be confused with the shear stress
dependent increase in endothelial permeability which occurs
at much lower levels of stress and is not associated with
any detectable structural changes either by light or by
electron microscopy (46).

DR. SCHWARTZ: Is it possible that much lower shear

stresses modify, for example, the process of vesicular
transport?

DR. DEWEY: I would suggest that either you or Dr. Fry

make a comment about your critical shear.

DR. FRY: The experimentally observed relationship

among hemodynamic stresses amd altered endothelial surface
permeability with and without associated structural changes
recently have been reviewed elsewhere (46). Turning specifically
to the question of "acute critical yield stress" (t c ),
this quantity represents the stress at which there is the
greatest conversion of structurally normal cells to structurally
abnormal cells. Thus tc represents the rheologic behavior
of a population of endothelial cells, not of an individual.
Very roughly speaking. tc is the stress at which half of
the endothelial cells in a given population will show structural
changes. It follows that many endothelial cells in this
individual's population have yielded at stresses below tc
and the hardier members won't yield until exposed to stresses
above tc. The value of tc appears to be relatively insensitive
to the duration of stress exposure; however, this should be
studied further. Although the group mean value of tc for
DOG the canine ventral thoracic aorta (23 animals) is around
FLUID MECHANICS OF ARTERIAL FLOW 95

400 dynes cm -2 , it is important to note that the range of

these values was from less than 200 to 990 dynes cm -2 (44,45).
Thus it appears that the (rheologic) strength of the endothelial
surface may vary widely even in normal animals and at the
same site. Moreover, subsequent studies, summarized elsewhere
(46), suggest that tc may vary considerably from site to
site in a given animal and may also vary significantly with
the metabolic state of the animal.

The acute critical shear stress is a term that refers

specifically to a level of stress that produces overt structural
changes in the endothelial
Acute Critical Shear Stress surface. As might be expected,
a massive increase in endothelial
permeability is associated with these structural changes.
However, at considerably lower levels of stress exposure that
are unassociated with any detectable histologic or ultrastructural
changes in the endothelial cells, one also sees a stress
dependent increase in permeability (45,47). In studies
specifically designed to study this shear dependent increase
in permeability, e.g., in an arteriovenous shunt preparation,
transmission electron microscopy suggests occasional changes
in the subendothelial region consistent with mild edema.
However, these observations are highly variable. In the
inverse situation, i.e., if transmission and scanning electron
microscopy are done at sites in the vascular system that are
normally more permeable to plasma substances, one commonly
finds subendothelial edema and also occasional small patches
of endothelial cell erosion. These structural changes are
never seen in regions of low or "normal" permeability. One
wonders if subendothelial edema associated with increased
permeability might not be one of the common mechanisms of
lowering the critical yield stress and promoting endothelial
erosion (46).

DR. SCHWARTZ: I wasn't questioning that Dr. Fry. I

was trying to suggest that perhaps critical shear could be
very much lower in terms of, say modification of transport
by some process other than endothelial injury.

DR. NEREM: Dr. Fry, relative to your studies ••• correct

me if I am wrong, I believe the yield stress is 400 dynes
per square centimeter. Was that the stress above which you
observed cell abnormality and was actual cell erosion associated
with a stress of about 1000 dynes per square centimeter?
 96
CHAPTER 2

DR. FRY: Not quite. Any stress equal to or in excess

of the tc for that animal would be associated with significant
cellular structural changes. These altered cells will
finally erode, provided the stress exposure is maintained
for a sufficient duration, say 2 hours. Studies in which
the stress was considerably in excess of tc ' of course,
would show erosion in a shorter period of time.

DR. NEREM: What was the lowest stress at which you

observed erosion?

DOG DR. FRY: The lowest value of tc in the 23 dogs was

less than 200 dynes cm- 2 .

DR. DEWEY: May I make a comment? With all due respect

to the concept of critical shear, I think one of the reasons
I brought it up was to promote discussion about it, because
in trying to estimate shear stresses particularly in coronary
arteries and the carotid arteries, really the ones that kill
us, you may get very high values from time to time. But it
seems to me that a concept like critical shear is not necessarily
the kind of thing we are looking for.

DR. FRY: I agree.

DR. CAREW: I believe both in the studies that I did,

which Dr. Dewey alluded to, and also in Dr. Nerem's and
Caro's studies, which bear the albumin transport across either
the aortic endothelium or, in the case of Nerem and Caro the
carotid endothelium, and in another study I did with Cronwright
on vessels in situ, there was no evidence that there was a
distinct threshold of shear stress below which permeability was
not a function of shear stress. There seemed to be a fairly
smooth continuation with increasing shear stress from the very
lowest levels. That could be very important.

DR. GREENE: Concerning the idea of a critical shear

stress at the wall, there are a great deal of data of critical
shear stresses for human
Critical Shear Stress erythrocytes, even if there are
for Erythrocytes little for human aorta. The data
have been summarized by Leverett
(48), and show that there is no such thing as a single critical
shear stress for an erythrocyte. Instead, there is a time-
shear stress domain such that cell rupture occurs rapidly at
exposures to 1500 dynes/cm 2 , but much more slowly when
stress level is in the region of 100 dynes/cm 2 •
FLUID MECHANICS OF ARTERIAL FLOW 97

In the absence of sufficient data, one might hypothesize that

aortic tissue responds similarly and that the 400 dynes/cm 2
value turns out to be critical only in the context of the
time frame of current experimentation.

DR. DEWEY: There is another type of instability mechanfsm

which may be operative here, but not, to my knowledge,
invoked previously in describing
"Gortler Instability" cardiovascular flows. It is
called a "Gortler Instability."
As you can see in Figure 2-17, a steady laminar viscous
flow over a curved wall can generate a series of longitudinally-
oriented vortex cells within the viscous boundary layer next
to the surface. One can calculate (49) the conditions under
which these vortex cells will become unstable and cause
turbulence in the boundary layer. My calculations suggest
that this type of instability should be generated in the
major arteries, and it would be interesting to set up some
experiments which are specifically designed to demonstrate
this effect.

The final topic I will cover returns to the problem of

arterial distensibility, a phenomenon which couples the
flow dynamics and the solid
Problems of Arterial mechanics of the arterial wall.
Distensibility As Dr. Kenyon pointed out so
clearly this morning, the
stretch of the artery depends on its geometry, tethering,
and elastic properties; and how these mechanical properties
are acted upon by the internal pressure and fluid shear
stress. But the internal fluid pressure and motion depend
intimately upon the stretch of the artery, so the whole
problem is strongly coupled. Several theoretical models
have been developed to represent the aorta and other major
vessels with axially varying compliance, and references to
this work can be found in MCDonald's book (1) the two
volumes edited by Bergel (50) and the book edited by Attinger
(29). One may briefly summarize these results as follows:
it is necessary to model the taper and axial variation of
compliance of the arterial tree if one wishes to predict the
temporal variations of flow and pressure at each point along
the arterial tree. But once the flow and pressure are
known, the velocity profile across the lumen, the wall shear
stress, and other flow parameters are the same as those
which would be observed in a rigid tube with the same temporal
variations of flow and pressure.
98 CHAPTER 2

Fig. 2-17: Cort1er instabilities in a viscous shear flow on a

curved wall.

But such a simplistic summary ignores one potentially

important variable, namely the effect of arterial distensibility
on the integrity of the endothelial layer. Circumferential
stretch clearly alters the morphology of the endothelial
cell, just as Fry (12) has shown that stretch alters the
cell permeability. Several years ago, Bergel (51) pointed
out that we know very little about the distensibility in and
around arterial bifurcations, and suggested that deformation
would be a maximum on the outside of the branch (i.e. the
side with the smallest radius of curvature). This was
emphasized by Dr. Kenyon this morning when he established
quantitative estimates for this effect.
FLUID MECHANICS OF ARTERIAL FLOW 99

BIBLIOGRAPHY

1. McDonald, D.A.: Blood Flow in Arteries. 2nd Ed.,

Williams and Wilkins, Baltimore, esp. pp. 92-95, 1974.

2. Mills, C.J., Gabe, I.T., Gault, J.H., Mason, D.T.,

Ross, J., Braunwa11, E., and Shilling, J.P.: Pressure-
flow relationships and vascular impedance in man.
Cardiovascular Research 4, 405-417, 1970.

3. Nerem, R.M., Seed, W.A., and Wood, N.B.: An experimental

study of the velocity distribution and transition
to turbulence in the aorta. J. Fluid Mech., 52,
137-160, 1972.

4. Clark, C., and Schultz, D.L.: Velocity distribution in

aortic flow. Cardiovascular Research 7, 601-613, 1973.

5. Ling, S.C., Atabek, H.B., Fry, D.L., Patel, D.J., and

Janicki, J.S.: Application of heated-film velocity
and shear probes to hemodynamic studies. Circulation
Research 23, 789-801, 1968.

6. Ling, S.C., Atabek, H.B., and Carmody, J.J.: Pulsatile

flow in arteries. In Proc. XII Int. App1. Mech.
(Me, He te nyi and Vincenti, W.G., Eds.) Springer-
Verlag, Berlin, pp. 277-291, 1969.

7. Fry, D.L.: Acute vascular endothelial changes associated

with increased blood velocity gradients. Circulation
Research 22, 165-197, 1968.

8. Gregg, D.E.: The Coronary Arteries, Charles C. Thomas,

Pub1., Springfield, Ill., 1965.

9. Nerem, R.M., Rumburger, J.A., Jr., Gross, D.R., Jam1in, R.L.,

and Geiger, G.L.: Hot-film measurements of coronary
blood flow in horses. Proc. Specialists Conf. on Fluid
Dynamic Aspects of Arterial Disease (R.M. Nerem, Ed.)
Ohio State University, pp. 28-31, 1974.

10. Wells, M.K., Winter, D.C., Nelson, A.W., and McCarthy, T.C.:
Estimated blood velocity profiles in the coronary
artery. Proc. 27th ACEMB, Philadelphia, 282, 1974.

11. Carew, T.E. III: Mechano-chemica1 response of canine

aortic endothelium to elevated shear stress in vitro.
Ph.D. Thesis, Catholic University of America,
Washington, D.C., 1971
100 CHAPTE R 2

12. Fry, D.L.: Responses of the arterial wall to certain

physical factors. In Atherosclerosis: Initiating
Factors, CIBA Foundation Symposium 12 (New Series)
Elsevier, Amsterdam, pp. 93-125, 1973.

13. Fronek, A., Johansen, K.H., Dilley, R.B., and Bernstein, E.F.:
Noninvasive physiologic tests in the diagnosis and
characterization of peripheral arterial occlusive
disease. Am. J. Surgery, 126, 205-214, 1973.

14. Hisland, M.B., Miller, C.W., McLeod, F.D., Jr.: Trans-

cutaneous measurement of blood velocity profiles and
flow. Cardiovascular Research, 7, 703-712, 1973.

15. Fronek, A., Coel, M., and Bernstein, E.F.: Quantitative

ultrasonographic studies of lower extremity flow
velocities in health and disease. Circulation, 6,
957-960, 1976.

16. Perroneau, P., Gisbertz, K.H., Steckmeier, B., Xhaard, M.,

Dalbera, A., and Bournat, J.P: In vitro and in vivo
experimental studies of pulsatile flow patterns in
curved and stenotic vessels. Proc. of Specialists
Meeting on Fluid Dynamic Aspects of Arterial Disease
(R.M. Nerem, Ed.) Ohio State University, pp. 5-8, 1974.

17. Gutstein, W.H., and Schneck, D.J.: In vitro boundary layer

studies of blood flow in branched tubes. J. Athero. Res.
7, 295-299, 1967.

18. Ferguson, G.G., and Roach, M.R.: Flow conditions at

bifurcations as determined in glass models, with
reference to the focal distribution of vascular
lesions. In Cardiovascular Fluid Dynamics, Vol. 2
(D.H. Bergel, Ed.), Academic Press, N.Y. pp. 141-156,
1972.

19. Attinger, E.O.: Flow patterns and vascular geometry.

In Pulsatile Blood Flow (E.O. Attinger, Ed.),
McGraw-Hill, N.Y. pp. 179-198, 1964.

20. Rodkiewicz, C.M., and Roussel, C.L.: Fluid mechanics in

a large arterial bifurcation. Trans. ASME, Series I
(Fluid Engineering), 95, 108-112, 1973.

21. Feuerstein, I.A., Elmasry, O.A., and Round, C.F.: Flow

pattersn and wall shear rates in a series of symmetric
bifurcations. (Abstract), Proc. 27th ACEMB, Philadelphia
Oct. 6-10, 278, 1974.
FLUID MECHANICS OF ARTERIAL FLOW 101

22. Freedman, R.W.: The measurement of wall shear rate in a

model of the human aorto-iliac bifurcation using an
electro-chemical technique. Ph.D. Thesis, M.I.T.,
1976,

23. Brech, R., and Bellhouse, B.J.: Flow in branching vessels.

Cardiovascular Research, 7, 593-600, 1973.

24. Kandarpa, K., and Davis, N.: Analysis of the fluid

dynamic effects on atherogenesis at branching sites.
J. Biomechanics, 9, 735-741, 1976.

25. O'Brien, V., Erlich, L.W., and Friedman, M.H.: Nonlinear

simulation of unsteady arterial flows in a branch.
Paper presented at the A.I.Ch.E. Annual Meeting,
Philadelphia, 1973.

26. Smith, C.S.: Flow studies in a model arterial bifurcation.

1st Prize Student Paper, presented at the 29th Annual
Conference on Engineering and Medicine and Biology,
Nov. 6-10, Boston, 1976.

27. Stehbens, W.E.: Turbulence of blood flow. Quart. J. Exper.

Path., 44, 110-117, 1959.

28. Stehbens, W.E.: Flow in glass models of arterial bifur-

cations and berry aneurysms at low Reynolds numbers.
Quart. J. Exper. Physiol., 60, 181-192, 1975.

29. Attinger, E.O. (Ed.): Pulsatile Blood Flow, Mc-Graw-Hill,

N.Y., 1964.

30. Fox, J.A., and Hugh A.E.: Static zones in the internal
carotid artery: correlation with boundary layer
separation and stasis in model flows. Br. Jour. Rad.
43, 370-376, 1970.

31. Dewey, C.F., Jr.: Qualitative and quantitative flow field

visualization utilizing laser-induced fluorescence.
AGARD Conference Proceedings No. 193 on Applications
of Non-Intrusive Instrumentation in Fluid Flow Research,
St. Louis, France, pp. 17-1 to 17-7, 1976.

32. Hugh, A.E., and Fox, J.A.: Precise localization of atheroma

and its association with stasis at the origin of the
interval carotid artery -- a radiographic investigation.
Br. Jour. Rad., 43, 377-383, 1970.

33. Lyne, W.H.: Ph.D. Thesis, University of London,

1970.
102 CHAPTER 2

34. B1ennerhasset, P.: Ph.D. Thesis, University of

London, 1970.

35. Dean, W.R.: Philosophical Mag. Series 7,2673,

1928.

36. Duncan, G.W., Gruber, J.D., Dewey, C.F., Jr.,

Meyers, G.S., and Lees, R.S.: Evaluation of carotid
stenosis of phanoangiography. New Engl. J. Med.,
293, 1124-1128, 1975.

37. Pitts, W.H. III, and Dewey, C.F., Jr.,: Spectral

and temporal characteristics of post-stenotic
turbulent wall pressure fluctuations. Trans. ASME
J. Biomech. Eng'g, 1976.

38. Flaherty, J.T., Ferrans, V.J., Pierce, J.E., Carew, T.E.,

and Fry, D.L.: Localizing factors in experimental
atherosclerosis. Atherosclerosis and Coronary Heart
Disease. (W. Likoff, B.L. Segal, W. Insull, Jr.,
and J.H. Moyer, Eds.) Grune and Stratton, N.Y.,
pp. 40-84, 1972.

39. Leung, D.Y.M., G1agov, S., Clark, J.M., and Mathews, M.D.:
Mechanical influences on the biosynthesis of extra-
cellular macromolecules by aortic cells. In:
Extracellular Matrix Influences on Gene Expression.
(H.C. Slavkin and R.C. Greulich, Eds.) Academic
Press, 633, 1975.

40. Ling, S.C., Atabek, H.B., Letzing, W.G., and Patel, D.J.:
Nonlinear analysis of aortic flow in living
dogs. Circulation Research, 33:198-212, 1973.

41. Atabek, H.B., Ling, S.C., Patel, D.J.: Analysis of

coronary flow fields in thoracotomized dogs.
Circulation Research, 37:752-761, 1975.

42. Pedley, T.J.: Flow in the entrance of the aorta.

Proceedings from a Specialists Meeting on Fluid
Dynamics of Arterial Disease, R.M. Nerem, (Ed.)
Ohio, pp. 20-23, 1974.

43. Ling, S.C., Atabek, H.B., Fry, D.L., Patel, D.J., and
Janicki, J.S.: Application of heated-film velocity
and shear probes to hemodynamic studies. Circulation
Research, 23:789-801, 1968.
FLUID MECHANICS OF ARTERIAL FLOW 103

44. Fry, D.L.: Acute vascular endothelial changes associated

with increased blood velocity gradients. Circulation
Research, Vol. XII, Feb., 1968.

45. Fry, D.L.: Certain histological and chemical responses

of the vascular interface to acutely induced
mechanical stress in the aorta of the dog.
Circulation Research, Vol. XIV, Jan., 1969.

46. Fry, D.L.: Hemodynamic forces in atherogenesis.

Cerebrovascular Disease, Raven Press, N.Y., 1976.

47. Carew, T.D.: Mechano-chemical response of canine aortic

endothelium to elevated shear stress In Vitro., Ph.D.
Thesis, The Catholic University of American,
Washington, D.C., 1971.

48. Leverett, L.B., Hellums, J.D., Alfrey, C.P., Lynch, E.C.:

Red blood cell damage by shear stress. 72nd National
Meeting, American Institutes of Chemical Engineers
St. Louis, May, 1972.

49. Lin, C.C.,: The Theory of Hydrodynamic Stability,

Cambridge U. Press, Cambridge, pp. 96-98. See
also papers by H. Gortler and G. Hammer 1 in ,
50 Jahre Grenzschictforshung (H. Gortler and W.
Tollmein, Eds.). Freidr. Vieweg & Sohn, Braunschweig,
1955.

50. Bergel, D.H.: (Ed.): Cardiovascular Fluid Dynamics.,

Vols. I and II, Academic Press, N.Y., 1972.

51. Bergel, D.H.: Comments on unpublished work by C. Mellon

and D.H. Bergel, Appearing in Atherosclerosis:
Initiating Factors, CIBA Foundation Symposium 12
(New Series), Elsevier, Amsterdam, pp. 159-161, 1973.

52. Fulton, W.M.F.: The Coronary Arteries, Charles C. Thomas,

Springfield, Ill., 1965.

53. Schultz, D.L.: Pressure and flow in large arteries. In

Cardiovascular Fluid Dynamics, Vol. 1 (D.H. Bergel, Ed.)
Academic Press, pp. 287-314, 1972.
54. Pitts, W.H. III, and Dewey, C.F., Jr.: Programmable
pulsatile flow apparatus for simulation of arterial
hemodynamics. Proc. San Diego Biomedical Symposium,
14, 119-124, 1975.
Chapter 3 CONTROL OF VASOMOTOR FUNCTION AND THE HEMODYNAMIC
CONSEQUENCES OF THE CONTRACTILE BEHAVIOR OF
ARTERIES

DR. COX: In the mammalian organism, a complex array of

control mechanisms exists to fulfill the temporal and spatial
demands of the body for the
Complex Array of transport services of the
Control Mechanisms circulatory system. These
control mechanisms appear to
involve almost all of the various elements of the cardiovascular
system: the heart, the microcirculation and the veins.

What about the large so-called conduit or elastic

arteries. Do they play any role in neural reflex mechanisms?
These are the vessels with which we are primarily concerned
at this workshop in relation to their role in atherogenesis.
Are they indeed passive conduits which simply respond to
hemodynamic events in a "programmed" manner or are they
actively involved in the control and regulation of cardiovascular
function?

The view has long been held that neural reflexes produced
at best only modest effects on large arteries (1). They
were not considered to be involved
Neural Reflexes in the control of peripheral
resistance and, therefore, to be
of no significance in the scheme of neural control. Yet,
the walls of larger arteries contain an abundant amount of
smooth muscle from 25 to 35% of wall volume. There is also
a significant amount of contractile protein present and
indeed larger arteries of some species are used as tissue
for pharmacological assay of vasoactive substances (2).
Teleologically, this contractile structure (the smooth
muscle cell) must serve some physiological purpose other
than being simply a synthetic structure for other wall
precursors (3).

We are now somewhat more sophisticated and realize

that there is more to neural control than the control of
peripheral resistance and
Control of Smooth cardiac output. There is now
Muscle Tone sufficient evidence to suggest
important physiological role (s)
for the control of smooth muscle tone in larger arteries (4,5,6).
105
106 CHAPTER 3

My plan for this presentation is to review current knowledge

of neural control mechanisms including the regional differ-
entiation of control as well as the control of both large and
small arteries, primarily under normal conditions but also for
some other physiological and pathophysiological conditions.
Finally, the possible significance of this control in light of
the objectives of this meeting will be discussed.

EFFECTOR MECHANISMS IN ARTERIES

1. Passive Mechanical Properties

Studies on the passive mechanical properties of arteries

(i.e., negligible smooth muscle tone) have been performed on
a variety of preparations
Arteries are Nonlinear including strips and intact
Viscoelastic Nonuniform segments. In general, the results
and Anisotropic of these studies are qualitatively,
if not always quanititatively
similar. Measurements of pressure-volume, pressure-diameter
and force-length relations on arterial samples show that
arteries are nonlinear, viscoelastic nonuniform and anisotropic.
Arteries are nonlinear because they do not obey Hooke's law.
They are viscoelastic because their responses are time
dependent. They are nonuniform because their properties are
different in different parts of the arterial system. They
are anisotropic because their properties are different in
different directions within the wall.

Tangential stress-strain relations from five arterial

sites are summarized in Fig. 3-1 and demonstrate the nonlinear
and nonuniform nature of arterial wall properties. The
nonlinear natures of these curves have been explained on the
basis of a nonuniform contribution of wall elements to load
bearing at different values of wall strain (7), (8), (9).
Wall stress is thought to be supported by elastin in the low
strain range, by elastin and collagen at intermediate strains
and by collagen alone at high strain.

The nonuniform nature of arterial wall properties is

also indicated in Fig. 3-1 by the wide variation in stress-
strain relations at different anatomical sites. One of the
principal contributors to this regional nonuniformity is the
anatomical variation in collagen and elastin content.
CONTRACTILE BEHAVIOR OF ARTERIES 107

6.Carotid, 1.98 a 0.08

• Iliac, 1.32 a 0.04
5 o MlIs~ntllr/c, 1.33.0.04
• Rllnol, 1.65.0.06
o Coronary, 2.75 c;O.l1

N
E 4
u
......
on
,..
G
c;
"tI

'"o

1.5 2.0
NORMALIZED EXTERNAL DIAMETER

Fig. 3-1: Passive tangential stress-stain relations of isolated

canine arterial segments in vitro. Abscissa is the diameter of
the ·segment normalized by dividing by the value of external
diameter at zero pressure. Smooth muscle in the arteries was
treated with cyanide, iodacetate and dinitrophenol to inhibit
active force development. Symbols are means and bars ± 1 SEM of
data averaged in transmural pressure steps of 10 mmHg from 0 to
250. Data at the top with symbols are the average values of ± 1
SEM of the ratio of collagen to elastin contents determined using
the same segments by chemical means. (Fisher and L1aurado, 1966).
108 CHAPTER 3

Since collagen fibers are stiffer than elastin, the ratio of

collagen to elastin content (C/E) has been suggested as an
empirical index of wall stiffness (10). While this relation
generally holds along the aorta (11), it is not valid in its
daughter branches as shown in Fig. 3-1. Obviously, other
factors must be of importance in determining arterial wall
mechanics, including, for example, the architectural coupling
and distriubtion of connective tissue elements as well as
the detailed recruitment of collagen fibers with increasing
wall strain (8).

The anisotropic nature of arterial wall properties are

the result of the ordered structure of the arterial wall at
the microscopic level. The various wall constituents are
not homogenously distributed but are structurally oriented.
Numerous morphological studies have described the organization
of arterial wall elements (12). Fig. 3-2 summarizes values of
anisotropic elastic moduli and Poisson ratios obtained from
a group of carotid arteries measured at in vivo length (13).
At low values of tangetial strain (actually extension ratio,
A a = P- values of tangetial (EO) and radial (Er) elastic
moduli R~e nearly equal and both are smaller than the axial
modulus (Ex). At higher values of tangetial wall strain, EO
becomes larger than Er and Ex while the latter two are nearly
equal. The Poisson ratios which relate deformation in one
direction to that in another show considerable variatio_!, with
ure and uer being nearly identical especially in the low strain
range.

The anisotropic elastic properties depend upon values

of wall strain in the various directions. The exact manner
in which the various wall strains interact to influence
anisotropic properties is not completely known, however.
The manner in which the various wall constituents contribute
to the various anisotropic sontants is also now known.

Viscoelasticity is manifest in arteries as a dependence

of the arterial wall response to a deformation upon both
the magnitude of the deformation
Viscoelasticity and the time course over which it
is applied. Viscoelasticity has
been studied using time domain transient responses or steady-
state frequency domain responses. Fig. 3-3 summarizes data
from Bergel (14) on the frequency dependence of the dynamic
modulus of several arterial sites studied in vitro at a mean
transmural pressure of 100 mmHg with a superimposed sinusoidal
pressure variation of different frequencies.
CONTRACTILE BEHAVIOR OF ARTERIES 109

40

(f)
:::>
...J
:::>
0
0
~ 20
U
t=
(f)
<t
...J
W

0
1.0
Q
~
<t
a::
Z
0
(f)
(f)
6
a.. 00

TANGENTIAL EXTENSION RATIO

Fig. 3-2: Variation of passive anisotropic elastic constants of

a canine carotid artery with tangential extension ratio in vitro.
Elastic moduli are tangential (open circles), radial (closed
squares) and axial (open triangles). Poisson ratios are Ur 9
(open circles), u9r (open triangles) and u9x (closed squares)
Units of elastic modulus are 10 6 dynes/cm 2 • Tangential extension
ratio is the ratio of mid-wall diameter or a given pressure
divided by the unstressed value.
110 CHAPTER 3

The value of dynamic modulus increases in magnitude between

o and 2 Hz and remains essentially constant up to 20 Hz.
The dynamic modulus ratio varies considerably at different
arterial sites as seen in Fig. 3-3. This regional variation of
viscoelasticity is thought to be related to the regional
variations in composition. The contribution of the various
wall components to arterial viscoelasticity is not completely
known at this time.

The viscous properties of the arterial wall arise from

at least two sources: inter - and intra-molecular cross-
linking of wall components and the interaction of contractile
proteins. The detailed deformation response of a coiled,
cross-linked macromolecular matrix to the application of a
strain and its removal are bound to be different, producing
a strain rate dependent trajectory in the stress-strain
plan.

The resistance to stretch in non-contracting smooth

muscle has recently been shown to be calcium dependent but
not dependent upon membrane
Calcium Dependence depolarization of intracellular
release of calcium. This has
been interpreted to suggest that most, if not all, crossbridges
are normally attached in smooth muscle and thus able to
resist deformation. When the muscle is stretched, the
breaking and subsequent reformation of cross-links is the
source of viscoelasticity. Such responses would be strain
rate dependent, i.e., amplitude and frequency of the strain.

The effects of the activation of smooth muscle on

arterial wall properties can be represented in two general
ways. One way is to consider
Effects of Smooth the artery to be an elastomer
Muscle Activation and describe the effects of
activation on its mechanical
properties. The other way is to consider the artery to be a
muscle and then describe the effects of activation of force
development and shortening. Both of these approaches have
been utilized in the determination of the contribution of
smooth muscle to arterial wall properties and behavior.

In general, activation of smooth muscle shifts the

tangential stress-strain curve of an artery to the left.
The magnitude of this effect varies with anatomical location
in the arterial system and depends upon a large variety of
factors too numerous to list here.
CONTRACTILE BEHAVIOR OF ARTERIES 111

o THOR x FEM
1·8 • ABOO • CAR

o 2 4 6 8 10 12 14 16 18 20
FREQUENCY Hz

Fig. 3-3: Frequency dependence of the passive dynamic modulus

ratio of several isolated canine artery segments in vitro.
Ordinate is the dynamic modulus divided by the static modulus.
A sinusoidal pressure of variable frequency was imposed with an
amplitude of 5-10 mmHg at a mean pressure of 100 mmHg (Bergel,
1961).
 112 CHAPTER 3

The data in Fig. 3-4 show the effects of maximum activation

by norepinephrine on the tangential stress-strain relations
of iliac arteries in vitro. At specific values of wall
strain, activation of smooth muscles produces an increase in
wall stress which is a strong function of strain.

As would be predicted from these results, substantial

effects on arterial wall elasticity are produced by smooth
muscle activation. For a number of years, a controversy
existed in the literature as to whether smooth muscle activation
produced an increase or decrease in elastic modulus (1).
The reason for this controversy was explained by Dobrin and
Rovick (16) who noted that the change in elastic modulus of
arteries following smooth muscle activation depended on
whether strain or pressure was the independent variable. As
shown in Fig. 3-5, the elastic modulus is increased at all
values of wall strain when the latter is the independent
variable. When plotted as a function of transmural pressure,
however, the elastic modulus is increased at low and decreased
at moderate to high values of pressure following maximal
activation.

In order to describe the effects of vascular smooth

muscle activation in terms of muscle mechanics, it is necessary
to determine both isometric contraction and isotonic shortening.
For intact arterial segments
Isometric Contraction in the form of cylinders, these
and Isotonic Shortening characteristics of muscle mechanics
translate into constant radius
(isometric) stress development and constant pressure (isobaric
constriction). These responses are shown diagrammatically in
Fig. 3-6. The steady state pressure-diameter points to the
right correspond to passive muscle while the points to the left
correspond to fully activated muscle. The region between
these two curves delineates the range of control of arterial
wall properties by the smooth muscle. Qualitatively similar
relations exist at all arterial sites studied to date.

As would be predicted from Fig. 3-6, these two forms of

FROG active response depend strongly upon the initial pressure or
diameter. The variation of the diameter response of vessels
in the frog mesentery to maximal norepinephrine activation
with the initial value of tangential wall stress (i.e.,
pressure) is shown in Fig. 3-7. (17). Stress is used in
this diagram to allow a more meaningful comparison among
different vessels from which these measurements were made.
CONTRACTILE BEHAVIOR OF ARTERIES 113

00 2
00
w
a:
t-
OO

...J
oCt
t-
Z
W
(!)
Z
oCt
I-

o
o

NORMALIZED EXTERNAL DIAMETER

Fig. 3-4: Relation of tangential wall stress and normalized

external diameter for passive conditions (closed squares) and for
maximal norepinephrine activation (open circles). Symbols are
means and bars ± 1 SEM of data averaged in 20 mmHg steps. Data
were obtained from isolated intact segments of canine iliac
artery (N-12) and averaged at specific values of transmural
pressure.
 114 CHAPTER 3

40 40

--
,t
N
E
~
CD
C
>-
'1:1
10
52
.....

J
CI)
::;)
.-l
::;)
0

ft/ +
0 I

lI
~ 20

J
20 /

(.)
t-
CI)
cl
.-l I
&oJ I

7r/ l
I

.-l
cl
t-

/+ If
Z /
I

&oJ I
~
UJ
0::
(.) t-O!4' .. / /
/1
~ ~ ___ >O<-~ O_~O.L • ' 1/
____ -0--0- - 0'

Ibo
I
0
0 1.0 1.4 1.8 00 200
NORMALIZED EXTERNAL DIAMETER PRESSURE (mmHg)

Fig. 3-5: Effects of smooth muscle activation by norepinephrine

on the static incremental elastic modulus of canine iliac
arteries. The panel at the left shows elastic modulus as a
function of normalized external diameter while the panel at the
right its variation with transmural pressure. Data are for
active (open circles) and passive (closed squares) conditons.
CONTRACTILE BEHAVIOR OF ARTERIES 115

250
H
Ii
~

0>
I
E
E
ii

·-1/·
I·i
w
a::
::)
(f)
(f)
W
g:
Ij
125
~.
, • CONSTANT.
RADIUS /

I• STRESS
DEVELOPMENT / •

•.............................................................•

o
I
.;
CONSTANT

.~
----.
PRESSURE CONSTRIC~

0.3 0.4 0.5 0.6

EXTERNAL DIAMETER (em)

Fig. 3-6: Schematic representation of the mechanics of arterial

smooth muscle. Curve and points to the right represent the pres-
sure-diameter relation for passive muscle while the points to the
right are data for active muscle. The mechanics of arterial smooth
muscle can be represented in terms of constant radius (isometric)
stress development or constant pressure (isobaric) constriction
responses to smooth muscle activation.
116 CHAPTER 3

50
4

40

30
10
<I
~

20

10 Dose I-Norepl' 250 "A-Sec

o ~------~------~--------~----~
o 123 4
INITIAL TANGENTIAL STRESS
( 10 5 dynes/ cm 2 )

Fig. 3-7: Dependence of diameter responses to norepinephrine

activation on initial tangential stress for arterial vessels in
the frog mesentery. The diameter response is normalized by
dividing by the initial control value of diameter. Norepinephrine
was applied topically by an iontophoretic method. The numbers
over the data points represent the number of observations per
data point (17).
CONTRACTILE BEHAVIOR OF ARTERIES 117

The diameter response is expressed as a percentage of the

initial diameter. The diameter response is a strong function
of the initial stress exhibiting a maximal response between
1 and 2 X 105 dynes/cm2 for this preparation. Rather
similar results have also been reported by a number of
investigators for mammalian arterial smooth muscle preparations
(18,19,20) •

Values of tangential wall stress developed in response

to smooth muscle activation show a strong dependence on wall DOG
strain as shown for iliac arteries in Fig. 3-8. For all
arterial sites, an optimum strain exists at which the stress
response to maximal activation is largest. Both above and
below this value of strain the stress response decreases.
The general characteristics of this relation are qualitatively
similar at all arterial sites which have been studied to
date (19,20,21). Both of these two manifestations of arterial
smooth muscle activation are consistent with a sliding
filament model of muscle function.

The results described in the previous section delineate

the limits of control of arterial wall properties by smooth
muscle. In vivo a number of
Density and Distribution factors will affect this
of Post Ganglionic relationship in such a manner as
Sympathetic Nerve Fibers to decrease the effective range
over which smooth muscle can
exert control of arterial wall properties. Of these factors,
probably the most important is the density and distribution of
the post ganglionic sympathetic nerve fibers.

Not only is there a nonuniform distribution of nerve fibers as

well as nerve activity to different vascular beds within the
body but also to the various series coupled vascular elements
in a particular vascular bed. Fig. 3-9 shows a diagrammatic
representation of the density of innervation of mesenteric
blood vessels in the rat (22). The distribution of nerve RAT
fibers is particularly dense to the small and medium sized
arteries in the mesenteric circulation as well as to the
terminal arterioles. Beyond the terminal arterioles adrenergic
innervation to precapillary arterioles has not been observed
(22). An analogous nonuniform distribution of innervation
density is also observed on the venous side as well.
118 CHAPTER 3

1.5
N
E
.,c
u
.....
>-
'1:J

'"....0
1.0
1&.1
U)
Z
0
Q.
en
1&.1
a::
U)
en 0.5
1&.1
a::
I-
en

o tf
o 1.0 1.4 1.8

NORMALIZED EXTERNAL DIAMETER

Fig. 3-8: Active tangential stress response to norepinephrine

for canine iliac artery segments as a function of normalized
external diameter. Symbols are mean and bars + SEM (20).
CONTRACTILE BEHAVIOR OF ARTERIES 119

pv

-
Fig. 3-9: Diagrammatic representation of the distribution of
sympathetic nerves to the various vessels in the mesenteric
circulation of the cat. Vessels are pa, principal artery, sa,
small artery, tat terminal artery; pea, precapillary arteriole;
c. capillary; cv, collecting venule; sv, small vein, pv,
principal vein (22).
 120 CHAPTER 3

As would be anticipated from this functional distribution

of sympathetic nerves there are considerable differences in
the response of various elements
Range of Control In Vivo within the mesenteric circulation
to both sympathetic nerve
stimulation and the intraarterial infusion of norepinephrine.
A comparison of such responses is shown in Fig. 3-10 (22).
While only minor differences exist in the maximum constrictor
response to nerve stimulation in innervated arterial segments,
responses to norepinephrine in larger arteries of the mesenteric
circulation are somewhat lower than responses in the other
arterial vessels. Precapillary arterioles respond to norepine-
phrine infusion with a constriction response that is equal
in magnitude to that produced in small arteries and terminal
arterioles. Therefore, while precapillary arterioles are
not innervated they are nonetheless responsive to circulating
plasma neurohumoral agents. The relative magnitude of these
constrictor responses (40-60% for nerve stimulation) are
qualitatively similar to those constriction responses which
have been determined for isolated arterial segments in vitro.
Responses of similar magnitude have also been recorded on
perfused segments of small metacarpal arteries, dorsalpedal
and anterior tibial arteries (23,24,25). It appears, therefore,
that this magnitude of constrictor response is a general
characteristic of small to medium sized arteries and terminal
arterioles in all regions of the arterial tree.

There are only a few studies in the literature that

have documented changes in the elastic modulus of arterial
segments in vivo with changes in vasomotor tone. In general,
at physiological levels of blood pressure, i.e., 100 mmHg,
activation of smooth muscle usually produces a decrease in
elastic modulus whereas inhibition of smooth muscle tone
produces an increase in elastic modulus (26,27). An example
DOG of a response of a canine femoral artery to the topical
application of acetylcholine and subsequently of norepinephrine
is shown in Fig. 3-11 (26).

Under normal circumstances, a tonic level of sympathetic

nerve activity exists to vascular smooth muscle. Vasoconstrictor
responses involve a further increase in vascular smooth
muscle tone and vasodilatation is produced by a withdrawal
of sympathetic tone. This is also illustrated in Fig. 3-11.
CONTRACTILE BEHAVIOR OF ARTERIES 121

Nerve stimulation Noradrenaline application

100 pa sa ta pea pa sa ta pea
18
80
25
60
III
CLI
~ 40
o
a.
III

~ 20
L-
o 28
o
~
U
L-
~
II>
C
o
u
E
::J
60 pv sv cv pv sv ev
E
x 12 16

ell]
40
'"
1:
15 13
20

o
Cl+J17
Fig. 3-10: Maximum constrictor responses to perivascular nerve
stimulation (1 to 6 Hz) and to topical norepinephrine
(10-10 to 10-5 g/ml). Bars represent ± 1 SEM. Letters refer
to vessel types in Fig. 3-9. Numbers above bars represent the
number of vessels of each type studied. Constrictor response
was normalized by dividing by the initial control value of
diameter (22).
 ...,...,

200
150
" ...........
.... .,.... - , bAn_ry!• .,~
' . . . . .'...-.-...... - - - _ _~...-;:.~
.. ...7.~ - 1.....
W~.. bW*""'"

I 50 -.52

13.~
5.0
l
8
D
mm
1Act. 10-3
6

12.3
144
J 42
5 min

Fig. 3-11: The effects of the topical application of acetylcholine (10- 3 g/ml)
and of norepinephrine (10- 4 ) on the diameter of a canine femoral artery in vivo.
Numbers along the diameter response represent values of dynamic elastic modulus
in 10 6 dynes/cm 2 (26). ()
I
l>
-l
"
m
:u
w
CONTRACTILE BEHAVIOR OF ARTERIES 123

The characteristics described above are generally

properties of all arteries in the vascular system. Quantitative
differences may exist between
Properties of different anatomical locations but
Receptor Sites the general characteristics are
similar. This includes sites
within the arterial system which contain specialized receptors
that provide afferent information to the central nervous
system, i.e., the baroreceptors. In general, a complex but
direct relationship exists between the transmural pressure
to which these receptor sites are exposed and the activity
of these afferent nerves. The two principal sites of arterial
baroreceptors are the carotid sinus and the aortic arch.
These are the sites that have been studied in the most
detail as well. It should be recognized that many other
sites exist within the chambers of the heart and the vascular
system which contain receptors sensitive to hemodynamic
variables (28).

Morphological studies have shown that the receptors in

the wall of the carotid sinus and aortic arch are primarily
located in the adventitial and medial adventitial border
region of the blood vessels. As the blood vessel wall deforms
in response to changes in transmural pressure, the receptors
respond with a qualitatively equivalent deformation. The
deformation of these receptors is responsible for the production
of generator potentials which then summated in the individual
axons generates action potentials (29). It is presumed
that a direct relation exists between the deformation of
receptors and the deformation of the wall of the blood
vessel in which they are situated. This relationship has
not been clearly demonstrated to date experimentally. As we
shall see subsequently there is some question about the
nature of this relationship.

Changes in transmural pressure at these baroreceptor

sites usually produce changes in nerve activity. A considerable
amount of information exists describing the relationship
between receptor site pressure and nerve activity both in
individual nerve fibers and in whole nerve bundles. There
are some limitations in evaluating data from whole nerve
recordings, however, such as the contribution of individual
nerve fibers within a bundle.
124 CHAPTER 3

Nevertheless, the relationship between nerve activity and

pressure is reasonably well established. An example of the
relationship between pressure
Threshold Pressure and within the aortic arch and carotid
Saturation Pressure sinus, and integrated whole nerve
activity is given in Fig. 3-12 (30).
The relationship between nerve activity and pressure has the
following general characteristics: There is a threshold
pressure below which no nerve activity exists. As pressure
is raised above threshold there is a nearly linear relationship
between frequency of nerve activity and transmural pressure.
There is a high pressure value at which nerve activity
saturates at a maximum level. Differences in the threshold
and saturation pressure of different individual receptors
produces a sigmoid curve such as shown in Fig. 3-12 rather
than a curve having distinct sharp changes in nerve activity/
pressure relations. The results shown in Fig. 3-12 indicate
that the threshold pressure for the aortic arch receptors is
higher than that for carotid sinus receptors and that both
receptors appear to saturate at about the same level of
transmural pressure.

The relationship between nerve activity and mean transmural

pressure at the receptor site is a strong function of the
presence of a pulsatile pressure component as shown in Fig. 3-13
(31,32). In general, the nerve activity at a particular
value of pressure above threshold increases monotonically
with the amplitude of a pulsatile pressure component. The
magnitude of the increase, however, decreases with increasing
sinus pressure. At an above saturation pressure there is,
of course, no effect of pulse pressure on nerve activity.
As a consequence of this relationship, an increase in mean
arterial pressure and pulse pressure which often occur
simultaneously produce a more effective reduction of efferent
sympathetic nerve activity.

On the other hand, the aortic arch baroreceptors have

been shown to be essentially unresponsive to pulsatile
pressure (33,34). It is
Baroreceptors Unresponsive not clear, at the present time,
to Pulsatile Pressure why the carotid sinus receptors
should be sensitive to pulse
pressure and the aortic arch receptors not. Clearly the
carotid sinus receptors are responding to wall deformation.
CONTRACTILE BEHAVIOR OF ARTERIES 125

100 Mea n r SE ( n = 8)
E • Aor lic Nerve
/.

.~" 11 Sinus Nerve

o"
~

75

50

25

t-
o
40 80 120 160 200

MEAN AORTIC PRESSURE (mm Hg)

Fig. 3-12: Comparison of integrated nerve activity as a function

of arterial pressure from carotid sinus and aortic nerves in the
dog (N;8). Threshold pressure was 70 mmHg for the carotid sinus
and 100 mmHg for the aortic arch. (30).
126 CHAPTER 3

.....I»
GO
0
.::;.
5

....
~

>
i=
0
c(

IAJ 4
>
a::
IAJ
Z

2 +-----.-----r---~~--~----~----~--
o 30 60 90 120 150 180

INTRASINUS PRESSURE (mm Hg)

Fig. 3-13: Comparison of the response of integrated carotid

sinus whole nerve activity to static and dynamic carotid sinus
pressure variations. Curves are plotted versus mean intrasinus
pressure at a variety of superimposed pulse pressures. (32).
CONTRACTILE BEHAVIOR OF ARTERIES 127

Therefore, the presence of a pulsatile component of pressure

should produce an equivalent pulsatile deformation of both
the sinus or the aortic arch wall. One would expect both
receptor sites to respond in a qualitatively similar fashion
to pulsatile pressure variations. This is an area that
requires further research to elucidate the reason for this
difference if, in fact, one exists.

The response of the carotid sinus wall to mean pressure

and to pulse pressure can be explained in terms of the
mechanical properties of
Responses of Carotid Sinus the carotid sinus or aortic
arch per se. The mechanical
properties of these two sites possess the same characteristic
as arterial sites, i.e., a nonlinear relationship between
pressure and diameter (4,35). With increasing transmural
pressure, the elastic modulus of the carotid sinus increases
as shown in Fig. 3-14 (35). Associated with the increased DOG
elastic modulus, a reduction of wall strain occurs associated
with a pulsatile pressure of given amplitude. Since nerve
activity is directly related to wall strain, the nerve
activity associated with a given level of pulse pressure
would produce a smaller increment in afferent nerve activity
as mean sinus pressure increases. This is, in effect, the
result shown in Fig. 3-13. One finds, therefore, that with
increasing mean pressure, pulsatile pressure becomes a less
effective stimulus of baroreceptor nerve activity.

As with other anatomical sites in the arterial system,

the carotid sinus and aortic arch also respond to variations
in activation of their smooth muscle. The walls of these
baroreceptor sites contain postganglionic sympathetic nerve
fibers with numerous norepinephrine containing vesicles. It
has been shown that in response to efferent sympathetic
nerve stimulation and local perfusion with norepinephrine or
relaxing agents such as phenoxybenzamine that baroreceptor
sites are normally tonically activated and respond to smooth
muscle activation (4). At physiological levels of arterial
pressure, activation of postganglionic sympathetic nerve
fibers to the carotid sinus produces a reduction in carotid
sinus diameter and a reduction in dynamic elastic modulus of
the carotid sinus (36). Similarly, perfusion of the aortic
arch with norepinephrine produces a vasoconstriction in the
aortic arch smooth muscle and a reduction in diameter. The
infusion of penoxybenzamine produces a relaxation of the
smooth muscle and an increase in aortic diameter (4).
128 CHAPTER 3

15 T

, ! ! I ! •

/~ 80 100 120 140 160 180

CAROTID SINUS PRESSURE mmHg

Fig. 3-14: Variation of the dynamic elastic modulus of the

canine carotid sinus with carotid sinus perfusion pressure.
Points are means and bars + 1 SEM (N=7) of data averaged at
specific values of pressure (35).
CONTRACTILE BEHAVIOR OF ARTERIES 129

Changes in the degree of activation of carotid sinus

smooth muscle has significant effects on its reflex responses.
The effects of the introduction of norepinephrine into the
isolated perfused carotid sinus changes the relationship
between mean sinus perfusion pressure and arterial blood
pressure as illustrated in Fig. 3-15. This effect is
manifest as a shift of this curve downward and to the left
with increasing concentration of norepinephrine in the
perfusate (37). Since no pressure pulsations were present.
the activation of smooth muscle simply reduced the diameter
of the carotid sinus. If the receptors in the sinus wall
were functionally connected in parallel with connective
tissue and smooth muscle elements, a reduction in diameter
should have produced a reduction in receptor deformation.

This would have produced a decrease in afferent nerve activity

and an increase in arterial pressure. To the contrary,
norepinephrine produced a
Norepinephrine Effect decrease in systemic pressure.
This would suggest that afferent
nerve activity was increased by the norepinephrine. If the
receptors are functionally connected in series with the
smooth muscle cells an increased contractile force by smooth
muscle in response to norepinephrine could increase receptor
deformation and could, therefore increase afferent nerve
activity. However, morphological studies indicate that the
receptors exist in the adventitia and not in the media of
the carotid sinus. One possible explanation is that in
addition to its affect of smooth muscle on carotid sinus,
norepinephrine may have a direct action on the receptors
themselves. This area requires further research in order to
elucidate the mechanisms involved in the action of norepinephrine
on the carotid sinus wall and on afferent receptor mechanisms.

There are studies in the literature that suggest a

physiological role for these properties of receptor site
smooth muscle activation. It has been demonstrated that the
pressor response to carotid occlusion in the cat is decreased CAT
if the postganglionic sympathetic fibers to the carotid
sinus are stimulated during the occlusion (38). Release of
norepinephrine from sympathetic nerve endings activating
carotid sinus smooth muscle would decrease both diameter and
elastic modulus of the sinus wall (Bagshaw and Peterson, 1972)
(36). The pulse pressure in the carotid sinus would then
produce an increase in wall strain thereby increasing afferent
nerve activity.
130 CHAPTER 3

210

180

150

60

30

o 30 60 90 120 150 180 210

CSP (mmHg)

Fig. 3-15: Variation of mean aortic blood pressure (P a ) with

mean carotid sinus perfusion pressure (CSP). The right carotid
sinus was perfused while the left was denervated. The upper
curve shows data for controlled perfusion, the middle curve for
perfusion with 1 u g/ml norepinephrine added and the lower curve
with 5 u g/ml added. The points and bars are means and ± 1 SEM
of data averaged at specific values of carotid sinus perfusion
pressure. The solid line through the origin is the line of
identify, i.e., PA = CSP.
CONTRACTILE BEHAVIOR OF ARTERIES 131

If norepinephrine also affects the receptors per se, this

sympathetic release of norepinephrine will also directly
activate these receptors and also contribute to an increased
afferent nerve activity. The role played by the sympathetic
innervation of the carotid sinus and aortic arch is not
completely clear at this time but it is possible that this
mechanism acts to increase the sensitivity of these receptor
sites, making them more effective in opposing changes in
arterial pressure.

NEURAL REFLEX CONTROL MECAHNISMS

The reflex responses produced by the carotid sinus

baroreceptor mechanism have been studied using a variety of
direct and indirect techniques.
Carotid Sinus Reflex These techniques include
Responses carotid artery occlusion (39)
carotid sinus nerve stimulation
(40) carotid or neck counterpressure (41), and perfusion of
the isolated carotid sinuses (42). Obviously, these various
preparations are limited in terms of the information they
provide as well as in their applicability to different
animals including man. I will restrict my comments to
studies made on isolated perfused carotid sinuses in anesthetized
animals.

The isolated perfused carotid sinus preparation originally

developed by Moissejeff (1926) (42) allows for the quantitative
description of the control of systemic hemodynamics by the
carotid sinus in a preparation where changes in systemic
hemodynamics do not affect the carotid sinus pressure, that
is, an open loop preparation. When perfusion pressure in
the isolated carotid sinus is increased, afferent nerve
activity increases. This results in an inhibition of efferent
nerve activity to the various elements of the cardiovascular
system which is nonuniform in terms of its spatial or anatomical
distribution. The net effect is that increasing carotid
sinus perfusion pressure produces a reduction in systemic
arterial pressure. The response of systemic circulation to
increased carotid sinus perfusion pressure is strongly
dependent upon the presence and magnitude of pulsatile
perfusion component increases. Around a given mean carotid
sinus perfusion pressure there is generally a decrease in
systemic resistance as illustrated by the results in Fig. 3-16
(43). That is, carotid sinus perfusion pressure is more
effective in producing systemic hypotension when the amplitude
of pulsatile variation is increased.
 w
100~ o~ ""
'G;' 20
00
..,c 80
oS:
u I 40
~
'"
'c;:,
CII 80
u
- r 60
c 60
~
'"
'"
'M
CII
L.
L. 40
..,
;:,
u

>
'"u'"
'f 20
41
~

>-.
'"
""
o
80 100 1 20 140 160 180 200
Carotid sinus-aortic arch mean pressure (mm Hg)

(")
Fig. 3-16: Effects of pulsatile perfusion pressure amplitude on the variation of I
systemic vascular resistance and pressure in the combined carotid sinus - aortic arch »
areas. Values of pulse pressure are given along the ordinate axis along each curve. --I
"m
(43). :JJ
w
CONTRACTILE BEHAVIOR OF ARTERIES 133

The relative contribution of the various elements of

the cardiovascular system to the reflex control of arterial
pressure is extremely nonuniform. The changes in arterial
pressure that occur with changes in carotid sinus perfusion
pressure are the result of changes in both peripheral resistance
and cardiac output as shown in Fig. 3-17. (44,45). With
the vagi intact, the aortic arch mechanoreceptors buffer the
reflex effects of changes in carotid sinus pressure to some
degree. Prior to vagotomy approximately 75% of the change
in arterial pressure is the result of changes in peripheral
resistance, with the other 25% resulting from changes in
cardiac output. Following vagotomy, the contribution of
changes in cardiac ouput increases to approximately 35% of
the change in arterial pressure. These changes in cardiac
output are the result of changes in myocardial contractility
as well as changes in preload (left atrial pressure) and
after load (aortic blood pressure) subsequent to changes in
carotid sinus perfusion pressure. The results in this
figure also illustrate the nonuniform effect of the aortic
arch mechanoreceptors on the system circulation. This
nonuniform contribution is indicated by the differences in
these two curves at given values of carotid sinus pressure.
At high carotid sinus pressure the aortic arch afferents
have a minor effect. At low values of carotid sinus pressure
their effect, on the other hand, is much larger.

The variation in peripheral resistance with carotid

sinus pressure is nonuniformly distributed throughout the
various vascular beds (45).
Variation in Peripheral A summary of variation of blood
Resistance with Carotid flow resistance in the celiac,
Sinus Pressure superior mesenteric, renal and
femoral arteries is shown in
Fig. 3-18. Following vagotomy, the unmasked effects of carotid
sinus baroreceptors is indicated by the closed symbols. The
largest variation of regional resistance with carotid sinus
perfusion pressure occurs in skeletal muscle (femoral artery).
Resistance in the splanchnic circulation and primarily in
the celiac artery, on the other hand, is much less sensitive
to changes in carotid sinus pressure than the changes in
peripheral resistance. This diagram also illustrates the
nonuniform contribution of aortic arch baroreceptors to
regional resistance control, both at various values of
carotid perfusion pressure and in the different vascular
beds. The effects of the aortic arch are largest in the
mesenteric circulation and smallest in the skeletal muscle.
Reasonably large effects also occur in the celiac and renal
circulation.
 to>
-I>-

w
w 250 u
a:: I- 4
:::> :::> ~ 7 I
(f)
(f)
t··!T"i , Il.
w I- ~ 1'.1
a:: :::>
Il. o t3 T\ 1
150'i1
.\\ - , a:
U -'
-'
T'',I
<t <t
a:: , 0 31- l ::i "l
w
I- 1\ 'i.,
~. ~
t/ Tvt-t~~<t5~>,
it
a::
<t ffi
0 I
l-f-l ~f..t
150 300 I Ill. I I
1o ro 150
f
o 300 00 150 300

CAROTID SINUS PRESSURE

Fig. 3-17: Variation of mean arterial pressure, cardiac output and peripheral
resistance with (mean) perfusion pressure in the isolated perfused carotid
sinuses before (open circles) and after (closed triangles) bilateral cervical
vagotomy in dogs (N=ll). Symbols are means and bars ± 1 SEM of data averaged
at specifc values of carotid sinus pressure.

(")
I
»
-c
-I
m
:0
w
CONTRACTILE BEHAVIOR OF ARTERIES 135

60 Ctlliac Mesenftlrlc
100

If)
E
~
U

• 50
III

•>-
c:
"0
It)
o
01~------~------~
o
o L-______ ~L_ _ __ _ _ _~

1&.1 150 300 o 150 300

(,)
Z
«
H"i.~mw.'
~ 100 Renal 500
en
en
1&.1
a::
...J
«
1
z 250
2
(!)
1&.1
a::
o L-______ ~L_ _ __ __ _~
O L---------~------~
o 150 300 o 150 300

CAROTID SINUS PRESSURE (mmHg)

Fig. 3-18: Variation of regional resistance with mean carotid

sinus perfusion pressure before (open circles) and after
(closed triangles) bilateral cervical vagotomy.
136 CHAPTER 3

The effects of baroreceptor reflexes on blood flow in

different vascular beds will be the result of the combined
changes in regional resistance and cardiac ouput. As indicated
by the data summarized in Fig. 3-19, significant differences
exist in the changes of regional blood flow with carotid
sinus perfusion pressure. In general, regional blood flow
in the celiac, mesenteric and renal beds remains reasonably
constant with changes in carotid sinus perfusion pressure
after vagotomy. On the other hand, blood flow in the femoral
artery significantly increases with carotid sinus pressure
both before and after vagotomy.

The relative distribution of cardiac output with activation

of carotid sinus afferents has also been demonstrated in a
slightly different manner by a number of investigators. For
example, with carotid sinus hypotension following carotid
occlusion skeletal muscle blood flow generally declines
while splanchnic and renal flow increase (46). In addition,
in response to stimulation of the carotid sinus nerve,
skeletal muscle blood flow increases while splanchnic and
renal flow decrease (40). These responses with an intact
closed loop function of the carotid sinus reflex support the
validity of these data obtained in this open loop preparation.

In addition to determining the effects of the carotid

sinus on regional resistance, we have also performed studies
to assess the effects of carotid sinus reflex on large
vessels (45).

This has been done by determining values of vascular impedance

from pulsatile pressure/flow measurements at the above
arterial sites. Vascular
Impedance impedance is conceptually analogous
to electrical impedance in a circuit
and represents the ratio of an a.c. pressure to an a.c.
flow. Values of vascular impedance can be determined from
pulsatile pressure and flow using either Fourier series or
time series analysis (47). By such techniques, pulsatile
pressure and flow are resolved into a number of harmonic
components at different frequencies. In this way, a frequency
spectrum of impedance can be determined It has been demonstrated
that the mechanical properties of large arteries close to
the measurement site of pulsatile pressure and flow primarily
contribute to the determination of the high frequency values
of impedance at that site (48). Accordingly, we have used
measurements of vascular impedance to determine the properties
of large vessels from the high frequency portion of the
impedance spectrum.
CONTRACTILE BEHAVIOR OF ARTERIES 137

700 Celiac 600 Mesenteric

500 400

c:
'E
l--+--t-l+-l--t-++-f
"-
E Of
0 300
or0 300
~
0
...J
Renol 200 Femoral
lL 500
...J
«
z
Q
(!)
~ ,f-f,f
w --~-+--~--+--~--t- -t

~~
a: 100
250 +--t~

00 00 300
300

CAROTID SINUS PRESSURE (mmHg)

Fig. 3-19: Variation of regional blood flow with carotid sinus

perfusion pressure. Symbols are as before.
138 CHAPTER 3

The results summarized in Fig. 3-20 (45) show the

variation of impedance spectrum in the ascending aorta,
renal, celiac and femoral arteries at three values of carotid
sinus perfusion pressure before and after vagotomy. At the
more peripheral sites not only is there a downward shift in
the d.c. value of impedance i.e., resistance, with increasing
carotid sinus pressure but also the a.c. portion of the
impedance spectrum is likewise decreased. When high frequency
values of impedance are averaged and plotted as a function
of carotid sinus pressure, results of the form shown in Fig.
3-21 are obtained (45). Associated with the inverse relationship
of regional resistance and carotid sinus pressure is a
similar but quantitatively different variation in regional
impedance. In the femoral artery, the variation of impedance
with carotid sinus pressure is smaller than that of femoral
resistance. On the other hand, the variation in tbe renal
artery is the same order of magnitude. These results suggest
that values of vascular impedance which once again reflect
the properties and geometry of large vessels are under the
reflex control of the carotid sinus.

The variations in impedance of the ascending aorta are

considerably different than those of the peripheral vascular
beds. These data are shown in Fig. 3-22. (45). While an
inverse relationship exists between aortic resistance and
carotid sinus pressure, aortic impedance shows a minimum at
normal values of carotid sinus pressure, both before and
after vagotomy. The physiological significance of this
minimum plays some role in the energetics of cardiac contraction,
perhaps by minimizing the pulsatile work of the heart (45).

While these results suggest that changes in the properties

of large arteries do occur they are certainly not definitive.
However, taken together with the results previously presented
on the effects of efferent sympathetic nerve activity and
catecholamines on the properties of large vessels it seems
reasonable to interpret these results to indicate that
neural reflexes exert a tonic activity on the smooth muscle
in large arteries. Furthermore, this tonic activity does
have a significant hemodynamic effect in large arteries by
virtue of its effect on vascular impedance.

There have been relatively few studies of the effects

of aortic arch baroreceptors on the reflex control of cardiovascular
function. In general, the results
Aortic Arch Reflexes have indicated a similar though
quantitatively smaller effect of
aortic arch afferents on the heart and the vascular system.
CONTRACTILE BEHAVIOR OF ARTERIES 139

AORTA
RENAL
Int.ct Cut I n t.ct Cut

.,
E 2
~
u
C1l
In
C1l
C
>. 2
"0

w
U ---~'
z
L . '- _ _- '
, !:-' !:-'_ _ _~' L . '_ _ _- ' ,

0 15 0 15 o IS 0 15
«
o
w
a.. FEMORAL
~ 105 CELIAC
Intoct Intoct
...J
«
z
o
o
W
0:::
2

o IS
~--'
o IS o 15
o 15

FREQUENCY (Hz)

Fig. 3-20: Regional vascular impedance spectra at four arterial

sites before and after bilateral cervical vagotomy. Data are
g~.ven for three values of carotid sinus perfusion pressure:
25 (solid square), 125 (open circles) and 250 mmHg (solid
triangles). Points are means and vertical bars + 1 SEM. (45).
140 CHAPTER 3

2 2
eli/file

w ..• . -; -+ -;
u
z
<l , 0'(

,
0 0'"
2 0
w 0
a. 2
~

..J
· ···i.
Rllnl1l
2
f{ FllmOrl1/

<l
Z
0

,.,--;-++-,
<.9
w
a:::
~
°0 2 0'[
0

CAROTID SINUS PRESSURE

Fig. 3-21: Variation of regional resistance (open circles) and

characteristic impedance (closed squares) with carotid sinus
perfusion pressure. The variables are all normalized by dividing
by normal control values of each quantity. Vertical bars are
+ 1 SEM. Data were obtained after sectioning the cervical
vagi. (45).
CONTRACTILE BEHAVIOR OF ARTERIES 141

1.8 Vagi intact

1.0

w
<.)
z 0 1:
<t
0 0 2
w
a..
~ 1.8 Vagi cut

<.)
l-
n::
0
<t 1.0

o1 L _ _ __ --1...-_ _ _- - - - '

o 2
CAROT ID SINUS PRESSURE

Fig. 3-22: Variation of aortic resistance (open circles) and

aortic impedance (closed squares) with carotid sinus perfusion
pressure. Variables were normalized as described previously.
(45).
 142 CHAPTER 3

As described above in previous sections, there are some

quantitative differences in the afferent characteristics of
aortic arch baroreceptors. Studies using isovo1umic contractions
of the isolated left ventricle have demonstrated that the
aortic arch mechanoreceptors do have a significant effect on
the contractility of the left ventricle (49). In addition,
other studies have indicated that the aortic arch receptors
do exert a significant effect on regional resistance at
least in the renal and skeletal muscle circulation (50,51). In
this control of regional circulations, there is a very
strong interaction between the aortic arch and the carotid
sinus baroreceptor mechanisms. This interaction is exemplified
by the results in Fig. 3-23 (50). This figure compares the
DOG control of the constant flow perfused canine hind limb and
heart rate by these two receptor mechanisms. The effects of
varying aortic arch perfusion pressure on these two variables
is a strong function of the perfusion pressure existing in
the isolated carotid sinuses. When carotid sinus pressure
is high and efferent nerve activity low, limb resistance is
essenta11y unresponsive to aortic arch perfusion pressure
while heart rate becomes strongly dependent upon the latter.

A number of attempts have also been made to analyze and

represent the interaction between carotid sinus baroreceptors
and other types of receptor mechanisms including chemoreceptors
and cardiopulmonary receptors. Activation of chemoreceptors
by hypoxia and hypercapnia produce an augmentation of vasomotor
reflex responses initiated in the aortic arch, carotid sinus
and cardiopulmonary areas (52,53,54). In addition, changes
in left atrial pressure also produce profound interaction
with carotid sinus induced reflex responses. Increases in
left atrial pressure antagonize the vasodilator response in
the renal and skeletal muscle vasculature associated with
carotid sinus hypertension. Hypotension in the left atrium
considerably augments the reflex vasoconstriction in the
renal and skeletal muscle circulations associated with
carotid hypotension (55). Most of the studies involving the
interaction of baro - and chemoreceptor mechanisms have
involved the constant flow perfusion of isolated vascular
beds such as the limb and the kidney. Such studies are
difficult to interpret because they artificially maintained
regional blood flow at a constant level. In contrast,
reflex alterations in cardiovascular function are associated
with changes in the distribution of cardiac output especially
to skeletal muscle.
CONTRACTILE BEHAVIOR OF ARTERIES 143

180

Cl 160 Carot id Pressures

:r
E ••- - - . 90
120 (mm Hg)
! 140
0-0---0
"'",.- - - I eIe 1 50
...
<II
:I
CII
120
CII
...
<II
0.
100
Jl
E
...J 80 o
)(

60 lC l5 II
)(
);

C 120
E
"!!I'CI
<II
100

....
-
Jl

<II 80
I'CI
.0::
60

40

l I I
o 40 80 120 160 200 240 280 320 360
Aortic Pouch Pressure (mm Hg)
Fig. 3-23: Variation of perfusion pressure in the isolated
constant flow perfused canine hindlimb and of heart rate with
pressure in the isolated aortic arch at three values of perfusion
pressure in the isolated carotid sinus. (50).
144 CHAPTER 3

The significance and quantitative aspects of the inter-

action of these reflex mechanisms under conditions in which
blood flow changes occur and are influenced by autoregulatory
responses as well are not clear at this time. This area
certainly represents one in which much more experimental
study is necessary before we understand the physiological
significance and role of these mUltiple baro- and chemo-
receptor reflex mechanisms in the regulation of cardiovascular
function.

It is important to realize that not only does the

function of the carotid sinus baroreceptor mechanisms interact
and depend upon the function
Hypothalamic Interactions of other peripheral vascular
receptor mechanisms but it is.
in addition. affected by higher central nervous system
structures. One in particular. the so-called hypothalamic
defense response has a significant effect on the regulatory
function of the peripheral circulation produced by the
carotid sinus (56). Stimulation of specific areas in the
hypothalamus produce cardiovascular responses which are
similar to those occuring in stress and exercise (57). It
is for this reason that the interaction between this hypothalamic
defense area and the medullary cardiovascular centers is of
importance. The data in Fig. 3-24 illustrate the cardiovascular
responses to hypothalamic defense area stimulation at different
levels of perfusion pressure in the isolated carotid sinuses
(57). At low levels of carotid sinus pressure in particular.
stimulation of the hypothalamus produces increases in arterial
blood pressure and cardiac output. The distribution of
cardiac output and the response of individual vascular beds
is nonuniform. Renal blood flow is significantly reduced
while skeletal muscle blood flow is considerably elevated by
hypothalamic stimulation. As carotid sinus pressure is
increased the magnitude of the response to hypothalamic
stimulation is attenuated. This illustrates the strong
interaction between the hypothalamus and the carotid sinus
on efferent neural outflow from the medullary centers. At
physiological levels of arterial pressure the responses
produced by hypothalamic stimulation mimic those that occur
in stress and exercise in man and experimental animals.

While the above quantitative description of the carotid

sinus reflex is based primarily upon animal experiments. the
function of this control
Comparative Aspects of mechanism is at least qualitatively
Neural Control similar in man. There are very
few quantitative studies in man
that can be used for comparison with experimental animal results
CONTRACTILE BEHAVIOR OF ARTERIES 145

A B c D E
:;:480[
~ 0 -----

=I~ 320[9 JOb.' ".

~ 0
"Pus . ,
• • &
JU. iJJJ I

320
=[~ -
0-----
ImmHg) 0

ASC£N>iNG
AORTIC
FLOW •
CARDIAC
aJTf'UT

RENAL
0- -
FLOW

FEMORAL
FLOW

MEART 120[
R.ttrE 240 ----- t----!
30 sec
~
• •

Fig. 3-24: Cardiovascular response to hypothalamic stimulation

at four values of carotid sinus perfusion pressure. Duration of
stimulation is indicated by the dark bars at the bottom of the
figure. Values of carotid sinus pressure were 60 (A and E),
120 (B), 180 (C) and 240 (D) mmHg. (57).
 146 CHAPTER 3

for obvious reasons. There are data from two types of

DOG experiments, however, that can be compared. The first is
the effects of vasoactive drugs on heart period and arterial
pressure, and the second is the effects of carotid sinus
nerve stimulation. Both of these maneuvers have been performed
in unanesthetized dog and man.

The reflex increase in heart period in response to the

hypertension following the intravenous infusion of phenylephrine
has been used as an index of baroreceptor sensitivity (58).
In normal supine humans, Bristow, et al. (59) found the
relation between heart period and diastolic blood pressure
during phenylephrine injection (IV) to be linear with a slope
HUMAN of 16.7 + 1.9 msec/mmHg, Higgins, et al. (60) performed
DOG similar studies in conscious dogs and found a slope of 22.4
+ 2.3 msec/mmHg in normal animals. While the slope for
dogs is higher than for man. the difference is not statistically
s ignif ican t •

In response to carotid sinus nerve stimulation, arterial

pressure and peripheral resistance both decrease. In man
these decreases averaged 23% and 16% respectively (61). In
the dog. these decreases were similar at 28% and 29%. respectively
HUMAN (40). Cardiac output was decreased significantly in man (by
DOG 8%) but unchanged in the dog. Forearm vascular resistance
decreased by 16% in man during carotid sinus nerve stimulation
as compared to 62% in the canine hindlimb. Heart rate is
significantly reduced in man (9%) but unchanged in dog
following a brief. transient decline. During exercise,
carotid sinus nerve stimulation produced a larger decrease
in arterial pressure in the dog (31% vs. 16%). Skeletal
muscle vascular resistance also shows a larger decrease in
the dog. These results suggest that carotid sinus reflex
responses in man are qualitatively similar to those of the
canine, with some quantitative differences.

The characteristics of the carotid sinus reflex described

above are those of the normal, healthy subject. Pathological
conditions are associated with
Pathological Changes in the substantial changes in the
Carotid Sinus Reflex characteristics of this reflex
mechanism. Perhaps the best
known of these changes is the baroreceptor resetting associated
with arterial hypertension (62).
CONTRACTILE BEHAVIOR OF ARTERIES 147

The afferent properties of the baroreceptor areas (carotid

sinus and aortic arch) are reset in established hypertension
in such a fashion that a higher transmural pressure is
necessary to produce a given level of afferent nerve activity
(62,63). Resetting usually involves a shift of the nerve
activity transmural pressure curve to the right. A decreased
receptor sensitivity (i.e., slope) and an increased threshold
pressure are usually observed in such conditions.

Similar changes in arterial baroreceptor properties

have been reported in experimental atherosclerosis (64) and
Vitamin D sclerosis (65). Aortic baroreceptor resetting,
decreased sensitivity and increased threshold have all been
described. A reduction in the distensibility of the receptor
areas has been suggested as at least a primary cause of
these observations.

A decrease in the baroreceptor reflex responses in

experimental heart failure in the dog has also been reported
(60). In this study, baroreceptor sensitivity was determined
from the phenylephrine induced heart period-arterial pressure
relation, and from bilateral carotid occlusion (carotid
sinus hypertension) responses. Similar conclusions have been
reached concerning the effects of heart disease in man. The
precise mechanism(s) responsible for this observation is not
known at this time.

Studies of changes in baroreceptor sensitivity in man HUMAN

and rat have shown a decrease associated with aging. These RAT
demonstrations have been made
Changes in Baroreceptor using the heart period-arterial
Sensitivity Associated pressure relation following
with Aging phenylephrine (66,67). We have
also compared baroreceptor
reflexes in young and old dogs using the isolated perfused DOG
carotid sinus preparation. Data showing the variation of
mean arterial pressure and peripheral resistance with
carotid sinus pressure are shown in Fig. 3-25 (68). These
results show that if anything, carotid sinus control of these
two variables is more sensitive in the older animals. A
question therefore exists concerning the validity of the heart
period arterial pressure relation as a general measure of
baroreceptor sensitivity.
148 CHAPTER 3

2 2
o Young

,.
~i:-'" r-:-;'t"- • Old

.. ,.! \
,.~

.....
'.
L~, .
W w ~9--.
a::: (,)
i""'~·O
:::>
en ~"i~ z
<:(
en I-
oL0 ,
~
w , en
a::: 2 en 2
-2
a.. -2 W
a:::
(,)
(,)

j;~
I- ~
a:::
0
a:::
0 /.~
,y• .0.\
<:( -I
-I <:(

rf .. ...
/"/ ?'
.!
-' .~.

.. \0."
d • '!><o~.
/- "'0\
00 !! ~I 00 ~
2
2

NORMALIZED CAROTID SINUS PRESSURE

Fig. 3-25: Variations in mean arterial pressure and peripheral

resistance with changes in carotid sinus pressure in the dog.
CONTRACTILE BEHAVIOR OF ARTERIES 149

Significant changes occur in the distribution of cardiac

output to fulfill temporal variations in the requirements of
the regional circulations in a
Carotid Sinus and Blood variety of normal and abnormal
Flow Distribution states. As shown above, an
increase in carotid sinus
pressure produces a redistribution of cardiac output with
blood flow decreasing in the renal and hepatosplanchnic
areas, and increasing to the extremities or skeletal muscle.
This pattern of blood flow changes is identical to those
reported in pressor responses to carotid sinus nerve stimulation
in the dog (40) to hypothalamic defense area stimulation
(69), to acute emotional stress (70,71), to central sciatic
nerve stimulation (72) and to intravenous norepinephrine
infusion (73). Also, a similar pattern of blood flow redistribution
occurs in moderate exercise in the dog which is associated
with an elevation in arterial pressure (74). Finally, essential
hypertension in man is also associated with the same change
in blood flow distribution compared with normal man (75).
In the case of acute, compensated stage of hemorrhagic shock
in the dog, the exact opposite blood flow response has been
reported (76).

Therefore, it appears that the distribution of cardiac

output is strongly influenced by neural control mechanisms.
These mechanisms can be viewed as representing the first
line of control of the peripheral circulation. Obviously,
other factors (such as hormonal, metabolic, etc.) interact
with neural mechanisms to determine the overall response in
a particular circumstance. However, the carotid sinus
reflexes are more than simply blood pressure controllers.

HORMONAL CONTROL MECHANISMS

There are a large number of hormonal and chemical

agents within the body which have significant effects upon
the cardiovascular system. Since it is not possible to deal
with all of these agents at this time, this discussion will
concentrate on only two of them; angiotensin and vasopressin.
These two hormonal agents have significant pressor effects
on the vascular system. These effects are mediated in
several ways and have a nonuniform effect on different
elements of the cardiovascular system.
150 CHAPTER 3

Angiotensin is formed in blood through the action of an

enzyme on a normally circulating plasma substrate. Through
a series of steps on octapeptide
Angiotensin angiotensin II is formed in
plasma. The enzyme responsible
for the conversion of the inactive precursor to the active
angiotensin II is renin. Renin is released from the kidney
by a variety of mechanisms including a reduction in renal
perfusion pressure, increased renal sympathetic nerve activity,
circulating catecholamines, and reduced sodium delivery to
the distal nephron in the kidney (77).

Once angiotensin is formed by the action of renin, the

hormone circulates to all parts of the cardiovascular system
where it has profound effects both on the heart and on
vascular smooth muscle (78). The action of angiotensin on
the cardiovascular system is mediated in several ways.
Angiotensin has a direct action on the membrane of cardiac
and vascular smooth muscle. In addition, angiotensin potentiates
the action of circulating catecholamines on these two structures
as well. Angiotensin also has a stimulatory action on
sympathetic ganglia. The effects of angiotensin on the heart
are diverse with a net effect resulting from the direct
action of angiotensin on cardiac muscle and its indirect
effects mediated through changes in arterial pressure and
venous pressure. In the intact animal, low doses of angiotensin
generally have no effect on cardiac output. At high doses,
however, cardiac output is generally decreased in response
to angiotensin which is primarily the result of the increased
afterload, that is, arterial blood pressure.

On the peripheral circulation, angiotensin has a diverse

action. Blood flow in almost all vascular beds is decreased
in response to low and high doses of infused angiotensin
(79). As a fraction of cardiac output, however, blood flow
to skeletal muscle and bone are significantly increased
during angiotensin infusion. Renal, skin and splanchnic
blood flow response to angiotensin is qualitatively similar
to that produced by the carotid sinus baroreceptor mechanisms
described above. As a result, the total effect of circulating
angiotensin on blood flow distribution depends in part on
the action of the changes in cardiac output that accompany
changes in angiotensin concentration in plasma. In addition
to its effect on small blood vessels, angiotensin has a
direct vasoconstrictor action on large vessels such as the
femoral artery. Again, this action is the combined effect
of the direct action of angiotensin on the muscle as well as
a facilitation of the effects of norepinephrine.
CONTRACTILE BEHAVIOR OF ARTERIES 151

Vasopressin is a peptide hormone released from the

posterior pituitary gland. It is released in response to
decreased extracellular fluid
Vasopressin volume, reduced arterial blood
pressure, increased plasma
osmotic pressure, and other stimuli such as pain, stress and
exercise (80). Hemorrhage is a very potent stimulus for the
release of pituitary vasopressin. In addition to its
action on the kidney where it promotes water retention,
vasopressin also has a significant action on the cardiovascular
system. When infused intraarterially at physiological
concentrations, vasopressin increases arterial blood pressure
and peripheral resistance in a dose dependent manner. In
addition, right atrial pressure is usually increased by
vasopressin administration. Cardiac output generally decreases
with the intravenous infusion of vasopressin primarily as a
result of an increase in cardiac afterload (81). The action
of vasopressin on the peripheral circulation is once again
relatively nonhomogeneous. In general, the fraction of
cardiac output delivered to the mesenteric and renal circulations
increases whereas iliac artery blood flow decreases as a
fraction of cardiac output. In absolute terms, all these
regional blood flows decrease significantly in response to
vasopressin.

Although both of these agents, angiotensin and vasopressin

have significant actions on the cardiovascular system, their
role in normal maintenance of blood pressure is not established.
In the case of angiotensin there is no evidence that angiotensin
contributes to the maintenance of arterial blood pressure in
the normal person. Studies which have employed the use of
angiotensin blocking agents or converting enzyme inhibitors
have illustrated that no significant effect on arterial
blood pressure occurs in normal man. While vasopressin
certainly has an effect on circulating blood volume by
virtue of its action on the handling of water by the kidney
there is no evidence to suggest that it plays a significant
role in the regulation of the distribution of cardiac output
or of cardiac output, per se. Vasopressin is primarily
involved in the regulation of blood volume.

PATHOGENIC SIGNIFICANCE OF CONTROL MECHANISM

From the above discussion, it should be apparent that

neurohumoral factors exert a significant degree of control
over arterial blood flow as well as the mechanical properties
of large arteries. What is the significance of this control in
the function of larger arteries as related to atherogenesis?
152 CHAPTER 3

It is only possible to provide some speculation on possible

answers to this question. By virtue of their action on
contractile proteins, neurohumoral factors can and/or could
effect the following: a) arterial geometry, b) arterial wall
mechanics, c) vasa vasorum function, and d) endothelial
permeability.

Changes in arterial wall diameter in response to neurohumoral

signals can be substantial. The limited information available
in the literature suggests that
Effects on Wall small changes in wall geometry can
Shear Stress produce significant changes in wall
shear stress, by virtue of its
effects on blood velocity profile at the wall (82,83).
Changes in wall shear stress could occur as a result of
reflex adjustments in wall geometry, blood flow rate and
cardiac frequency. All of these factors can effect the
velocity gradient at the blood-wall interface, and therefore
wall shear stress. A direct link between wall shear stress
and atherogenesis has been suggested by several investigators
(84,85).

Neurohumoral control factors also modify the arterial

wall mechanical properties. The strain energy density
stored within the arterial wall depends upon its stress-
strain relations, i.e., its mechanical properties. Activation
of smooth muscle in large arteries produces an increase in
the strain energy density in the physiological pressure
range. Fry (86) has hypothesized that an increase in the
strain energy density of the arterial wall can facilitate
macromolecular transport across the endothelium. Again,
this provides a direct link with a proposed atherogenic
mechanism.
The outer layers of the media and the adventitia of
large to medium sized arteries have a vascular supply from
the vasa vasorum. This
Possible Role of microcirculatory network has
Vasa Vasorum been described as having
characteristics of small blood
vessels elsewhere, including its own "wall" containing
innervated smooth muscle. Do the vasa vasorum participate
in neurohumoral control? Specifically, is blood flow in the
vasa vasorum affected by neurohumoral agents? I am not
aware of any evidence in the literature on this subject.
Yet, the vasa vasorum has been suggested to playa role in
atherogenesis (87,88), especially in the relation between
hypertension and atherosclerosis (89).
CONTRACTILE BEHAVIOR OF ARTERIES 153

The role of the vasa vasorum in the normal function of the

arterial wall is not completely understood, nor is its
relation to the arterial pathology. Much more work is
needed in this area to clarify these relationships.

Recent studies have shown that the endothelial cells of

large arteries contain the contractile proteins actin and
myocin as well as the regulatory proteins, tropomyosin and
troponin (90). It has been suggested that activation of
these contractile proteins by angiotensin. epinephrine.
cholesterol. prostaglandin E. and serum triglycerides
produces an increase in endothelial permeability and the
influx of macromolecules (91.92). Thus. another potential
link exists between neurohumoral factors and atherogenesis
in the contraction of endothelial cells. It should be
pointed out that while this phenomenon is controversial,
critical experiments remain to be performed. Thus. while
endothelial cells contain contractile material. it has not
been demonstrated to date that the myosin possesses ATPase
activity. Furthermore. it has not been demonstrated that
these agents have actions on endothelial cells per set or on
interendothelial cell junctions.

These speculations indicate that several important

effects of neurohumoral agents exist on the arterial wall
and various determinants of its function. Thus. neurohumoral
factors may play an important modifying role in the response
of the arterial wall to its environmental state and in
particular to the events initiating atherosclerosis.

DR. DEWEY: One of the interesting questions that comes

up with regard to the carotid sinus control is that the
control arising from changes in configuration of the arterial
wall could change as the properties of the carotid sinus
itself change. For example during the course of disease. I
could imagine that in a calcific artery the amount of stretch
which would activate the nerves in the carotid sinus would
change quite dramatically from the normal giving false
signals to the regulatory functions of the body.

DR. COX: That is quite correct. There is very little

information in the literature about this subject especially
with regard to man. There are. however. at least three
conditions which have been identified. associated with
changes in baroreceptor function - one. of course. is hypertension
and I am sure we are all very familiar with baroreceptor re-
setting in both experimental and natural forms of hypertension.
 154 CHAPTER 3

But recently Jennifer Angell-James has demonstrated that in

experimental atherosclerosis (64) and Vitamin D sclerosis
RABBIT (65) in the rabbit there are changes in aortic arch baroreceptor
afferent function. I stress afferent function because that
is as far as she went. She was able to show that there are
shifts in the relationship between nerve activity and pressure
within the aortic arch. Basically, the shift is in such a
fashion as to reduce the sensitivity of the aortic arch
baroreceptors. At a given level of transmural pressure
there is a smaller nerve activity so that one would expect
an increased arterial pressure. In the case of the experimental
atherosclerosis study, she did find an elevation of arterial
pressure in these animals over the controls, so there is at
least some evidence that altered baroreceptor function
exists in atherosclerosis. I don't really know of any
direct evidence of the effects of arterial wall pathology on
carotid sinus reflex in man other than that which can be
inferred from indirect information. I don't know any information
on the carotid sinus reflex in atherosclerosis, although
that may change next year.

DR. CARO: I would like to comment that there is

strong evidence that the endothelial cells provide an important
resistance for the transport of macromolecules between the
blood and the artery wall, and suggest evidence that deformation
of them in some way enchances that transport. The important
point in the present context is that a decreased distens-
ibility of the artery wall will diminish the deformation
suffered by the endothelial cells; their deformation will
largely become that due to shearing stress.

DR. SCHWARTZ: I wonder if Dr. Cox would speculate on

the possible clinical implications of a non-deformable or
rigid carotid sinus. What intrigues me specifically is the
possible role that loss of the baroreceptor mechanism might
have in submitting the intercerebral circulation to a very
high peak of systolic ejection pressures. Could such peak
pressure be important in the development of cerebrovascular
disease, and, in particular, intercerebral hemorrhage?

DR. COX: With regard to the control of blood pressure

there is no doubt that changes associated with atherosclerosis
will be of significance. There was a paper by Heath et al.,
(93) who showed some information on the morphology of
the human carotid sinus in atherosclerosis and there were
certainly some very impressive looking lesions in the carotid
sinus.
CONTRACTILE BEHAVIOR OF ARTERIES 155

Winson et al., (94) studied force-length relations of

strips of carotid sinus from normal and from atherosclerotic
individuals. There is a very
What Sets Arterial Pressure definite increase in stiffness in
the latter. There is no doubt
that it is going to effect carotid sinus reflex control of
blood pressure. Fortunately for us there is considerable
redundancy in neural control mechanisms. My feeling would
be that if the carotid sinus were the only site of disease,
the other arterial mechanoreceptor sites would have the
capacity to regulate arterial pressure. Probably the individuals
would lose their ability to minimize fluctuation in arterial
pressure so that a given degree of activity would produce
much larger arterial pressure responses. For example,
exercise, orthostatic changes or what have you. The other
interesting thing is this whole idea of baroreceptors resetting.
If baroreceptors can reset with chronic alterations either
in pressure or chronic alterations in mechanical properties
of the wall of the receptors, then what is it that determines
arterial blood pressure? Mechanoreceptors are really following
the lead of this controller which sets the level of arterial
pressure. Engineers like to think of control systems in
terms of a set point and if we translate this to man, we
think of something in the brain that determines blood pressure
and it is the dial on the temperature control in a room, or
it is the speedometer on the automobile. We don't know the
analogous function of the body that sets arterial pressure.

DR. CHIEN: How were the impedence measurements made?

What was the method?

DR. COX: Blood flow was measured with electromagnetic

flow meter and pulsatile pressures were recorded with strain-
gauge manometers. The pressure and flow data were subjected
to time series analysis, Fourier series. From Fourier
series analysis you get a series of harmonics at integral
multiples of the heart rate. For each of these mUltiples,
we take the ratio of pressure and flow and use that information
to compute impedance.

DR. CHIEN: The other question is, if there is any

separation of flow in the carotid sinus as Dr. Dewey showed
yesterday in the model, how would this affect the baroreceptor
reflex?

DR. COX: Our experiments were performed using a low

flow. It is creeping flow so that there is no separation.
156 CHAPTER 3

As far as the physiological question is concerned in vivo

there is no information available on flow separation in the
carotid sinus. One of the things that I did not mention is
that the carotid sinus nerve activity is not only sensitive
to changes in diameter of the sinus wall, but also to longitudinal
deformation of the sinus as well. One could certainly
visualize the formation of an atherosclerotic plaque on the
free border of the sinus producing a very severe restraint
to the longitudinal motion of the wall.

This would certainly have an effect on the carotid sinus

reflex. Basically, it would reduce sensitivity because it
will not allow for longitudinal deformation to activate the
nerve endings, and there would be a smaller amount of afferent
nerve activity for a given arterial pressure pulse.

DR. KENYON: In terms of long term pressure regulation,

a stiff carotid sinus means the stretch activity of the cell
is diminished ••• What role do you see for kidney pressure
regulation in terms of it being able to overcome the deficiency
in the carotid sinus?

DR. COX: This is one of the arguments of the Guytonian

school. Alan Cowley in Arthur Guyton's group has demonstrated
that you can denervate the carotid sinuses, you can cut the
aortic arch afferent nerves, and you don't produce very
large changes in arterial blood pressure (95). If you
average pressure over 24 hours, you get essentially normal
values of arterial pressure. In other words, you do not get
neurogenic hypertension. What you do get is a very unstable
animal and any kind of perturbation produces very wide
fluctuations in arterial pressure, but the mean over 24
hours is pretty much unchanged. What basically they say,
therefore, is that there are other factors that regulate the
arterial pressure and I think, this group would say it is
the venous return. The latter is primarily determined by
extracellular fluid volume and that is in part one of the
functions of the kidney. I think I will shield my own prejudice
and not go any further on this question.

DR. MALLIANI: This first comment I would like to make

concerns the classic scheme of the neural control of the
circulation to which you referred. In an interdisciplinary
meeting like this one, I think that it should be clearly
pointed out that such a scheme, although useful for some
operational purposes, reflects a great deal of oversimplification.
CONTRACTILE BEHAVIOR OF ARTERIES 157

This has been amply demonstrated in recent years. For

instance, in the case of the neural reflex mechanisms contributing
to the hypertensive state, the problem has been mainly
studied from the pOint of view of resetting of baroreceptive
reflexes. However, we know from very recent work of Mancia
et al., (96) that hypertensive patients do not necessarily
have a reduced efficiency of the baroreflexes with regard to
blood pressure control. Indeed, I believe that additional
spinal sympathetic reflexes, of which I shall speak may have
a relevant importance. Another point concerns in general
the variability of properties of cardiovascular receptors.

For example, the experiments of Recordati and co-workers

have recently confirmed that vagal atrial receptors of type
A and B can profoundly
Two Types of Atrial modify their firing characteristics
Receptors under difficult experimental
states (changes in atrial loads,
in inotropic states, etc.). However, it seems that atrial
receptors of type B mainly signal static and dynamic changes
in atrial wall passive tension (97) while type A atrial
receptors appear to be mainly influenced by the active
tension developed by atrial muscle during contraction (98).
Thus, in conclusion, it is likely that any given receptor as
a sensor covers a wide spectrum of possibilities for detecting
the static and the dynamic components of mechanical stimuli.

DR. COX: I would agree with most of what you said. I

am not sure that I completely agree with the separation of
atrial receptors into types for the following reasons.
These receptors are mechanoreceptors and originally investigators
tried to relate their afferent activity to atrial pressure
changes. You have to remember that there are changes in the
deformation of the atrium other than those associated with
atrial pressure changes alone. for example, with movement of
the heart and of the great veins. So one has to be careful.
There is in fact a recent paper by Armour from Loyola (99)
in which he showed that even the receptors that appear to
possess higher afferent activity in diastole were responding
to deformation. He could probe around the atrium mechanically
until he found the receptor site and show by pushing on it
that it would respond. Again, this is a complex business
and that is not exactly a physiological demonstration of
control.

DR. STONE: The topic under discussion is the dynamics

of arterial flow which includes the fluid and mechanical
properties of the flowing medium and the vessel wall.
 158 CHAPTER 3

In the approach to this problem it is important to consider the

dynamic properties of the vessels as influenced by the autonomic
nervous system. In some physiological textbooks, (100) the
pressure-volume characteristics of various vascular segments are
shown to be influenced by the presence of catecholamines. The
pressure at any volume has been found to be larger in both
arteries and veins in the presence of the catecholamines. This
would imply that the vessel has become stiffer by the action of
the agent on either the smooth muscle cells or the matrix of the
vessel. The catecholamine norepinephrine is felt to be the
primary transmitter released by most of the sympathetic nervous
system efferent terminals. Norepinephrine has also been shown
to reduce the elastic modulus of blood vessel by Dobrin (101) and
McDonald (102) and to influence the velocity of wave-propagation.
The latter factors will impact on the fluid mechanics of the
circulatory system. The effect of the autonomic nervous system
on the dynamics of arterial flow
Role of Autonomic must be considered. The cerebral
Nervous System and coronary circuits are of
particular interest in this regard
since only in recent years has information concerning the role
of the autonomic nervous system in control of these circuits
become available.

Early studies by Penfield (103) and Chorobski and Penfield

(104) demonstrated the presence of a dense nerve apparatus
surrounding the large cerebral
Cerebral Circulation vessels. These authors found
through nerve degeneration studies
that the nerves originated in the vicinity of the carotid artery
as the artery enters the cranium. At this time the authors could
not make a distinction between the nerves as to whether the
nerves were part of the sympathetic or parasympathetic nervous
system. Subsequently several authors (105,106,107) have shown
that the sympathetic nervous system innervates the major cerebral
vessels to at least the size of pial arteries (40-100 u dia.).
These nerves form a rich ground plexus in the adventitia of
the vessels and terminate in the tunica media of the vessel wall.
MONKEY In the monkey, a very consistent pattern of distribution of
the postganglionic sympathetic fibers was apparent. The basilar,
middle cerebral and anterior cerebral arteries were more densely
innervated than the posterior cerebral and vertebral arteries.
An example of the sympathetic postganglionic fiber can be seen
in Fig. 3-26. The cerebral vessel was stretched on to a glass
slide and exposed to paraformaldehyde vapor causing the
catecholamine in the nerves and vesicles to produce a fluorescence
when properly excited.
CONTRACTILE BEHAVIOR OF ARTERIES 159

Fig. 3-26: Sympathetic Post Ganglionic Fibers.

 160 CHAPTER 3

The or1g1n of many of the postganglionic sympathetic fibers

has been found to be the superior cervical ganglion (107).
Removal of the ganglion resulted in a loss of most of the
fibers on the larger cerebral vessels but in the small
vessels some fluorescence was still present. This could
indicate that there might be an intracranial source of
postganglionic sympathetic fibers. Hartman (IDS) and Itakura
(109) have shown that the small arterioles and possibly the
capillaries may receive a postganglionic sympathetic innervation
from an intracranial source. The intracranial source may be
the nucleus locus coerulius of the brain (IDS). These
findings imply that the sympathetic neural control extends
to the level of the capillary in the brain.

The parasympathetic nervous system has been shown to

innervate the arteries of the brain (110,111) also. A
representative example of the cholinergic innervation can be
seen in Fig. 3-27. The distribution of the parasympathetic
fibers is very similar to that
Source of Parasympathetic of the sympathetic fibers. The
Fibers parasympathetic nerves are found
in the adventitial layer of the
vessel wall and have been shown in cross section to enter the
tunica media. The posterior cerebral and vertebral arteries in
MONKEY the monkey were found to have a less dense innervation than
the other large vessels. Fibers that are apparently parasympathetic
have been shown to be present on the pial arteries. The origin
of the parasympathetic fibers found in associaiton with the
cerebral vessels has not been clearly identified. Suggestions
(lID) have been made that the fibers arise from the seventh
cranial nerve and enter the cranium via the internal carotid
artery. An intracranial source of these fibers cannot be
ignored.

The intimate association of the parasympathetic and

sympathetic fibers in the adventitia of the blood vessels
may imply an interaction at the postganglionic terminal. A
schematic representation of this association can be seen in
Fig. 3-2S. The transmitter released from either type of
postganglionic fiber must diffuse to the target cells, i.e.,
the vascular smooth muscle cells. Sympathetic postganglionic
fibers are felt to be vasoconstrictor while the postganglionic
parasympathetic fibers have been reported to cause vasodilation
of the vessels. Autonomic nerve fibers that transmit information
to the brain center (afferent fibers) have not been described
in relation to the cerebral vessels with the exception of
the extracranial carotid bifurcation region.
CONTRACTILE BEHAVIOR OF ARTERIES 161

Fig. 3-27: Cholinergic Arterial Innervation.

162 CHAPTER 3

Sympathetic Postganglionic Fiber

arasympathetic Postganglion ic F b
i er

8 wo"

Fig. 3-28: Schematic representation of possible sympathetic -

parasympathetic wall interaction.
CONTRACTILE BEHAVIOR OF ARTERIES 163

The real question to be asked in relation to the autonomic

innervation of the cerebral vessels is the physiological
role of the nerves in the dynamic response characteristic of
the cerebral circulation. There are several ways to demonstrate
the role of the sympathetic nervous system but the parasympathetic
system has been much more difficult to characterize.

A method to determine the effect of the sympathetic

nervous system in cerebral circulation would be to measure
the change in resistance and flow during stimulation or
removal of the superior cervical ganglion. Stimulation of
the sympathetic nervous system to the cerebral vessels can
cause an 80% reduction in cerebral flow and a tremendous
increase in resistance (112). Removal of the superior
cervical ganglion in the monkey resulted in a 25% increase
in cerebral inflow (113). From these studies it was clear
that the sympathetic nervous system when activated could
cause cerebral vasoconstriction but more important the
removal of the sympathetic innervation resulted in an increase
in cerebral flow.

The latter point would indicate a tonic sympathetic

activity to the major cerebral vessels. The finding of a
tonic sympathetic activity to the vessels would indicate a
possible role in the ability of the cerebral circulation to
regulate cerebral flow when arterial driving pressure is
changed. A comparison of the ability of the cerebral vascular
bed to maintain flow over a pressure range was made in a
group of animals with one superior cervical ganglion removed.
A schematic representation of these results can be seen in
Fig. 3-29. As arterial pressure was changed the cerebral
flow in the intact side remained constant and began to
decline as the arterial pressure approached 80 mm Hg. The
sympathectomized side began with an increased cerebral flow
and the relationship between flow and pressure was the same
as the intact side. The unique characteristic of the cerebral
vascular bed to maintain flow constant over a large pressure
range was not affected except that cerebral flow was greater
after sympathectomy. If the major site of pressure drop was
in vessels larger than the pial arteries, this would indicate
that the role of the sympathetic nervous system was predominantly
on the larger cerebral vessels. A 50% reduction in pressure
across vessels larger than 0.5 mm in diameter has been found
by in situ pressure measurements (114).
164 CHAPTER 3

c
,"e
50

E 40
Sympathectomy
~
51
u.. 30 Intact
"0
0
0
iii 20
0~
.a 10
.,f
u
0
40 80 120 160
Mean Arterial Pressure

Fig. 3-29: Adaptation of cerebral blood flow to changes in

arterial pressure.

60
- c Intact
e "e
.a,
~ E 40
u-
c ~ Sympathectomy
.,-0
"-

g'~ 20
o 0
62
m
o 10 70
30 50
Change in Arterial pC02

Fig. 3-30: Changes in cerebral blood flow related to arterial

p C02·
CONTRACTILE BEHAVIOR OF ARTERIES 165

The cerebral circulation is very sensitive to changes

in the arterial carbon dioxide pressure. Is the sympathetic
nervous system involved in
Arterial pC02 the response to carbon dioxide?
Previous studies (115,116) have
shown that when the arterial pC02 was increased above normal.
cerebral blood flow would increase and conversely if arterial
pCOZ was decreased cerebral blood flow decreased also. This
direct relationship can be seen schematically in Fig. 3-30
to increases in arterial pCOZ. In the intact animal, the
slope of the relationship between the increases in cerebral
blood flow and the increase in arterial pCOZ was approximately
1 (117). Bilateral removal of the superior cervical gang1ions
resulting in near total sympathectomy reduced the sensitivity
of the cerebral vessels to changes in arterial pC02. Since
there was a tonic sympathetic vasoconstrictor tone found in
the cerebral vasculature. the change in pCOZ through either
a direct effect or by an increase in hydrogen ion concentration
inhibited the release of sympathetic transmitter resulting
in dilation. Removal of the sympathetic nervous system to
the cerebral vessels reduced the effect of pCOZ. Recent
evidence would indicate that the change in hydrogen ion
concentration may be the mechanism of cerebral vasodilation
instead of a direct action of pCOZ on vascular smooth muscle
(118). The role of the sympathetic nervous system in the
dynamics of cerebral blood flow now seems very clear. This
vascular bed certainly responds to the local needs of the
tissue through local control, but the total inflow appears
to be regulated by the autonomic nervous system. The mechanism
outlined above would furnish the cerebral vascular bed with
a central neural control and a fine local control mechanism.

The dynamic characteristics of the sympathetic innervation

to the cerebral vessels was demonstrated in another study
(119), by stimulation of the
Differential Effect of cerebellum of the brain.
Sympathetic Nervous Stimulation of the cerebellum
System on Circulation result in a tremendous increase
in mean arterial pressure in
the monkey. The increase in mean arterial pressure was found MONKEY
to be a result of a vasoconstriction of the splanchnic
circulation indicating an increase in sympathetic nervous
system activation. Cerebral blood flow was measured in
intact monkeys and in monkeys with the sympathetic innervation
of the cerebral vessels removed.
166 CHAPTER 3

In the intact animals cerebral blood flow increased with

stimulation of the cerebellum and cerebral resistance decreased.
Removal of the sympathetic nerves to the cerebral vessels
reduced the flow response and the decrease in cerebral
vascular resistance. This study demonstrated a differential
effect of the sympathetic nervous system on the circulation.
A vasoconstriction in the peripheral vessels and a vasodilation
of the cerebral vessels was found. It was clear from this
data that the vasodilation found in the cerebral vascular
bed may also have been in part due to an increase in the
activity of cholinergic vasodilator nerves.

In summary. the dynamic characteristics of the cerebral

vascular bed must encompass the autonomic nervous system and
its role in the minute to minute adjustment of flow to the
tissue. The sympathetic nervous system has clearly been
shown to influence cerebral flow and the presence of the
parasympathetic nervous system has been demonstrated while
its role at present in regulation is not known.

Coronary Circulation

In the coronary circulation, the vessels have been

found to be innervated by efferent autonomic fibers as well
as afferent autonomic fibers particularly sympathetic efferent
fibers. The presence of both an afferent and efferent
innervation suggests very strongly the possibility of local
reflexes and feedback control that may influence flow in
large vessels as well as the distribution of flow across the
myocardial wall.

Sympathetic afferent fibers ar1s1ng from the coronary

arteries or near coronary arteries have been demonstrated
(120). These fibers course
Activation of Afferent with the sympathetic nerves to
Receptors the heart only the afferent
fibers do not synapse in the
paravertebral ganglion but have their cell bodies in the dorsal
root ganglion of the spinal cord. The afferent receptors
may be mechanoreceptors and activated by a decrease in
pressure in the vessel or the suggestion has been made that
some may be chemoreceptors and activated by a reduction in
oxygen supply to the myocardium. The exact nature of the
stimulus that activates these receptors is not known.
Efferent nerves from both the sympathetic and parasympathetic
nervous system have been identified in the adventitia of the
coronary vessels (121) as described for the cerebral vessels.
CONTRACTILE BEHAVIOR OF ARTERIES 167

The efferent sympathetic nerves were found to be rather

dense in the adventitia of all major coronary vessels.
Removal of the stellate ganglion caused a loss of the sympathetic
nervous system innervation of the vessels. A gradient in
the degree of cholinergic innervation was found. The left
anterior descending coronary artery was more densely innervated
than either the left circumflex artery or the right coronary
artery. The cholinergic innervation was found to course at
two distinct levels in the vessel wall. One group of fibers
could be removed by stripping the adventitia of the vessel
while a second group adhered to the surface of the tunica
media. Section of the vagus nerve failed to eliminate the
cholinergic innervation as expected. The fibers running
along the coronary vessels must be postganglionic parasympathetic
fibers arising from the intrinsic cardiac ganglia.

The role of the autonomic innervation of the coronary

vessels has been much debated over the last few years.
Until recently the major
Debate over Role of determinant of coronary flow
Autonomic Innervation had been felt to be local factors
of Coronary Vessels liberated when oxygen need was
greater than supply (122). The
rich autonomic innervation and the possibility of reflexes were
not counted as being important in the coronary vascular bed.
Stimulation of the cardiac sympathetic nerves resulted in a
vasoconstriction (123) while activation of the parasympathetic
nerves to the heart caused a vasodilation (124). If the
local control of coronary flow were the dominant factor,
changes in coronary blood flow would not be expected to
occur until abnormal levels of oxygen supply had been achieved.
The question of sensitivity of the central system must be
answered. In studies in conscious dogs, the inspired oxygen DOG
was lowered from a normal 20% to 5%. Fig. 3-31 shows the
transient response of arterial oxygen saturation, coronary
flow, and the maximum rate of rise of left ventricular
pressure to the exposure to 5% inspired oxygen. It should
be noticed that coronary flow has increased before there was
a significant decrease in arterial oxygen saturation. This
would clearly indicate that some neurogenic influence had
caused the coronary dilation and increase in coronary flow.
Blockage of beta-adrenergic receptors did not change the
transient response to hypoxia nor did drugs that blocked the
effect of adenosine. Thus neurogenic regulation of coronary
flow may be as important as local factors in the response of
coronary flow to decreases in oxygen supply.
168 CHAPTER 3

100

ART~'AL
OKY ~
SIn'URAT~

(°41

SY"I8O\. ·s·
70

60
70

CORONARY LEfT
BLOOD FLc:NI VENTRlCUlM
fP,*
((rn/lee I 50
(~/IIKI
SYI48a.. " r ·

20

?4 il 40
' lfII! ISfC )

Fig. 3-31: Changes in left ventricular pressure and coronary blood

flow during Hypoxia.

..... .....
..........2
.... ....
Mean .... ......
Coronary
Flow
0-

Fig. 3-32: Typical coronary reactive hyperemia response. The

portion labeled A is the flow deficit a nd the portion lab e led B
is the payback following release of occlusion. See text for
explanation of Figure.
CONTRACTILE BEHAVIOR OF ARTERIES 169

Another approach to defining a possible role of the

autonmic nervous system in the control of coronary flow is
to challenge the system by very brief (10 sec.) occlusion of
a coronary artery. Following the 10 second occlusion of a
coronary artery, the flow rapidly increases above the control
level and returns to control level over some time period.
Fig. 3-32 Curve 1 shows a typical normal response of coronary
flow to a brief occlusion. A comparison of the area B to
area A provides an estimation of the overpayment of coronary
flow to the occlusion. This response is termed reactive
hyperemia and has been described previously (125,126). The
normal reactive hyperemia was 476% in a group of dogs (127).
In this same group of animals, the left stellate ganglion
was removed and the experiment repeated. Curve 2 of Fig. 3-
32 shows representative results of this experiment. The
reactive hyperemia was increased to 622%. Beta-adrenergic
blockage (Curve 3) reduced the hyperemic response to 390%.
Alpha-adrenergic blockage mimicked the response of left
stellate ganglionectomy. From these studies, we concluded
that there was a tonic Alpha-adrenergic vasoconstrictive
tone on the coronary vessels that did not permit a maximum
vasodilation with brief (10 sec.) occlusions of the coronary
artery. Beta-adrenergic blocade resulted in an opposite
effect on the hyperemic response and may have been caused by
a direct depression of myocardial metabolism by this blocking
agent.

The distribution of coronary flow across the myocardial

wall may also be influenced by the sympathetic nervous
system. Various drugs have
Dynamics of Coronary been shown to effect the
Circulation must Include distribution of myocardial
A Neurogenic Component flow (128). Removal of the
left stellate ganglion was
found to increase endocardial blood flow when all other
conditions were the same as compared to the control conditions.
These studies suggest that the control of coronary blood
flow and distribution is influenced by the autonomic nervous
system.

In the coronary vascular bed, the larger arteries

course across the surface of the heart and give off branches
which dive into the myocardial wall. The intramural vessels
are subjected to the tension generated by the contracting
myocardium and thus some degree of neural control of these
vessels would be important.
170 CHAPTER 3

It is entirely possible that the various vascular segments

in the myocardium have different adrenergic receptors on the
smooth muscle cells thus amplifying the control of flow. It
seems clear that sympathetic activation of the myocardial
cells and coronary circulation may be entirely separate thus
a high degree of coronary control can be achieved. As in
the case of the cerebral circulation, the dynamics of the
coronary circulation must include a neurogenic component.

DR. WOLF: Dr. Stone's demonstration of increased

cerebral blood flow in association with increased blood
pressure raises an old story that was started by Cohnheim
and picked up by Heinbecker several years ago. They held
that the biological significance of essential hypertension
was to increase cerebral blood flow.

Aorta and Sympathetic Reflexes

DR. MALLIANI: I shall try to analyze some of the

problems that may originate from the fact that the vessels
are richly innervated structures and, in particular, I shall
examine some new experimental data concerning the aorta.

During recent years it has been amply demonstrated that

afferent sympathetic fibers with cardiovascular sensory
endings can mediate cardiovascular
Reflexes in Neural Control reflexes. Moreover, as many of
of Cardiac Functions these afferents have a spontaneous
impulse activity and, in addition,
are extremely sensitive to hemodynamic stimulii, we advanced
the hypothesis that they may participate in the tonic regulation
of the circulation (129). Thus, in the case of the neural
control of cardiac functions, as far as reflexes arising
from the heart are concerned, one should distinguish vago-
vagal, vago-sympathetic, sympatho-sympathetic and sympatho-
vagal reflexes (129).

The same complexity holds true for reflexes arising

from the aorta. Afferent vagal fibers innervating the aorta
constitute one of the classic afferent pathways subs erving
cardiovascular neural control. However, the aorta is also
possessed of a very abundant sympathetic sensory innervation.
Fig. 3-33 shows the activity of an afferent sympathetic
nerve fiber with its receptive field located in the proximal
third of the descending thoracic aorta. The calculated
conduction velocity was 5 m/sec (hence the fiber belonged to
the group A S (130). It is clear that the fiber was excited
by any increase in aortic pressure, however obtained.
CONTRACTILE BEHAVIOR OF ARTERIES 171

m':~:] ........... ","" .................................. --............r!:'HO

o , ! ,, ! ! I I I r 5 ! I ! I I I I I ! , ! I ' ! ! r 5 l I 1 I L j I ! , , ! I ! 1 ! , , ! 1 ' I ! [0
I ' , I
a j ! ! :
!
:

]. .
. . . . . , .....
...............................................................................................
t ! , , l ' " l 11 I I , I ' ,
...;''-~~~~::;;;::[
I I I ' , I I c", I I I I I I 1 '

b I ll lllll lil jillII Pi iif. i i! ! p!i !

]:::::::::::::;:~~~~~~~[
r
I I 'III I ! 1 "
11 I 1 I 1 I I ' I I [ I I ( I ' I 'I I f ', I I I 1 I I I,' ,' 't' ! ' , 1

" 11 I ':
! , ! ! , , " " ,

I ", i .;
c ,1 1I 1,,,1,, I ;
-:'I i '"
' ,' I "
' I '
. ;I : I: i '
I " : t

]~=:: :::::::::::::::::::::::;::::z: ~:;.=::.=[

I I I I t , It,
l ' I It! t ' , ' ' , ' , 1,' , ',. ' , 1. 1., ' I ' , I I I i I I I I • t II I I 1I I I I
d. \ :' ' ,.
t ,

] ::==========::::~~:::~:::~:::.:::~::::::::~:::~:::~:::~:::~:::~:::~:::~:::~::::::~:::~:::~:::~:::~~~~~~~~~~~~~~;~::~~ [

i :i:;
I ~ , I I I I I I II t It 1 I I I I I I I I I 1 I I I , 1 , I I I I 1 I I II I I , l l J, :t,
iii ii i
!

e ! i i
t
Fig. 3-33: Activity of an afferent sympathetic nerve fiber
Group A ) with receptive field located in the proximal third of
the descending thoracic aorta. Calculated conduction velocity
5 m/sec. Tracings in a,b,c,d and e represent, from top to bottom,
the endotracheal pressure (inflations upwards) the aortic and
femoral arterial pressures, the e.c.g. and the nervous recording.
a, control. b, occlusion of the descending throacic aorta
(indicated by the diverging blood blood pressure traces). c,
effects of an I.V. injection of 2 ug angiotensin, performed just
before the beginning of the record. d, effects of a reflex blood
pressure rise produced by occluding the right carotid artery
(occlusion indicated by the bar). e, effect of an abrupt increase
in venous return produced by releasing, at the arrow, an occlusion
of the inferior vena cava. (from ref. 131, by courtesy of
J. Physiol.).
 172 CHAPTER 3

In Fig. 3-34 the impulse activity is displayed of an

afferent sympathetic fiber belonging to the group C (130)
(conduction velocity of 1 m/sec). That fiber was also
studied post mortem by stretching the aortic wall with a
balloon after the animal had been killed by bleeding. When
the balloon was inflated (Fig. 3-34c and following parts of
the figure) the highest impulse frequencies were attained
during the rising phase of the pressure stimulus, followed
thereafter by an adapted discharge (Fig. 3-34 e and h). A
dynamic component of the stimulus was also proven by reaching
the same level of absolute pressure at different speeds
(Fig. 3-34 c and d) or from different initial pressures
(Fig. 3-34 e and f). Finally a response which might be
attributed to receptor fatigue and/or mechanical alterations
of the vascular wall was observed: in Fig. 3-34 h, a pressure
rise produced a much higher activation of the impulse activity
than that that illustrated in Fig. 3-34 where the same step
of pressure was applied after the stimulus had been sustained
for 100 sec. and just released for a few seconds. These
data were published recently and more details can be found
in the original paper (131).

In order to investigate the potentiality of reflexes

arising from sympathetic aortic receptors we stimulated them
by means of a mechanical stretch (132). We used a special
aortic cannula which made it possible to gradually stre~ch
the aortic walls, without interfering with aortic blood
flow. In Fig. 3-3Sa, the effects can be seen of such a
stretch of the thoracic aorta, performed in a vagotomized
CAT cat, with an intact central nervous system and with both
common carotid arteries occluded. The reflex effects,
mediated through an excitation of the sympathetic outflow,
consist of an increase in arterial blood pressure, myocardial
contractility and heart rate. The effects of propranolol
administration are illustrated by Fig. 3-3Sb: aortic
stretch, in these conditions, induced a smaller but clear
increase in arterial and left ventricular dP/dtmax increased
only very slightly and slowly during the period of stimulation.

On the basis of our understanding of blood pressure

regulation (Fig. 3-36, upper diagram), the stretch of a
reflexogenic area (R), simulating
Feedback Mechanisms a rise in arterial blood pressure,
produces a reflex decrease in
systemic blood pressure through an inhibition of the sympathetic
(8) discharge affecting vascular (V) resistances and thus
blood pressure (BP).
 (')
a b o
Z
200 ] ] [200 --I
mmHgJ [ mmHg ::0
»
(')
o ] I" I ] ,I ' I ] I 0 --I
• , • t I I '! :,1 ; I ; I: "I '~ r .
I', !; ; r
m
OJ
c d m
I
[200
»
<
:mHg (5
, [ ::0
~ i' : ,i :I o
e
- f g
"
»
::0
--I
m
::0
m
, ],[ (Jl
~. - - I'I"'!,.
:! I " ', : !l:; I! ;1
l : I ."

h 0.2

'~~:;~irp:r;;r;~lj (i" ~!: "I' !]

:i Il'il ' 1"1
:;Ii.! ! r Iii;[
L ~ j 1tIL ~ ~ :i: I f _j i .j !

~
Fig. 3-34: Activity of an afferent sympathetic nerve fiber (Group C) with receptive field
located in the distal third of the aortic arch. a, control. b, occlusion of the descending
aorta. c,d,e,f,h and i, effects of stretching aortic wall by distending a latex balloon
located in the distal part of the aortic arch (see text). g, electricl stimulation of the
left inferior cardiac nerve activating the fiber. Approximate length of the fiber, 5 cm.
Calculated conduction velocity 1 m/sec. Tracings in a and b as in Fig. 3-33; c,d,e,f,h, and
i, top tracing; pressure applied to the distending balloon; bottom tracing: nervous
activity (from ref. 131, by courtesy of J. Physiol.). 'rl
W
174 CHAPTER 3

CAP..f'-f' ..... .............

~
A
~~

-J'.J' ~
B
_ _ [200
50
_ _ [200

50
LvpM
LV LL
,,%,',1
________ [225
HR -----[ 115

Fig. 3-35: Effects of stretching the thoracic aorta in a vago-

tomized cat with an intact central nervous system and both common
carotid arteries occluded. CAP = carotid artery pressure (recorded
proximal to the point of occlusion), FAP = femoral artery pressure,
LVP = left ventricular pressure (all in mmHg) , LV dP/dt = rate of
change of left ventricular pressure (mmHg/sec), and HR = heart
rate (beats/min.). On the left of each section are fast-speed
records of each variable. The aortic stretch is indicated by a
bar. A: Control; B: After administration of propranolol (1 mg/kg,
iv). (from ref. 132 by courtesy of Circulation Res.).

R +

C~BP
+ ,-
I
,
I
I
1I _____ -
R ._-----'
+

Fig. 3-36: Schema of suggested mechanisms underlying nervous

control of blood pressure regulation. For details see text.
(from ref. 133, by courtesy of Clin. Sci.).
CONTRACTILE BEHAVIOR OF ARTERIES 175

AP
mmHIL A B
178 110

..
t' 100 .:
~
\

-, '.
., ....
180
.g
eo
..
,
'"
.;. ....
;0 ."
, ,-
1U
.. so
. 0:•.
,4'

, .1 -
0'

' &"

10 N, . 70 ::: ..
,,
.
00

, .... , so
~.,.
'
cI' ~ . '"

.. ;... ,"
75 " AD
4.S 5.0 7.0 7.5 8 .0 mm

Fig. 3-37: Effect of stimulation of cut central end of leN (A)

and decentralized left thoracic sympathetic chain (B) on diastolic
pressure (AP) and diameter (AD) relationship in 2 animals with
both carotid arteries occluded. 0 Before stimulation; • stimu-
lation; and 0 after stimulation. (from ref. 135 by courtesy of Am.
J. Physiol.).
176 CHAPTER 3

In this closed loop a negative feedback assures a tonic

control of the system. Vice versa, our experiments, in
which supraspinal inhibitory mechanisms were not operative,
have revealed sympatho-sympathetic reflexes which seem to
exhibit positive feedback characteristics (Fig. 3-36, lower
diagram), as the stretch of aortic walls produced an excitation
of the sympathetic outflow. We have advanced the hypothesis
that these sympathetic reflexes may contribute to the tonic
maintenance of the effects of an increased central command
(c, in both diagrams in Fig. 3-36 and thus to the pathogenesis
of arterial hypertension (133).

However, it should be pointed out that electrophysiological

techniques have detected in the sympathetic efferent discharge
both excitations and inhibitions induced reflexly by a
similar aortic stretch (134): the coexistence of these
inhibitory components in spinal sympathetic reflexes is
represented by the broken line in the lower diagram in Fig. 3-36.

Spinal sympathetic reflexes can also modify the mechanical

properties of the thoracic aorta (135). In Fig. 3-37a the
effect is shown of the electrical stimulation of the cut
central end of the inferior cardiac nerve (eN) on the aortic
diastolic pressure (AP) and diameter (AD) relationship. It
can be seen that during a sympathetic reflex the diastolic
aortic diameter is reduced for any given pressure. Such a
reflex response is similar to that obtained with direct
electrical stimulation of sympathetic efferents (Fig. 3-37b).

It is important to realize that sympathetic reflexes

can therefore arise from the aorta and can be distributed
back to the aorta. Moreover, reflexes that modify the
aortic pressure-diameter relationship are also likely to
modulate the sensitivity of aortic mechano-receptors and the
characteristics of the reflexes initiated by them.

In the hypothesis that these concepts may provide some

understanding for the interpretation of some pathophysiological
sites, one should consider, however, that pathophysiological
alterations often represent the result of mechanisms progressing
for years and that it is impossible to ask an acute experiment
for more than a trace in the world of possibilities. On the
other hand, a phenomenon which may appear of a non-relevant
magnitude during the short span of an experiment, may indeed
constitute the core of a fundamental dynamic process.
CONTRACTILE BEHAVIOR OF ARTERIES 177

To end UP. I would like to say that most of this complexity.

in my opinion. will never be solved unless multidisciplinary
approaches are devised. In other words, we should be ready
not only for multidisciplinary meetings, but for a real
multidisciplinary strategy of research.

DR. COX: Where did you measure carotid blood flow?

DR. STONE: Internal carotid.

DR. COX: Do you feel all those changes represent

intracranial blood flow?

DR. STONE: Yes. Many years ago we compared three

things. Xenon 133 washout. internal or middle cerebral flow
directly with internal carotid. We found that directionally
all of these things changed in the same direction with
interventions. The magnitude of changes ••• was certainly
different in the three places but directionally they were
always constant.

DR. COX: In your cerebellar stimulation, you had a

decreased carotid resistance that is internal carotid again?

DR. STONE: Yes.

DR. COX: What about peripheral resistance?

DR. STONE: Peripheral resistance has gone up.

DR. COX: Did you measure that?

DR. STONE: We have not measured it. but other people

have.

DR. COX: This is not a cardiac output?

DR. STONE: No, cardiac output goes up a little bit.

but. we have not measured that either. The reports in the
literature say that cardiac output goes up slightly. Although
there is a tremendous vasoconstriction in the limbs, the
increased arterial pressure seems to be from an increase in
splanchnic resistance. One of the questions that obviously
comes out of that is what is the carotid sinus doing?

DR. COX: Was it not a chronic output response primarily?

DR. STONE: Not entirely.

 178 CHAPTER 3

DR. COX: If it were entirely a cardiac output response

it might explain why you were getting more blood flow.

DR. STONE: I don't think so.

DR. CAREW: I want to address my question to Dr. Stone.

I thought the information about the stellate ganglion was a
fascinating observation and I wondered if you would expand
on that a little bit, by telling us to what extent you think
the regional distribution of flow in the heart is determined
by neurogenic factors. Could you also give us more details
as to how this was measured, I presume by microspheres, and
whether that increase in endocardial flow is a relative
increase and epicardial flow an absolute increase and whether,
for example, heart rate and blood pressure were controlled
during the experiments.

DR. STONE: Yes, we did measure the distribution of

flow with microspheres, they were 15 microns in diameter and
measure distribution fairly well. We actually directly
measured a circumflex flow in these studies. These were
DOG anesthetized dog studies where we did the distribution. We
held heart rate constant and measured MV02 across the heart
during the experiment both control and after removal of the
left stellate ganglion. We also measured contractility. So
all these parameters were controlled or were constant and
what we had in those conditions was that a greater increase
in the endo-epi ratio with a constant inflow, and no change
in flow through the circumflex which would indicate a reduction
in epicardial flow and an increase in endocardial flow.

DR. BJORKERUD: I am not too familiar with mechanics

and this may be a very silly question. If you increase the
pressure in the aorta, could there be two alternative effects
on the coronary circulation --either a constriction or a
dilatation with either an increased flow or a decreased
flow? Is that right? -- Of course what I am getting at is
the possible balance between the cholinergic and the sympathetic
effects.

DR. STONE: On the coronary vessels?

DR. BJORKERUD: Yes, and if there is, that will then

explain why trained people behave differently from untrained
people, in a number of cardiovascular aspects, for instance,
the increase of heart rate in response to physical exercise.

DR. STONE: As far as the interaction on the coronary

vessels, I cannot answer that because no one really looked
at a situation where you get large changes in flow and what
CONTRACTILE BEHAVIOR OF ARTERIES 179

that vasculature is going to do in the absence of cholinergic

fibers. So, unfortunately, I cannot answer that. My feeling
is that it may be a balance between the two but that is just
a feeling.

DR. BJORKERUD: Because there is a significantly lower

mortality in patients with heart infarction treated with B-
receptor blocking agents in comparison to untreated patients.
That is thought to be due to reduced incidence of arrhythmias
but it could, of course, be due to other factors which are
more related to what you have been talking about.

DR. STONE: Well, yes and also the reduced overall

energy requirements in the myocardium. We have shown that
many arrhythmias are directly coupled to the sympathetic
nervous system and part of it is probably due to some of the
electrical properties as well as possible flow properties.
You can't separate the two.

DR. CHIEN: If I understand you correctly, the sympathetic

innervation to the brain can cause both vasoconstriction and
vasodilatation because if we look at the control data comparing
the denervated and innervated sides, the sympathetic nerves
seem to be causing vasoconstriction; but when you look at
the P02 experiments, or the cerebellar stimulation experiments,
the innervation seems to be causing vasodilation. Do you
think there are two separate receptors such as alpha and
beta receptors that are mediating these different changes in
the cerebral circulation?

DR. STONE: Owman and Edwidsson (136) have shown very

well that there are both receptors present in the vasculature
of the brain. We are seeing a tonic constriction in the
brain that if removed, there occurs a dilation on that side.
Certainly there is also a cholinergic innervation, as the
Russians have shown. It may well produce vasodilation as
well, as Penfield showed back in the 30's. So how that
enters I don't know but with P02 there is also a direct
effect on the vasculature. I don't know what the relationship
at this point between the sympathetic and cholinergic innervation
would happen to be but, yes, I would make the assumption
that a withdrawal of a tonic activity would cause vasodilation
or a decrease in resistance.

DR. NEREM: Were you saying that the primary innervation

of the coronary system is the extramural vessels and thus
that primary neurogenic control is in these extramural
vessels?
180 CHAPTER 3

DR. STONE: No. We have looked at vessels as small as

we could visibly dissect them and found that the distribution
of the innervation decreases with vessel size. So that I
will not say where the major site of resistance is in the
coronaries. The brain studies have shown that the major
site of pressure drop is in vessels larger than .5 millimeter
which excludes the pial circuit. So here we are talking
about relatively large small arteries, let us say. In the
coronaries the picture is not quite clear.

BIBLIOGRAPHY

1. Gow, B.S.: The influence of vascular smooth muscle on

the viscoelastic properties of blood vessels.
Cardiovascular Fluid Dynamics. D.H. Berge (Ed.)
Vol. II. London and New York Academic Press,
pp. 65-110, 1972.

2. Furchgott, R.F.: The pharmacology of vascular smooth

muscle. Pharmacol. Rev. 7:183-265, 1955.

3. Wissler, R.W.: The arterial medial cell, smooth muscle or

multifunctional mesenchyme? J. Atheroscler. Res.
8 :201-213, 1968.

4. Aars, H.: Effects of altered smooth muscle tone on

aortic diameter and aortic baroreceptor activity in
anesthetized rabbits. Circ. Res. 28:254-262, 1971.

5. Cox, R.H.: Determinants of systemic hydraulic power in

unanesthetized dogs. Am. J. Physiol. 226:579-587,
1974.

6. Pagani, M., Schwartz, P.J., Bishop, V.S. and Malliani, A.:

Reflex sympathetic changes in aortic diastolic
pressure-diameter relationship. Am. J. Physiol.
229:286-290, 1975.

7. Roach, M.R. and Burton, A.C.: The reason for the shape
of the distensibility curves of arteries. Can. J.
Biochem. Physiol. 35:681-690, 1957.

8. Wolinsky, H. and Glagov, S.: Structural basis for the

static mechanical properties of the aortic media.
Circ. Res. 14:400-413, 1964.

9. Apter, J.T.: Correlation of viscoelastic properties with

microscopic structure of large arteries. Circ. Res.
21:901-918, 1967.
CONTRACTILE BEHAVIOR OF ARTERIES 181

10. Fischer, G.M. and Llaurado, J.G.: Collagen and elastin

content in canine arteries selected from functionally
different vascular beds. Circ. Res. 19:394-399, 1966.

11. Harkness, M.L.R., Harkness, R.D. and McDonald, D.A.: The

collagen and elastin content of the arterial wall in
the dog. Proc. Roy. Soc. London, Series B. 146:
541-551, 1957.

12. Benninghoff, A.: Blutgefasse und Herz. In: Handbuch der

mikroskopischen Anatome. Berlin, Springer-Verlag.
Vol. VI., pp. 1-225, 1930.

13. Cox, R.H.: Anisotropic properties of the canine carotid

artery in vitro. J. Biomechanics 8:293-300, 1975.

14. Bergel, D.H.: The dynamic elastic properties of the

arterial wall. J. Physiol. 156:458-469, 1961.

15. Siegman, M.J., Butler, T.M., Mooers, S.U. and Davies, R.E.:
Crossbridge attachment, resistance to stretch, and
viscoelasticity in resting mammalian smooth muscle.
Science 191:383-385, 1976.

16. Dobrin, P.B. and Rovick, A.A.: Influence of vascular

smooth muscle on contractile mechanics and elasticity
of arteries. Am. J. Physiol. 217:1644-1651, 1969.

17. Gore, R.W.: Wall stress: a determinant of regional

differences in response of frog microvessels to
norepinephrine. Am. J. Physiol. 222:82-91, 1972.

18. Speden, R.N.: Muscle load and constriction of the rabbit

ear artery. Am. J. Physiol. 248:531-533

19. Dobrin, P.B.: Isometric and isobaric contraction of

carotid arterial smooth muscle. Am. J. Physiol.
214:561-565, 1968.

20. Cox, R.H.: Mechanics of canine iliac artery smooth

muscle in vitro. Am. J. Physiol. 230:462-470, 1976.

21. Herlihy, J.T. and Murphy, R.A.: Length-tension relation-

ship of smooth muscle of the hog carotid artery.
Circ. Res. 33:275-283, 1973.
182 CHAPTER 3

22. Furness, J.B. and Marshall, J.M.: Correlation of the

directly observed responses of mesenteric vessels of
the rat to nerve stimulation and noradrenaline with
the distribtuion of adrenergic nerves. J. Physiol.
239:75-88, 1974.

23. Davis, D.L. and Dow, P.: Intraluminal pressures and

rate and magnitude of arterial constrictor responses.
Am. J. Physiol. 227:1149-1157, 1974.

24. Davis, D.L. and Baker, C.H.: Arterial segment constriction

under constant-pressure and constant in-flow perfusion.
Am. J. Physiol. 227:1149-1157, 1974.

25. Abboud, F.M.: Control of the various components of the

peripheral vasculature. Fed. Proc. 31:1126-1239, 1972.

26. Gow, B.S.: Viscoelastic properties of conduit arteries.

Circ. Res. 26 and 27, Suppl. 11:11-113-11-122, 1970.

27. Cox, R.H.: Pressure dependence of the mechanical properties

of arteries in vivo. Am. J. Physiol. 229:1371-1375,
1975.

28. Shepard, J.T.: Intrathoracic baroreflexes. Mayo Clin.

Proc. 48:426-437, 1973.

29. Koushanpour, E. and Kelso, D.M.: Partition of the carotid

sinus baroreceptor response in dogs between the
mechanical properties of the wall and the receptor
elements. Circ. Res. 31:831-845, 1972.

30. Pelletier, C.L., Clement, D.L. and Shepard, J.T.: Comparison

of afferent activity of canine aortic and sinus nerves.
Circ. Res. 31:557-568, 1972.

31. Koushanpour, E. and Kelso, D.M.: Partition of the carotid

sinus baroreceptor response in dogs between the
mechanical properties of the wall and the receptor
elements. Circ. Res. 31:831-845, 1972.

32. Koushanpour, E.: Quantitative analysis of whole nerve

action potentials recorded from the carotid sinus
baroreceptors. J. Electrophy. Tech. 3:39-45, 1975.

33. Angell-James, J.E. and Daly, M. de B.: Comparison of the

reflex vasomotor responses to separate and combined
stimulation of the carotid sinus and aortic arch
baroreceptors by pulsatile and non-pulsatile pressures
in the dog. J. Physiol. 209:257-293, 1970.
CONTRACTILE BEHAVIOR OF ARTERIES 183

34. Ninomiya, I. and Irisawa, H.: Aortic nervous activities in

response to pulsatile and nonpulsatile pressure.
Am. J. Physiol. 213:1504-1511, 1967.

35. Bagshaw, R.J.: Pressure dependence of the carotid sinus

elastic modulus in the dog. M.LT. J. Life Sci.
5:43-48, 1975.
36. Bagshaw, R.J. and Peterson, L.H.: Sympathetic control of
the mechanical properties of the canine carotid sinus.
Am. J. Physiol. 222:1462-1468, 1972.

37. Alarcon, J.E., Campbell, K.B. and Peterson, L.H.: Effect

of norepinephrine on the carotid sinus. The Physiologist
17: 171, 1974.

38. Wurster, R.D. and Trobiani, S.: Effects of cervical

sympathetic stimulation on carotid occlusion reflexes
in cats. Am. J. Physiol. 225:978-981, 1973.

39. Cox, R.H., Bagshaw, R.J., Detweiler, D.K and Peterson, L.H.:
Effects of aging on the carotid sinus control of
canine arterial hemodynamcis. IRCS Hed. Sci.
3:293, 1975.

40. Vatner, S.F., Franklin D., Van Citters, R.L. and

Braunwald, E.: Effects of carotid sinus nerve
stimulation on blood-flow distribution in conscious
dogs at rest and during exercise. Circ. Res.
27:495-503, 1970.

41. Bjurstedt, H., Rosenhamer, G. and Tyden, G.: Cardiovascular

responses to changes in carotid sinus transmural
pressure in man. Acta Physiol. Scand. 9:497-505, 1975.

42. Moissejeff, E.: Zur Kenntnis des Carotissinus-reflexus.

Z. Ges. Exptl. Med. 53:696-704, 1926.

43. Angell-James, J.E. and Daly, M. de B.: Effects of graded

pulsatile pressure on the reflex vasomotor responses
elicited by changes of mean pressure in the perfused
carotid sinus-aortic arch regions of the dog.
J. Physiol. 214:51-64, 1971.

44. Schmidt, R.M., Kumada, M. and Sagawa, K.: Cardiac output

and total peripheral resistance in carotid sinus
reflex. Am. J. Physiol. 221:480-487, 1971.
184 CHAPTER 3

45. Cox, R.H. and Bagshaw, R.J.: Baroreceptor reflex control

of arterial hemodynamcis in the dog. Circ. Res.
37:772-786, 1975.

46. Bond, R.F. and Green, H.D.: Cardiac output redistribution

during bilateral common carotid occlusion. Am. J.
Physio1. 216:393-403

47. O'Rourke, M.F. and Taylor, M.G.: Vascular impedance of

the femoral bed. Circ. Res. 18:126-139, 1966.

48. Taylor, M.G.: The input impedance of an assembly of

randomly branching elastic tubes. Biophysical J.
6:29-51, 1966.

49. Levy, M.N., NG, M.L. and Zieske, H.: Cardiac and
respiratory effects of aortic arch baroreceptor
stimulation. Circ. Res. 19:930-939, 1966.

50. Hainsworth, R., Ledsome, J.R. and Carswell, F.: Reflex

responses from aortic baroreceptors. Am. J. Physio1.
218:423-429, 1970.

51. Pelletier, C.L., Edis, A.J. and Shepard, J.T.: Circulatory

reflex from vagal afferents in response to hemorrhage
in the dog. Circ. Res. 29:626-634, 1971.

52. Ott, N.T. and Shepherd, J.T.: Modifications of the aortic

and vagal depressor reflexes by hypercapnia in the
rabbit. Circ. Res. 33:160-165, 1973.

53. Pelletier, C.L. and Shepherd, J.T.: Effect of hypoxia

on vascular responses to the carotid baroref1ex.
Am. J. Physio1. 228:331-336, 1975.

54. Mancia, G.: Influence of carotid baroreceptors on vascular

responses to carotid chemoreceptor stimulation in the
dog. Circ. Res. 36:270-276, 1975.

55. Mancia, G., Shepherd, J.T. and Donald, D.E.: Role of

cardiac, pulmonary, and carotid mechanoreceptors in
the control of hind-limb and renal circulation in
dogs. Circ. Res. 37:200-208, 1975.

56. Bagshaw, R.J., Lizuka, M. and Peterson, L.H.: Effect of

interaction of the hypothalamus and the carotid sinus
mechanoreceptor system on renal hemodynamics in the
anesthetized dog. Circ. Res. 25:569-585, 1972.
CONTRACTILE BEHAVIOR OF ARTERIES 185

57. Kumada, M., Schramm, L.P., Altmansberger, R.A. and

Sagawa, K.: Modulation of carotid sinus baroreceptor
reflex by hypothalamic defense response. Am. J. Physiol.
228:34-45, 1975.
58. Symth, H.S., Sleight, P. and Pickering, G.W.: Reflex
regulation of arterial pressure during sleep in man:
A quantitative method of assessing baroreflex
sensitivity. Circ. Res. 24:109-121, 1969.

59. Bristow, J.D., Brown, E.B., Cunningham, D.J.C.,

Howson, M.G., Peterson, E.S., Pickering, T.G. and
Sleight, P.: The effect of bicycling on the
baroreflex regulation of pulse interval. Circ. Res.
28:582-592, 1971.

60. Higgins, C.B., Vatner, S.F., Eckberg, D.L. and Braunwald, E.:
Alterations in the baroreceptor reflex in conscious
dogs with heart failure. J. Clin. Invest. 51:715-724,
1972.

61. Epstein, S.E., Beiser, G.D., Goldstein, R.E., Stampfer, M.,

Wechsler, A.S., Glick, G. and Braunwald, E.:
Circulatory effects of electrical stimulation of the
carotid sinus nerves in man. Circ. 40:269-276, 1969.

62. McCubbin, J.W., Green, J.H. and Page, I.H.: Baroreceptor

function in chronic renal hypertension. Circ. Res.
4:205-210, 1956.

63. Angell-James, J.E.: Characteristics of single aortic and

right subclavian baroreceptor fiber activity in
rabbits with chronic renal hypertension. Circ. Res.
32:149-161, 1974.

64. Angell-James, J.E.: Arterial baroreceptor activity in

rabbits with experimental atherosclerosis. Circ. Res.
34:27-39, 1974.

65. Angell-James, J.E.: Pathophysiology of aortic baroreceptors

in rabbits with vitamin D sclerosis and hypertension.
Circ. Res. 34:327-338, 1974.

66. Pickering, T.G., Gribbin, B.and Sleight, P.: Comparison of

the reflex heart rate response to rising and falling
arterial pressure in man, Cardiovascular Res.
6:277-283, 1972.
186 CHAPTER 3

Rothbaum, D.A., Shaw, D.J., Angell, C.S. and Shock, N.W.:

Age differences in the baroreceptor response of rats.
J. Gerontology 29:488-492, 1974.

68. Cox, R.H., Fronek, A. and Peterson, L.H.: Effects of

carotid hypotension on aortic hemodynamics in the
unanesthetized dog. Am. J. Physio1. 229:1376-1380,
1975.

69. Forsyth, R.P., Hoffbrand, B.K. and Melmon, K.L.: Hemodynamic

effects of angiotensin in normal and environmentally
stressed monkeys. Circ. 44:119-129, 1971.
70. Forsyth, R.P. and Harris, R.E.: Circulatory changes during
stressful stimuli in rhesus monkeys. Circ. Res.
26 and 27 Suppl. 1:1-13-1-20, 1970.

71. Caraffa-Braga, E., Granata, L. and Pinotti, 0.: Changes

in blood-flow distribution during acute emotional
stress in dogs. Pflugers Arch. 339:187-205, 1973.

72. Fell, C.: Changes in blood flow distribution produced by

central sciatic nerve stimulation. Am. J. Physio1.
214:561-565, 1968.

73. Hoffband, B.I. and Forsyth, R.P.: Regional blood flow

changes during norepinephrine, tyramine and methoxamine
infusions in the unanesthetized rhesus monkey.
J. Pharmaco1. and Exp. Ther. 184:656-661, 1973.

74. Vatner, S.F., Higgins, C.B., White, S., Patrick, T.

and Franklin, D.: The peripheral vascular response
to severe exercise in untethered dogs before and
after complete heart block. J. Clin. Invest.
50:1950-1960, 1971.

75. Brod, J.: Essential hypertension haemodynamic observations

with a bearing on its pathogenesis. The Lancet
2:773-778, 1960.

76. Kaihara, S., Rutherford, R.B., Schwentker, E.P. and

Wagner, H.N., Jr.: Distribution of cardiac output
in experimental hemorrhagic shock in dogs. J. Appl.
Physiol. 27:218-222, 1969.

77. Davis, J.O.: Viscoelastic properties of conduit arteries.

Am. J. Med. 55:333-350, 1973.

78. Oparil, S. and Haber, E.: The renin-angiotensin system.

New Eng. J. Med. 291:389-401, 446-457, 1974.
CONTRACTILE BEHAVIOR OF ARTERIES 187

79. Forsythe, R.P., Hoffbrand, B.I. and Melmon, K.L:

Hemodynamic effects of angiotensin in normal and
environmentally stressed monkeys. Circ. 44:119-129,
1971. --

80. Ganong, W.F.: Medical Physiology, 5th Ed., Lange Med.

Publications, Los Altos, California

81. Schmid, P.G., Abboud, F.M., Wendling, M.G., Ramberg, E.S.,

Mark, A.L., Heistad, D.D. and Eckstein, J.W.:
Regional vascular effects of vasopressin: plasma
levels and circulatory responses. Am. J. Physiol.
227:998-1004, 1974.

82. Lutz, R.J., Cannon, J.N. and Monroe, R.E.: Shear stress
measurements in model arteries during steady and
pulsatile flow. Fluid Dynamic Aspects of Arterial
Disease. Columbus, Ohio, State University Press,
pp. 5-8, 1974.

83. Nerem, R.N., Rumberger, J.A., Jr., Gross, D.R., Hamlin, R.L.
and Geiger, G.L.: Hot-film measurements of coronary
blood flow in horses. Fluid Dynamic Aspects of Arterial
Disease. Columbus, Ohio, State University Press,
pp. 28-31, 1974.

84. Fry, D.L.: Responses of the arterial wall to certain

physical factors. Atherogenesis: Initiating Factors.
Ciba Foundation Symposium, Amsterdam, Associated
Scientific Publishers. pp. 93-125, 1973.

85. Caro, C.G. and Nerem, R.M.: Transport of l4C-4-Cholesterol

between serum and wall in the perfused dog cornmon
carotid artery. -'-""':"""---'-
Circ. Res. 32:187-205, 1973.

86. Fry, D.L.: Certain chemorheologic considerations regarding

the blood vascular interface with particular reference
to coronary artery disease. Circ. 39 and 40, Suppl.
4:IV-38-IV-59, 1969.

87. Nakata, Y., Shionoya, S., Matsubara, J. and Shinjo, K.:

An experimental study on the vascular lesions caused
by disturbance of the vasa vasorum and the periaortic
vein. Jap. Circ. J. 36:945-951, 1972.

88. Glagov, S.: Mechanical stresses on vessels and the non-

uniform distribution of atherosclerosis. Medical
Clinics of North America. 57:63-77, 1973.
188 CHAPTER 3

89. Sacks, A.H.: The vasa vasorum as a link between hypertension

and arteriosclerosis. Angiology 26:385-390, 1975.

90. Blose, S.H.: Contractile proteins and cytoplasmic filaments

in cloned venous endothelial cells. Fed. Proc.
35:234, 1976.

91. Shimamoto, T.: New concept of atherogenesis and treatment

of atherosclerotic diseases with endothelial cell
relaxant. Jap. Heart J. 13:537-562, 1972.

92. Robertson, A.L. and Khairallah, P.A.: Effects of angiotensin

II and some analogues on vascular permeability in the
rabbit. Circ. Res. 31:923-931, 1972.

93. Heath, D., Smith, P., Harris, P., and Winson, M.: The
atherosclerotic human carotid sinus. J. Path.
110:49-58, 1973.

94. Winson, M., Heath, D. and Smith, P.: Extensibility of the

human carotid sinus. Cardiovasc. Res. 8:58-64, 1974.

95. Cowley, A.W., Jr., Laird, J.F. and Guyton, A.C.: Role of
the baroreceptors in daily control of arterial blood
pressure and other variables in dogs. Circ. Res.
32:564-576, 1973.
96. Mancia, G., Ludbrook, J., Ferrari, A., Gregorini, L.,
Zahchetti, A.: Baroreceptor Reflexes in Human
Hypertension. Circulation Res. 43:170-177, 1978.

97. Recordati, G., Lombardi, F., Bishop, V.S., Malliani A.:

Response of type B atrial vagal receptors to changes
in wall tension during atrial filling. Circulation Res.
36:682-691, 1975.
98. Recordati, G., Lombardi, F., Bishop, V.S., Malliani, A.:
Mechanical stimuli exciting type A atrial receptors in
the cat. Circulation Res. 38:397-403, 1976.
99. Armour, J.A.: Physiological behavior of thoracic
cardiovascular receptors. Am. J. Physiol. 225:
177-185, 1973.

100. Guyton, A.C.: Textbook of Medical Physiology. W.B.

Saunders, Philadelphia, PA, 5th Ed., 1976.

101. Dobrin, P.D. and Rovick, A.A.: Influence of vascular

smooth muscle on contractile mechanics and elasticity
of arteries. Am. J. Physiol. 217:1644-1651, 1969.
CONTRACTILE BEHAVIOR OF ARTERIES 189

102. McDonald, D.A.: Blood Flow in Arteries. William &

Wilkins, Co., Baltimore Md., 2nd Ed., 1969.

103. Penfield, W.: Intracerebral vascular nerves. Arch.

Neurol. Psychiat. (Chic). 27:30-44, 1932.

104. Chorobski, J. and Penfield, W.: Cerebral vasodilator

nerves and their pathway from the medulla oblongata.
With observations on the pial and intracerebral
vascular plexus. Arch. Neural. Psychiat. 28:
1257-1289, 1932.

105. Falck, B., Nielsen, K.C. and Owman, C.: Adrenergic

innervation of the circulation. Scand. J. Clin.
Lab. Invest. Suppl. 102, VI:B, 1968.

106. Nielsen, K.C., and Owman, C.: Adrenergic innervation

of pial arteries related to the circle of Willis in
the cat. Brain Research 6:773-776, 1967.

107. Hernandez-Perez, M.J. and Stone, H.L.: Sympathetic

innervation of the circle of Willis in the macaque
monkey. Brain Research 80:507-511, 1974.

108. Hartman, B.K.: Immunofluorescence of dopamine-beta-

hydroxylase. Application of improved methodology to
the localization of the peripheral and central nora-
drenergic nervous system. J. Histochem. Cytochem.
21:312-332, 1973.

109. Iwayama, T., Furness, J.B and Burnstock, G.: Dual

adrenergic and cholinergic innervation of cerebral
arteries of the rat. An ultrastructural study.
Circ. Res. 26:635-646, 1970.

110. Forbes, H.S., Schmidt, C.F. and Nason, G.I.: Evidence of

vasodilator innervation in the parietal cortex of
the cat. Am. J. Physiol. 125:216-219, 1939.

ll1. Denn, M.J. and Stone, H.L.: Cholinergic innervation of

monkey cerebral vessels. Brain Research 113:394-399,
1976.

ll2. D'Alecy, L.G., and Feigl, E.O.: Sympathetic control of

cerebral blood flow in dogs. Circ. Res. 31: 267-283,
1972.
190 CHAPTE R 3

113. Hernandez-Perez, M.J., Raichle, M.J., Stone, H.L.:

The role of the sympathetic nervous system in cerebral
blood flow autoregulation. Stroke 6:284-292, 1975.

114. Stromberg, D.D. and Fox, J.R.: Pressures in the pial

arterial microcirculation of the cat during changes in
systemic arterial blood pressure. Circ. Res. 31:
229-239, 1972.

115. Raichle, M.E. and Stone, H.L.: Cerebral blood flow

autoregulation and graded hypercapnia. Proceedings
of the 5th International Symposium on Cerebral Blood
Flow Regulation (S. Karger) pp. 1-5, 1972.

116. Reivich, M.: Arterial pC02 and cerebral hemodynamics.

Am. J. Physiol. 206: 25-35, 1964.

117. Betz, E.: Cerebral blood flow: Its measurements and

regulation. Physiol. Rev. 52: 595-630, 1972.

118. Stone, H.L., Raichle, M.E., Hernandez-Perez, M.J.: The

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

119. McKee, J.C., Denn, M.J. and Stone, H.L.: Neurogenic

cerebral vasodilation from electrical stimulation of
the cerebellum in the monkey. Stroke 7: 179-186,
1976.

120. Malliani, A., Parks, M., Tuckett, R.P., Brown, A.M.: Reflex
increases in heart rate elicited by stimulation of
afferent cardiac sympathetic nerve fibers in the cat.
Circ. Res. 32:9-14, 1973.

121. Denn, M.J. and Stone, H.L.: Autonomic innervation of dog

coronary arteries. J. Appl. Physiol. 41:30-35, 1976.

122. Berne, R.M.: Regulation of coronary blood flow. Phsyiol.

~ 44:1-29, 1964.

123. Feigl, E.O.: Sympathetic control of coronary circulation.

Circ. Res. 20:262-271,1967.

124. Feigl, E.O.: Parasympathetic control of coronary blood

flow in dogs. Circ. Res. 25:509-519,1969.

125. Olsson, R.A., Gregg, D.E.: Metabolic responses during

myocardial reactive hyperemia in the unanesthetized
dog. Am. J. Physiol. 208:231-236, 1965.
CONTRACTILE BEHAVIOR OF ARTERIES 191

126. Eikens, E., Wilcken, D.E.L.: Myocardial reactive

hyperemia and coronary vascular reactivity in the
dog. Circ. Res. 33:267-274, 1973.

127. Schwartz, P.J. and Stone, H.L.: Tonic influence of the

sympathetic nervous system on myocardial reactive
hyperemia and on coronary blood flow distribution
in dogs. Circ. Res. 41:51-58, 1977.

128. Fortnun, N.J., Kaihara, S., Becker, L.C., Pitt, B.:

Regional myocardial blood flow in the dog studied
with radioactive microspheres. Cardio. Res. 5:
331-336, 1971.

129. Malliani, A., Lombardi, F., Pagani, M., Recordati, G.,

and Schwartz, P.J.: Spinal cardiovascular reflexes.
Brain Res. 87:239-246, 1975.

130. Burgess, P.R., and Perl, E.R.: Cutaneous mechanoreceptors

and nociceptors. In: A. Iggo (Ed.), Handbook of
Sensory Physiology, Vol. 2, Somatosenory System,
Berlin, Springer-Verlag, pp. 851, 1973.

131. Malliani, A. and Pagani, M.: Afferent sympathetic nerve

fibers with aortic endings. J. Physiol. 263:157-169,
1976.

132. Lioy, F., Malliani, A., Pagani, M., Recordati, G., and
Schwartz, P.J.: Reflex hemodynamic responses
initiated from the thoracic aorta. Circ. Res.
34:78-84, 1974.

133. Malliani, A., Lombardi, F., Pagani, M., Recordati, G.,

and Schwartz, P.J.: Spinal sypathetic reflexes in
the cat and the pathogenesis of arterial hypertension.
Clin. Sci. Mol. Med. 48:259s-260s, 1975.

134. Pagani, M., Schwartz, P.J., Banks, R., Lombardi, F. and

Malliani, A.: Reflex responses of sympathetic
preganglionic neurones initiated by different
cardiovascular receptors in spinal animals. Brain
~ 68:215-225, 1974.

135. Pagani, M., Schwartz, P.J., Bishop, V.S., and Malliani, A.:
Reflex sympathetic changes in aortic diastolic
pressure-diameter relationship. Am. J. Physiol.
229: 286-290, 1975.
Chapter 4 HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS

DR. WEINBAUM: I would like to discuss how fluid

mechanics may be changing the transport of macromolecules
and augmenting vesicle transport
Models for Vesicle and also how endothelial cells
Transport establish intracellular channels
with typical 100 to 200 angstrom
spacing. I think it would be useful to show what models for
vesicle transport have been developed and to show how these
models tie in with the experimental information which is
just now being gathered. Let us look at the first three
figures.

I should mention that all these figures were made when

I was at Imperial College working with Dr. Colin Caro. Dr.
Clifford Lewis of the Zoology Department at Imperial prepared
the EMs I will now show.

Figure 4-1. The interesting feature observed in the

first figure is that when an artery is in a relaxed state,
with no transmural pressure, but stretched back to in vivo
length, numerous blebs are noticed all around the periphery
of the endothelial cell. You are able to easily remove
these blebs by applying transmural pressure. Another feature
which I think will be critical in understanding how vesicle
motion can be enhanced, is what is happening to the nucleus
when mechanical stresses are applied. The nucleus obviously
occupies a large fraction of the volume of an endothelial
cell. It is not known whether a tethering exists between
the nucleus and the plasmalemma membrane of the endothelial
cell. I think it is reasonable to assume that this tethering,
if it exists, is not rigid, but rather flexible. In other
words, if you can produce a slushing motion of the cytoplasm
the nucleus will also move. To turn the nucleus around, one
can hypothesize that you may need very long exposure to
stress, as in Dr. Fry's experiment. To change the orientation
of the endothelial cell, one must change the flow direction.
The time scale for this type of realignment of the nucleus
is several weeks. These long time scale changes should have
little effect on vesicle transport rate.

Figure 4-2. In Fig. 4-2 notice that the blebs are

gone. There is no flow here and the artery is at 100 mmHg
transmural pressure.
193
194 CHAPTER 4

Fig. 4-1: Low magnification (13,000 X) electron micrograph of

canine carotid endothelium in relaxed state. Note numerous blebs
in plasmalemma when there is no transmural pressure. (Courtesy
of C. Lewis). (Reduced 30% for purposes of reproduction.)
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 195

Fig. 4-2: High magnification (45,000 X) electron micrograph of

canine carotid endothelial cell with 100 mmHg transmural pres-
sure. Note involuted appearance of nucleus and long tortuous
extracellular channel o Blebs have disappeared. (Courtesy of
C. Lewis). (Reduced 20% for purposes of reproduction.)
196 CHAPTER 4

Transmural pressure removed all the blebs. The other interesting

feature is the involuted appearance of the nucleus. I have
looked at higher magnification Ems from Dr. Pa1ade's group
and have studied the cytoplasm surrounding the nucleus. In
particular, I wished to see if there is a tethering between
the nucleus and plasmalemma in that region. The other
feature that is beautifully shown here are the long tortuous
intracellular channels. The question I raised this morning
was how the extracellular channels form? The information
that is of great interest to us is the magnitude of the
force required to spread the channel apart. It is difficult
to perform an experiment where one could quantify this, but
I will tell you my current thoughts. In many endothelial
cells, the ciliary body, the gall bladder, arterial endothelium,
you find out that the extracellular channel has a strikingly
uniform spacing that runs between 100 and 200 angstrom units
when there is no flow and which widens to perhaps 1000 A
when there is active transport with a large water flow
accompanying it. There are also localized regions where the
spacing diminishes to 20A or less. This latter spacing
determines the cutoff between the size macromolecules that
traverse the intracellular clefts and the size molecules
that cross the endothelium through vesicle transport. It is
reasonable to hypothesize that there are attractive Van der Waals
forces between the two lateral membranes on each side of the
extracellular channel. We also know that the exterior
surfaces of cell membranes have small negative charges
which provide repulsive forces. Those two forces in balance
establish the equilibrium spacing. The surface charge can
be redistributed, that is, there are localized regions where
you can have more or less surface charge. This is the
reason I believe the equilibrium spacing can be diminished
locally. I also have seen experiments with lecithin and
related experiments with low molecular weight dextrans where
you can change the polarization of molecules comprising the
phospholipid bilayer or the intervening medium. This also
causes a change in the equilibrium spacing. The equilibrium
spacing by itself does not tell you anything about the
absolute magnitude of the forces. All you know is that the
two forces are equal because they are in balance. To determine
the magnitude of each force, one is interested in the departure
from the equilibrium spacing. One wishes to measure the
additional separation caused by a force that can be quantified.
The extracellular channel water flux creates a pressure
force which can be calculated from hydrodynamic theory if
the flow rate is known. These two measurements, the flow
rate in a channel, and the change in channel spacing should
provide a reasonable estimate of the magnitude of the forces
that hold the channel together.
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 197

Ems showing this change in spacing are available --unforuntately,

the people who are most active in this research: Dave Tormey
and Jared Diamond, are in California and hard for me to
visit. The Ems these investigators have taken of the gall
bladder beautifully show the changes in channel spacing as a
function of the flow rate.

Figure 4-3. The next figure shows a flattened endothelial

cell at 100 mmHg pressure. The flow direction is from top
to bottom. It is evident from
Vesicular Transport this figure that the endothelial
cell can deform very easily due
to hydrodynamic forces at its surface. The nucleus here
looks symmetric; it really isn't if you look carefully. The
principal transport of vesicles does not occur in the nuclear
or perinuclear region because of the large resistance offered
by the endoplasmic reticulum which is concentrated in these
areas. I believe that the important vesicular transport
occurs in the periphery of the cell which has relatively
little volume. Small movements of the nucleus could easily
generate sizeable velocities in the cell periphery. The
fluid response time for these changes is almost instantaneous.
Thus, if one can establish a cytoplasmic circulation to
augment the Brownian diffusion of vesicles, in the periphery
of the cell, one could have a nice mechanism for increasing
or changing the macromolecule transport rate. Another thing
which is very nicely illustrated in this figure is the two
phase structure of the
Macromolecule Transport underlying tissue which is
comprised of smooth muscle cells
and interstitial space. The endothelium and the underlying
tissue are ultrastructural opposites from the standpoint of
macromolecule transport. The macromolecules cross the
endothelium through vesicles, since the extracellular channels
are too narrow. The macromolecules, once they are released
by the vesicles at the tissue front, can easily pass through
the large gaps between adjacent smooth muscle cells since
these gaps are 1000 A or more.

Figure 4-4. I wish I could show you some of Dr. Palade's

electron micrographs of vesicles interacting with the plasmalemma
membranes. The electron micrographs
Vesicles Interacting with show the various stages during the
Plasmalemma Membranes attachment and with rupture
processes. Fig. 4-4 is a schematic
showing the various steps in the vesicle attachment process.
198 CHAPTER 4

Fig. 4-3: Electron micrograph of flattened endothelial cell

in canine carotid artery with flow in lumen. Transmural
pressure 100 mmHg. 17,000 X (Courtesy of C. Lewis). From
Weinbaum and Caro (1).
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 199

=r+==
75 A

~ 350A

(b)
(a)

-
(c) (d)

Fig. 4-4: Sketch of proposed sequence of events leading to

the attachment of free vesicle to plasmalemma. (a) X 200 ~
region of strong hydrodynamic interaction, (b) indentation of
plasmalemma due to Van der Waals force interaction, (c)
configuration before formation of vesicle neck and (d) attached
vesicle. From Weinbaum and Caro (1).
200 CHAPTER 4

When a vesicle approaches to within roughly one radius of

the plasmalemma it starts to experience a very strong hydrodynamic
interaction. At even closer spacings one encounters the
Van de Waals force interaction that we talked about earlier.
While exterior surfaces of membranes have a small negative
surface charge, it appears that the interior surface is
devoid of charge. You have to look at many pictures to
catch a vesicle in the transitory step before attachment.
If the plasmalemma had a negative surface charge on its
interior surface, you would expect an equilibrium spacing
like one observes in the intracellular channel. It does not
exist. There is no fluid gap between the vesicle and the
plasmalemma prior to reattachment. The configuration where
the vesicle and the plasmalemma abut against one another
forming a double membrane is unstable.

What appears to happen is that the protein heads of the

bilayer membrane flow radically outward leaving the lipids
behind in the middle of the newly formed vesicle attachment
stalk. It is difficult to estimate the time scale of this
process or how long the lipid diaphragm remains in the
vesicle neck. The next question is how an attached vesicle
ruptures? The most likely hypothesis is that this is due to
Brownian motion. We are currently trying to develop a theory
to predict how far an attached vesicle will zig-zag in its
attached state.

DR. NEREM: I am confused on one point. What were you

implying what the long channel was?

DR. WEINBAUM: The long intracellular channel.

DR. NEREM: The other question is: Do you want to say

anything about the relative importance of pressure versus
shear stress in enhancing some type of vesicular transport?

DR. WEINBAUM: I don't know the answer to this. We are

going to have to perform more sophisticated EM studies to
show how a nucleus moves inside a cell. We need to find out
what kind of internal circulation is established in the
periphery of the endothelial cell due to hydrodynamic forces
at its surface.

DR. BJORKERUD: I think one should also then try to

account for some other factors when trying to make a hypothesis
of how vesicles fuse with the plasma membrane.
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 201

I mean, if this were so, the cells would have difficulty in

discerning between mitochondria, vesicles of other types,
etc. The membrane would pick up things rather unselectively.
People have discussed the matter and it has been shown that
concommitantly with fusion or before fusion of vesicles with
the plasma membrane, there is a rearrangement of the protein
part of the membrane, i.e., of the intramembranous particles.
These are rearranged when a vesicle will form. This is
thought to reflect some selective mechanism by which the
plasma membrane can identify structures. As a consequence,
something which should not leave the cell does not fuse with
the membrane. I think you must also incorporate these
features into your theory before we would be able to accept
it.

DR. WEINBAUM: The vesicle's membrane is undoubtedly

the same phospholipid bilayer as the plasmalemma membrane.
The two membranes are indistinguishable. The vesicles are
passive bodies. There are many other bodies in the interior
of the cell which have the specialization you are talking
about, but the vesicle membrane is just extra membranous
material which is identical in structure as far as I know to
plasmalemma.

DR. BJORKERUD: It is very probable that the outer leaf

of the membrane has different phospholipid distribution as
compared to the inner one which would be difficult to reconcile
with your theory.

DR. WEINBAUM: No, the vesicle membrane is inside out

compared to the plasmalemma.

DR. COLTON: When the vesicle fuses to the surface, the

inside is like the outside of the plasma membrane.

DR. WEINBAUM: That is the reason I think there is a

difference in surface charge, between the exterior surface
of the vesicle and the exterior surface of the plasmalemma.

DR. COLTON: But then you need a separate mechanism

which is not charge dependent which prevents all the other
organelles in the cell from going through and fusing with
the plasmalemma.

DR. KENYON: I would like to ask you about your hypothesis

on the junction characteristics during pressurization and
how you are going to calibrate, essentially, this hydraulic
resistance because if you don't calibrate you still don't
get, I believe, the pressure difference across the endothelial
surface.
202 CHAPTER 4

DR. WEINBAUM: In a gall bladder epithelium you have a

tight junction at the luminal surface. The gall bladder
water flux can be arrested by poisoning the active transport.
Jared Diamond (1) has shown that the active transport--in
the gall bladder is electrically neutral. The chloride ions
are pumped and the Na ions follow to satisfy electroneutrality.
The water movment is passive and follows the local osomotic
gradient produced by the ion pumps. By collecting exuded
fluid and estimating the geometry of the cell layer, you can
make an estimate of the flow rate in the extracellular
channel. With this information, I can calculate the pressure
distribution in the channel.

DR: KENYON: Which are you considering known--the

dimension of the channel or the flux?

DR. WEINBAUM: I only know the dimension at the base of

the channel. I would like to determine the distorted geometry
of the channel during flow using a lubrication theory approach.

DR. COLTON: Are you suggesting that you are going to

do this with endothelial cells?

DR. WEINBAUM: No, not with endothelial cells, with gall

bladdGr epithelium. While it is not the same endothelium
cell layer, I expect the membrane molecular force to be
similar.

DR. CAREW: I would again like to bring up the point

that not all vesicles are equal with regard to this model.
I would also point out that
Not All Vesicles Are Equal the total picture may not be
as simple as all vesicles
having the same sort of Van der Waals interaction in attaching.
There is some experimental evidence for there being differences
in attachment as a function of what is inside the vesicle.
Some work done by the Steins in Jerusalem, (2) when they
were on sabbatical in La Jolla, with regard to the binding
of endocytic vesicles, primarily lysosomes, in cultured
fibroblasts, showed that they could block the fusion of the
primary lysosome with the endocytic vesicle by pre-incubating
the cells with Concanavalin A and that suggests that what
may be found on the inner membrane of the vesicle could
sufficiently alter the outer membrane leaflet so that you
may have different fates for different sorts of vesicles and
it raises a question of what is responsible for the attachment
of vesicles to one body or the other, to what extent, are
the phospholipids important, to what extent are the li~olipids
important, and membrane fluidity?
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 203

How do things change with what is carried in the vesicle?

DR. WEINBAUM: I agree with all the things that you

have just said. There are many intracellular bodies which
are membrane-bound. The very striking feature about plasmalemma
vesicles is their homogeneity in size. If you look at the
statistical distribution of plasmalemma vesicles sizes you
will find that they are seven hundred angstroms with a
standard deviation of less than fifty angstroms.

DR. SCHWARTZ: Be careful, because they have looked

largely at capillary vesicles and there are very real differences
between arterial and capillary endothelium.

DR. WEINBAu~: I agree. The internal structures of

capillary and arterial endothelial cells are quite different.
The vesicles themselves appear to be quite similar.

DR. FRY: Hemodynamic forces interact with the arterial wall

in two ways: 1, to convey blood borne substances and cellular
elements to and from the wall,
Effects of Hemodynamic and 2, to exert a mechanical force
Forces on Endothelial directly on the endothelial
Surface surface and its subjacent tissues.
This presentation concerns only
the latter, the effects of hemodynamic forces on the endothelial
surface. For simplicity these forces may be resolved into two
components, a perpendicular force field, "pressure stress,"
and a field of force acting parallel to the surface, a
"shear stress." Pressure stresses are related mostly to the
arterial blood pressure whereas shear stresses are related
mostly to the drag of the adjacent blood flow.

The response of the endothelial surface to shear stress

appears to depend on the magnitude of the stress and also
upon the stability of the stress field. At relatively
physiologic levels of stress exposure the endothelial cells
are seen to be elongated with elliptically shaped nuclei
having major semiaxes that are oriented in the direction of
the adjacent streamlines of the blood flow (3). If the
adjacent streamlines are changed by certain experimental
surgical techniques, the patterns of endothelial cell orientation
are seen to realign to the new pattern of streamlines within
a matter of weeks (4). Thus many of the structural features
of the normal endothelial surface appear to be determined,
at least in part, by the adjacent hemodynamic forces.
204 CHAPTER 4

The permeability of the endothelial surface to macromolecules

in the blood also appears to be sensitive to the adjacent
hemodynamic stress fields.
Permeability of Permeability increases monotonically
Endothelial Surfaces with actuely increased levels
of stress exposure (5,6,7).
Surface permeability is also increased in regions of turbulence
(5).
Under conditions of exposure to chronically elevated
shear stress, such as in an artery supplying a surgically
induced arteriovenous fistula, the permeability is seen
gradually to decrease over a matter of weeks. Thus it
appears that under chronic conditions compensatory mechanisms
are apparently invoked which tend to return the endothelial
permeability to normal values (8).

The levels of shearing stress considered up to this

point, although increased, are unassociated with any significant
structural evidence of endothelial damage. However, if
shearing stress is elevated to levels ranging from 200 to
700 dynes/em 2, endothelial structural changes become evident
even by light microscopy. These consist of cellular swelling,
deformation, and vacuolization
Acute Critical Yield Stress of the cytoplasm followed by
progressive endothelial cell
disintegration and erosion. The level of stress in an
individual that is associated with the occurence of these
changes has been defined as the acute critical yield stress
of the endothelial surface for that individual at that site
(5,9). Erosion of the endothelial surface is associated
with a massive influx of plasma substances into the intimal
surface as well as the deposition of platelets, leucocytes,
and fibrin on the exposed intimal surface. It is of interest
to note that small areas of endothelial erosion can be found
in the arterial trees of normal animals (8) suggesting that
this process may be a recurring event throughout the lifetime
of an individual. Whether these sites represent regions
that have been exposed momentarily to an intense hemodynamic
force or whether the acute yield stress of the endothelial
surface at these sites has been lowered by certain metabolic
or hemodynamic events (8) remains to be explored.

Pressure stress can also be shown to alter endothelial

surface permeability. Pressure can increase the permeability
of the arterial wall by stretching the endothelial surface
(3) and also by increasing the driving pressure across the
wall to increase the flux of plasma substances (10).
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 205

In principle, an increased pressure stress may also influence

the endothelial surface permeability indirectly by producing
a deformation of the normally streamlined conduit geometry
thereby predisposing the normally stable streamlined flow
patterns to unstable patterns at various points along the
arterial trees. To the extent that shear stress patterns that
are changing in magnitude and directions are associated with
chronically endothelial surface
Altered Endothelial permeability, this indirect effect
Permeability in Experimental of increased pressure on flow
Atherogenesis stability may be important (3,8).
Thus a variety of hemodynamic
events can be shown either directly or by inference to alter
endothelial structure and permeability. Data presented
elsewhere (3,8,11) suggest that altered endothelial permeability
plays one of the central roles in experimental atherogenesis
and by inference perhaps also in human atherosclerosis.

DR. MANSFIELD: Our interest in the endothelial cell,

its growth in tissue culture, and its function as an in vivo
lining for artificial vascular prosthesis is an outgrowth of
our work with a left ventricular assist device and the problems
of thrombus generation associated with the use of that device.
(12-17) •

During our evaluation of several anticoagulation techniques

designed to reduce thrombus in these devices, we have used a
model for thrombus generation
Anticoagulation Studies which is shown in Fig. 4-5.
With this technique, we remove
a segment of the descending aorta in the calf and replace it
with a non-porous, prosthetic graft containing an inside
lining which we wish to study. The graft is attached to the
ends of the aorta using continuous suture (2-0 Ticron)
techniques currently used in clinical vascular surgery. One
of the most promising materials we have used has been
polypropylene microfiber. With this lining surrounded by a
non-porous polyurethane back and a dacron sleeve to permit
suturing to the ends of the individual aorta, we have studied
thrombus generation with and without anticoagulants.

When anticoagulants were not utilized in this model

with the polypropylene microfiber surface, thrombus, rapidly
deposited on the surface up to
Thrombus Generation depths of 1 to 3 millimeters was
found after 7 days implantation.
See Fig. 4-6. Frequently, thrombus hung downstream into the
lumen of the aorta, but where the aortic endothelium was intact
there was no thrombus attachment.
206 CHAPTER 4

Fig. 4-5: Thrombus generation model used in these studies. A

segment of the descending aorta in the calf is removed and
replaced with a vascular prosthesis CG). Commonly used clinical
vascular surgical techniques are used to anastomose the graft
and the aorta (AO). The graft, shown in cut-out, is composed
of an external dacron sleeve CDS) used for sutering, a non-
porous polypropylene backing tube (NPP) , and an inner lining
surface which is varied according to the study in progress.
Flow in the aorta is shown by the arrow. Because of its
location in the descending aorta, emboli which may originate
from the graft can be evaluated by inspection of the kidneys for
infarcts. This serves as a sensitive indicator of distal emboli.
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 207

Th

Fig. 4-6: Drawing of thrombus generation on aortic tubular

graft, 7 days after implantation. No anticoagulants used.
Thrombus (TH) originates only from the bare surface of the
graft (polypropylene microfiber) and hangs downstream into
the lumen of the aorta. Embolic infarcts found in the kidneys.
NPP: non-porous polyurethane backing, DS: dacron sleeve,
AO: aorta, G: graft.
208 CHAPTER 4

Placing the graft in the descending aorta permitted us to

evaluate the incidence of peripheral emboli which could be
seen as multiple infarcts within the kidney. Multiple
infarcts were found when no anticoagulation was used.

After evaluating 11 different anticoagulation protocols

in nearly 100 calves, we found the most effective protocol
was a combination of coumadin, persantine, and aspirin (18).
With this anticoagulation protocol, initiated 5 days before
surgey, there was minimal thrombus generation on the surface
of the aortic prosthetic graft after 7 days implantation.
See Fig. 4-7. Imperfections in the surface were still
apparent, however, and the thrombus depths of 0.1 to 0.3 mm
were seen. We found islands of whitish material on the
thrombus surface which we have termed "white caps." On
histologic section these were found to be accumulations of
platelets and masses of white cells with interwoven fibrin
strands. These were rarely seen after implantations of more
than 3 to 4 weeks.

The anticoagulation protocol has been satisfactory for

the long-term function of left ventricular assist devices
in calves for periods up to
Endothelial Cells 3 months in our hands. However,
to accomplish the ultimate goal
of no thrombus deposition we turned to the area of endothelial
cell linings for prosthetic devices. When we began, (17) the
techniques for serial passage of viable endothelial cells in
tissue culture had not been developed.

We initially utilized cells from the inferior vena cava of

calves and more recently have used endothelial cells from
the carotid artery of the
Technique for Endothelial calf. The technique we developed
Tissue Culture for obtaining these cells is
shown in Fig. 4-8a & b. The
blood vessel is washed with Hanks Salt Solution (calcium and
magnesium free) cut longitudinally and exposed to 0.02%
versene-trypsin (0.25% solution) (pH 8.0). After 25 minutes
the incubating solution is gently pipet ted over the endothelial
surface and the dislodged cells collected. Complete tissue
culture medium was added to the cell suspensions as a source
of calcium and serum protein to inhibit the action of versene
and tryspin. Following centrifugation the cell pellet was
resuspended in complete tissue culture medium and seeded
into plastic flasks. Using this technique we have now had
mUltiple cell lines which have had over 25 serial passages
in tissue culture.
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 209

Fig. 4-7: Drawing of thrombus generation found on aortic

tubular graft using coumadin, persantin and aspirin anti-
coagulation protocol. Graft implantation for 7 days.
Maximum thrombus depth 0.3 milimeters. No renal embolic
infarcts found. TH: thrombus layer, NPP: non-porous
polyurethane backing layer, DS: dacron sleeve, AO: aorta,
G: tubular graft.
210 CHAPTER 4

2.

6.

Fig. 4-SA: Technique for obtaining cells for tissue culture.

A: 1. The excised vessel segment is washed with calcium and
magnesium-free Hanks Balanced Salt Solution to remove residual
blood. Carotid artery cut into three sections after the
surrounding adventitial tissue had been dissected away.
2. The segments are placed endothelial surface up in a petri
dish containing 0.02% versene solution in calcium and magnesium-
free phosphate buffered saline (pH 7.0). After 5 minutes at
room temperature, the specimens are transferred to a second
petri dish containing 0.02% versene tryps in (0.25%) solution
(pH 8.0). After 25 minutes at 37 0 C in a humidified atmosphere
of 5% C02 and 95% air, the incubating solution is gently
pipetted (3) over the endothelial surface, and the dislodged
cells collected (4).
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 211

7.

9.

Fig. 4-8B: These cells were then transferred to flasks for

further growth. When a satisfactory number of cells had been
grown, they were centrifuged into a pellet (6). When approx-
imately 35 X 10 6 cells were available they were resuspended and
gently distributed evenly over the inside lining of the aortic
tubular graft (8). The cells were cultured for 7 days before
implantation in animals (9).
212 CHAPTER 4

The endothelial cells grow well in plastic as well as on

glass surfaces.

We feel it is significant that the surface anatomy of

venous endothelium is significantly different from the
surface anatomy of arterial
Venous Endothelium Different endothelium. In culture, the
from Arterial Endothelium arterial endothelium has grown
more aggressively than venous
endothelium and by preference we have used arterial sources
for endothelial cells in most of our work. Fig. 4-9a is a
scanning electron micrograph of venous endothelium taken
from the inferior vena cava of a calf. Fig. 4-9b is a
similar scanning electron micrograph of arterial endothelium.
The prominent microvillae and rounded endothelial cell
configuration of the venous endothelium is in contrast to
the less prominent and often absent microvillae in the
arterial endothelium and the obvious longitudinal alignment
of the cells which line the artery. The characteristics of
multiple microvillae on venous endothelium and usually a
single or occasionally two or three projections from the
arterial endothelium are characteristics which are maintained
during serial passage in tissue culture.

As we became more familiar with the technique of handling

endothelial cells and growing them in tissue culture, we
began to seek surfaces which could be used as artificial
vessels or linings for assist devices on which the cells
would grow. Our experience has shown that polypropylene
microfiber, parylene c-coated (Type B) from Union Carbide
Corporation has been an excellent substrate for endothelial
cell growth. Figure 4-10 is a scanning electron micrograph
of the microfiber surface. We have found the endothelial
cell to be rather fastidious as to the substrate to which it
likes to adhere. It prefers a relatively regular surface and
has difficulty in covering areas where there are large gaps
between supporting structures. Polypropylene microfiber
of identical chemical composition and surface characteristics
if physically teased apart during manufacture often leads to
poor endothelial cell coverage.

To test the adherence characteristics of the endothelial

cells to the underlying prosthetic substrate we devised a
technique of applying a
Adherence of Endothelial jet of fluid onto the surface
Cells to Prosthetic of the cells. The fluid passes
Materials through a tapered glass nozzle
and is directed vertically
toward the cells.
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 213

Fig. 4-9A: In vivo venous and arterial endothelium. A.

Scanning electron micrograph of Bovine venous endothelium.
Surface microvillae are numerous and cellular arrangement is
cobblestone in appearance (X 1180).
214 CHAPTER 4

Fig . 4-9B: Scanning electron micrograph of Bovine arterial

endothelium showing single or double projections from the
surface of each cell. These projections have a different
configuration than the microvillae as seen on venous endothelium
in Fig. 4-5A. (X 1599).
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 215

Fig. 4-10: Scanning electron micrograph of polypropylene

microfiber, parylene C coated (Type B). Microfiber distribution
is random, the fibers are fairly uniform in diameter. (Approx-
imately 1 micron). eX 2210).
216 CHAPTER 4

By maintaining a constant pressure and velocity and therefore

force of the fluid hitting the cell surface from sample to
sample, it is possible to compare qualitatively the adherence
characteristics of each of the substrates on which the cells
have been grown. See Figure 4-11. Vital staining of the
cells after shear testing demonstrates which cells have
survived and over how great a surface area the cells have
disrupted from the substrate below. Our calculations, which
can only be considered approximate, would suggest that we
have used shear stresses (which vary over the surface in a
predictable way from the center of the fluid jet) between no
stress to approximately 600-800 dynes per centimeter squared.

Once we had learned to line discs for study in the shear

apparatus, we proceeded to line tubular grafts which had the
polypropylene microfiber surface. This was done in tissue
culture prior to surgical implantation into the thrombus model
shown in Figure 4-5. Figure 4-12 is a histologic section of
endothelial cells pre-lining the polypropylene microfiber
surface prior to implantation into the aorta. Note the
monolayer and ability of the cells to conform to the surface
irregularities.

When the endothelial cell lined tubular grafts were

implanted in the aorta without the use of anticoagulants, we
found no thrombus where the
Endothelial Cell Grafts endothelium remained intact.
Figure 4-13 is a drawing of
the findings of such a graft implanted for 4 weeks in the
descending aorta of the calf. We found a monolayer of viable
endothelial cells (these were autologous cells obtained from
the carotid artery of the same calf into which the graft was
later implanted) lining the tubular graft at the time the
animal was sacrificed. Minute bits of thrombus were found at
the anastomotic junctions, but nowhere else in these animals.
(Compare to Figure 4-6). Obviously, the endothelial cell
linings were able to withstand the shear of the descending
aorta, remain viable, and were as effective in preventing
thrombus as the normal aortic endothelium above and below
the graft. Figure 4-14 is a scanning electron micrograph of
the endothelial surface of a 4 week tubular implant. One
can see that the endothelial cells have aligned in the
direction of flow and have remained as a confluent monolayer.
No such flow alignment was present at the time of implantation.
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 217

Fig. 4-11: Shear testing apparatus. A disc supporting the

substrate and a covering of endothelial cells (S: shear tested
side, c: control side) is placed on a support stand within the
beaker . A glass nozzle is mounted immediately above the surface,
and fluid at a known pressure and velocity impinges on the
endothelial cell surface at right angles for 30 seconds. The
disc is then removed, vital staining done to determine cell
distribution and cell dea th. As can be seen, an area of
significant staining has occurred on the shear tested side of
the disc in comparison to the control side.
218 CHAPTER 4

MF
"' ..

Fig. 4-12: Histologic section of endothelial lined poly-

propylene microfiber. Note the monolayer of endothelial
cells and ability to conform to surface irregularities.
Line drawing defines the microfiber (M) and the endothelial
cell (N) (X 670). (Reduced 30% for purposes of reproduc-
tion. )
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 219

NPP

os

Fig. 4-13: Aortic tubular graft pre1ined with autologous

arterial endothelium in tissue culture. No significant
thrombus was found on the surface of this graft. No anti-
coagulants were used. Minute amounts of thrombus were seen
at the anastomotic junctions. An endothelial monolayer
covered the polypropylene microfiber lining of this graft, and
cell orientation was parallel to that of blood flow through
the vessel. E: endothelial monolayer, NPP: non-porous
polyurethane backing, DS: dacron sleeve, G: aortic tubular
graft, AO: aorta.
220 CHAPTER 4

Fig. 4-14: Scanning electron micrograph of autologous

endothelial cell-lined tubular aortic graft after 4 weeks
implantation. The cells have aligned in the direction of flow
and no thrombus is evident. The surface remains a monolayer
of endothelial cells. (X 4609).
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 221

In addition to preventing thrombus generation, we found

that viable endothelial cells lining the prosthetic graft
have a significant influence on the tissue reaction at the
anastomotic junctions between the graft and the normal aorta.
Figure 4-15 is a drawing of the findings at 4 weeks of a
control graft surface (no cell lining, no anticoagulants).
There has been a continuous in-growth of pannus (19) which
has begun to digest and organize the thrombus present on the
surface of the polypropylene microfiber. We have found this
pannus continues to progress for at least 2 centimeters from
the anastomotic margins in experiments lasting up to 4
months. Macrophage activity within the thrombus itself
precedes the in-growth of pannus over the surface of the
thrombus and leads to actual thinning of the thrombus. In
contrast, the endothelial cell lined grafts had no such
pannus in-growth and aortic endothelium to graft attachment
occurred with minimal underlying tissue reaction. No macrophage
involvement beyond the anastomosis was seen. It appears
that viable endothelium has the ability (even after multiple
serial passages in culture) of communicating a message to
normal arterial endothelium inhibiting a hyperactive healing
response to an implanted prosthetic material. This may be a
form of in vivo contact inhibition.

We have been interested in the method by which endothelial

cells attach to their underlying substrate, be it sub-
endothelial tissues in the wall
Role of Microfilaments in of a vessel, or the glass or
Endothelial Cell Attachment plastic vessels in tissue culture.
In tissue culture there are no
sub-endothelial tissues present and all cellular components
must therefore be of endothelial cell origin. Dr. Schwartz
showed slides which suggested there were filaments, very
fine microfilaments, if you will, under the endothelial cell
and suggested that these might playa role in endothelial
cellular attachment to underlying tissues. We have found
similar structures under endothelial cells in tissue culture.
Figure 4-16 demonstrates these very fine filaments, far
finer than retraction fibrils in dimension. These transmission
electron micrographs suggest that there may be physical
connections between these filaments and structures within
these bovine endothelial cells, possibly the endoplasmic
reticulum. Certainly our results demonstrate that whatever
the mechanism, endothelial cell attachment to suitable
substrates is more than adequate to withstand the shear
stresses within the descending aorta of the calf. In
addition to the bovine endothelial cell we have been able to
grow and serially passage human endothelial cells in tissue
culture also.
222 CHAPTER 4

Fig. 4-15: Drawing of findings of aortic tubular graft after

4 week implantation. No pre-implant cellular linings and no
anticoagulants were used in this study. The cutaway section
demonstrates pannus (P) ingrowth over the previously deposited
thrombus (TH) on the microfiber surface lining this control
graft. Diagrammatic representation of macrophage activity
leading the pannus ingrowth is seen as white dots in the
thrombus. No such pannus activity was seen when grafts were
prelined with autologous endothelium. See Fig. 4-13. NNP:
non-porous polyurethane backing, DS: dacron sleeve, G: graft,
AO: aorta.
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 223

This is technically much more difficult and it is only

rarely possible to obtain a true endothelial cell of human
origin to serially passage in culture. It is interesting to
note that when fibroblasts of either human or bovine origin
are studied in a similar manner, no such microfilamentous
structures between the cell and the underlying substrate
have been seen in our studies. Figure 4-17 shows typical
Weibel-Palade bodies seen after serial passage of human
endothelial cells in tissue culture. We have found that the
Weibel-Palade bodies are not as frequently encountered
within endothelial cells of animals as compared to humans.
Structures within the cell shown may be similar to microfilaments
seen in bovine endothelial cells. Fig. 4-16.

The ability to provide a monolayer of true vascular

endothelial cells on the surface of a porous microfiber
substrate has suggested
Transport Mechanisms to us the possibility of using
across Endothelial Cell this approach to study transport
Linings mechanisms across endothelial
cell linings. The polypropylene
microfiber aortic tubular grafts have polyurethane backing to
support them. However, it seems conceivable to us as is
shown in Fig. 4-18, that it may be possible to line such a
fabric with endothelial cells support-mounted between two
rings. The ring becomes the connection between two fluid
media-containing chambers. Ionic and molecular movement
between the chambers which was passed through the endothelial
cell lined disc could provide valuable information with
regard to endothelial cell transport mechanisms. While our
laboratory does not have the biochemical capability to
perform such studies, we hope this suggestion will encourage
others more capable in that area to consider this possible
approach. Difficulties, of course, will be in proving that
there is 100% coverage of the microfiber with endothelial
cells at the time the experiments are performed.

Although there are similarities in the endothelial

cells of human and of bovine origin, there are also differences.
Endothelial cells of human origin
Human vs. Bovine can be separated out to individual
Endothelial Cells cells and can survive alone. We
have been unable to get bovine
arterial endothelial cells to survive as single cells.
They require other endothelial cells in contact with them to
survive.
224 CHAPTER 4

Endo

Fig. 4-16: Transmission electron micrograph of Bovine endo-

thelium grown in tissue culture on plastic. A section is taken
at the junction between the cell and the plastic and demonstrates
the microfilaments discussed in the text. (X 7600). (Reduced
30% for purposes of reproduction.)
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 225

WPB
o WPB

Fig. 4-17: Transmission electron micrograph of human endothelial

cells after serial passage and culture demonstrating Weibel-
Palade bodies characteristic of endothelial cells.
226 CHAPTER 4

SR fabric SR
dlac

Fig. 4-18: Design concept for the study of endothelial cell

transport mechanisms. Now that endothelial cells can be grown
as mono1ayers covering a porous microfiber, we suggest the
use of this t ec hniqu e to study endothelial cell transport
mechanisms. By supporting the micro fiber between discs
connecting two chambers, it should be possible to study the
ionic and molecular movement between the chambers and across
the cells.
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 227

In addition. the bovine arterial cells are very dominant in

cell culture. They will overgrow fibroblast cultures and
replace them. an exceedingly unusual occurrence in tissue
culture work. We have also seen in bovine cultures that
endothelial linings may pile up to 2 or 3 cells deep. Once
they have been implanted in the aorta of the autologous
calf. upon removal they are found to be monolayered. The
fate of the other cells which were previously present is
unknown. We have not seen this pile up of endothelial cells
in human culture. If the antithrombogenic capability of the
endothelial cell is not dependent on cell viability but
rather on a surface structure or a bio-chemical matrix, it
is exciting to think of the possibilities should that matrix
or surface be generated on the inside of a vascular prosthesis.
If indeed monocellular linings which are 100% confluent can
be grown on porous backing. studies of transport across
endothelial cell linings may give us new insight into the
development of arteriosclerotic vascular disease.

DR. DEWEY: Referring· to the differences in the shear

resistance that you measured between bovine and human endothelial
cells. did you actually obtain quantitative values for the
shear stresses which would remove endothelial cells from the
surface?

DR. MANSFIELD: It is extremely difficult to get a

precise number. You can get a ball park figure and I might
say you would be plus or minus a factor or two. We are
talking in general; they did seem to shear in the range of
four to six hundred dynes per centimeter squared. as we
would calculate it.

DR. DEWEY: Do these values refer to human or bovine

cells?

DR. MANSFIELD: No. the humans would be the upper limit

of that and the bovine at the lower limit. There is a
distinct separation between the two.

DR. SCHWARTZ: I would like to make a comment. I

thought that was an excellent presentation. Dr. Mansfield.
There are a lot of questions I would like to ask but I guess
I will have to ask some of them later. I was very intrigued
first of all with the polypropylene Type V. This presumably
does not permit any cell growth from the adventitial surface
into the luminal surface?
228 CHAPTER 4

DR. MANSFIELD: No. none at all. It is completely

separated from the adventitia by a one millimeter backing of
polyurethane which is absolutely solid.

DR. SCHWARTZ: So then it is completely solid. The

influence of the combination of coumarin. persantin. and
aspirin on thrombosis in the graft is particularly interesting.
I wonder if you might tell us why you chose persantin as
part of the package. and how much you gave?

DR. MANSFIELD: Surely. We have studied heparin.

coumarin. persantin. coumadin. persantin, aspirin and dextran
and several other anticoagulants. The best results were
obtained from a combination of coumadin, persantin and
aspirin, a combination found useful by Harker and Schlichter
in prosthetic valve surgery. (20). The population doubling
time for the human from saphenous vein origin is approximately
from 2 to 96 hours. The doubling time for the bovine calf
arterial endothelial cell is approximately 32 hours. ·See
Table 1.

We have further studied the interaction of platelet-

rich plasma and the surface of tissue cultured fibroblasts
and endothelial cells of
Effects of Platelet-rich bovine origin. When platelet-rich
Plasma on Cultured Cells plasma is added to tissue cultured
fibroblasts of bovine origin,
clumping, pseudopod formation, and evidence of a release
reaction can be seen. See Figures 4-19 and 20. If under the
influence of versene fibroblasts contract and separate on a
plastic surface in culture, and platelet-rich plasma is
added, the platelets adhere to the fibroblast cell margins
and surface, but do not adhere to the exposed area of the
plastic.

When an endothelial cell culture is exposed to platelet-

rich plasma of similar platelet concentration we have not
seen any attachment to the endothelial cell surfaces but
merely a pooling of non-creative platelets in the valleys
between the nuclei of the endothelial cells. If versene is
added to the culture and endothelial cells retract (leaving
fibrils between cells) the platelets adhere to these retractile
processes. Attachment and some pseudopod formation is seen
but the release reaction has not been identified in our
studies. (21). See Figures 4-21 & 4-22. We have not seen
platelets adhere to areas between endothelial cells under
the influence of versene, except where there are retraction
fibrils.
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 229

TABLE 1

Cell Characteristics Human Bovine

Source Adult Vein Yearly Artery

Attachment Individual Clump

Single-Cell Viability High (90%) Low (30-40%)

Density Dependent Yes No

Dominance Non-Dominant Dominant

Population Doubling Time 96 Hours 32 Hours

Maximum Passages 7 25

Weibel-Palade Bodies Present Present

Extra Cellular Microfilaments Present Present

230 CHAPTER 4

Fig. 4-19: Artist's concept of reaction between platelet-rich

plasma and Bovine fibroblasts in culture. Note clumping,
pseudopod formation and evidence of release reaction on the
surface of the fibroblast.
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 231

Fig. 4-20: Scanning electron micrograph of reaction between

platelet-rich plasma and Bovine fibroblasts in culture. Note
clumping, pseudopod formation and evidence of release reaction
on the surface of the fibroblast.
232 CHAPTER 4

Fig. 4-21: Artist's concept of reaction seen between p1ate1et-

rich plasma and endothelial cell treated with versene. Retrac-
tion fibrils are prominent, platelet attachment is only to these
retraction fibrils (RF). Attachment and pseudopod formation
has been seen, but no release reaction has been identified in
these studies. P: platelet.
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 233

("'''
,. ,...
It. f'
~
('\"
ft"' ...
~ : ~ ;: ~ ': ~: .. r
,.. ft" 1'\ 1'\
f\
~

Fig. 4-22: Scanning electron micrograph of platelet-rich plasma

and endothelial cells on plastic treated with versene to create
retraction. Note platelets are attached only to the retraction
fibrils and not to the body of the endothelial cells. Pseudopod
formation is seen, but no evidence of release is apparent. RF:
retraction fibrils, P: platelets (X 1465). (Reduced 30% for
purposes of reproduction.)
234 CHAPTER 4

See Figure 4-23. Thus. we have not been able to determine

whether the microfilament processes seen on transmission
electron microscopy have an affinity for platelet attachment
or not. Occasionally, with transmission microscopy there
has been the suggestion that attachment has occurred in the
area of microfibril distribution, but we have not been able
to consistently document this finding. We have some evidence
to suggest in early experiments that the ability of the
endothelial cell to avoid triggering platelet release reactions
is not dependent on metabolic integrity. Endothelial cells
fixed with glutaraldehyde did not appear to activate platelets
in studies similar to those described above. This work and
the other tissue culture work which I have described has
been the work of Ms. Arlene Wechezak. who is the tissue
culture supervisor at our laboratory.

Thus the ability to obtain and serially passage true

endothelial cells in tissue culture opens a vast area for
future research in blood vessel disease. We can now utilize
these cells totally separate from their subendothelial structures
for studies of their specific metabolic requirements and
capabilities. There appears to be some potential for the
use of these cells as anti thrombogenic linings for vascular
prosthetic devices in the future. In the calves when we
were using 400 mgs of persantin a day orally and 1.2 grams
of aspirin a day with coumadin we maintained their prothrombin
time at 1.7 and two times their control. This was done for
four days prior to surgery and also after surgery. (18).

DR. STONE: What is the half life of the normal arterial

endothelial cell?

DR. MANSFIELD: I believe that depends on its location

within the arterial tree, most certainly. Turnover seems
greatest at bifurcations. (22,23). Tritiated thymidine
studies have shown very active turnover in bifurcations
whereas, for example, in the descending aorta at the mid
points it is much longer. I can't give you a specific time
in that regard.

DR. WEINBAUM: I have a question as to when you will

get the re-endothelialization. I look at it as a dual
problem. The one which we
Approximation of Normal have dealt with, as you have so
Arrangement of Endothelial beautifully shown today, is the
Cells in Grafts development of the fibrous
attachments underneath. The
other aspect which is what I have been interested in, is the
ability of two endothelial cells to form intercellular
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 235

()

Fig. 4-23: High power scanning micrograph of portion of

Fig. 4-22. Note the attachment of platelets to retraction
fibrils with pseudopod formation, but no evidence of release
reaction. We have been unable to identify microfilaments with
scanning electron microscopy. RF: retraction fibrils. P:
platelets. (X 7540). (Reduced 30% for purposes of reproduction.)
236 CHAPTER 4

channels with a typical spacing of IOO-200A. The force

balance between two adjacent endothelial cell membranes
appears to be a balance between Van der Waals attractive
forces and a repulsive force due to a small surface charge
on the exterior surface of the endothelial cell. The repulsive
force is the same elastic bilayer effect that Dr. Shu Chien
has been studying in rouleaux formation between red cells.
Although I don't know the magnitude of the forces that
determine the equilibrium spacing, I do see a way of getting
an approximate answer to this problem. In epithelial cell
layers where you have active transport with lots of water
movement, electron micrographs show that there is a change
in spacing in the extracellular channel due to this water
movement. One can get some estimate of the pressure build
up at the closed end of the channel, thus some idea of the
magnitude of the forces that develop between two adjacent
endothelial cells across the intercellular channel. I
wonder whether you know or have some idea of the magnitude
of the forces present in your endothelial grafts and if
surface electric effects might be present. Unless the
endothelial cell is able to develop this small negative
surface charge, you can't establish an extracellular channel.

DR. MANSFIELD: Well, we certainly don't have any

answers as to why it goes on but it certainly does. We have
seen areas within an endothelial lined graft where the ~ells
are gone and thrombus has formed. When this occurs, the
endothelial cells from the tissue culture origin will grow
up over the thrombus and cover it. It is interesting that
they do not organize the thrombus underneath. This is in
contrast to the anastomosis where the macrophage comes in
from the wall of the aorta itself organizing and thinning
the thrombus layer. The endothelium itself apparently does
not have that capability but merely covers it over.

DR. BJORKERUD: For how long a time did you study that
phenomenon? How long did you follow the re-endothelialized
thrombus?

DR. MANSFIELD: We have seen that at eight weeks and at

twelve weeks.

DR. COLTON: On the plastic dishes did you in fact get

tight junctions?

DR. MANSFIELD: Yes.

DR. COLTON: Have you seen filamentous material bridging

to endothelial cells?
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 237

DR. MANSFIELD: We have not seen it in bridging. We are

now trying to find various ty'pes of junctions between the
endothelial cells in a tissue culture as compared to what is
seen for example, on a Hautchen preparation. We have had no
meaningful results so far, but it is an area that needs and
deserves a tremendous amount of work and effort. Microfilaments
have been described I believe, at least once before, in
smooth muscle cells near the basement membrane area.

DR. COLTON: Have you tried growing endothelial cells

on a substrate which is microporous?

DR. MANSFIELD: The discs that you saw in the slides do

not have the polyurethane backing and yet the cells grow
very well on them. The problem that you get into is that if
you have something that is too porous the cells can't bridge
the gap and you get an incomplete endothelial covering.

DR. COLTON: That may make a very interesting substrate

for transport studies.

DR. MANSFIELD: Exactly. That is what I was suggesting.

DR. CARO: One question to Dr. Mansfield and one question

to Dr. Fry. Continuing the present line of discussion, does
versine cause the cells to retract?

DR. MANSFIELD: Yes.

DR. CARO: What does it do? Do the junctions between

the cells open?

DR. MANSFIELD: It may tend to do that. The cells pull

away from the substrate leaving fibrils that are attached to
the plastic where the entire cell body was attached before.

DR. CARO: Do you have any explanation of how this

comes about?

DR. MANSFIELD: The versine binds the calcium but we

can't give you any further reason than that. One of the key
techniques in obtaining endothelial cells that survive,
rather than to strip them physically from the lining of a
carotid artery, is to use versine first instead of trypsin
and we used modifications of the two. Trypsin tends to
damage the cells except at very low concentration. Since
versine worked, we used it.
238 CHAPTER 4

DR. SCHWARTZ: Concerning Dr. Mansfield's use of versine,

I suspect that this effect reflects the removal of calcium
and the need for a calcium
Role of Calcium dependent ATPase for cell
adhesion to a surface.

DR. CARD: You are implying that the cell is becoming

spherical and it is the attachment that is weakened by the
versine.

DR. SCHWARTZ: No, I am simply suggesting that the

attachment ultimately is an energy dependent process, and
that a calcium dependent ATPase is involved in the energy
system.

DR. MANSFIELD: I think there is no question that it is

an energy dependent system because by any means that you
kill or damage cells, they do separate from the plastic or
whatever their connection is. The difference is that when
the fibroblast separates it leaves nothing. The endothelial
cell separates and leaves these microfilaments which are
incredibly reactive with platelets, far more so than any
other surface we have studied.

DR. WERTHESSEN: I am recalling the work we did with

Florey and Poole a number of years ago. (24). We used the
DeBakey prosthesis in a baboon's aorta. It provides a blood
clot. It was very disappointing work as far as I was concerned
because what I wanted out of that experiment was the ingrowth
of individual cell populations. With that I could have
disentangled which one synthesized the lipids. But the
conclusion of the study was that all of the cells were
pluripotential. I won't go through all of the data. I
should like to ask Dr. Mansfield whether in the experiments
he described earlier he tried to raise the transplanted
endothelium on a clot or put a clot into his fibers?
DR. MANSFIELD: No, we did not.

DR. WERTHESSEN: I regret that, since you could have settled

an issue. In one of our experiments a capillary grew into
the lumen of the aorta from the outside of the clot and then
spread out. All the cells were in the sheet of tissue over
the prosthesis lumen. This was true also of the ingrowth
tissue from the edge. Your findings surprised me.

DR. MANSFIELD: We actually have seen that now in

humans. We have designed filamentous dacron prostheses which
do accelerate transmural healing. We have now seen almost
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 239

total endothelialization in an actual femoral graft, after

three years in a human, which does not occur otherwise. It
would appear that this was not fall-out seeding, but rather
transprosthetic ingrowth with what turned out to be true
endothelial cells.

DR. WERTHESSEN: Then did you see smooth muscle

underneath?

DR. MANSFIELD: What looks like smooth muscle underneath,

yes.

DR. DEWEY: Could you expand that statement about the

sub-endothelial material?

DR. MANSFIELD: We are talking about a very porous

prostheses which when inserted has to be pre-potted in a very
specific way. We end up with a very high concentration of
heparin, trying to remove the thromboplastin on the surface.
We have found that there is ingrowth with the endothelial
cells lying on top of what appears to be compacted fibrin and
on occasion what appear to be smooth muscle cells, again which
have to have come from either the clot itself or outside. As
compared to the type of graft we are using here which is,
of course, not graft at all, it is totally impervious. This
makes a very major difference. We have done this also in the
baboon, but the baboon will heal a non-porous graft with
endothelium completely in six weeks. So it doesn't help us
a bit.

DR. CARO: I would like to ask Dr. Fry a question,

naemely, why he thinks that fluctuating shearing stresses
affect the permeability of
Regional Variation of the wall differently from steady
Macromolecular Permeability ones. I am not sure of the
significance of his results. He
has shown that wall permeability is initially increased and
then falls with time in his arterio-venous shunt but I am
not convinced that this is a normal situation or how to relate
it to the normal situation. Thus there are normally fluctuating
shearing stresses in arteries and blood flow can normally be
turbulent in the aorta. Nevertheless, it is generally agreed
that there are, in the steady state, regions where arterial
macromolecule permeability is relatively high and others
where it is relatively low.

DR. FRY: I doubt that I can convince you since I don't

have the necessary hard data. Rigorously designed experiments
to evaluate endothelial surface permeability directly as a
function of acute changes in magnitude as opposed to changes
240 CHAPTER 4

in direction of the shear stress are extremely difficult to

do and as yet have been technically unsuccessful. Similar
studies to evaluate these responses in the chronic situation
are even more difficult. The A/V shunt preparation seems at
present to be the best approach to evaluation of the chronic
response to changes in magnitude of the stress. As was
noted, it appears that with chronically elevated exposure,
compensatory processes are occurring in the endothelial cell
surface that end to return the permeability toward normal
values.

DR. CARO: What is the evidence that the wall was

normal, especially during the early stage when permeability
in the arterio-venous shunt was increased?

DR. FRY: The chronically exposed endothelial surface

when viewed by scanning electron microscopy is not "typically
normal" in the sense that endothelial cellular hyperplasia
and cellular elongation in the direction of flow have occurred.
Surprisingly, however the ultra structure by transmission
electron microscopy is completely normal.

DR. STONE: I would like to ask Dr. Fry a question.

You intimated that there is a tremendous difference with
elevated pressures. What about the effect of sustained
elevated flow rates on critical shear stress?

DR. FRY: You mean the tissue response to the increased

shear stresses in the A/V shunt preparations?

DR. STONE: Yes.

DR. FRY: The flow rates are increased many fold in the
arteries supplying the shunt. We have never seen endothelial
erosion in either the carotid or iliac preparations suggesting
that these elevated flows are associated with stresses less
than the acute critical shear stress at these sites. Acutely
one sees an occasional endothelial surface bleb by scanning
electron microscopy and questionable subendothelial edema by
transmission electron microscopy. Both disappear chronically
and except for the aforementioned endothelial cell hyperplasia
and elongation in the direction of flow, the wall appears
essentially normal by ultrastructural criteria.

DR. STONE: Is there intimal thickening in that A/V shunt?

DR. FRY: In normolipemic animals one sees surprisingly

little intimal fibromuscular hyperplasia in response to this
chronically increased stress exposure. In hyperlipemic animals,
however, intimal thickening is much more prominent.
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 241

BIBLIOGRAPHY

1. Diamond, J.M.: The mechanism of isotonic water transport.

J. Gen. Physiol., 48:15, 1964.

2. Stein, 0., Weinstein, D.B., Stein, Y., Steinberg, D.:

Binding, internalization and degradation of low density
lipoprotein by normal human fibroblasts and by fibro-
blasts from a case of homozygous familial hypercholes-
terolemia. Proc. Nat'l Acad. of Sciences, 73:14-18,
1976.

3. Fry, D.L.: Responses of the arterial wall to certain physical

factors. In: Atherogenesis: Initiating Factors.,
pp. 93-125, Assoc. Sci. Pub., Amsterdam, 1972.

4. Flaherty, J.T., Pierce, J.E., Ferrans, V.J., Patel, D.J.,

Tucker, W.K., and Fry, D.L.: Endothelial nuclear
patterns in the canine arterial tree with particular
reference to hemodynamic events. Circ. Res. 30: 23-33,
1972.

5. Fry. D.L.: Certain histological and chemical responses of

the vascular interface to acutely induced mechanical
stress in the aorta of the dog. Circ. Res. 24:93-108,
1969.

6. Carew, T.E.: Mechano-chemical Response of Canine Aortic

Endothelium to Elevated Stress In Vitro. Ph.D. Thesis,
The Catholic University of America, Wash., D.C., 1971

7. Caro, C.G. and Nerem, R.M.: Transport of l4C_4 cholesterol

between serum and wall in the perfused dog common
carotid artery. Circ. Res. 32:187-205, 1973.
-'--....;;...;--'-'-':...;....

8. Fry, D.L.: Hemodynamic forces in atherogenesis. Cerebro.

~ 77-95, 1976.

9. Fry, D.L.: Acute vascular endothelial changes associated

with increased blood velocity gradients. Circ. Res.
22:165-197, 1968.

10. Fry, D.L.: In preparation.

11. Fry, D.L.: Aortic Evans Blue dye accumulation: Its

measurement and interpretation. Am. J. Physiol.:
Heart and Circ. Physiol., Vol. 1, No.2, pp.
H204-H222, 1977.
242 CHAPTE R 4

12. Mansfield, P.B.: Tissue cultured endothelium for

vascular prosthetic devices report to medical
devices applications program. NIH-NHL-7l-2060-l
National Heart, Blood, and Lung Institute of
HEW. Public Health Service, Bethesda, MD., 1971.

13. Mansfield, P.B.: Tissue cultured endothelium for

vascular prosthetic devices report to medical
devices application program. NIH-NHL-7l-2060-l
National Heart, Blood, and Lung Institute of
HEW. Public Health Service, Bethesda, MD., 1972.

14. Mansfield, P.B.: Tissue cultured endothelium for

vascular prosthetic devices report to medical
devices applications program. NIH-NHL-7l-2060-l
National Heart, Blood, and Lung Institute of
HEW. Public Health Service, Bethesda, MD., 1973.

15. Mansfield, P.B.: Tissue cultured endothelium for

vascular prosthetic devices report to medical
devices applications program. NIH-NHL-7l-2060-l
National Heart, Blood, and Lung Institute of
HEW. Public Health Service, Bethesda, MD., 1973-1975.

16. Mansfield, P.B., Wechezak, A.R. and Sauvage, L.R.:

Preventing thrombus on artificial vascular surfaces:
True endothelial cell linings. Trans. Amer. Soc.
Artif. Int. Organs. 21:264-272, 1975.

17. Wechezak, A.R.: and Mansfield, P.B.: Isolated and growth

characteristics of cell lines from bovine venous
endothelium in vitro. 9:39-45, 1973.

18. Mansfield, P.B., Sauvage, L.R. and Smith, J.C.: Factors

influencing thrombus generation in artifical hearts.
Symposium on Coronary Artery Medicine and Surgery
(February, Surgery: Concepts & Controversies. N.Y.)
Appleton-Century-Crofts, pp. 968-985, 1975.

19. Sauvage, L.R., Berger, K., Wood, S., Yates, S., Smith, J.C.,
Mansfield, P.B.: Interspecies healing of porous
arterial prostheses: Observations 1960-1974
Arch. Surg. 109:698-705, 1975.

20. Harker, L.A. and Schlichter, S.J.: Studies of platelet

and fibrinogen kinetics in patients with prosthetic
heart valves. N. Eng. J. Med. 283:1302-1305, 1970.
HYDRODYNAMIC EFFECTS ON ENDOTHELIAL CELLS 243

21. Wechezak, A.R., Mansfield, P.B., and Way, S.A.: Platelet

interation with cultured endothelial cells following
in vitro injury. Artery 1:507-517, 1975.
22. Poole, J.C.F.: Regeneration of aortic tissues in fabric
grafts of the aorta. Symp. Zool. Soc. London
11:131-140, 1964.

23. Wright, H.P.: Endothelial Turnover. Thromb. Diath.

Haemorr. Suppl. 40:79-84, 1970.

24. Poole, J.e., Sanders, A.G. and Florey, H.W.: The

regeneration of aortic endothelium. J. Path. Bact.
75: 133, 1958.
Chapter 5 METABOLIC ACTIVITIES IN THE ARTERIAL WALL

DR. SMITH: The review has been prepared in collaboration

with my husband, Dr. R. H. Smith.

A fundamental requirement for all cells is energy

production. In most cells in the presence of a normal oxygen
supply the energy pathway
Energy Production in is through the metabolism of
Arterial Wall glucose to pyruvate, which goes
into the Krebs cycle and through
the cytochrome system (Fig. 5-1, left side): this is the
pathway of oxidative phosphorylation and produces a large
amount of energy - 32 molecules ATP and 417 kilo calories.
In the absence of molecular oxygen glucose is again converted
to pyruvate but then the c~ll has to regenerate NAD by
reduction of pyruvate to lactate, which cannot be metabolised
further (Fig. 5-1, right side). This anaerobic glycolysis
produces less energy - only 2 molecules of ATP and 32 kilo
calories and lactate accumulates in the tissue (1).

As soon as oxygen is readruitted lactate production

ceases and the system switches over to the aerobic pathway
again, and this is known as
Pasteur Effect the Pasteur Effect (1). It
was first described by Pasteur
in fermentation studies and is the normal pathway of
behavior in, for example, skeletal muscle.

There have been numerous reports that artery wall does

not normally follow the oxidative phosphorylation pathway,
but follows the glycolytic pathway even in the presence of
molecular oxygen, but more recent work suggests that this is
an artifact. The standard biochemical procedure for any
metabolic study is to cool the tissue to (C, which is
supposed to stab1ize everything. In 1970 Scott and co-
workers (2) in Albany, did an experiment in ~hich, instead
of precooling the tissue, they kept it at 37 C and found a
five or six fold increase in oxygen consumption (2,3,4). It
appears that arterial wall is extremely susceptible to what
they called "cold shock," and cooling the tissue to 40 C
seems irreversibly to damage the oxidative phosphorylation
pathway. There have now been several studies in which the
artery was maintained carefully at 3~C and some results are
245
246 CHAPTE R 5

GLUCOSE

~
ATP)
ADP
glucose - 6 - P
ATP) I
ADP ..
2 x glyceraldehyde - 3 - P

+4 x ATP r 4ADP) J C2NAD+~

L 4ATP" ""';
2NADH
2 x pyruvate
--~~- ~
2 x lactate

+ 30 x ATP

30 ADP

Net Gain 32 x A TP + 417 k.cal Net Gain 2 x ATP + 32 k.cal

AEROBIC ANAERO BIC

OXIDAT IVE PHOSPHORYLATION GLYCOL YSIS

Fig. 5-1: Energy produ ction from glucos e.

METABOLIC ACTIVITIES IN ARTERIAL WALL 247

TABLE l. Oxygen and glucose utilization by artery

TISSUE Pre- llmo1/s wet tissue/hr (37°C)

Treatment UPTAKE Lactate
Oz Glucose produced

PIG AORTA

4°C 1.0 5.0

Scott et a1.
Normal
1970. (Ref.Z)
37°C 5.7 11.3

RABBIT AORTA

Endothelium 37°C 8.7 13.6 3.4

intact Morrison, Berwick,
Orci & Wine grad •
Not intact 37°C 5.1 5.8 5-6 1976. (Ref.5)

BOVINE 37°

MESEN- aerobic 6 3
Arnqvist & Lundholm
TERIC. anaerobic 1976. (Ref.6)
14
248 CHAPTER 5

summarized in Table I. It is clear that under these conditions,

the artery is utilizing oxygen rapidly and producing relatively
small amounts of lactic acid.
Furthermore, Morrison, Berwick, Orci, and Winegrad (5) found
that the state of the tissue is extremely important. If
they were able to keep the endothelium apparently structurally
intact, there was a much higher oxygen uptake than if the
endothelium was damaged, although they could see no damage
in the smooth muscle cells.

The Pasteur effect has now been demonstrated in artery

wall; Table II shows that in arteries maintained at 37 0 C
changing from aerobic to anaerobic conditions increases
lactate production four fold. (5,6). In the context of
this meeting it is important to notice that lactate production
was approximately doubled in stretched artery (6).

Thus it appears that earlier findings indicating primary

dependence on glycolysis were probably the result of artifacts.
Under normal conditions of oxygen
Oxygen Tension with supply ATP appears to be generated
Artery Wall by oxidative phosphorylation,
and the balance between glycolysis
and oxidative phosphorylation depends on oxygen supply, as
in other muscles; this raises the question - does the oxygen
tension in arterial wall fall to levels at which the glycolytic
pathway becomes the major energy source? There seem to be
very few studies on the oxygen tension within the wall.
Niinikoski et al., (7) measured the 02 tension in rabbit
aorta by passing a very fine electrode through the wall from
the adventitia inwards in an intact anesthetized animal
(Fig. 5-2, top). The p02 remained constant through the
adventitia then fell as the electrode approached the central
media, then it rose again, and there was a surprisingly
large difference of oxygen tension between the blood and the
immediate subendothelial area. I have tried to relate these
studies on the actual oxygen tension in the wall to Arnqvist
and Lundholm's (6) figures for lactic acid production at
different oxygen tensions (Fig. 5-2, bottom). In the immediate
sub-endothelium the oxygen tension was about 40 rnrnHg, and at
that level there is already significant lactic acid production,
but of course the artery is to some extent damaged. In the
central media the oxygen tension was about 20 rnrnHg, and at
that level there will be quite a large lactate production.
Extrapolation back from the intima (dotted line in Fig. 5-2
top) indicates what might happen in the aorta of a larger
animal. This cuts the point of zero oxygen at a distance of
about 300 to 350 M from the endothelial surface.

DR. STONE: Does this extrapolation ignore any vasa

vasorum of the vessel wall?
METABOLIC ACTIVITIES IN ARTERIAL WALL 249

Adventitia "C
0
0
co
80
Profile of 02 tension across
rabbit aorta.
40 (Niinikoski et al)
p02
mm Hg
,
20 ,,
,,
",
0
I I I I
0 100 200 300
Distance Advanced (!1) from Adventitia

., Lactate production
~ in stretched bovine
.;:; 20
...., mesenteric artery
(Arnqvist & Lundholm)
3:
'"
"0
E
::t 10
.,

I
"C
(.)
:::J
"C
0

....,
IS.
~ 0
(.)

'"
-l
0 20 40 60 80 100
p02 mm Hg

Fig. 5-2: Diagram relating the data of Niinikoski et al.,

(ref. 7) on oxygen tension within arterial wall with the data
of Arnqvist and Lundholm (6) on lactate production in stretched
artery at different oxygen levels.

(Reproduced by kind permission of the authors and the publishers

of Atherosclerosis).
250 CHAPTER 5

DR. SMITH: I assume that the vasa stops at the adventitial-

medial junction the rabbit does not have a vasa vasorum in
the media.

DR. STONE: What I am saying, your extrapolation to a

larger aorta .••

DR. SMITH: Perhaps you would leave that for the moment,
because I will be talking about it later.

A different approach to the oxygen tension was made by

Kirk and Laursen (8). They took sheets both of pure media and
of intima inner media and measured the diffusion coefficient for
oxygen across these sheets. Surprisingly, the mean coefficient
for intima for both layers tended to increase with age. They
then utilized the equation developed by A.V. Hill for skeletal
muscle (9) to calculate the critical diffusion distance for
oxygen - that is, the greatest depth to which oxygen will
penetrate. This is a function of the diffusion coefficient, the
rate of utilization of oxygen by the tissue and the p02.

Greatest thickness to which 02 penetrates = b =~2ky/a where

k diffusion coefficient
a rate of oxygen consumption
y concentration of oxygen

Using the very low figure for oxygen consumption obtained with
cold-shocked arteries they calculated the critical diffusion
distance of about 900 to 1000p
02 Diffusion and which would mean that the wall is
02 Requirements unlikely to become anoxic. However,
using instead the much higher
oxygen consumption figures obtained with intact and non-cold
shocked vessels, the critical diffusion distance is about 350fl
which is close to the figure that we obtained from the extra-
polation in Fig. 5-2. This is clearly an area which is extremely
important and in which more data is urgently required on larger
sized blood vessels. Moss et al., (10) found continuous decrease
in p02 from adventitia to lumen in the femoral arteries of dogs,
but the diameter of the electrode tip was l25}l and the authors
state that the area sensed was large in relation to the vessel
wall thickness.

If we now try to relate this to the adult human aortic intima,

and assume from Fig. 5-2 that significant lactate production starts
at a depth of about 100}l it can be seen in Fig. 5-3, that in
35% of samples of apparently lesion-free intima lactate production
may be increased in the deepest one-third of the tissue, in about
10% in more than half, and in another 10% the deepest layers of
the intima may be totally anoxic.
METABOLIC ACTIVITIES IN ARTERIAL WALL 251

40

30
'"
QI
Q.
E
:i
....0
20
...8.c.,
.
2l
QI
Q.
10

o
< 100 100-199 200-299 >300 ~M

Thickness from lumen to first fragmentary elastic lamina

Fig. 5-3: Thickness of apparently normal intima in 54 subjects

aged 30-69.
252 CHAPTER 5

One might postulate that anoxia is pertinent to fibrous

plaque development in the fourth decade. In developing
lesions there must be areas that are totally anoxic. Thus,
these very crude calculations support the concept that
oxygen tension may be a critical factor in atherogenesis,
modifying both the development and regression of lesions.
(11,12,13,14).

Furthermore, haemodynamic factors must have a profound

influence on diffusion of 02 into the wall; calculations
indicate that in regions of decelerated flow and flow
separation there will be decreased 02 transport (15).

One of the major consequences of oxygen deficiency is,

as we have seen, production of lactic acid. Can a significant
amount of lactic acid accumulate
Accumulation of Lactic in the tissue? Kirk and Laursen
Acid (8) measured the diffusion
constants for lactid acid, and
A.V. Hill (9) gave an equation for calculating the concentration
of a metabolite in tissue. yl = -ab 2 /2k + ab 2 /k + Yo

yl steady state concentralio~ of

the metabolite

a rate of production of the

metabolite

b distance from the surface

k diffusion constant in the tissue

Yo concentration of the metabolite

in the surrounding fluid (plasma)

In anoxic stretched artery lactate was produced at a

rate of 24 p mol/g. wet tissue/hour (Table 2); using this
figure the concentration of lactic acid at 300p from the
lumenal surface would be 33 milligrams per 100 grams of wet
tissue and at 400p which would correspond with an early
gelantinous thickening, it would be nearly doubled to 51 mg.

So it looks as if significant accumulation of lactic

acid could occur, and this seems to be an area where actual
experimental information is urgently needed.

At 800}l the center of a developing fibrous plaque, it

would increase more than three fold to 174 mg/100 g wet
tissue.
METABOLIC ACTIVITIES IN ARTERIAL WALL 253

TABLE 2. Magnitude of the Pasteur effect

02 content Lactate production

~mol/g. wet tissue/hr.
7. of gas
mixture

Resting aorta 1. 1·
957. 3.6 3.3
07. 13.8 13.5
1. Morrison, Berwick et al.
1972. (Ref. 5)

2. Arnqvist & Lundholm

1976. (Ref.6)

Stretched aorta

157. 4.5
107. 7.0
57. 11.5
07. 24.0

15% 02 =100mg Hg (p02 of arterial blood}

 254 CHAPTER 5

One consequence of increased lactic acid in the tissue

is a reduction of pH, and the cell lysosomes seem to be
particularly susceptible to
Effect of Lactic Acid low pH, which increases their
Accumulation on Lysosomes fragility or leakiness. The
lysosornes contain numerous
hydrolases (16), including cathepsins, and B-glucuronidase
and hyaluronidase which degrade the glycosaminoglycans. A
cathepsin Bl with collagenolytic activity has also been
reported (17). In heart muscle there seems to be a population
of lysosomes with varying enzyme content and slightly varying
density. They may become filled with material such as lipid,
detergent or silica which alters their density (16,18).

I could not find any studies on the effects of hypoxia

on lysosomes in arteries, but there is a large amount of
experimental information on
DOG Hypoxia infarcted heart muscle. In dog
hearts Brachfeld (19) found an
increasing lactate concentration within 6 seconds of coronary
artery occlusion, and in 5 minutes it had increased l8-fold;
pH fell by 0.4 units within 2 minutes. Large intracellular/
extracellular pH gradients have been demonstrated (20,21),
intracellular hydrogen ion concentration reaching twice the
extracellular level whereas intracellular bicarbonate was
only half the extracellular level. The effects of coronary
occlusion on the distribution of lysosomal enzymes in heart
muscle have been studied by several groups (22,23,24,25) and
typical results are shown in Tables 3 and 4. Four hours
after coronary occlusion there was significant release of
acid phosphatase into the supernatant fraction, so there
appears to be a significant increase in lysosomal leakiness.
Of course, there are other changes in the cell following
myocardial infarction, including a decrease in glycogen
(Table 4) but work on isolated lysosomes suggests the pH is
one of the most important factors in maintaining lysosomal
integrity. Thus, if heart muscle is a valid model for
arterial smooth muscle, hypoxia leads to the accumulation of
lactic acid, producing a fall in pH which mediates leakage
of lysosomal enzymes, which cause general intra- and extracellular
damage to arterial tissue.

How do early human lesions seem to tie in with this

concept of anoxia as an initiating factor? The early human
lesions fall into two groups:
Early Lesions and Their fatty streaks consisting of cells
Relation to Low Oxygen laden with lipid of which the
Tension largest single component is
cholesterol ester; and
METABOLIC ACTIVITIES IN ARTERIAL WALL 255

TABLE 3. Activity of lysosomal enzymes in dog left ventricular

muscle

Acid P-ase ~ -Glucuronidase

SHAM-OPERATED
Supernatant fraction 20.1 1.2
Particulate fraction 4.7 0.9
Particulate 1 supernatant 0.23 0.75

4 HOURS POST CORONARY OCCLUSION

Supernatant fraction 18.1 1.8
Particulate fraction 2.4* 0.2*
Particulate 1 supernatant 0.13 0.13

* p<0.001

TABLE 4. Relation between tissue glycogen and pH and lysosomal

stability in infarcted heart muscle from dogs

Time after Glycogen: pH Acid P·ase Activity

Coronary Occlusion Il 9 1100 mg wet tissue Ratio: Particulate I
Supernatant

Control 60.4 7.0 0.26

I hour 32.9 6.5 0.15

2 hours 17.0 6.4 0.10

4 hours 14.6 6.3 0.06

256 CHAPTER 5

proliferative lesions which are mounds of thickening of smooth

muscle cells and collagen fibers. These include a wide range
of morphological variants, and in the early stages may contain
very little lipid.

Looking first at fatty streaks, does it seem reasonable

to postulate that they could be induced by reduced oxygen
tension? In our last fifty aortas
Fatty Streaks I measured the distance from the
lumenal surface to the middle of
the main band of fat filled cells in the fatty streaks. Looking
at them macroscopically they appear to be superficial lesions
and this was amply confirmed by microscopic measurement. In
all but one lesion the most superficial fat filled cells lay
immediately under the endothelium; in 40% all the main band
of fat filled cells within 50p of the intimal surface (Fig. 5-4)
- in fact, in half of these lesions all fat filled cells were
within 25fl of the intimal surface. In a further 47% of lesions
the middle of the main band extended to 50JU. There was just a
small group in which the cells extended deeper, but in all of
these the fat-filled cells also extended all the way to the
surface. In a study on early lesions in cholesterol-fed Rhesus
monkeys, Stary also found that the fat-filled cells were
extremely superficial (26). Increased oxygen consumption in
arteries made atherosclerotic by cholesterol feeding has
been reported (reviewed in references 27,28,29). This is
probably the result of increased fatty acid oxidation and not
increased glucose utilization by the Krebs cycle. There seems
to be no information on the spontaneous human fatty streaks,
and results of metabolic studies in post mortem material are
unlikely to be valid.

Thus, it seems difficult to postulate that oxygen shortage

could be a factor in the development of fatty streaks, but it
might be a factor in the destruction of the fat-filled cells.
In 18 raised fatty/gelatinous or fatty/fibrous plaques contain-
ing numerous scattered fat-filled cells in the cap, no intact
fat-filled cells occurred at a depth greater than about 400 fl.

However, I think the situation looks different in the

gelatinous and fibrous lesions. Fig. 5-5a shows the
gelatinous tail of a developing
Gelatinous and Fibrous plaque stained with the Verhoeff
Lesions and Van Gieson stain for elastin
and collagen, and it demonstrates
several characteristic features.
METABOLIC ACTIVITIES IN ARTERIAL WALL 257

en
CI)
a. 30

-
E
It!
en
0
....c
CI)
20
...u
8?

10

o
< 25 25-50 >50 pM

Distance from lumen to middle of

main band of fat-filled cells

Fig. 5-4: The depth of the centre of the main band of

fat-filled cells in fatty streaks from 50 aortas.
258 CHAPTER 5

There is massive proliferation of smooth muscle cells and

thick collagen bundles rather widely spaced. The lesions
contain large amounts of intact LDL, fibrinogen and other
plasma macromolecules, and in the deep layers what Daria Haust
calls an "area of insudation" can be seen (30,31,32).
Three extreme forms of proliferative lesion are encountered:
closely packed smooth muscle cells with a few thin collagen
strands between, close collagen bundles with few cells
between and the lesions with "loose" collagen bundles shown
in Fig. 5-6. The quantity of plasma macromolecules is roughly
proportional to the looseness of the structure. These
lesions become prominent in the fourth and fifth decades; by
this age there is well developed diffuse intimal thickening
and the deep layers of the intima may be hypoxic.

DR: COLTON: Dr. Smith, before going on give me a typical

dimension or what was there before the lesion developed?
This could tell me what has grown.

DR. SMITH: This lesion was about 1000 in its

maximum thickness. It is the tail end of a fibrous plaque
which was about an inch long. In this age group normal
intima has an average thickness of about 200 fl ..

DR. COLTON: Are there cells?

DR. SMITH: Yes, there are. There are variable numbers

of smooth muscle cells lying along the collagen bundles.

The central area of these thick lesions must be hypoxic

and at least two features are compatible with response to
hypoxia. First, lysosomal damage with release of hydrolases
into the tissue matrix would be compatible with separation
of the connective tissue elements, and with slight decrease
in glycosaminoglycans, thereby allowing plasma components to
accumulate (33,34).

DR. COLTON: Is there an identifiable necrosis?

DR. SMITH: I don't know what you mean by 'identifiable

necrosis.' In this higher powered section of the insudation
area Fig. 5-5B the collagen fibers seems to thin out and
disappear. One can interpret this in two ways: Daria Haust
suggests that they are "dissolving" in the insudate, but one
could also postulate that they are growing out of the insudate.
So, I just don't know, one can only speculate (30).

DR. STONE: Dr. Smith, a question. Could this still

be collagen material, young collagen, down here that is not
staining?
METABOLIC ACTIVITIES IN ARTERIAL WALL 259

Fig. 5-5: Gelatinous "tail" of a large fibro/gelatinous plaque

from a male aged 58 (sub-arachnoid haemorrhage; aorta obtained
4 hours after death).

A. Low power view of the whole "tail"; two surface layers,

each about 10 uM thick, were peeled off for analysis and their
strip planes are indicated by large arrows. The lesion shows
typical thick collagen bundles with wide spaces between, and
an "area of insudation" in the base. Collagen bundles showed
pale, diffuse sudanophilia, and the cholesterol content was
2.lmg/100mg lipid extracted dry tissue compared with 3.4mg
in adjacent normal intima.
260 CHAPTER 5

I IOO"JI I

Fig. 5-5: Gelatinous "tail" of a large fibro/gelantinous plaque

from a male aged 58 (sub-arachnoid haemorrhage; aorta obtained
4 hours after death).

B. Edge of the insudation area at higher magnification.

METABOLIC ACTIVITIES IN ARTERIAL WALL 261

THE CONCENTRATIONS OF 'SOLUBLE' AND 'INSOLUBLE' PLASMA DERIVATIVES

(mg 1100mg dry tissuel

Fatty Streak Adult Normal Gelatinous WhitePla~

'Tail' Edge Centre
Cap Lower

Lipoprotein 1.2- 5.2 13.4- 6.0 2,9 3.7

Fibrinogen 2.5 2.2 6.6- 6.S- 1.4 4.0
'Fibrin' 2.4 2.0 3.5 9.7- 2.1 10.3-
Residual
Cholesterol 25.6- 3.S 4.2 27.2- 11.2 97.S-
Immobilized
Lipoprotein 1.0 O.S 1.6 1.0 5.7

( « Significantly different from normal I

Fig. 5-6: Data from 4 fatty streaks, 12 samples of normal

intima, 23 gelatinous "tails" of plaques, 9 plaque edges and
9 central caps, and 37 samples of atheroma lipid from the
lower layers of plaques.
 262 CHAPTER 5

DR. SMITH: I don't know. I can't answer that.

DR. STONE: But they would give different reactions

with this particular stain. Young and old collagen, as I
have heard described.

DR. SMITH: I find very variable staining with the Van

Gieson stain. Some collagen bundles stain rather dimly,
while adjacent strands are bright pink, and I have no idea
what the differences are. Yes, I am quite certain that
different collagens in the lesion stain differently but
whether they are young collagen or old collagen, or what the
differences are, I have no idea. The large amount of collagen
is the second feature of these lesions that is compatible
with hypoxia and lactate production. There are now at
least four studies showing that lactic acid specifically
increases collagen, and proco11agen proline hydroxylase
activity in cell cultures (35,36, 37,38).

It appears to be a specific effect of lactic acid and not of

low pH or of oxygen deficiency. In vivo studies on granulomas
and wound healing are, however, contradictory (39,40) and
this seems to be an area that requires further investigation
in terms of smooth muscle cells and the artery wall. Both
these in vivo studies indicate that cell protein synthesis
HAMSTER is decreased by low oxygen tension and with CHL-F Chin~se
hamster cells in culture it decreased growth rate (41).
Thus, it seems un1ke1y that hypoxia can be implicated in
smooth muscle cell proliferation.

In addition to the risk of hypoxia, the smooth muscle

cells are subjected to mechanical stress, which may be a
determining factor in
Influence of Mechanical their metabolic pattern. I
Stress on Proliferation showed a slide from Leung et a1.,
and Synthesis (42) demonstrating the differential
development of collagen and elastin
in the developing pulmonary and aortic truns (Fig. 2-16). They
have also done some elegant culture experiments in which
smooth muscle cells were grown on elastin membranes; then
some of the membranes were subjected to cyclical stretching;
whereas others were agitated in the medium without stretching.
Compared with agitation, stretching increased total protein
synthesis by 80% and collagen synthesis by 200%, but there was
no increase in DNA synthesis (43).
METABOLIC ACTIVITIES IN ARTERIAL WALL 263

In these experiments Types I and III collagen synthesis were SHEEP

stimulated to the same degree but in a study on chondrocytes,
those grown in spinner cultures produced a different type of
collagen from monolayers (44). In short-term organ cultures
of pregnant sheep myometrium, R.H. Smith and R. Palmer
studied 3H-proline incorporation and conversion to 3H-
hydroxyproline. In gravid (stretched) areas the ratio
hydro/pro was higher than in unstretched areas, indicating
that a higher proportion of the protein synthesized was
collagen (unpublished observations). So there seems to be a
growing body of information suggesting that mechanical
stresses have a considerable influence on collagen synthesis,
and this may be a self-magnifying system. Stretching increases
collagen synthesis, but it also increases lactic acid production,
and the lactic acid again stimulates collagen synthesis.

A different approach to stress factors has been made

through studies on hypertension. In both spontaneous and
experimental hypertension Wolinsky reports proliferation of
SMCs, and a roughly parallel increase in the amount of
collagen and elastin (45,46). However, in experimental
hypertension produced by renal ischaemia it is possible that
the observed response is to the reninangiotension system and
not to the mechanical factors. In experimental coarctation
of the aorta in dogs Hollander et al., found significant DOG
increases in sulphated MPS and 35S0 4 incorporation in the
hypertension segments above the coarctation, significant
decreases below it, and no change in lipid content (47).
Fernandez and Crane (48) examined the very early response of
small arteries rate to constriction of the aorta between the
renal arteries. There was no response in vessels distal to
the constriction, and there was a positive response in
animals subjected to nephrectomy, suggesting that the renin-
angiotension system was not a primary factor.

By 24 hours, small arteries proximal to the constriction

showed "plasmatic vasculosis," with progressive fragmentation
of elastin, SMC necrosis and deposition of fibrin, but there
was no increase in 3H-thymidine labelling. By 4 days, there
was "florid necrotising arteritis" and a marked increase in
3H-thymidine labelling which continued to increase up to 8
days, (the longest time of observation). Unfortunately the
authors do not report on the large vessels, and made no
comment on them.
264 CHAPTER 5

Nevertheless, this experiment raises the question of whether

the observed response in acute experimental hypertension is
truely related to the mechanical factors, or if it is in
reality a secondary response to increased insudation of
plasma macromolecules, in particular, the increased conversion
of fibrinogen to fibrin within the wall. This seems to be an
area requiring further elucidation.

Plasma macromolecules seem to enter the intima in

relatively large amounts. In normal human intima from the
fourth decade upwards in
Plasma Macromolecules normotensive subjects the
in Intima concentration, on a crude
volumetric basis, of low density
lipoprotein (LDL) is approximately the same in the intima and
in the plasma, and the concentration of intimal fibrinogen
is 1/2 - 1/3 the plasma concentration (59). In the hypertensive
individual, intimal LDL may be doubled. The steady state
concentration of plasma macromolecules in normal intima
seems to be a function of their concentration in the plasma
and their molecular weight (50). In a number of samples of
normal intima we examined the relative concentrations of LDL
(molecular weight 2 x 10 6 ), d-macroglobulin (molecular
weight 820,000), HDL (molecular weight 200,000) and albumin
(molecular weight 68,000). They were quantified in terms of
the volumes of the patient's own plasma from which they were
derived, to correct for the different concentrations in
different subjects.

Taking LDL as a hundred, the retention of the other

proteins was calculated as percentage of LDL retention.
This gave a linear relation between retention and molecular
weight; thus, relative to the plasma concentration, LDL was
retained to the greatest extent while there was least retention
of albumin (50). This strongly suggests that the accumulation
depends on retardation by molecular sieving rather than on
specific binding of LDL. In the gelatinous lesions there is
also a linear relationship and the slope is steeper; whether
that simply depends on the thicknesss, I don't know. The
actual concentrations in normal intima are shown in Table 5.
where the concentrations in plasma and intima are compared.
The figures are for a cube of tissue, so the actual concentration
in the extracellular space must be very high. Albumin is
still the largest component in intima but relative to its
concentration in plasma it is retained to a very small
extent. There is a large amount of LDL and fibrinogen
(Table 5); fibrinogen is behaving as if it had a molecular
weight of about a million.
METABOLIC ACTIVITIES IN ARTERIAL WALL 265

TABLE 5. Concentration of plasma proteins in normal intima

(Volumetric Basis).

INTIMA PLASMA

Mg./100cc Mg. /lOOml.

Albumin 725 4,500

LDL 500 500

Fibrinogen 200 325

a 2- macroglobulin 130 285

HDL 90 300
266 CHAPTER 5

The protease inhibitor, d2-macroglobulin is also present in

quite high concentration. thus there is a very interesting
mixture in the intima. and this must influence the cells and
modify the extracellular matrix to a considerable extent.
In the gelatinous lesions the concentrations of LDL and
fibrinogen may increase up to three-fold compared with
adjacent normal intima. so that the concentration of LDL is
substantially greater than the concentration in plasma (49).
Thus. these SMCs are in a localized hyperlipoproteinaemic
environment even in a subject with normal lipoprotein levels.

DR. CAREW: Would you. just for reference. talk about

the concentration •..•

DR. SMITH: The figure in Table 5 is milligrams total

LDL. Intimal LDL is measured against plasma LDL standards.
In normal intima LDL cholesterol is about 30% of the total
cholesterol.

DR. COLTON: Okay. So it is protein but in fact what

is in the wall could be aproprotein and low cholesterol.

DR. SMITH: No, it is not, because we electrophorese

the mobile lipoprotein out of the tissue and stain the
immunopeak with a lipid stain; so we are not staining aproprotein
which is not carrying lipid.

DR. CAREW: Is it conceivable or certain that this is

all intact LDL or fragments or what?

DR. SMITH: On two dimensional electrophoresis it

migrates close to the patient's LDL. We did some recovery
experiments in which we cut the immuno-peaks out of the
agarose gel and measured the ratio of cholesterol to peak
area, and compared this for the patient's plasma LDL and the
intimal LDL and they came out approximately the same. Now I
say this rather guardedly because when we looked at the
cholesterol ester fatty acids they were very odd. but then
so were the cholesterol ester fatty acids that we recovered
from the immunopeaks from the patient's plasma. But it is
my impression that this is intact LDL.

I am not going to discuss the effect of cholesterol

feeding on atherogenesis in intact animals. There is a vast
literature on this type of
Effects of Plasma Fractions experimental atherosclerosis, and
on Aortic Smooth Muscle I do not know whether the
Cells in Culture atherogenic agent is in fact
cholesterol.
METABOLIC ACTIVITIES IN ARTERIAL WALL 267

TABLE 6. Effect of plasma fractions on cell cultures

Growth Mitosis Lipid Uptake

Dzoga e t al. (1974) -
Primary exp1ants.
(51)
5% hyper1ipaemic monkeX serum ++ ++ ++
LDL ++ ++ ++
VLDL 0 0 0

HDL 0 0 0

Low-lipid infranatant 0 0 0

Ross (1975) - Sub-cultures grown to zero growth in 1%

(52) normal monkey serum.

Normal monkey serum. Growth

10/ serum
0 +++
5% reconstituted serum ++
Low-lipid infranatant +

LDL + ++
HDL + +
 268 CHAPTER 5

Maybe Dr. Werthessen will talk about this later. There

have been a number of studies on the effect of lipoproteins
on smooth muscle cells in culture and some of the results
reported by Dzoga (52) and by Ross (53) are summarized in
Table 6. In primary explants, Dzoga (52) found that compared
with normal monkey serum, hyperlipaemic monkey serum stimulated
growth, mitosis and lipid uptake. In her system, the isolated
LDL seemed to be responsible and there was no stimulation
relative to normal monkey serum by VLDL, HDL or the lipid
poor protein infranatant. Ross (53) seems to get slightly
different results; he works with subcultures, and I don't
know how much difference this makes. They are grown to zero
MONKEY growth in one percent normal monkey serum, the supplements
of monkey serum are added. Serum and reconstituted serum
stimulated growth but he did not get full stimulation without
both LDL and the low lipid infranatent. He suggests that
this contains a low molecular weight protein derived from
platelets which is in fact the stimulating factor. It is,
interesting that Ronnemaa, Juva and Kulonin (54) also found
stimulation of collagen synthesis by the low lipid infranatant
and not by isolated LDL.

How do these cultures relate to the artery wall? I

think this is one of the major questions that we will have
to answer soon. Table 7 shows the concentration of LDL
cholesterol in intima and in the culture medium, and there
is an order of magnitude difference. So either the cultured
cells are very much more sensitive to plasma component or in
the intact intima the cell has some sort of micro-environment
around it which insulates it from the high concentrations of
LP in the whole tissue, or the rate of change of LP level
may be a critical factor.

DR. COLTON: Are you suggesting that people are making

experiments in culture media that is too low in LDL cholesterol?

DR. SMITH: As I understand it, at the concentrations

found in intima the cultures die. I am not suggesting
anything; I am just asking the question, what is the relation
between the cell in culture and the cell in intima?

DR. CAREW: May I comment. We put in a low density

lipoprotein at concentrations upwards of serum concentration
and they don't die in short term. This is a very interesting
problem, though, because we measured degradation rates in
culture of low density lipoproteins which are sufficient to
account for all of the LDL degradation in vivo and serum
concentration or medium concentrations of LDL as low as 1%
plasma.
METABOLIC ACTIVITIES IN ARTERIAL WALL 269

TABLE 7. Concentration of LDL cholesterol in intima and culture

medium

Normal intima 190 mg/100 cc wet tissue

Gelatinous
thickenings 380 " " "

Culture Medium (Dzoga et al., 1974, 52)

Normal monkey LDL 4mg/100ml.

Hyperlipaemic LDL 30mg/100ml.

 270 CHAPTER 5

That, of course, raises the question of what is the relationship

between the cultured smooth muscle cells and the cells in
the artery. I think this is a very important point. I
would just like to ask another question which refers back to
your previous slide indicating the concentrations of LDL,
HDL in the intima and I think these are the data determined
from saline extracts and then electrophoresed saline extracts ••.

DR. SMITH: No, this is from direct electrophoresis

actually from the tissue.

DR. CAREW: Well, what is the possibility that not all

the LDL comes out of here so that the actual concentration
in this tissue is perhaps even considerably larger than this
because of that amount which would be collagen or elastin or
smooth muscle cell membranes which is not released by the
electrophoresis?

DR. SMITH: In normal intima, most of the LP comes out,

but in certain lesions a lot more can be released by incubation
with proteolytic enzymes, and I will be talking about that
shortly.

DR. COLTON: In the normal case could there be more LDL

that has not come out? Could it be bound and not liberated
by electrophoresis?

DR. SMITH: We get very little more out of normal

intima after incubation with chondroitinase or proteolytic
enzymes, so in normal intima I don't think there is much
more to come out but, as I say, in some of the lesions there
is a lot more to come out. I will talk about that later.

DR. SCHWARTZ: Have you added HDL to these tissues to

see if in fact you might use that to unblock the LDL binding?

DR. SMITH: No, I haven't.

Studies in vivo on the effects of hyperlipidaemia on

collagen synthesis have produced contradictory results
(Table 8). Fuller et al., (54)
PIG The Effect of Hyperlipidaemia found that in weanling mini-pigs
on Collagen Synthesis cholesterol feeding stimulated
protocollagen proline hydroxylase
activity equally in atherosclerotic lesions and in adjacent
lesion-free segments of aorta. In contrast, St. Clair et al., in
(55) young pigeons found no difference between aortas from
PIGEON controls and cholesterol-fed birds, and no significant
difference between lesion-free segments and lesions.
 s::
m
-I
»
III
TABLE 8. Effect of hyperlipaemia on collagen synthesis in vivo o
r
Lesion Lesion-free c=i
Authors. Species »
("')
Proline hydroxylase activity -I
<:
Fu ller e t. al. Cholesterol fed 620* 647<>. =l
~ m
1972 (54) en
Control 294~' 1958
z
»:::tJ
(* 8 Same area, but no lesion)
-I
m
:::tJ
St.Clair et al. Pigeons Cholesterol fed
1975 (55) »
r
Young Fatty streak 5128 4831
Plaque 5233 ~r
r
Control 4417

Old Spontaneous plaques 3976 4767

Collagen synthesis.
McCullagh & ~ Cholesterol fed
Ehrhart
1974. (56) Abdominal aorta 977 187

Femoral artery 1954 186

Control
Abdominal aorta 176

femoral artery 97

~
'"......
 272 CHAPTER 5

These preparations included adventitia as well as media, and

as the intima is a very small proportion of the total tissue,
changes occurring there might be obscured; furthermore, the
lesions contained very little collagen. McCullagh and
DOG Erhart (56) measured collagen synthesis in dogs fed semisynthetic
diets supplemented with cholesterol and found up to 10-fold
stimulation of collagen synthesis in the very fibrous lesions,
but no changes in lesion-free segments. Clearly this is an
area that needs further clarification.

In addition to acting directly on SMCs, plasma macromolecules

may react within the intima with each other, or with the
connective tissue components, and
Interactions between thereby change the extracellular
Plasma and Intimal environment, and thus influence
Macromolecules cellular metabolism. We have
investigated the relation between
soluble fibrinogen and insoluble fibrin, or fibrin-like material
("fibrin") in intima (49). The distribution of "fibrin" is
highly correlated with morphology (Fig. 5-6). In normal
intima there are approximately equal concentrations of
fibrinogen and fibrin, whereas in gelatinous lesions and the
gelatinous peripheries of plaques, there is a very high
concentration of fibrinogen with only a small increase in
"fibrin." At the edge of developed plaques, there is a
consistent and abrupt increase in the concentration of
insoluble "fibrin," suggesting an increase in thrombin
activity in these localized areas. High levels of "fibrin"
are also found in the atheroma lipid pool in the middle of
developed fibrous plaques. I believe that much of this
"fibrin" is derived from fibrinogen which is converted to
"fibrin" within the intima. Fibrinolysis also seems to
occur in some lesions, but it is not clear if this is the
result of plasmin or lysosmal protease activity. In either
case, the activity might be modified by the protease inhibitors,
2-macroglobulin and I-antitrypsin, which are present in
substantial concentrations in intima.

In answer to the question: Does all the lipoprotein

come out of intima on electrophoresis? Recently we have
found an immobilized lipoprotein fraction that can be released
by incubating the residual tissue, after electrophresis,
with plasmin or crude collagenase (57). The amounts released
after a single incubation with plasmin are shown in Fig. 5-6
(bottom line); in normal intima and low lipid gelatinous
lesions the lipoprotein released is only 10-15% of the
mobile lipoprotein fraction, and is not increased by repeating
the incubation.
METABOLIC ACTIVITIES IN ARTERIAL WALL 273

It is present in highest concentrations in areas of maximum

extracellular lipid accumulation; in these areas more lipoprotein
is released on repeating the incubation with plasmin. and
its total concentration may be two or three times greater
than the concentration of mobile LP. This suggests that
immobilization of the LP may be a first step in the irreversible
disposition of lipid from plasma LDL. Incubation with
chrondroitinase ABC releases only small amounts of lipoprotein.
and maximum release is obtained with plasmin; the simplest
explanation of these findings is that lipoprotein is in some
way bound to fibrin (57.58). This idea receives some support
from the work of Camejo et al. (59). who have found a lipoprotein
protein complex in intima-medial extracts

Lipoproteins form ionic complexes with glycosaminoglycans

(GAG) in vitro and it has frequently been suggested that
this also occurs within the intima (60,61,62); our data on
the relation between molecular weight and relative retention
(50)does not really support the concept that this is the
mechanism of LP accumulation in intima. Fibrinogen can also
form complexes with GAG (63) and GAG - fibrinogen - lipoprotein
complexes have been described. There is no definite proof
that any of these complexes actually exist in the wall.

Thus. there are high concentrations of LDL and other

plasma macromolecules in normal intima, and in the gelatinous
precursors of fibrous plaques
Stimulation of Smooth the concentrations are subtantially
Muscle Cell Proliferation higher than in plasma. It seems
in Intima reasonable to postulate that, as
in cultures, (52) this will
stimulate the smooth muscle cells to proliferate and accumulate
lipid. Unfortunately, there does not seem to be any constant
relationship either between intimal lipoprotein concentration
or proliferation of intracellular lipid accumulation. This
is clearly demonstrated in Table 9 which compares the concentration
of lipoprotein in normal intima and adjacent gelatinous
thickenings in three patients with different levels of serum
cholesterol and blood pressure. In all patients the lesion
contained twice as much lipoprotein as adjacent normal
intima, but there was more lipoprotein in the normal intima
of Patient 2 than in the lesion of Patient 1. Patient 3 had
a high serum cholesterol and hypertension. and his normal
intima contained more lipoprotein than the lesion of Patient
2. and twice as much as the lesion in Patient 1. None of
the lesions contained a significant number of cells with
intracellular fat droplets.
274 CHAPTER 5

TABLE 9. "Lipoprotein-bound" cholesterol in normal intima

and early fibrous lesions

Serum Intimal lipoprotein

Subject cholesterol Blood Cholesterol
sex & age mg/lOOml Pressure mg/100mg dry tissue

1) F.32y 130 130 Normal Lesion

80 1.3 3.2

2) M.6ly 267 130 3.8 6.2

80

3) M.49y 332 260 220 6.4 12.5

130 120
METABOLIC ACTIVITIES IN ARTERIAL WALL 275

Thus, a high concentration of intimal lipoprotein does not

in itself constitute a lesion: possibly it provides a
stimulating environment for proliferation of smooth muscle
cells that are already altered in some way, or possibly the
high level found in gelatinous lesions is secondary to pre-
existing proliferation, and not an initiating factor.

Large, raised lesions with massive proliferation of

smooth muscle cells and collagen but not significant increase
in cholesterol are frequently found in human arteries,
(64.65,32); thus cholesterol accumulation is not an obligatory
factor in lesion development. Recently, Benditt and Benditt
(64) have made a very important contribution to the problem
of smooth muscle cell proliferation by suggesting that the
fibrous plaque is comparable to a benign tumor, and the
cells are transformed. As a result of X-chromosome inactivation
in some heterozygous Negro women, the tissue consists of a
mosaic of cells in which the X-chromosome is derived from
either the father or the mother. Using the A or B iso-
enzymes of glucose - 6 - phosphate dehydrogenase (G-6PDH) as
cellular markers, Benditt showed that normal intima contained
both isoenzymes but fibrous plaques contained only one iso-
enzyme (either A or B), suggesting that they are monoclonal
in origin. These findings have been confirmed and extended
by Pearson et al. (66) and the combined results from the two
groups are summarized in Table 10.

In normal intima 97% of the samples contained both A

and B isoenzymes. In the fibrous plaques 85% of the samples
examined were monoclonal and they were equally divided
between A and B isoenzymes; most fatty streaks have an AB
configuration and are not monoclonal lesions. It is difficult
to integrate this idea into our present concepts of atherosclerosis.
But it is an idea that we cannot dismiss and that we must
try to reconcile with existing information.

Benditt has commented that the risk of transformation

by agents such as viral or chemical mutagens is increased
where there is increased rate of mitosis. In a gelatinous
lesion there may well be an increased rate of mitosis; and
it also contains an enormous amount of low density lipoprotein
which might carry a mutagenic steroid. Thus one can postulate
that cells are undergoing increased mitosis in a micro-
environment which is greatly enriched in LDL carrying the
mutagen. I will leave it to Dr. Werthessen to tell us what
sort of mutagenic steroid might be involved; Lee, Werthessen
and co-workers (67) have isolated an extremely atherogenic
fraction from cholesterol, and increased mitotic activity is
an early response to cholesterol feeding (67,68,69).
 276 CHAPTER 5

TABLE 10. Distribution of isoenzymes of glucose-6-phosphate

dehydrogenase in samples of aortic intima from heterozygous women

ISOENZYME DISTRIBUTION
Number of sam121es Percentas;e
No. of
l2atients AB A B AB A or B
Lesion-free intima 20 156 3 97.5 2.5

Fibrous plaques 12 9 25 25 15.3 84.7

Fatty streaks 11 23 2 3 83.2 17.8

Combined data of Benditt & Benditt (1973) (64), and Pearson et al. (1975) (66).
METABOLIC ACTIVITIES IN ARTERIAL WALL 277

In the centers of fibrous plaques that have accumulated

a pool of extracellular atheroma lipid it appears there is
either irreversible precipitation
Effect of Cellular Enzymes of LDL or the apoprotein has been
Components of the degraded or split off the lipid.
Extracellular Environment There is also loss of cells,
decrease in the concentration of
GAG, and loss of the architecture of the collagen bundles,
although it is not clear if there is a quantitative decrease.

In preliminary studies we have found a rapid loss of

intact LDL from the amorphous lipid centers of plaques that
have been minced, and then incubated; the level falling to
about a quarter of the control level in 3-4 hours incubation
(70). Increased activities of lysosomal enzymes, particularly
-glucuronidase, have been reported in atherosclerotic aortas
(71). Increased cathepsin D has been reported by several
authors (reviewed by Zemplenyi (72) and there seems to be
little doubt that lysosomal enzymes, released as a result of
increased lactic acid production, or of cell death in the
hypoxic centers of larger plaques, play an important role in
the later stages of plaque development.

THE ACCUMULATION OF LIPID IN THE FAT-FILLED CELLS OF FATTY STREAKS

One reason that I have hardly mentioned lipid metabolism

is that it has been covered in several recent and expert
reviews (27,28,29). There are
Irrelevance of Induced also other reasons: first, and
Hypercholesterolaemia most important, I do not now
believe that the primary factor
in the genesis of occlusive atherosclerotic lesions in humans is
disordered lipid metabolism of the artery wall. Second,
most of the studies of arterial lipid metabolism have been
made on cholesterol-fed animals and I question if the response
to this acute cholesterol stress is really relevant to the
average human subject. It may be relevant to subjects with
familiar Type II hypercholesterolaemia, but according to
Oliver's calculations (73) in Britain these patients account
for only one in every 400 coronary deaths. An example of
this acute cholesterol stress is given by Day and Proudlock
(74) who found increased cholesterol esterifying activity in
aortas of rabbits fed cholesterol for only three days. RABBIT
During the three days, serum cholesterol rose from 78 -
4l5mg/100 ml. It is difficult to imagine any situation in
which a patient's cholesterol could increase more than 5-
fold in three days.

The lesions that do appear to be primary disorders of

arterial SMC lipid metabolism are fatty streaks.
278 CHAPTER 5

Fatty streaks in children and adolescents consist of clusters

of cells filled with lipid, of which cholesterol ester is the
major component. All the evidence
Intracellular Lipid suggests that most of the cholesterol
Accumulation in Fatty comes from plasma lipoprotein, but
Streaks the ester that accumulates is not
like plasma cholesterol ester, sa
either the cells take up only free cholesterol and esterify it -
mainly with oleic acid - or they take up whole lipoproteins,
hydrolyse the cholesterol esters and re-esterify them.

There is now very detailed information on the metabolism

of cholesterol and LDL by fibroblasts and SMCs in culture
and also by macrophages. Goldstein and Brown (75) have
recently summarized their extensive studies on lipoprotein
uptake, and discussed them in relation to the intracellular
lipid accumulation in fatty streaks.

It is clear that the results still do not explain what

is happening; indeed, the data from the cultures suggest
that all the cells in the body should be filled with cholesterol
ester. The problems of translating from cultures to intact
animals are further emphasized by Weinstein, Carew and
Steinberg (76) who calculate that the rate of degradation of
LDL protein per kg smooth muscle is 120 times greater in
cultures than in the whole animal. The fat-filled cells of
fatty streaks consistently deplete their immediate vicinity
of lipoprotein reducing it to about a quarter of the concentration
in adjacent normal (Fig. 5-6) (31,32). Possibly they are
behaving more like cells in culture, and have somehow
escaped a control that is acting on adjacent cells.

If, as seems possible, smooth muscle cells and fibroblasts

constantly degrade LDL protein without accumulating cholesterol,
then cholesterol must be transported
Reverse Transport of out of cells as well as into them,
Cholesterol and failure of this "reverse
transport" may account for
cholesterol ester accumulation. There have been several
studies showing transport of cholesterol out of cells and
HDL or derivatives of HDL seem to be particularly effective
cholesterol receptors. (76,77,78,79).
METABOLIC ACTIVITIES IN ARTERIAL WALL 279

,.
"'C
<II

"0,.
c:
20

15
r~o
/~~~o~~ ... Girls

Boys

- I
CII
u
n:I 10
.:I.. I

./
'"
ftj
:~ 5
c:
"#.

0
Holman, 1961
16 Boys
180

"0... 170

~
....CII 6 Boys
'"CII
"0

\A
s:. 160
u
E
...
:I ~q
~ 150 \ f> 12 Girls
<>
140
5 10 15 20 25 Age

Virginia Lee, 1967

Fig. 5-7: Comparison of data from Holman (80) on the extent

of fatty streaking with data from Lee (81) on serum cholesterol
levels in three groups of children studied longitudinally over
several years.
280 CHAPTER 5

When one considers the relation between accumulation of

intracellular cholesterol and plasma cholesterol levels, the
idea of a failure to transport cholesterol out of the cells
appears particularly attractive. Figure 5-7 combines the
data from Holman on the percent of the intimal surface involved
in fatty streaking in young people (80) which the data of
Virginia Lee on serum cholesterol levels followed longitudinally
in children (81). These are groups of children in which she
measured the cholesterol every six months for five or ten years.
The extent of fatty streaking was lowest when serum cholesterol
was highest. During the 8 - 18 age span, when fatty streaking
was increasing at the maximum rate, these childrens' cholesterol
was drifting down. At the end of the adolescent growth
spurt when the normal adult rise in serum cholesterol starts,
the fatty streaking has levelled out. I think this diagram
is worth pondering by anybody who is thinking of feeding
cholesterol to rabbits.

EDITORIAL NOTE: From evidence developed since this

colloquium the angiotoxic properties of oxydation products
of cholesterol, notably 25 hydroxy cholesterol have been
established and newly developed data indicate that the
oxygenated sterols, normally present in very small concentration
in man, perform a regulatory function through negative
feedback of cholesterol synthesis.

Kandutsche A.A., Chen H.W., Heiniger, H.C., Biological

Activity of some Oxygenated Sterols. Science: 201: 498-501,
1978.

DR. DEWEY: Dr. Smith, I would like to make a comment

relative to one of the figures on the lipoprotein cholesterol
levels in the intima for a normal individual with elevated
blood cholesterol, also the one with elevated cholesterol
and the one with hypertension. One of the characteristics
of the hypertensive state is that the excursion between
systolic and diastolic pressure goes up more proportionally
with blood pressure. The significance of this effect depends
on a lot of things including the potential effects of trans-
endothelial transport during the cardiac cycle according to
the mechanism that Dr. Kenyon suggested yesterday. When we
are comparing a hypertensive and a normal tensive individual
and their propensity to exhibit disease, I think that the
pressure excursion as well as the absolute pressure level
are important.

DR. SMITH: Unfortunately, I have not looked at the

data in terms of systolic-diastolic differences.
METABOLIC ACTIVITIES IN ARTERIAL WALL 281

The hypertensives always have more lipoprotein in the

intima than normals and the average is just about double.
But whether this is a direct result of the hypertension or
whether it is because the hypertensio~ does something to the
wall, so that the wall retains more, I have no idea.

DR. COX: I have a question about the points that you

made with regard to the collagen synthesis and its relationship
to stretch. I want to comment on
Collagen and Elastin some work being done in our
Synthesis Increased in laboratory by Dr. Grace Fisher.
Cholesterol-fed Rabbit She is currently studying the
differences in collagen and
elastin synthesis in vitro using segments of thoracic and
aortic arch in normal rabbits and rabbits fed one of those RABBIT
non-acceptable high cholesterol diets. Basically, what she
has found in vitro is that both collagen and elastin synthesis
are elevated in the cholesterol fed rabbit. within, both
instances. a higher rate of synthesis in the arch compared
with the thoracic aorta. This, however, was done in vitro
in which there was no stretch on the vessel. The question
with regard to the studies that you mentioned, is it possible
that these cells have memory in the sense that having been
subjected to elevated stretch, do they still retain the
characteristics of higher rat~s of collagen synthesis or do
they turn off quickly? My question basically relates to the
time course of response of collagen synthesis both in the
uterus and in these cultured smooth muscle cells. Is there
any information that you know of?

DR. SMITH: I don't think there is any information on

this, and it is the sort of information that is required. I
suppose that the stretch induces proline hydroxylase synthesis,
but how rapid this is and how long the additional proline
hydroxylase lasts, I didn't find any data on this.

DR. COX: One might expect in the aortic arch compared

to the thoracic aorta. a larger stretch over the cardiac
cycles associated with pulse pressure. If there is memory
in the form of a protein or something of that sort, one
might expect to see this difference in the arch and in the
descending aorta.

DR. SMITH: I suppose it depends on the turnover rate

of the enzyme.
282 CHAPTER 5

DR. COX: The other comment I had was in making this

connection to neurohumoral control in regard to your suggestion
about this hypoxic or anoxic zone beyond 300p in the arterial
wall. Activation of smooth muscle cells to produce a contraction
in blood vessels, would be expected to increase the thickness
of the arterial walls so that the fraction of the wall that
is "within a hypoxic zone" would increase through muscle
activation.

DR. SCHWARTZ: I enjoyed Dr. Smith's presentation very

much and I have several very brief questions and comments.
One relates specifically to lactate accumulation; could Dr. Smith
say just where this lactate is, what level of pH change
would result, and what evidence have we that lactate might
interfere with the lysosomal stability. I notice that the
pH values you gave are in fact not nearly as low as the pH
optima for the lysosomal enzymes.

DR. SMITH: I don't know if there is really any adequate

reply to this. As far as I can see from the literature on
myocardial infarction, not of
Lysosomal Stability course, on aorta, progressive
and pH loss of lysosomal stability occurs,
below about pH 7. Probably, as
you say, this is not down to the pH optimum for the enzymes but
then they have significant activity even at pH 6 or so. So
I think this is an area that requires a lot of further work.
I think it is a very interesting area. The calculated
concentration of lactic acid in a plaque at 800~ from the
surface is l75mg/100m1. At this concentration, in 25% serum
the pH was 4.8 and in 50% serum it was 5.9. With 50mg/10Oml
lactic acid - the calculated concentration in an early
thickening of 400p - the pH in 25% serum was 6.9, and in
50% serum was about 7.4.

DR. SCHWARTZ: Going back quite a few years, I recall

some studies on the effect of hypercholesterolemia on oxygen
transport. I am not sure what the current status of this
work is, but it is my understanding that under certain
circumstances hypercholesterolemia may modify oxygen transport
across membranes. Is this being re-examined, and furthermore,
is A.V. Hill's theoretical data applicable in the hypercholesterol-
emia situation?

DR. SMITH: In the only papers that I managed to find

the results seemed very equivocal, and I did not really
understand what they were telling us. I did not find any
clear and convincing data on what hypercholesterolemia is
doing to oxygen transport but one's guess would be that it
wouldn't improve it.
METABOLIC ACTIVITIES IN ARTERIAL WALL 283

DR. SCHWARTZ: Concerning the apparent discrepancy

between data from Russell Ross and Dr. Dzoga's group
in Chicago it is my understanding that the Chicago group
have shown that LDL prepared without any platelet contamination
still produces smooth muscle cellular proliferation.

DR. SMITH: I did forget to make a comment there that

this monkey hyperlipemic LDL is not really a large amount of MONKEY
normal LDL, it is a highly abnormal molecule containing a
gross cholesterol overlay, and I think one might think of it
as being like a floating beta of a type III hyperlipidemia -
it is not a normal LDL.

DR. WOLF: Relating to the lactic acid concentration

and the evidence of hypoxia, do we have evidence that the
collagen excess or increase is real collagen or is it defective
collagen in some way? And I am reminded of a note or page
25 volume I of our Lindau Conference in which it was pointed
out that Stetten had shown that the synthesis of hydroxyproline
requires the presence of molecular oxygen and the suggestion
was made that abnormal collagen was being synthesized in the
presence, if indeed there was a presence of low oxygen
tension (82).

DR. SMITH: This idea that hydroxyproline could only be

synthesized at a high 02 tension has been shown by Langness
and Udenfriend (36) to be untrue.

DR. LEE: This is just a short comment on the G-6-PD

studies. Dr. Benditt of Seattle reported that one out of
three or four black families
Benditt's Neoplastic is heterozygous and has two
Theory of Atherogenesis variants, A and B, of G-6-PD
while black males and all other
non-black people are homozygous and have only one variant,
A or B, of G-6-PD. He found that atherosclerotic lesions
of these heterozygous black females showed only one variant
of G-6-PD and concluded that atherosclerotic lesions develop
from a single cell, thus monoclonal in origin (64). A group
at Johns Hopkins carried out a similar study in heterozygous
females and confirmed that well developed fibrous lesions
had one variant of G-6-PD (83). However, they found both
variants of G-6-PD in fatty streaks and concluded that
fibrous lesions and fatty streaks have different origins.
We have been interested in this subject for the past couple
of years and studied a large number of aortas from autopsied
black females.
284 CHAPTER 5

Essentially, we confirmed the findings of the Seattle and

Baltimore groups but came to a somewhat different conclusion.
When we examined very small lesions, we found both variants
A and B, of G-6-PD but well developed lesions showed one
variant, A or B. It seems to us that when a lesion begins
to develop smooth muscle cells multiply in a random fashion
until it reaches a certain size. At this stage, the lesion
will show both variants of G-6-PD. After the lesion reaches
a certain size, the fittest cell would dominate within the
lesion and multiply faster than other cells, as in a bacterial
culture; and at a later stage the lesion will consist of
this particular type of cells. At this stage, the lesion
will show one variant. Thus, a well developed lesion will
show a monoclonal pattern. Our conclusion is that an athero-
sclerotic lesion begins to develop from cells in a random
fashion but it takes a monoclonal pattern after it reaches a
certain size.

DR. COLTON: According to Benditt's theory there is

high mitotic activity in those regions that have been looked
at from a fluid mechanic viewpoint such as near bifurcations.
The blue regions also have high mitotic activity. So we can
add another speculative hypothesis which in fact totally
bypasses fluid mechanics.

DR. DEWEY: I think it would be interesting to do a

quantitative calculation of the rate at which one would need
to increase the mitotic activity of one particular cell type
so that over a substantial period of time, the net volumetric
growth of lesion material would exhibit a preponderance of
the hyperactive cell type. If there already is a substantial
amount of growth and turnover activity in the tissue under
normal circumstances, we may be talking about the small
difference between two large numbers, which does not strike
me as being the same as cancerous-type mitotic activity.

DR. STONE: What is the critical oxygen concentration

for the vessel wall and what is the answer to my question
about the vasa vasorum.

DR. SMITH: I was showing you the thickness of normal

intima in the human aorta; normal intima may be anything up
to 300 microns; the media in the human is about 1100 M in
thickness and the vasa are supposed to penetrate only the
outer third of the vessel. So, in a human aorta you have
theoretically a large block that could pretty well be anoxic.
The vasa go in about 350 M.
METABOLIC ACTIVITIES IN ARTERIAL WALL 285

DR. STONE: Let me make sure. Are you saying "anoxia"

or "hypoxia."

DR. SMITH: Hypoxia, and I think if the Hill calculation

is correct there will be a quite definitely anoxic area. We
require a repetition of the oxygen probe data in a dog, or
better still, in a bovine or something with a thick vessel.
Those were two studies that I put together myself. One was COW
the study on the oxygen tension across the rabbit aorta in RABBIT
vivo; that was the top curve (Fig. 5-2). And the other was
in vitro work on the effect of different oxygen tensions on
lactic acid production.

DR. COX: You have to be careful about the vasa vasorum

because they are very variable. There are reports in the
literature that the vasa vasorum extend all the way to the
intimal-medial junction in some places, especially in the
main pulmonary artery and in the ascending aorta of man.
There could be a lot of variability.

DR. MANSFIELD: I think we have to be very careful in

terms of extrapolating from a p02 value as to the viability
of the tissues that we are talking about. If in the dog or DOG OR
cat, you excise a segment of the aorta so it is totally CAT
separated on the ends, sew it back and encapsulate it in an
absolutely non-porous backing, you will find that all the
exterior wall is dead up to about 500 microns from the
intima. Now if you take fibroblasts and line the inside of
a non-porous tube, thrombus deposits on this and when that
thrombus reaches the level somewhere between 500 and 800
microns, the fibroblasts will die and you can show cell
death at that time. So these are two indications that we
are in the ball park in terms of the limits of diffusion
depth which leads to cell death. We do not see that in the
human normal aorta, regardless of what your p02 values are,
it is unlikely that anoxia is occurring or we should see the
histological changes of cell death and destruction. Rather
there may be various degrees of hypoxia depending on what
the oxygen requirement is within the wall itself. I think
it is very important that we are cautious in terms of saying,
"What is going on in the cell?" When the cell is obviously
alive, and even though it may be quite removed from either
side of lining of the vessel wall itself.

DR. SMITH: I quite agree with you, and I never seem to

see medial necrosis.
286 CHAPTER 5

DR. SCHWARTZ: Just a very brief comment to remind us

of Seymour Glagov's important work, which indicates that
approximately ZO medial lamella can occur without any vasa
vasorum, a finding which suggests that diffusional nutrition
is of considerable importance in the inner media (84).

DR. COLTON: With regard to those tabulations one must

be very careful. Hill's calculation assumed zero order
kinetics. The available evidence suggests that oxygen
uptake kinetics follows a Michaelis-Menten form. When pOZ
is low enough to become first order, the rate is not constant
but decreases as the pOZ decreases. One needs to know what
is the critical pOZ at which necrosis occurs. I don't think
there are good data for this.

DR. KENYON: There is always interesting speculation on

the effects of increased demand due to mechanical stressing
on the wall and I think that in the arch of the aorta and
especially with unusually large pressure differences there
are large dynamic strains which could produce anoxia. This
could be quite important because the smooth muscle cells
presumably are attempting to synthesize the material and any
additional barrier to diffusion you may get could be a self-
destructive mechanism.

DR. LEE: Of course, we do not see necrosis in a normal

vessel no matter how thick the vessel is, for example the
aorta. However, when lesions develop and the thickness of
the intimal lesion exceeds the thickness of the media we
begin to see necrosis in the center of the lesion. I was
told some years ago that one simple reason to have central
necrosis in these thick lesions is that the distance to the
center of the lesion is too great for efficient supply of
oxygen. I don't know whether that is true or not.

DR. SMITH: Well, I don't know that they expect an

answer. When I was looking at the depth of the fat filled
cells from the surface in
Mixed Lesions fatty streaks I also had a number
of mixed lesions which I called
mixed lesions. They were prolifer~tive lesions in which there
were also fat filled cells in the cap of the lesion, and I
noticed that there seemed to be a critical depth below which
fat filled cells deeper than about, I think, three to four
hundred microns from the surface seemed to disappear and I
think it is possible that these fat filled cells come to
grief once they get to a certain critical depth, and disintegrate.
METABOLIC ACTIVITIES IN ARTERIAL WALL 287

DR. BJORKERUD: We have not emphasized the fact that

mechanical factors and injury could be beneficial in some
situations and function as feed-
Growth Factors of back mechanism. If we assume, for
Arterial Tissue instance, that a segment of an
artery is mechanically insufficient
the mechanical factors may injure the wall. The injury will
promote entry of nutrients or other factors will stimulate
growth in the tissue. The growth will reinforce the wall so
that injury no longer occurs. What I would like to ask is:
Is it really that clear, what factors we have stimulating growth
of the arterial tissues? Because as far as I can understand
there are many factors.

DR. SCHWARTZ: I can't really make any definitive comments,

but let me respond at least superficially to some aspects of
cell growth. There are many factors which appear to enhance
cell growth or proliferation in tissue culture. Gospodarowicz
has reviewed this field (85); growth factors include insulin,
pituitary factor, a platelet factor described by Russel Ross
of Seattle, (86) low density lipoprotein, and prostaglandin,
Fzx to mention but a few. It will, I believe be some time
before we are in a position to place specific factors into a
pathological perspective.

DR. BJORKERUD: I think the notion in some of Benditt's

work that the fibrous plaque might be similar to a benign
tumor might be unnecessarily specific. It would be quite
enough if the control of growth by one of these factors was
lost.

EDITORIAL NOTE: Thomas and associates at Albany (W.A. Thomas,

J.M. Reiner, R.A. Florentin, K. Janakidevi and K.T. Lee: Arterial
Smooth Muscle Cells in Atherogenesis: Births, Deaths and Clonal
Phenomena, Proceedings of IVth Inter'l Symposium on Atherosclerosis
in Tokyo, Japan, 1976) have offered an alternate to Benditt's
monoclonal or neoplastic theory of the origin of atherosclerotic
plaques. They observed that the monotypism ( a simple variant of
the G-6-PD) gene in fibrous arteriosclerotic lesions of hetero-
zygous women) was characteristic only of the very thick lesions.
The presence of both variants in thinner fibrous plaques in the
same vessel led them to the inference that there was not an
original cell transformation but rather a selection process that
endowed the progeny of one variant with greater survival capability
than the other. Hence one variant would appear exclusively in
older lesions. In experiments on pigs they were able to strengthen
their evidence that diet-induced lesions were of multiple cell
origin but that one or anther cell type often showed superior
survival over several generations of mitosis.
288 CHAPTER 5

Secondly. in connection with the oxygen concentration in

the tissue it has been shown in cultured smooth muscle cells.
that they lose. some of their capacity to cope with LDL when
cultured at low oxygen pressure. This seems to be a very
important observation. Finally. I think Colin Adams showed
that a certain degree of intimal thickening. some figures
indicate 100-200 microns. the underlying cells seem to be
injured (87).

DR. SMITH: I seem to visualize it as a mid-medial loss.

but you think it was as near the surface as 100 microns.

DR. BJORKERUD: Yes. I might be wrong but Benditt

disusses that.

DR. SMITH: We really do not have accurate information on

the pathways of energy production in artery. It looks as if
the earlier results indicating primary dependence of glycolysis
are probably artifacts. and that under normal conditions of
oxygen supply ATP is generated by oxydative phosphorylation.
and that balance between glycolysis and oxydative phosphorylation
depends on oxygen supply as in other muscles.

However. the results may also indicate that artery is

abnormally sensitive to damage. and switches over to glycolysis
and lactic acid production in response to rather minor injury.
Thus. although the endothelium was damaged in Morrison. et aI's ••
preparation (5) they could see no evidence of damage to the
smooth muscle cells. which suggests that minor disturbance of
the endothelium may influence the metabolism of the whole wall.
METABOLIC ACTIVITIES IN ARTERIAL WALL 289

BIBLIOGRAPHY

1. Lehninger, A.L. Biochemistry. Worth Publishers,

N. Y., 1970.

2. Scott, R.F., Morrison, E.S. and Kroms, M.: Effect of

cold shock on respiration and glycolysis in swine
arterial tissue. Am. J. Physiol. 219, 1363-1365,
1970.

3. Scott, R.F., Morrison, E.S., and Kroms, M.: Aortic

respiration and glycolysis in the pre-proliferative
phase of diet-induced atherosclerosis in swine.
J. Atheroscler. Res. 9, 5-16, 1969.

4. Morrison, E.S., Scott, R.F., Kroms, M., and Frick, J.:

Glucose degradation in normal and atheroslerotic
aortic intima-media. Atherosclerosis 16, 175-184,
1972.

5. Morrison, A.D., Berwick, L., Orci, L., and Winegrad, A.I.:

Morphology and metabolism of an aortic intima-media
preparation in which an intact endothelium is
preserved. J. Clin. Invest. 5, 650-660, 1976.

6. Arnqvist, H.J. and Lundholm, L.: Influence of oxygen

tension on the metabolism of vascular smooth muscle;
demonstration of a Pasteur effect. Atherosclerosis
25, 245-253, 1976.

7. Niinikoski, V., Heughan, C. and Hunt, T.K.: Oxygen

tensions in the aortic wall of normal rabbits.
Atherosclerosis 17, 353-359, 1973.

8. Kirk, J.E. and Laursen, T.J.S.: Diffusion coefficients

of various solutes for human aortic tissue, with
special reference to variation in tissue permeability
with age. J. Gerontol. 10, 288-302.

9. Hill, A.V.: The diffusion of oxygen and lactic acid

through tissues. Proc. Roy. Soc. (London) Ser. B.,
104, 39-96, 1928-29.

10. Moss, A.J., Samuelson, P., Angell, C. and Minken, S.L.:

Polarographic evaluation of transmural oxygen
availability in intact muscular arteries. J. Atheroscler.
~, 8, 803-810, 1968.
290 CHAPTER 5

11. Kjeldsen, K., Wanstrup, J. and Astrup, P.: Enhancing

influence of arterial hypoxia on the development
of atheromatosis in cholesterol-fed rabbits.
J. Atheroscler. Res. 8, 835-845, 1968.

12. Kjeldsen, K., Astrup, P., and Wanstrup, J.: Reversal of

rabbit atheromatosis by hyperoxia. J. Atheroscler.
~, 10, 173-178, 1969.

13. Vesselinovitch, D., and Wissler, R.W.: Experimental

atherosclerosis in rabbits - the effect of oxygen
and/or cholestyramine on its reversibility.
~ 38, Suppl. VI, 198, 1968.

14. Vesselinovitch, D., Wissler, R.W., Dzoga, K., Hughes, R.H.,

and Dubien, L.: Regression of atherosclerosis in
rabbits; Pt. 1, Treatment with low-fat, hyperoxia
and hypolipidemic agents. Atheroclerosis 19,
259-275, 1974.

15. Zemplenyi, T.: In: The Smooth Muscle of the Artery

(Eds. S. Wolf & N.T. Werthessen) . Advan. Exp. Med.
~ 57, 302, 1975.

16. Peters, T.J.: Lysosomes of the cardiovascular system.

Prog. in Cardiol. 4, 151-164, 1975.

17. Burleigh, M.C., Barrett, A.J. and Lazarus, G.S.: Cathepsin

Bl: a lysosomal enzyme that degrades native
collagen Biochem. J. 137, 387-398, 1974.

18. de Duve, C., and Wattiaux, R.: Functions of lysosomes.

Ann. Rev. Physiol., 28, 435-492.

19. Brachfeld, N.: Maintenance of cell viability. Circa

Suppl. 4, Vols. 39-40, 202-214, 1969.

20. Robin, E.D., Wilson, R.J., and Bromberg, P.A.: Intra-

cellular acid-base relations and intracellular
buffers. Ann. N.Y. Acad. Sci. 92. 539-546, 1961.

21. Reijngoud, D.J., and Tager, J.M.: Measurement of

intralysosomal pH. Biothem. Biophys. Acta.
297, 174-178, 1973.
METABOLIC ACTIVITIES IN ARTERIAL WALL 291

22. Ravens, K.G. and Gudbjarnason, S.: Changes in the

activities of lysosomal enzymes in infarcted canine
heart muscle. Circl. Res. 24, 851-856, 1969.

23. Ricciutti, M.A.: Myocardial lysosome stability in the

early stages of acute ischaemic injury. Am. J.
Cardiol. 3D, 492-497, 1972.

24. Ricciutti, M.A.: Lysosomes and myocardial cellular

injury. Am. J. Cardiol. 3D, 498-502, 1972.

25. WeIman, E.: Lysosomal changes during anoxia in guinea-

pig heart. Biochem. Soc. Trans. 2, 746-748.

26. Stary, H.C.: Coronary artery fine structure in Rhesus

monkeys: the early atherosclerotic lesion and its
progression. Primates in Medicine 9, 359-395, 1976.

27. Whereat, A.F.: Atherosclerosis and metabolic disorder

in the arterial wall. Exp. Mol. Path. 7, 233-247,
1967.

28. Portman, O.W. and Illingworth, D.R.: Arterial metabolism

in primates. Prim. Med. 9, 145-223.

29. St. Clair, R.W.: Metabolism of the arterial wall and

atherosclerosis. Atherosclerosis Revs. I, 61-117,
1976.

30. Haust, M.D.: The morphogenesis and fate of potential

and early atherosclerotic lesions in man. Hum.
Pathol. 2, 1-29, 1971.

31. Smith, E.B. and Slater, R.S.: Lipids and low-density

lipoproteins in intima in relation to its morphological
characteristics. In: Atherogenesis: Initiating
factors. Ciba Symp. No. 12 (NS) 39-52, 1973.

32. Smith, E.B. and Smith, R.H.: Early changes in aortic

intima. Atherosclerosis Revs. I, 119-136, 1976.

33. Smith, E.B.: Acid glycosaminoglycan, collagen and

elastin content of normal artery, fatty streaks and
plaques. In: Arterial Mesenchyme and Atherosclerosis
(Eds. W.D. Wagner and T.B. Clarkson) Advan. Exper.
Med. & BioI. 43, 125-138, 1974.
292 CHAPTER 5

34. Lindner, J.: Regressive und progressive arterielle

Reaktionen bei Atherosklerose: 5. Veranderungen
im extracellularen Kompartment. Verh. Dt. Ges.
inn. Med. 78, 1166-1175, 1972.

35. Green, H. and Goldberg, B.: Collagen and cell protein

synthesis by an established mammalian fibroblast
line. Nature. 204, 347-349, 1964.

36. Langness, U. and Udenfriend, S.: Collagen proline

hydroxylase activity and anaerobic metabolism.
In: Biology of fibroblast (E. Kulonen &
J. Pikkarainen) Academic Press, 373-377, 1973.

37. Levene, C.I. and Bates, C.J.: The activation of

protocollagen proline hydroxylase and its effect
on collagen synthesis in culture 3T6 fibroblasts.
Ital. J. Biochem. 24, 36, abs. 1975.

38. Schwarz R., Colarusso, L. and Doty, P.: Maintenance of

differentiation in primary cultures of avian
tendon cells. Exp. Cell Res. 102, 63-71, 1976.

39. Chvapil, M., Hurych, J. and Mirejovska, E.: Effect of

long-term hypoxia on protein synthesis in granuloma
and in some organs in rats. Proc. Soc. Exp. BioI.
Med. 135, 613-617, 1970.

40. Hunt, T.K. and Pai, M.P.: The effect of varying ambient
oxygen tensions on wound metabolism and collagen
synthesis. Surg. Gyn. Obs. 135, 561-567, 1972.

41. Bedford, J.S. and Mitchell, J.B.: The effect of hypoxia

on the growth and radiation response of mammalian
cells in culture. Brit. J. Radiol. 47, 687-606,
1974.

42. Leung, D.Y.M., Glagov, S., Clark, J.M. and Mathews, M.B.:
Mechanical influences on the biosynthesis of
extracellular macromolecules by aortic cells.
In: Extracullular Matrix Influences on Gene
Expression. (Eds. H.C. Slavkin & R.C. Greulich).
Academic Press 633-645, 1975.

43. Leung, D.Y.M., Glagov, S., and Mathews, M.B.: Cyclic

stretching stimulates synthesis of matrix components
by arterial smooth muscle cells in vitro. Science
191, 475-477, 1976.
METABOLIC ACTIVITIES IN ARTERIAL WALL 293

44. Norby, D.P., Ma1emud, C.J. and Sokoloff, L.: Modulation

of phenotypic expression of collagen synthesis by
lapine articular chondrocytes in spinner and mono-
layer cultures. Fed. Proc. 35, 714 (abs) 1976.

45. Wolinsky, H.: Effects of hypertension and its reversal

on the thoracic aorta of male and female rats.
Circulation Res. 28, 622-637, 1971.

46. Wolinsky, H.: Long term effects on hypertension on the

rat aortic wall and their relation to concurrent
aging changes. Circulation Res. 30, 301-309, 1972.

47. Hollander, W., Kramsch, D.M., Farmelant, M. and Madoff, I.M.:

Arterial wall metabolism in experimental hypertension
of coarctation of the aorta of short duration.
J. Clin. Invest. 47, 1221-1229, 1968.

48. Fernandez, D. and Crane, W.A.J.: New cell formation in

rats with accelerated hypertension due to partial
constriction. J. Path., 100, 307-316, 1970.

49. Smith, E.B., Alexander, K.M. and Massie, I.B.: Insoluble

"fibrin" in human aortic intima: quantitative studies
on the relationship between insoluble "fibrin" soluble
fibrinogen and low density lipoprotein. Atherosclerosis
23, 19-39, 1976.

SO. Smith, E.B. and Crothers, D.C.: Interaction between plasma

proteins and the intercellular matrix in human aortic
intima. Protides of the Biological Fluids, 22,
315-318, 1974.

51. Fisher-Dzoga, K., Chen, R. and Wissler, R.W.: Effect of

serum lipoproteins on the morphology, growth and
metabolism of arterial smooth muscle cells. Advan.
Exp. Med. BioI., 43, 299-311, 1974.

52. Ross, R.: The smooth muscle of the artery. Advan. Exp.
Med. BioI. 57, 64-79, 1975.

53. Ronnemaa, J., Juva, K., and Kulonen, E.: Effect of

hyperlipidemic rat serum on the synthesis of collagen
by chick embryo fibroblasts. Atherosclerosis, 21,
315-324, 1975.
294 CHAPTER 5

54. Fuller, G.C., Miller, E., Farber, T. and Vanloon, E.:

Aortic connective tissue changes in miniature pigs
fed a lipid-rich diet. Connective Tiss. Res.
1, 217-220, 1972.

55. St. Clair, R.W., Toma, J.J. and Lofland, H.B.: Proline
hydroxylase activity and collagen content of pigeon
aortas with naturally-occurring and cholesterol-
aggravated atherosclerosis. Atherosclerosis, 21,
155-165, 1975.

56. McCullagh, K.G. and Ehrhart, L.A.: Increased arterial

collagen synthesis in experimental canine atherosclerosis
Atherosclerosis, 19, 13-28, 1974.

57. Smith, E.B., Massie, I.B. and Alexander, K.M.: The

release of an immobilized lipoprotein fraction from
atherosclerotic lesions by incubation with plasmin.
Atherosclerosis, 25, 71-84, 1976.

58. Smith, E.B.: Arterial wall and lipoproteins - steady state

aspects. Proc. IV International Symposium on
Atherosclerosis, Tokyo, 1976. In Press.

59. Camejo, G., Lopez, A., Vegas, H., and Paoli, H.: The
participation of aortic proteins in the formation of
complexes between low density lipoproteins and
intima-media extracts. Atherosclerosis, 21, 77-91,
1975.

60. Tracy, R.E., Dzoga, K., and Wissler, R.W.: Sequestration

of serum low density lipoproteins in the arterial
intima by complex formation. Proc. Soc. Exp. BioI.
Med., 118, 1095-1098, 1965.

61. Bihari-Varga, M. and Vegh, M.: Quantitative studies on

the complexes formed between aortic mucopolysaccharides
and serum lipoproteins. Biochem. Biophys. Acta., 144,
202-210, 1967.

62. Srinivasan, S.R., Dolan, P., Radhakr ishnamur thy , B. and

Berenson, C.S.: Isolation of lipoprotein-acid
mucopolysaccharide complexes from fatty streaks of
human aortas. Atherosclerosis, 16, 95-104, 1972.

63. Anderson, A.J.: The formation of chondromucoprotein-

fibrinogen and chondromucoprotein-B-lipoprotein
complexes. Biochem. J., 88, 460, 1963.
METABOLIC ACTIVITIES IN ARTERIAL WALL 295

64. Benditt, E.P. and Benditt, J.M.: Evidence for a

monoclonal origin of human atherosclerotic plaques.
Proc. Nat. Acad. Sci. U.S.A., 70, 1753-1756, 1973.

65. Panganamala, R.V., Geer, J.C., Sharman, H.M. and

Cornwell, D.G.: The gross and histologic appearance
and the lipid composition of normal intima and
lesions from human coronary arteries and aorta.
Atherosclerosis, 20, 93-104, 1974.

66. Pearson, T.A., Wang, A., Solez, K. and Heptinstall, R.H.:

Clonal characteristics of fibrous plaques and fatty
streaks from human aortas. Am. J. Pathol., 81,
379-387, 1975.

67. Lee, K.T., Imai, H., Werthessen, N.T., and Taylor, C.B.:
Necrogenic agent obtained from cholesterol used in
dietary experiments. In: Atherosclerosis III
(Eds. G. Schettler and A. Weizel) 344-347, 1974.

68. Florentine, R.A., Nam, S.C., Lee, K.T., Lee, K.J.,

and Thomas, W.A.: Increased mitotic activity in
aortas of swine. Arch. Path., 88, 463-469, 1969.

69. Stary, H.C. and McMillan, G.C.: Kinetics of cellular

proliferation in experimental atherosclerosis.
Arch. Path., 89, 173-183, 1970.

70. Smith, E.B. and Massie, I.B.: Destruction of endogenous

low density lipoprotein in incubated intima.
Atherosclerosis. In press.

71. Miller, B.F. and Kothari, H.V.: Increased activity of

lysosomal enzymes in human atherosclerotic aortas.
Exp. Mol Path., 10, 288-294, 1969.

72. Zemplenyi, T.: Vascular enzymes and the relevance of

their study to problems of atherogenesis. Med. Clin.
of North Amer., 58, (no.2), 293-321, 1974.

73. Oliver, M.F.: Dietary cholesterol, plasma cholesterol

and coronary heart disease. Brit. Heart J., 38,
214-218, 1976.

74. Day, A.J. and Proudlock, J.W.: Changes in aortic

cholesterol-esterifying activity in rabbits fed
cholesterol for 3 days. Atherosclerosis, 19,
253-258, 1974.
296 CHAPTER 5

75. Goldstein, J.L. and Brown, M.S.: Lipoprotein receptors,

cholesterol and metabolism and atherosclerosis.
Arch. Patho1., 99 181-184,1975.

76. Weinstein, D.B., Carew, T.E. and Steinberg, D.:

Uptake and degradation of low density lipoprotein
by swine arterial smooth muscle cells with inhibition
of cholesterol biosynthesis. Biochim. Biophys. Acta,
424, 404-421, 1976.

77. Stein, Y., Glangeaud, M.C., Fainaru, M. and Stein, 0.:

The removal of cholesterol from aortic smooth muscle
cells in culture and Landschutz ascites cells by
fractions of human high density apo1ipoprotein.

78. Brown, M.S., Faust, J.R. and Goldstein, J.L.: Role of

the low density lipoprotein receptors in regulating
the content of free and esterified cholesterol in
human fibroblasts. J. C1in. Invest., 55, 783-793,
1975.

79. Werb, Z. and Cohn, Z.A.: Cholesterol metabolism in the

macrophage, III. Ingestion and intracellular fate
of cholesterol and cholesterol esters. J. Exp. Med.,
135, 21-44, 1972.

80. Holman, R.L.: Atherosclerosis - a pediatric nutrition

problem? Am. J. C1in. Nutr., 9, 565-569, 1961.

81. Lee, V.A.: Individual trends in the total serum cholesterol

of children and adolescents over a ten-year period.
Am. J. C1in. Nutr., 20, 5-12, 1967.
82. Stetten, M.R.: Some aspects of metabolism of hydroxyproline,
studied with aid of isotopic nitrogen. J. Bio1.
Chern. 181, 31, 1949.

83. Pearson, T.A., Wang, A., Solez, K., and Heptinsta11, R.H.:
Clonal characteristics of fibrous plaques and fatty
streaks from human aortas. Am. J. Path. 81, 379-387,
1975.

84. Wolinsky, H., G1agov, S.: A lamellar unit of aortic

medial structure and function in mammals. Circ.
~, 20:99-111, 1967.
METABOLIC ACTIVITIES IN ARTERIAL WALL 297

85. Gospodarowicz, D., Moran, J.S.: Growth factors in

mammalian cell culture. Ann. Rev. of Biochem.
45:531, 1976.

86. Ross, R., and Glomsett, J.A.: The pathogenesis of

atherosclerosis. N. Eng. J. Med. 295-369, 1976.

87. Adams, C.W.H. and Bayliss, O.B.: The relationship

between diffuse intimal thickening, medial enzyme
failure and intimal lipid deposit in various human
arteries. J. Athero. Res., 10:327, 1969.
Chapter 6 TRANSPORT OF PROTEIN AND LIPID INTO THE ARTERIAL WALL

DR. COLTON:* We have recently carried out experimental

measurements of the distribution of labeled albumin and of
labeled low density lipoprotein (LDL) across the rabbit RABBIT
thoracic aorta in vivo following intravenous injection
(1,2). We have also developed a theoretical model of labeled
protein transport across the arterial wall, and we have
begun to use it to analyze our experimental results (3). In
this mini-review, I shall first discuss our experimental
procedures and then summarize our results. Next I will
discuss the basis for the mathematical model we have developed.
Since our efforts in this area are still in progress, I will
provide some representative examples of comparison between
theoretical prediction and experimental data, and I will
conclude with a brief discussion of the kind of insights
this type of analysis provides.

The experimental procedures employed are illustrated in

Fig. 6-1. LDL was isolated from freshly drawn human blood
plasma and was iodinated with
Experimental Procedure Na 125 I using the iodine mono-
chloride method. Free iodine
was removed by ultracentrifugation followed by three
sequential dialyses against saline containing EDTA. The
integrity of the labeled LDL was verified by a variety of
techniques. The labeled LDL was filtered and injected
intravenously into normal conscious New Zealand White rabbits, RABBIT
(3.5 to 5.0 kg). The fraction of label in the injectate
which was precipitable by trichloroacetic acid (TCA) averaged
about 99% for albumin and 97% for LDL. A blood sample was
taken about five minutes after injection and at specified
times thereafter for plasma radioactivity and total cholesterol
determinations. Plasma lipoproteins were isolated by ultracen-
trifugation and assayed for radioactivity at the end of
several experiments. The animals were sacrificed after 10
min., 30 min., 4 hr., 24 hr., or 67 hr., and the descending
thoracic aorta was immediately excised, opened longitudinally,
rinsed, and frozen to prevent further diffusion. The time
from sacrifice to freezing was 3 to 6 min.

Samples of frozen aorta were sectioned (20~m thickness)

parallel to the intimal surface on a refrigerated microtome.
* The work reported here was done in collaboration with
R.L. Bratzler, K.A. Smith and R.S. Lees.
299
 125 II. 131 . Co)
I and/or I A bl umin, LDL
g
I ~

J
~ /
.._
/7I.
.... . ./ . ~
'" .
"~
•... .
.:. , ,...". ",

~
-~;;';;:horaCiC Aorta

Tissue Slice (20~ thick)

\. J ( ..--Advenlilia----... ~/
Medio~
\-

-1-

It (')
:J:
»
Fig. 6-1: Illustration of experimental procedure used to measure the distribution of
"-im
::tl
l25I-labeled proteins across the rabbit thoracic aorta in vivo. en
TRANSPORT OF PROTEIN AND LIPID 301

Slicing proceeded from the adventitial side. After every

slice, the upper and lower edges of the knife blade were
cleaned with absorbent paper to prevent radioactive contamination
of subsequent sections. Each slice and the associated paper
were placed in 10-mm diameter, precooled test tubes and set
aside for radioassay. Slices exhibiting visible blood
contamination were discarded. Positions of the intimal
surface and the medial-adventitial border were noted, and
the total thickness (L) between them was estimated to within
+ 20pm. Similarly, the distance (x) from the intimal
surface to the midpoint of each tissue slice was noted and
the location of the slice designated by x/L. The volume of
each slice was estimated from its known thickness and measured
cross sectional area. We have measured aortic wall thickness
as a function of applied pressure when the aorta is stretched
to its normal in vivo dimensions. The medial thickness in
the related state (in which it was frozen for this study)
was 2.4 times greater than under in vivo conditions. Each
tissue slice, injectate, and plasma sample was extracted at
least twice with 10% (w/v) TCA to remove non-protein-bound
radioactivity before radioassy in a gamma counter. All
slices were counted for 10 min., and almost all had counting
rates in excess of 10 cpm above the background rates, which
ranged between 40 and 70 cpm. Virtually identical procedures
were employed using 125I-albumin prepared from rabbit serum
albumin obtained commercially.

Figure 6-2 contains two graphs which show the concentrations

of labeled albumin and labeled LDL in plasma, divided by
their initial values, as a
Results function of time. The initial
rapid decay and subsequent slower
decline for each solute could be fitted well with a double
exponential function. The concentration of plasma label
dropped to about 45% of its initial value after 24 hours
with albumin, whereas LDL decreased to about 25%. After 67
hours, the relative plasma concentration for LDL was about 5
to 10%.

Figures 6-3 and 6-4, taken from reference (2), compare

the concentration profiles and the mean concentration in the
media as a function of time for both l25I-albumin and
l3ll-LDL. The results obtained with 13lI-LDL were qualitatively
similar to those obtained with l25I-albumin. Up to 4
hours, transmural concentration profiles of TCA-precipitable
radioactivity for both solutes had steep gradients near the
intimal surface, moderate gradients near the medial-adventitial
border, and were relatively flat in the middle of the media.
 CHAPTER 6
302

Fig. 6-2: Relative TCA-precipitable 125I-labeled protein con-

centration in plasma as a function of time following injection
of labeled albumin (uppergraph) and LDL (lower graph). Cp is the
TCA-precipitable plasma concentration (counts/min per cm3 plasma),
and Cpo is the value about 5 minutes after injection. Sold
curves were fitted to data by nonlinear regression analysis.
(Composite figure taken from references 1 and 2).
TRANSPORT OF PROTEIN AND LIPID 303

24

20
r<l
o 10 MIN 30 MIN
x 16 AVERAGE AVERAGE
N L. p.m N L. p'm
.4 207±37 . 5 206±55
'- 12
.4 234±32 5 252 ±32

z
o
r
<t
0::
r
z
w
u
z
o 24
u

~ 20 \
N
4 HR
AVERAGE
L. p'm N
24 HR
AVERAGE
L. p.m
r 16
.5 210±39 • 5 237 ±37
w 5 258±29 • 3 252 ± 23
> 12
r
<t
--.J

:~~~t~-
W
0::

o 0.2 0.4 0.6 0 .8 1.0 0 0 .2 0.4 0.6 0.8 1.0

NOHMALIZED DE P T H. x/ L

Fig. 6-3: Comparison between profiles of relative tissue con-

centration, CT/C p , for 125I-LDL (~) and 125I-albumin (0) for
lO-min, 30-min, 4ohr, and 24-hr experiments. CT is the con-
centration of TCA-precipitable tissue radioactivity (counts/min
per cm 3 wet tissue). Cp is the TCA-precipitable plasma con-
centration (counts/min pgr cm3 plasma) 5 minutes after injection.
x is the distance from the intimal surface to the midpoint of the
tissue slice, and L is the distance between the intimal surface
and the medial-adventital border. Error bars represent standard
error of the mean. N is the number of rabbits studied (Taken
from reference 2).
304 CHAPTER 6

10

,..,
Q
• ALBUMIN
w
:J
(f) U
><
(1.0
8
• LDL

(f) "-
i= Iu~
6
w
~ Z
f- 0
<I i=
...J<I
w a:: 4
a:: f-
z Z
W
<I
W u
:::E z 2
0
u

2 3 4 5 6 7 8 9
(TIME) 1/2 ( HR )1/2

Fig. 6-4: Mean relative labeled LDL and labeled albumin

concentrations in the media at each time period. The values
were obtained by numerical integration of the data in Fig. 6-3.
(Taken from reference 2).
TRANSPORT OF PROTEIN AND LIPID 305

The results were consistent with entry of l3lI -l abeled

protein into the media from both the luminal and adventitial
sides. After 24 hours, the steep intimal gradient disappeared.
Concentration levels were otherwise comparable to those at 4
hours. TCA-soluble tissue radioactivity slowly increased
with time, suggesting that some labeled protein may have
been catabolized to TCA-soluble fragments by aortic tissue.
The rate of accumulation of TCA-precipitable radioactiviy
was initially rapid (measurable concentrations of both
solutes were found throughout the media after only 10 minutes)
and decreased with time. The relative concentration levels
and the rate of influx was greater for labeled albumin than
for labeled LDL, thereby suggesting that the transport
mechanism(s) involved may be in part dependent upon molecular
size. The very low values for relative tissue concentration
are noteworthy. The concentration of l3lI-LDL in tissue
ranged from 10- 3 to 10- 2 times its initial concentration in
plasma; the mean value reached a maximum of about 5 x 10- 3
at four hours following injection.

We have begun to develop theoretical models of the

arterial wall to aid in the interpretation of experimental
data. To date, these models
Theoretical Models have been compared only with
the LDL data. These models
are based on fundamental mass transport principles and, at
different levels of complexity, incorporate mechanisms
believed relevant to transport of LDL into the wall. The
physical basis for the theoretical models is presented
schematically in Figure 6-5. The description is applicable
to the transport of LDL in rabbit aorta; it may also apply
to other plasma proteins and to other large blood vessels as
well. It is likely that plasma LDL is capable of penetrating
intimal endothelium by transendothelial vesicular transport.
Diffusion and/or convection of LDL through the junctions
between endothelial cells is possible but not likely under
normal conditions. Both mechanisms are included in the model
for the sake of generality. Another possible route of LDL
entry into the aortic media is by transport across capillary
endothelium present in the adventitia. LDL which has gained
access to the media will diffuse in a direction consistent
with its local concentration gradient. Since there exists a
pressure difference across the rabbit aortic wall, a, filtration
flow may also affect LDL movement by convecting entrained
solutes radially toward the adventitia. LDL transport across
the arterial wall may also be influenced by interactions
with components of the wall. For example, LDL will likely
bind to components of the tissue, such as glycosaminoglycans
elastin, and collagen.
306 CHAPTER 6

Cn"vpt'llt\n and Diffusion

Endothelium
Internol elastic ~,..f]!!i\;i!I_~~~->~~
membrane

I~ Convection
I
... Diffusion - - - Binding, Cellular Permeation, and Degradation

Fig. 6-5: Conceptual illustration of arterial wall showing modes

of LDL transport and reaction believed to occur in the rabbit
thoracic aorta.
TRANSPORT OF PROTEIN AND LIPID 307

Evidence from tissue culture studies suggests that aortic

smooth muscle cells may be capable of interiorizing and
degrading LDL. Thus, smooth muscle cell metabolism may be
an important determinant of LDL removal rates. Another
mechanism for clearance from the arterial wall is drainage
into the lymphatic system. In short, the concentration of
LDL in the arterial wall is likely to be effected in a
complicated fashion by the plasma LDL concentration, the
transport properties of both intimal and capillary endothelium,
and the rates of diffusion, convection, binding, and cellular
uptake and reaction within the arterial wall, as well as
transport into the lymphatic vessels.

The partial differential equations which describe

transport in the media are summarized in Figure 6-6. We
have assumed that the media can be represented by a continuum
so that species conservation equations can be derived for
transport of labeled solute. Transport across the intimal
endothelium, and between the media and the capillaries and
lymphatic vessels in the adventitia, is represented by the
boundary conditions shown in Figure 6-7. This represenation
is physically reasonable for rabbits because the "vasa
vasorum" are contained solely within the adventitia and do
not penetrate the media itself. In many larger mammals,
including humans, the vasa vasorum penetrate well into the
media, for which case the analysis would have to be extended.

Within the media, labeled protein can be found in at

least three forms: (a) in the extracellular fluid in a
freely diffusible state; (b)
Transport of Labeled in the extracellular fluid but
Protein bound to the wall components
(e.g., glycosaminoglycans and
cell surfaces); and (c) within the cells. The latter category
may itself represent a variety of forms. We have assumed that
the binding reaction is reversible and that it leads to
formation of a non-diffusible complex. The total concentration
of TCA-precipitable labeled protein is the sum of the concentrat-
tions of each of the various forms. The species conservation
relation is obtained by applying a material balance for
freely diffusible labeled protein over a differential volume
element within the media. The various terms within that
relation represent accumulation of free and bound solute,
convection of free solute, diffusion of free solute, and
permeation of free solute into the cells. Separate conservation
relations are also written for accumulation of bound and
intracellular solutes. The kinetics of the binding reaction
are assumed to be linear.
308 CHAPTER 6

lC f V a"C f PaE • Ef
D:-r - --::2 -E1(Cf--C.)
ax Efax - f Ei
diffusion convection permeation free bound
into cells accumulation
k2
kl (C f - - C )
kl b

binding reaction
( re ve rs i b1e )

intracellular permeation
degradation into cells

concentration (moles/unit volume of media) of solute

(f) free in extracellular space
(b) bound in extracellular space
(i) intracellular
CT = Cf + Cb + Ci
volume fraction (diffusion space/unit volume of media)
available to solute in (f) extracellular and (i) intra-
cellular phases
rate constants for binding to extracellular components
rate constant for intracellular degradation
effective diffusion coefficient in media
superficial fluid velocity in media
diffusive permeability (length/time) of medial cell membran
a specific surface area (cell surface area/cell volume) of
medial cells
t time
x spatial coordinate

Fig. 6-6: Conservation of species relationships which describe

transport of a tracer solute in the media.
TRANSPORT OF PROTEIN AND LIPID 309

x = 0 (Intimal Endothelium)
Cf VC f aC f
VC p( 1 - RE) + KE(C p - E )
f Ef
- D-
ax

transport across vesicul ar convection diffusion

junctions transport in media in media

x =L (Medial-Adventitial Border)
VC f Cf Cf
-Y-f (1 - RL) + KL( -E - C)
L + KC( -E - CP )
f f

transport across vesicular vesicular convection diffusion

junctions transport to transport from in media in media
(lymphatic) lymphatics capillaries
t = 0

t > 0

apparent mass transfer coefficients (length/time) for

vesicular transport across (E) intimal endothelium,
(L) lymphatic, and (C) capillary endothelium cells

K NVK, where N = vesicle flux [nunDer/(time x unit intimal

surface area )), v = vesicle volume, K = partition coefficient
phenomenological rejection coefficients for simultaneous
convection and diffusion across junctions of (E) intimal
endothelium and (L) lymphatics.
tracer concentration in (p) plasma and (L) lymphatic vessels

Fig. 6-7: Boundary conditions and initial conditions for

transport of a tracer solute across rabbit arterial wall
following injection of tracer.
310 CHAPTER 6

The permeation rate into the smooth muscle cells is assumed

to be a linear function of the concentration difference.
Finally, a linear kinetic expression is assumed for the rate
of disappearance of protein label by catabolic processes
within the cells. These hypotheses are reasonable for the
case of tracer kinetics which applied to our experiments.

The boundary conditions at the intimal endothelium are

derived from a balance which equates convective and diffusive
(including pinocytotic) transport across the endothelium to
similar processes in the immediately adjacent media. We
have combined convective and diffusive phenomena in the
intracellular junctions and have expressed the transport
rate therein in terms of an effective rejection coefficient.
A comparable relation holds at the medial-adventitial interface;
it includes both convection and diffusion into the lymph but
only diffusion from the capillaries. By considering diffusion
as the dominant mode of transport between capillaries and
lymphatic vessels, we have implicitly assumed that lymph
formation in the adventitia results from the sum of transmural
volume flow and from a very small difference between a large
local fluid filtration and an almost equally large fluid
reabsorption, either in the same capillary or in different
capillaries. Although this assumption may not be valid, the
predicted profiles are not sensitive to this aspect of the
boundary condition at the medial-adventitial border.

The equations to be solved are shown in dimensionless

form in Figure 6-8. Among the independent dimensionless
parameters, the most notable are the Fourier number, which
is a measure of the elapsed time following injection divided
by the relaxation time for diffusion in the media; the
Peclet number, which is a measure of the relative rates of
convection and diffusion in the media; three Biot numbers,
each of which is a measure of the diffusive mass transfer
resistance in the media divided by the diffusive resistance
at a specific boundary; two Thiele moduli, each of which is
a measure of the relaxation time for diffusion in the media
divided by the relaxtion time for a specific reaction; and
an internal permeation-diffusion modulus which is a measure
of the relaxation time for diffusion through the media
divided by that for permeation into smooth muscle cells.
By use of Laplace transforms, we have obtained an analytical
solution to the general problem when all parameters are
assumed to be independent of position. We have also obtained
an analytical solution for the case when plasma concentration
changes with time, through the use of Duhamel's superposition
integral, as shown in Figure 6-9.
TRANSPORT OF PROTEIN AND LIPID 311

DIMENSIONLESS RELATIONS

a26 f a6 f f2
an 2 - Pe ail - v (6 f - V6 i )

a6 b
a:r

f2
- ¢. 6. + -( 6 f - v6.)
1 1 V 1

6f
o pe6 p {1 - RE) + B (6 --)
E P Sf

Peef 6f 6f
n= ~f (l - RL) + BL ( sf - 6 p) + BC{ sf - 6 p ) [Pe6 f -a6f]
an-
T =0 6p

T > 0

Dimensionless Parameters

Concentration 6.
J
C.
CPo
I·
=.....:L. J= f , b ,1.
L,p
Internal
Permeation-Diffusion
Dt
Time (Fourier No.) T =:z Binding Equilibrium
L

X
Distance n =[ Volume Fraction Ratio

Peclet No. Biot Nos. j=E,L,C

j=b,i
Thiele Moduli n=l ,3 Rejection Coefficients RE, RL

Fig. 6-8: Equations from Figs. 6-6 and 6-7 expressed in terms
of dimensionless parameters.
312 CHAPTER 6

6p = 1, all T

1j! (r" T) 8 f + 8 b + 0. , f(r, , , ... )

ep 8 p (T)

T de (~)
8 T h,T)
cp-
(T
,~( r, T) + f li-'(r,T-U ~ d~
0
o d~

(T 1
8Th)
c;- f
0
8 T(1l,T)dll
0

Fig. 6-9: For a step change in plasma concentration, the

solution to the equations in Fig. 6-8 is given by "'V ("1') T).
Integration across the wall provides the solution for the
mean concentration in the wall, 0T (\").
TRANSPORT OF PROTEIN AND LIPID 313

This permits us to analyze the effect of the time-varying

plasma concentration levels which actually occurred in our
experiments (Figure 6-2), thereby permitting comparison
between theoretical prediction and experimental data.

Even though many simplifications were made in deriving

this theoretical model, it is nevertheless complicated and
contains many adjustable parameters. A wide range of these
parameter estimates have been investigated in order to find
one or more combinations which yield reasonable agreement
between theoretical prediction and experimental data.
Specific combinations of parameters for three illustrative
examples are tabulated in Figure 6-10. Two of these are
compared with data from the 10-min and 3D-min LDL experiments
in Figure 6-11. Theoretical prediction is plotted for cases
1 and 2 which are based upon the occurence of only diffusion
and convection (no metabolic phenomena) in the media. Near
the intimal surface, the predicted profiles for case 1
approximate the experimental data nicely, but agreement is
less satisfactory for case 2. Prediction for case 2 compares
favorably with experimental data in the center and outer
media but not near the intima where the measured concentration
gradient is much steeper than that predicted by theory.
Numerous parameter combinations have been tried to improve
the correspondence between prediction and data. However, it
was not possible to fit the data precisely over the entire
thickness of the aorta with a single set of parameters. The
two values of the LDL diffusion coefficient in cases 1 and
2, which differ by a factor of about 6 and may bound the
true value, are approximately a factor of 100 lower than the
diffusion coefficient in solution. Such a large reduction
may not be surprising in view of the microporous and heterogenous
nature of the arterial wall.

At times longer than 30 minutes, the concentration

profiles cannot be described by any single combination of
parameters which includes diffusion and convection only, as
shown in Figure 6-12. The predicted profiles for case 2 are
clearly unsatisfactory at 4 hours and longer. Conversely,
when binding, cellular permeation. and intracellular degradation
are included in the model (case 3), the agreement between
theoretical prediction and data is much better. These findings
suggest strongly that metabolic phenomena play a very important
role in the transport of LDL in the arterial wall.
314 CHAPTER 6

Example Cases
Variable
Parameters 2 3

D (cm 2/sec) 1.2 x 10- 9 7.5 x 10- 9 7.5 x 10- 9

KC (em/sec) 1.1 x 10- 8 3.3 x 10.- 8 3.3 x 10- 8
KL (em/sec) 4.1 x 10- 7 3.3 x 10- 6 3.3 x 10- 6
------------- --- - --------

Pe o. 1 0.1 0.1
0 0 0.3
0 0 10
0 0 10
m 0 0 0.2

BE 0.3 0.05 0.05

BL 8 10 10

0.2 0.1 0.1

BC
T 10 min 0.0075 0.05 0.05
30 min 0.0225 0.15 0.15
4 hr 1.2 1.2
24 hr 7.0 7.0
67 hr 19 19

" Diffusion and

¥ I

Binding, Cellular
Convection Only Permeation, and
Intracellular
Degradation Included
Fixed Parameters:

L 96 ]Jm E: = 0.42 v o

KE = 1.6 x 10- 8 em/sec R = 0 RE =

L

Fig. 6-10: Parameter values used in illustrative examples.

TRANSPORT OF PROTEIN ANO LIPID 315

o IOmn LDL DATA

.t:. 30min LDL DATA
-THEORY

Q
x
'(J
~g
u
5l:irn
0::0
I-
Z
w
o IOmin LDL DATA
u .t:. 30min LDL DATA
§~ - THEORY

~
1-0
~
l:i
-l
~Q!~~~=-~~~~~~~~
o 0.2 0.4 0.6 0.8 10
NORMALIZED DISTANCE. X/L

Fig. 6-11: Comparison between theoretical prediction and

experimental data for l25I-LDL distribution in rabbit thoracic
aorta. Theoretical prediction includes diffusion and convection
only. Parameters of case 1 (Fig. 6-10) are used on the lower
graph; parameters of case 2 apply to the curves on the upper
graph.
 40 1 w
<>-

10MIN
I 24 HR
•
.. 30 MIN ~ ~ •
• 67 HR
." .. 4 HR

30 1-" """-
""
'{
2 0 I\-
"
4HR~
" "\,
I 0I " I VI I I .., --" ""-
".., ..
o LI_ _~~~-L __~__~-L__~__L-~__~
o 0.2 0 .4 0 .6 O.B 1.0 0 0.2 0.4 0 .6 0 .8 1.0
NORMALIZED DEPTH, x/L (')
I
l>
Fig. 6-12: Comparison between theoretical prediction and experimental data.
-I
"m
diffusion and convection only (case 2); binding, cellular permeation, and intra----
:JJ
cellular degradation also included (case~ Q')
TRANSPORT OF PROTEIN AND LIPID 317

Further analysis and refinement of the model parameters

is underway. Nevertheless, even at this early stage we have
been led to several important findings with respect to the
transport of 131I - LDL in the arterial wall of normal rabbits,
First the endothelium is the dominant barrier to transport,
and permeation rates across it are consistent with uptake
solely by vesicular transport. Secondly, the solute concentration
profiles demonstrate a predominatly diffusive, rather than
convective, character. Although it appears not to playa
major role in transendothelial transport in normal rabbits,
convection may play some role in the removal of LDL from the
media to the lymphatic vessels. Quantitative estimation of
transport parameters showed that the mass transfer resistance
of the endothelium may be as much as 20 times larger than
that of the entire media, despite the fact that the thickness
of the endothelial lining is less than one percent of the
medial thickness. Thus, a very large concentration drop
occurs across the layer of normal endothelial cells, even
under steady state conditions. These findings suggest that,
to a first approximation~ the LDL concentration to which
medial smooth muscle cells are exposed in the normal rabbit
is less than five percent of their plasma concentration. An
increased permeability of the endothelial barrier could
therefore lead to a major change in the milieu of the medial
smooth muscle cells in terms of the concentration of LDL and
other plasma constituents to which they are exposed.

DR. DEWEY: I want to bring everybody's attention to

some recent work by Dr. Michael Gimbroni. His group, at the
Peter Bent Brigham Hospital
Local Control of Cellular in Boston, has been studying
Functions various kinds of chemical
synthesis processes in the
endothelial cell itself (4). One of the findings that I am
not in a position to evaluate but I think is very interesting,
is that the angiotensin level in the endothelial cell is
governed by chemical processes which are occurring primarily
in the cell itself and are not determined by the circulating
levels in the blood. If that is true, it means that there
may be vaso-active compounds which do, in fact, locally
control the state of the endothelial cell and its properties
and by inference, also control transport to the subendothelial
space. I can conceive that if you upset the balance of the
angiotensin by increasing its plasma concentration, that you
could, in the initial instance, change the character of the
endothelial barrier. After some time, the cell itself and
the autonomous chemical processes in the cell could then
compensate for this change.
318 CHAPTER 6

But in the initial instance when you change the angiotensin

concentration in the plasma, you would have essentially a
wave of infusion of material across the endothelial barrier
which would, in fact, appear as a "bump" at a later time
after the autonomic processes took over the restored ordinary
transport across the endothelial cell. If the barrier to
diffusion were broken down on a short time scale, on the
order of minutes, there would be a large transendothelial
infusion after which other processes would have taken over
which would produce a lower concentration at the surface but
produce the high concentration observed further into the
arterial wall.

DR. SCHWARTZ: Could there be an interaction of epinephrine

with angiotensin in these frightened animals?

DR. COX: Yes, as a matter of fact, the effects of

angiotensin are mediated in part by augmenting the effects
of circulating catecholamines.
Interaction of Angiotensin That is one of their actions
with Circulating Catholamines as well as their direct action.
In addition, there are liberated
catecholamines and these certainly produce further augmentation
of any kind of a response.

DR. BJORKERUD: Rabbits are rather specific in this

respect. You can induce medial sclerosis just by taking the
animal and putting it in an upright position or turning it
upside down. They are very sensitive to stress, so a number
of things could happen by rather trivial manipulation as
e.g. blood sampling. We sampled blood each hour in a series
of rabbits and many of the animals developed severe medial
sclerosis.

DR. STONE: What anesthesia did you use on those

rabbits?

DR. COLTON: Sodium pentobarbitol.

DR. STONE: The anesthesia may add a further complicating

factor. It will potentiate a norpinephrine effect.

DR. CAREW: I would like to come back to your challenge

to the tissue culturists. Being a tissue culturist, I would
suggest conversely that there are certain methodologic things
which we have been able to get a handle on in tissue culture
which are a little bit more difficult to handle in the other
situation.
TRANSPORT OF PROTEIN AND LIPID 319

And that is the question of how good tracer either the 1-131
labeled albumin or the iodide
Problems of Isotopic labeled LDL is? It is a difficult
Labeling thing to test, how good a tracer
is? If it is not an adequate
tracer, if it is bound in either greater or lesser quantities
than is the tracee, then all of the suppositions with regard
to what are the dominating factors may be influenced. I would
cite, particularly, some evidence -- I can't remember the
author at the moment - but, having looked at 1-131 labeled
albumin uptake by macrophages, he found that in halogenated
albumin preparations, uptake of the albumin was enhanced
many fold over that which was apparently more physiologically
labeled. In our own studies on labeled LDL, what we do is
present low density lipoprotein at a given concentration to
cells and then vary the ratio of labeled/ unlabeled and make
sure that the tracer is an adequate tracer and in most cases
that is the case. But there are occasional cases where
apparently the labeled material is denatured and it is taken
up to a considerably lesser degree than the unlabeled material.
So I would just offer that as a caution, I am not sure that
I have an answer as to how it could be tested in the in vivo
situation.

DR. COLTON: We have exercised great care to ensure

that our labeled LDL was not denatured. Chromatography of
labeled LDL on Sepharose 4B and on Sephadex G-200 resulted
in a single radioactive peak, indicating that neither aggregates
nor smaller labeled molecules were present. Immunoelectrophoresis,
double immunodiffusion and paper electrophoresis also failed
to indicate the presence of any contaminating proteins.
Extraction into chloroform-methanol indicated that less than
4% of the TCA-precipitable radioactivity in the injectate
was associated with lipid.

DR. WERTHESSEN: I would like to take the Chairman's

prerogative here to make a comment. I suggest that you use
a built in isotope when you use a tracer. A very good one
is Carbon 14, biologically and chemically synthesized into
the desired molecules. This can be very easy to do if you
don't mind using a lot of 14 C acetate.

Popjacks' classic method of the 40's to make labeled

cholesterol by feeding an egg laying hen labeled acetate has
often been repeated. A good hen can lay a dozen eggs and
also provide a liver containing labeled lipids (5).

DR. COX: Most often our animals are humans.

320 CHAPTER 6

DR. WERTHESSEN: Well, it becomes more difficult in

your case. But in your tissue culture work, you can use
tissue in vitro to build the tracer you want. And you can
get very high counts via these preparations.

DR. COLTON: You are talking of building it into what

part of the molecule?

DR. WERTHESSEN: Just let the appropriate tissue build

the tracer that you want from an appropriately labeled
substrate.

DR. COLTON: We have been interested in developing a

double label which has 1251 on the apoprotein and 14 C on the
cholesterol ester so as to avoid the artifactual results
associated with exchange of free cholesterol, by sequentially
contacting the LDL with red cells. We have succeeded in
doing this in vivo, and produced 99+% of the label on the
cholesterol ester. The problem is that the yield is low and
the cost is prohibitive for the large amounts needed in our
experiments.

DR. SCHWARTZ: I think that Dr. Carew's comment is

really very timely. One has to be cautious with the isotopic
cholesterol as a lipoprotein label, not only because of
unesterified and cholesterol exchange, but also because of
the possibility that cholesterol esters might also exchange.

DR. CARO: We have Ralph Dell with us at the moment

from Columbia University, and he has reported the same
problem, studying the uptake of radioactively labeled albumin
by rabbit vessels in vivo. His results suggest that denatured
label and free iodine are interfering very significantly.

DR. WEINBAUM: We have done, as you might guess, a lot

of modeling. I want to describe how we are going to use the
vesicle data that Colin Schwartz
Steady State Vesicle and Dr. Palade's group have
Diffusion Model collected. Those of you who are
interested in the details of the
theoretical model are referred to our recent paper in the
Journal of Fluid Mechanics (6). Fig. 6-13 is a schematic of
the mathematical model 0 A typical transendothelial ditfusion
distance is about 3200 A. The vesicle diameter is 700 A.
The vesicle diffusion problem is complicated by the two
features which I mentioned yesterday. When a vesicle comes
near a wall it undergoes a strong hydrodynamic interaction.
 -l
:0
»
z
en
""tl
o
:0
LUMINAL SURFACE -l
o
o
"T1
VAN DER WAALS FORCE RANGE (15- 100A ) ""tl
x=o x=o :0
o o
___ _-_-::-::_-..:-::-x =t -l
200A_ m
o
Z
350A-! »
kSICLE z
DIFFUSION 0
RELE::~Y o
r
¥" DISTANCE ""tl
o
2500b FREE VESICLE
o
350 A

o \VESICLE RELEASE
2850~ ATTACHED VESICLE
_~ x=I-Y
3000A_ - - - - - - -_ ___
- -II'" x -_ I - t
o
3200A
"- x=!
x -1 r- 2OOA LSLUMINAL SURFACE
Fig. 6-13: Schematic illustration of mathematical model for the Browncain diffusion of a
plasmalemma vesicle across an endothelial cell. Dashed line plane of releases; unequal
dashed line effective range of Van der Waals forces. From Weinbaum and Car a (4-1)

w
""
322 CHAPTER 6

The effective hydrodynamic resistance is a sensitive function of

its spatial position within the endothelial cell. The Van der
Waals force on a vesicle as it approaches the plasmalemmas has
been treated as the sum of all binary interactions between
molecules distributed over the surface of the plasmalemma and
the vesicle. In this first paper, we consider only the steady
state problem. This corresponds to an in vitro experiment with
no disturbances present. The key result one wishes to obtain
from the steady state model is the ratio of the time a vesicle
stays attached to the plasmalemma to the time it takes to
diffuse model. The diffusion time can be determined from
time-dependent vesicle diffusion experiments by looking at
labeled vesicle density profiles at short time intervals after
fixation. The attachment time is then determined from the
ratio provided by the steady state theory. Figure 6-14 shows
the steady state vesicle concent~ation profile near the plasma-
lemma for vesicles released 200 A from the luminal surface.
This is the length of a typical vesicle neck. Note that the
vesicle concentration rapidly approaches zero for small
spacings. The attractive Van der Waals force very quickly draws
the vesicle toward the plasmalemma for small fluid gaps.

Figure 6-15 shows the vesicle concentration profile in

the interior of the cell for vesicles released at the luminal
surface. There are several different curves each one corres-
ponding to a different effective range of the Van der Waals
forces. The number of free vesicles is the area under each
of the curves. The crucial data that we need to determine
from the EM studies is the ratio of the number of attached
to free vesicles. Vesicles are also migrating from the
abluminal surface towards the lumen. The concentration
profile for these vesicles is just the reverse of the
curves I have shown. In Figure. 6-15, is the number of
vesicles released at the luminal membrane per unit area
per second. 0 R is the vesicle number flux per unit area per
second that cross to the other side of the cell. The
probability 0R/0 of a vesicle crossing the cell is a function
of the size of the vesicles, where they are released, and
the effective Van der Waals force distance. The next
result ta/tn is the ratio of how long a vesicle stays attached
to how long it is free. Na and Ng are the numbers of attached
and free vesicles per unit endothelial surface. F is the
interaction function obtained from the theory. The ratio
ta/tn depends on Na, Nf and F.
TRANSPORT OF PROTEIN AND LIPID 323

1.2

1.0

0.8

cDo/(Aa)
0.6

0.4

ro"
0.2

0 ----
0 0.5 1 1.5 2 2.5 J
Xho

Fig. 6-14: Dimensionless vesicle concentration profile cDo/Aa near

plasmalemma in region where Van der Waals forces are important.
Eo effective Van der Waals force range X/EO scaled normal coordi-
nate. Dashed curve concentration profile if no Van der Waals forces
present; solid curve concentration profile modified by Van der Waals
forces. Theory in Weinbaum and Caro (4-1).
 324 CHAPTER 6

0 . 40 1 o
y (200 Al
I"-- x

0 . )0
C Do

0 . 20

0 . 10

0 .0
o 0.1 0.2 0.) 0 .4 0·5 0.6 0.7 0.8 1.0
y

Fig. 6-15: Dimensionless concentration profile cDo/£¢ for free

vesicles outside of region of Van der Ivaals force interaction.
X = Y plane of vesicle release and Van der Waals cut off distance
defined in Weinbaum and Caro (4-1). X normal coordinate scaled
relating to transendothelial diffusion distance £. Curves shown
are for vesicles released at luminal surface. Curves for vesicles
released at albuminal plasmalemma are a mirror image.
TRANSPORT OF PROTEIN AND LIPID 325

Based on capillary data for Na and Nf and the effective Van der
Waals force range between roughly 75 and 15 angstom units, I
obtained ratios of ta/tD somewhere between 1/10 and 1/14. Thus,
for capillaries the diffusion time is roughly four to ten times
as long as the attachment time. I believe that the values of Na
and Nf are very different for arterial endothelium. Thus, the
ratio ta/tD will be substantially different.

DR. WERTHESSEN: Which way?

DR. WEINBAUM: I think the attachment time is going to be

much longer, that is why when we look at freeze fracture EMs,
you see such a high density of attached vesicles. I think you
will find many more attached vesicles than free.

DR. SCHWARTZ: Hy impression is that one normally sees

considerably more free vesicles than caveolae. The relative
numbers may depend on whether one is looking at enface prepara-
tions or transfer sections.

DR. WEINBAUM: We are going to have to do extensive number

counts in the future for Na and Nf in arterial endothelium. The
other thing which we have already done is to combine the diffusion
model, just described, with the model for the underlying tissue.
This is an alternate way of obtaining the value of tD. The first
approach which I have already mentioned is to use time dependent
EM tracer studies for vesicle labeling. These studies would
be more reliable, however they are not yet available. We have
developed a model for the underlying tissue which is very much
like the one Dr. Colton has already described in detail. In
Figure 6-16 I will describe the basic differences between the
models. Dr. Colton's model corresponds to an in vivo situation
where the correct boundary condition at the adventitial surface
is very complicated and one also has to worry about the vasa
vasorum. For these reasons we have treated the simplest
experimental situation an in
An In Vitro Model in vitro isolated artery segment
with an outer bathing solution at
essentially zero concentration. The model assumes a homogeneous
distribution of interstitial fluid with a dispersed cellular
phase of smooth muscle cells. The volume fraction of each
phase was determined by measuring with a polar planimeter, the
area actually occupied by smooth muscle cells as compared
to interstitial fluid cross-sectional area.
326 CHAPTER 6

0.3 0 r----------------------------------------------------,
,
0.25 " " , "~ ~-ta
tD
for_a_
N

Nf
0.714

d. 0.20 ..... .........

-_._-
TR ta
A.
-7
rR .....
. . . . . .......... ..... .....
cj; 1J ¢
' ...... ............
- .......... ..... .....
---
0.15

...... .....
0.10
I
I
-~----------- t N ---
_a_ f o r _a = 0.417 --
1-25 ~ tD Nf
I
0.005 0.01 0.015 0.02 0.025 0.03 0.035
E
Fig. 6-16: Solution for fraction ¢R/¢ of vesicles released at
luminal front that cross cell and attach at abluminal plasmalemma.
ta/tD ratio of vesicle attachment to diffusion time. Na/Nf ratio
of the member of attached to free vesicles. Numbers based on
capillary endothelium. £ Van der Waals cut off distance. From
Weinbaum and Caro 4-1.
TRANSPORT OF PROTEIN AND LIPID 327

DR. COLTON: Do you have estimates for the volume fraction

of smooth muscle cells in the arterial wall?

DR. WEINBAUM: I do have those numbers. They are in the

paper I mentioned before (6).

DR. CARD: Bill Keating, and I haven't got the reference,

says that his results showed a marked variation of the ratio of
interstitial to extracellular volume proceeding from the intima
toward the adventitia in an artery. The volume of the ratio
increases proceeding radially (7).

DR. WEINBAUM: It varies. The cellular area can be as

high as 85% in some regions of the intima and can decrease to
60% in the media. One of the things that we plan to examine
more carefully is the spatial density distribution of smooth
muscle cells. Dr. Colton's model is very general. It introduces
many phenomenological coefficients which have to be determined
experimentally. Convection is not present and the complicated
in vivo boundary condition at the adventitial surface is
eliminated. It is only a three parameter model. One is the
volume fraction of the two phases just discussed, and the
other two are the dimensionless transport coefficients. Now
this was very appealing to me becaue I could immediately
use uptake experiments available in the literature like
Dr. Fry's (8) or Sif1inger, Parker and Carols (9). To
determine the two transport coefficients you need two sets
of data, one for the normal artery and the other for an
artery specimen with the endothelium removed. Unfortunately
in these in vitro experiments, we don't have concentration
profiles like those Dr. Colton has obtained in his in vivo
studies (1) This is something we shall want to do in the
future. Dr. Harold Weyland at Cal. Tech. has been measuring
the interstitial fluid diffusion coefficient using other
techniques than those discussed here. These measurements
provide an independent check on one of the two transport
coefficients determined by curve fitting the theoretical
model results to Dr. Fry's or Dr. Carols data. The values
obtained are in rather good agreement.

DR. MALLIANI: The electrostatic charges which are on

the walls of the vessel may be changing continuously. As
far as I know, the e1ectro-
Electrostatic Charges static charges are dependent
also on the electrogenic activity
of the contractile elements. Thus, in vivo, the active muscle
tension, which is also modulated by reflexes, may modify these
electrostatic properties.
328 CHAPTER 6

DR. WEINBAUM: We are gOing back to the very simplest

possible experimental situation to try and understand the
dynamics of vesicle motion. The things that you mention
that happen in vivo may well be more important than anything
that happens passively. I don't know.

DR. MALLIANI: The point is that the wall of a vessel

which undergoes active contractions may attract more vesicles,
or less vesicles. In short, the electrostatic properties may
change in a more complicated manner with respect to what can
be simulated in vitro.

DR. WEINBAUM: The capillary experiments are all, by the

way, in vivo experiments. This includes most of the previous
work done by easley-Smith, Palade, Shea and Karnovsky, and by
all of these other people who have been doing EM studies on
vesicle transport.

DR. DEWEY: One very brief comment. It is indeed a

problem to get diffusion coefficients for large molecules
through various solutions like interstitial fluid. There is
a technique which has been developed over the last few years
which has been quite successful in obtaining this type of
data and that is using a Laser-Doppler technique to measure
Browning diffusion, thereby obtaining the diffusion coefficient.
Assuming that one could extract the interstitial fluid it
should be possible to get accurate numbers -- certainly
within a factor of two -- for those diffusion coefficients.

DR. STONE: May I add a comment to that. There are

data similar to that if you will accept one small possible
difference. Guyton and Kurt Winderhelms have used implanted
plastic balls to obtain tissue pressure measurements
(10). The fluid inside the ball has been shown to be very
representative of interstitial fluid.

DR. COLTON: The parameter one wants is not the diffusion

coefficient in interstitial fluid which should be very close
to the value in water -- rather, one would like to have the
value in the extracellular space in the presence of collagen,
elastin, and glycosaminoglycans. For that kind of situation
the Laser-Doppler technique would run into problems because
in that technique one would observe the movement of all the
macromolecules at the same time, including those which make
the extracellular matrix.
TRANSPORT OF PROTEIN AND LIPID 329

I also have a comment for Dr. Weinbaum. Even in the in

vitro experiments with albumin, there is the possibility
that intracellular degradation plays a role. Our data
suggest it occurs in vivo and significantly influences the
concentration profiles four hours after injection.

DR. WEINBAUM: The experiments that we have done so far

have lasted roughly, at most, two or three hours. Professor
R. Pfeffer, my colleague
Kinetic Model at City University, has been
looking into development of
kinetic models which are almost identical to the ones
Dr. Colton talked about. The reason I was worried about the
comparison with the Evans blue dye experiments is that we
know that the dye does not stay attached to the albumin
molecules and a kinetic model is needed to describe this
reaction. The albumin itself is inert. I discussed this
with Derek Bergel and he felt that it was a reasonable
assumption.

DR. SCHWARTZ: I just want to make a point to

Dr. Weinbaum, because it could be helpful. A number of
people have done fairly thorough stereomorphometric studies
on the vessel wall. Ross Gerrity, one of my colleagues at
McMaster University has done this with respect to the rat
aorta (11). These data permit one to define the relative
volumes of all the components present. I suspect this would
give one a better approximation for theoretical transport
studies rather than to assume a totally homogeneous media
for diffusion.

DR. SCHWARTZ: Now I should like to review some aspects

of endothelial structure and function, with particular
reference to permeability.
Endothelial Structure There are a number of processes
and Function Regarding involved in transport; one that
Permeability Transport is particularly interesting is
the likely possibility that
particles of differing size are taken up in plasmalemmal
vesicles and then transported bi-directionally across
the vessel, some being discharged on the albuminal surface
of the cell. This process has been termed vesicular transport.
Siminescu, Siminescu and Palade (12) have demonstrated that
vesicles may coalesce to form a continuous patent channel
across the endothelium, providing a variation on the theme
of vesicular transport.
330 CHAPTER 6

Another mechanism may combine both vesicular and junctional

transport. The overall importance of junctional transport,
however, remains uncertain. Junctional transport is theoretically
feasible, as demonstrated by the passage of horseradish
peroxidase, but the extent of intercellular transport of
this probe appears to depend not only on dose, but duration
of exposure. Simple diffusion of molecules across the
endothelium may also occur. If this process is important, it
is likely to be restricted to small molecules such as ions
and water. These and other aspects of transport have been
recently reviewed in some detail elsewhere (13).

I would like to further discuss aortic permeability,

employing the Evans Blue model to illustrate selected facets.
When the protein-binding azo
Evans Blue Model dye Evans Blue is injected
intravenously, the dye exhibits
a specific uptake pattern in the aorta in and around
the intercostal ostia and the main brachiocephalic branches,
and on the flow dividers at the aortic trifurcation. If one
studies radioiodinated albumin uptake at two hours, for
example, in blue versus white areas, there is a significantly
greater uptake of the l25I-albumin into aortic tissue from
blue as distinct from white areas. Additionally there is
indeed a significantly greater l25I-albumin uptake in the
thoracic part of the aorta relative to the upper or lower
abdominal segments (14).

Interestingly, if one studies the distribution of the

isotope across the aortic wall, activity is greatest in the
innermost intima and media and least in the outer media.
With albumin this transmural gradient exhibits a biphasic
curve, with a "secondary" peak at some 500)J from the endothelial
surface. At all levels across the aortic wall isotopic
activity is greater in blue than in white areas.

DR. WEINBAUM: Where are the vasa vasorum in the last

figure?

DR. SCHWARTZ: This is an interesting point. The vasa

vasorum are probably not all that far from the secondary
peaking and in all likelihood about 500p from the endothelium.

DR. CARD: Also, at what time •••

DR. SCHWARTZ: Those were two hour studies.

TRANSPORT OF PROTEIN AND LIPID 331

DR. FRY: Did you pressure rinse these animals to clear

the capillary beds in the outer media of the residual radiolabeled
plasma?

DR. SCHWARTZ: No. These were not pressure rinsed, so

there could be some vasa vasorum contamination. The serial
sections were rinsed carefully, however.

DR. CHIEN: Were these uptake measurements made at

physiologic pressures?

DR. SCHWARTZ: These were all in vivo studies.

Essentially similar data emerged for fibrinogen. We

used fibrinogen as a convenient and much larger molecule.
Again the studies were conducted
Transmural Gradients for on a 2-hour period. Fibrinogen
Fibrinogen and Albumin probably behaves like a molecule
1 x 106MW although its actual
molecular weight is much less. It is interesting that the
transmural gradient of fibrinogen is very much steeper than
that of albumin with no tendency to the secondary peak.

DR. CARO: Again at two hours?

DR. SCHWARTZ: Again at two hours. I would be interested

in your comments on why we should get this difference in the
slope of the transmural curves for albumin and fibrinogen.
Isotopic cholesterol also exhibits a distinct transmural
gradient with the activity greatest in the intima and least
in the outer media. The cholesterol distribution probably
does not to any great extent reflect lipoprotein movement
across the vessel, but rather a physico-chemical exchange of
unesterified cholesterol. Approximately 80 to 90 percent of
the label across the aortic wall is associated with unesterified
cholesterol and only 10% of the label is associated with
cholesterol ester. Additionally, it is interesting that the
gradients are associated with unesterified cholesterol, the
transmural distribution of cholesterol ester activity showing
no slope differences across the aortic wall.

DR. DEWEY: Inasmuch as the gradient is so dramatic at

the luminal side, and it is well understood that these are
difficult sections to do, it might be that in certain of
these experiments the gradient prior to sectioning is in
fact even steeper than shown in these pictures.

DR. SCHWARTZ: Yes, this is possible.

332 CHAPTER 6

DR. DEWEY: Another way of saying it is that the penetration

from the luminal side may be incredibly small.

DR. SCHWARTZ: On the contrary, the data suggest

that penetration from the lumen is greater than entry from
the adventitia. We did another set of experiments to examine
the possible problem of non-specific binding to the endothelial
surface. The endothelium was removed by doing Hautchen
preparations. With fibrinogen, only 5% of the total activity
of that particular area was removed with the endothelial
monolayer, leaving in excess of 90% of the label in the
underlying media.

DR. CAREW: I have two questions really. First, a

methodological question -- how was the cholesterol presented?

DR. SCHWARTZ: The cholesterol in these experiments was

presented initially by adding the isotopic cholesterol in a
very small volume of ethanol to a large volume of plasma,
mixing, and allowing the mixture to stand for 60 minutes.
By this time the cholesterol had exchanged with the plasma
lipoprotein cholesterol; so the cholesterol was in fact
introduced as lipoprotein cholesterol. Particulate cholesterol
was not present in the preparation, as the mixtures were
filtered prior to injection. Isotopic cholesterol was also
introduced by mouth. The resulting uptake patterns were
similar to those obtained by intravenous administration.

DR. CAREW: The next question probalby relates to Dr.

Dewey's question that it is surprising that the gradient
appears steeper with time. Do you have any explanation for
its being steeper between the first and second sections?

DR. SCHWARTZ: I have no explanation. I would certainly

be interested in any comments you may have on these data

DR. FRY: The concentration gradient as measured by

this technique might be expected to rise with time in the
subendothelial region particularly in the presence of an
intact endothelial and basement membrane barrier. The gradient
as measured is proportional to the difference in the average
concentration of radiolabeled substance between successive
tissue sections. Since the initial concentration distribution
across the wall would be very nonlinear with the distance
(x) across the wall, i.e., very convex toward the x axis,
and only gradually assume a more linear configuration with
the approach to the steady state, the average concentration
in the first slice will increase with time more rapidly than
in the second slice.
TRANSPORT OF PROTEIN AND LIPID 333

Accordingly the difference in the average concentration between

the two successive slices (the measured gradient) will increase
with time while the true gradient may in fact be decreasing with
time.

DR. CHIEN: The gradient is shown here on a linear scale.

In order to make a fair comparison of different slopes, one
should really make semilogarithmic plots.

DR. SCHWARTZ: That is a good point.

DR. CHIEN: If a semilog plot is made, I don't think

there is that much difference in gradients.

DR. SCHWARTZ: To recapitulate, in the normal, healthy,

six to eight week old pig one can identify areas of spontaneously
differing aortic permeability to proteins with recourse to
perfusion with a trypan blue dye (Evans blue). Areas of greater
permeability, or blue areas, show a variety of fine structural
features (15). The endothelium is relatively cuboidal, and
the intercellular junctions tend to be short and abut end to
end without any interdigitations. The subendothelial space is
considerably thickened and edematous, and contains many cells,
many of which are undifferentiated. In the subendothelial
space there is copious scattered elastin and collagen, with a
floccular matrix or low electron density.

By contrast, areas of lesser permeability to the blue dye

(white areas) differ significantly. The subendothelial space
is thin and inconspicuous containing some floccular material.
Cells are present in the subendothelium, most of which have
the appearance of smooth muscle cells, while only a few are
undifferentiated. In the white areas, the endothelial cells
are characteristically thinner and elongated with long
interdigitating intercellular junctions. Cell death or cell
injury is seen more frequently in blue relative to white areas,
the dead cells occurring with a frequency of 2.9 in blue and
0.7% in white areas, respectively (15).

DR. SMITH: In the very young endothelium do you get

this cuboidal type of cell? I am wondering whether there
is a rapid turnover in the blue area so that, in a sense, there
is young endothelium. Would that account for the thicker cells
or not?

DR. SCHWARTZ: I think it could, Elspeth. Generally

speaking, younger cells tend to be more rounded and larger.
334 CHAPTER 6

Whether the turnover difference is sufficient to account for

the consistent morphologic differences observed between blue
and white areas is, I think speculative. It is more likely
that the more rounded, less polarized endothelial cells of blue
areas are reflective of altered hemodynamic stress in these
areas (15).

DR. LEE: It is interesting to note that in the blue area

endothelial cell death or injury occurs more frequently than
in the white area.

DR. SCHWARTZ: This is one of the important questions

which has not yet been resolved. I will come back to permeability
data in terms of ultra-structure in a minute. We obviously
considered the possibility that cell death might provide the
ultrastructural basis for an "ultra large pore" which could
contribute significantly to the observed permeability differences.
We have not been able to obtain evidence that these sites of
injury, death or ablative loss are indeed areas of increased
permeability. It is possible that platelets and proteins
which adhere to the surface may provide an impenetratab1e
barrier, thus preventing influx at these sites. At present
we have no evidence pro or con the "ultra large pore" resulting
from cell injury or death.

DR. LEE: We denude the artery by removing the endothelial

cells by mechanical means to accelerate the development of
atherosclerosis in various vessels. We presume that denuding
endothelial cells and thus removing the barriers between the
bloodstream and subendothelial space results in increased
diffusion of blood constituents and acceleration of development
of atherosclerosis.

EDITORIAL NOTE: Platelet adhesion to the endothelium

free surface has been observed. The platelet "mytogenic
factor" or "growth stimulating factor" described by Ross and
others may be involved in this process (16).

DR. SCHWARTZ: Over these spontaneously-occurring sites

of cell death or cell loss, we sometimes find an attenuated
degranu1ated platelet. How these platelets will influence
permeability at these sites is unknown.

DR. BJORKERUD: I think there is some evidence that

these are large pores, or super-large pores. I am referrring
to perfusion staining experiments with markers.
TRANSPORT OF PROTEIN AND LIPID 335

If a solution of uncomplexed Evans blue is used, we can

actually see that the cells that take up Evans blue also
pass Evans blue into the media. This indicates that injured
cells represent large "pores." I am also referring to
Shimamoto's experiments with coal particles in suspension
where he could see that coal particles actually pass into
sub-endothelial space at places with more frequent dead
cells.

DR. SCHWARTZ: With scanning electron microscopy, white

area endothelium exhibits regular contours, with a significant
degree of polarity in the direction of blood flow (Fig. 6-17)
Blue area endothelium on the other hand, exhibits a less
discrete polarity (Fig. 6-18). The cells appear more cuboidal
and raised into the lumen. Their surface is distinctly rougher
with many projections and blebs. These differences are entirely
consistent with the silver-stained !lautchen preparations of
endothelium derived from areas of spontaneously-differing
permeabili ty.

Another striking difference between blue and white

areas relates to the surface coating or glycocalyx over the
endothelium. Blue area endothelium stained with Ruthenium
Red reveals a dense inner glycocalyx layer closely applied
to the plasma membrane and a somewhat fibrillary, less dense
layer luminal to this. Ruthenium-staining material is present
in some but not all the plasmalemmal vesicles. This surface
coat is some 2-5 times thicker over the endothelium of white
areas. The role of the glycocalyx is of considerable interest.
It has to be part of any theoretical consideration of the way
in which molecules corne into contact with the cell surface.
It may present a simple chemical barrier, or it may function
like a gel and thus modify the ways in which molecules are
oriented at the surface. The glycocalyx may also control
entry to vesicles and may have a significant role in
determining the nature of receptor sites, or immunologic
reactions at the cell surface.

DR. CARO: Dr. Schwartz, does this also lie in a similar

thickness over cell junctions?

DR. SCHWARTZ: It is certainly present in the outer part

of the junction.

DR. CARO: What I had in mind is that it is going to

present a barrier there to the transport of small molecules
into the intercellular junctions. They would have to get
through that barrier.
336 CHAPTER 6

Fig. 6-17: Scanning electron micrograph of the endothelial

surface of a white area in the aortic arch of a normal six-week-
old pig. Endothelial cells (E) are ovoid in shape, with smooth
outlines and few surface projections, and exhibit a distinct
polarity in the direction of blood flow. Compare with Fig. 1-16.
X 2,000.
TRANSPORT OF PROTEIN AND LIPID 337

Fig. 6-18: Scanning electron micrograph of a blue area from the

aortic arch of a normal six-week-old pig. Endothelial cells (E)
are more cuboidal and raised than in white area, with less
distinct polarity (compare Fig. 1-15). Numerous surface
projections are visible (arrows) which are not as frequent in
white areas. X 2,000.
338 CHAPTER 6

DR. SCHWARTZ: We ask ourselves questions about junctions

in terms of transport phenomena. Now I would like to show
you a little of the recent data that is being obtained in our
laboratory on the transport of ferritin particles across
arterial endothelium. We have examined ferritin transport
at one, five and fifteen minutes after a single intravenous
pulse in the young, healthy pig. Transport of this particulate
probe is, in the aorta, exclusively vesicular, and there is
no evidence of transendothelial junctional transport.
Preliminary quantitative results of these studies are now
available. We have examined both blue and white areas, and
have quantitated the number of plasmalemma vesicles per
square micron of endothelium in cross-section, the percentage
of vesicles that contain ferritin, and the average number of
ferritin grains per vesicle. Blue-white differences in the
transendothelial transport of ferritin appear to be due to
the number of active vesicles. In terms of the Evans Blue
model a valid question could relate to the nature of the
biological control mechanisms for vesicular transport.

DR. DEWEY: Dr. Schwartz, I would like to ask you if

there are any chemical mechanisms that you or other people
have identified that would cause the transformation of the
endothelial cells from the type that you see in white areas
to the conformation you observe in blue.

DR. SCHWARTZ: Could you repeat the question? Are you

referring to the shape of the cells?

DR. DEWEY: You mentioned that they were in the two

different areas, either cuboidal or spear-like and this
difference in shape is clear from the scanning electron
microscope and all the other evidence. But are there any
chemical substances which would cause this?

DR. SCHWARTZ: The reason for moving to a particulate

probe and not to horseradish peroxidase specifically in this
set of experiments is explicit. We wanted to be able to
precisely quantitate at an ultrastructural level where the
probes are going, at one minute, before there is any significant
back diffusion, and at 5 and 15 minutes after injection. A
time sequence set of studies such as this is a reasonable
way to construct a dynamic picture of transport ultrastructurally.
We have avoided horseradish peroxidase for another reason,
namely the histamine release which accompanies this probe in
some species.
TRANSPORT OF PROTEIN AND LIPID 339

DR. DEWEY: I have already referred to the work of Dr.

Michael Gimbroni who found evidence for very substantial
chemical activity within the endothelial cell itself. There
is more than one vasoactive substance being produced, and
the list includes factors like angiotensin. To the extent
that chemical activity might be influenced by fluid shear
stress mechanisms or whatever you want to evoke, the chemical
activity in itself might cause the shape changes. So this
might be an indirect effect of some chemical variation.

DR. BJDRKERUD: If the Evans Blue-albumin complex is

transported through vesicles, then you are showing the same
parameter with ferritin as an object for transendothelial
transport.

That is, blue areas might be blue because of a larger

vesicular transport of Evans Blue-albumin complex which may
also apply to ferritin in your blue areas.

DR. MEYER: Are there some structural differences in

the internal elastic membrane between the white and blue
areas or some other structural differences in light micrography?

DR. SCHWARTZ: I am unable to say if there are any

structural differences.

DR. SMITH: My comment really fits in with that quite

well; coming back to the gradient, in the blue area you have
a very thick subendothelial layer. Now if you plot your
gradient in terms of where the internal elastic lamina
comes, it is going to come in a different position in the
blue areas than in the white areas. And I wonder how that
will affect the gradient. And thinking again along the
lines of this thickness, if you have an endothelial cell
which is twice as thick and you have the same number of
vesicles taking the same length of time, in a thick cell
won't you have, at any given moment, twice as many loaded
vesicles actually visible in the cell? In the thin cell
they would have got across and discharged where as in the
thick cell they would only have got half way across. Or is
that not true? In terms of the data that are involved at
this point of time. Because the number of vesicles with or
without particles is the same in both blue and white areas.

DR. SCHWARTZ: Well, that is not true.

DR. SMITH: Dh, I see. So the total number of vesicles

in the whole cell is the same, not just the same number on
the surface.
340 CHAPTER 6

DR. SCHWARTZ: In terms of the data we have, the number

of vesicles with or without ferritin grains is similar in
both blue and white areas.

DR. CAREW: You may have just answered part of the

question I was going to ask. It basically relates to whether
all vesicles are the same. You mentioned that you made a
distinction between vesicles and inclusions and I was wondering
if you could give us some size distribution of the types of
vesicles versus inclusions that you see.

DR. SCHWARTZ: After preliminary observation it became

apparent that it would be wise to separate the large vacuole
from the smaller vesicle and this we have done. I have
already commented also on the fact that some but not all
vesicles contain Ruthenium Red-staining material.

DR. CAREW: Is that your only criterion?

DR. SCHWARTZ: No, there are other possible distinctions.

It is just possible, for example, that some of the larger
vesicles or vacuoles are
Some Large Vesicles or lysosomal, or that they are
Vacuoles may be Lysosomal formed from the fusion of
pinocytotic vesicles. We have
ultrastructural evidence for the latter. The larger vacuoles
can be five to ten times the diameter of the vesicles.

DR. CAREW: Because much of the discussion has been

focused on the idea that all vesicles are identical, I
really wonder whether all
Possible Specificity for parts of the cell membrane are
Certain Types of Endocytic alike and in that regard there
Processes is some preliminary evidence
from Dr. Norman Miller in our
laboratory in La Jolla. His evidence suggests that the
internalization of low density lipoprotein and
sucrose may, in fact, be independent of one another,
indicating that there may be some specificity for certain
types of endocytic processes to the exclusion of non-specific
types.

DR. SCHWARTZ: This is a very important observation.

If you look closely at, say, Ruthenium Red staining, you
will note that not all vesicles contain glycocalyx-type
material. One could speculate that perhaps the amount or
nature of the surface containing glycoproteins may have
something to do with whether certain vesicles take up certain
types of macromolecules or not.
TRANSPORT OF PROTEIN AND LIPID 341

DR. CARO: What was the thickness of your sections? Is

the endothelial cell a large fraction of anyone section, or
only a very small fraction?

DR. SCHWARTZ: The first section was one hundred microns

thick and thereafter sections were cut at 200 microns.

DR. CARO: So really whether the cell is fat or thin

makes very little difference.

DR. SCHWARTZ: Certainly. The first section comprises

endothelium, SES, and inner media.

DR. CARO: You showed a very nice slide of ferritin.

I anxiously looked to see
Ferritin whether it was held up in
the subendothelial space.

DR. SCHWARTZ: Yes. I doubt if endothelial thickness

per se will account for the gradient differences.

DR. CARO: I asked because in my unpublished studies

with 14 C sodium acetate, in the isolated perfused dog
common carotid artery, the transport rate was about two
orders of magnitude higher than for proteins. It may well
be that a structure beyond the subendothelial space provides
a substantial barrier for low molecular weight materials,
whereas the endothelium provides the main barrier for higher
molecular weight materials. A preliminary account of these
experiments is contained in Cara, 1973 (17) The fact that
you don't see Ruthenium Red in all the vesicles could just
possibly be explained by the fact that some vesicles unload
again. They have been equilibrating with the fluid in the
subendothelial space and not with the luminal fluid.

DR. SCHWARTZ: Not specifically being held up in the

subendothelial space but rather within cells in the intima;
what we have found are significant differences in the blue
and white areas in the rate at which the subendothelial
cells phagocytose the ferritin particles. Not all the
subendothelial cells are phagocytic, and there may be more
than one cell type in the subendothe1ium.

DR. COLTON: I would like to second Dr. Caro's idea

concerning barrier properties of the subendothelial space.
In some uptake experiments in vitro with denuded aorta, we
have seen the kind of growth of the concentration on the
surface which would only be consistent with surface resistance,
one whose magnitude is much smaller than the endothelium but
nevertheless significant.
342 CHAPTER 6

DR. CHIEN: I will go over very quickly some of the

ultrastructural studies we have done on the common carotid
dog. Scanning electron
Ultrastructural Studies microscopic pictures of the
on Common Carotid Artery common carotid artery fixed at
zero transmural pressure (done
with collaboration of Dr. Mary Lee at Columbia) show many
parallel grooves inside the artery along its longitudinal
axis. (Fig. 6-19). Endothelial nuclei can be seen along the
grooves. Sometimes, the endothelial surface has a velvety
appearance, with very tiny foldings. Pictures taken with
the artery fixed at lOOmm Hg transmural pressure gradient
show that grooves are no longer visible, but the nuclei are
still present. The comparison between the zero pressure and
the lOOmm Hg pressure indicates that the surface geometry of
the arterial lumen, as seen by scanning electron microscopy,
varies considerably with the transmural pressure.

We have performed some preliminary studies on the

uptake of 125 1 labelled albumin by isolated canine common
carotid arteries at different transmural pressures. We
determine the arterial uptake of l25I-albumin following 15
minutes of incubation with labeled serum. The result of
radioactivity per unit artery weight was calculated as a
ratio to the serum activity. The uptake at lOOmm Hg was
higher than at zero mm Hg by approximately 60%, and this
difference is statistically significant. If the uptake
results are expressed in terms of activity per unit inner
area of the arterial lumen, then the data between the two
pressure levels are more comparable.

Now I would like to mention some of our studies on

transmission electron microscopy (done in collaboration with
Dr. Kung-ming Jan), which agrees
Transmission Electron with Dr. Schwartz's results. The
Microscopy transmission EM pictures show a
large number of plasmalemma
vesicles in the endothelial cell (Fig. 6-20). There are
many vesicles facing the intercellular junction. Thus, in
agreement with what Dr. Schwartz said, materials may be
transported by the vesicle and discharged into the intercellular
junction. With the use of horseradish peroxidase, Dr. Jan
showed their uptake into the plasmalemma vesicles of the
carotid artery endothelium. Wherever there are filaments in
the endothelium the vesicles are not present. Hence, macromole-
cular movements to the basal side of these filaments probably
are achieved by vesicle transport via adjacent areas.
 -i
:0
l>
Z
en
"'C
o
:0
-i
o
"'C
"
:0
o
-i
m
Z
l>
z
o
r
"'C
o

Fig. 6-19: Scanning electron micrograph of luminal surface of canine cornmon carotid
artery fixed at zero transmural pressure (M.M.L. Lee and S. Chien, submitted to Anat.
Rec. for publication) (80Ox) (Reduced 40% for purposes of reproduction.) w
:..
w
 w
~
~

Fig. 6-20: Transmission electron micrograph of canine common carotid artery showing
endothelial cells, internal elastica and smooth muscle cell. The internal elastica has a
(")
dark appearance because of tissue treatment with low molecular weight galloyglucose I
(tannic acid) after osmication. Note the vesicles in the endothelial cells and also in the »
-0
smooth muscle cell. 40,OOOX (Courtesy of Dr. Kung-ming Jan.) (Reduced 40% for --!
m
purposes of reproduction.) :0
O'l
TRANSPORT OF PROTEIN AND LIPID 345

Dr. Jan has performed freeze fracture studies on carotid

endothelium with Dr. Maia Simionescu in Dr. George Palade's
laboratory at Yale University.
Freeze Fracture Studies By freezing the specimen and
then fracturing it, one can
expose the interior of the endothelial plasmalemma membrane
just underlying the luminal surface. The vesicles can
appear as either a protrusion or a depression, depending
upon where the fracture plane is. The fractured surface is
nearly covered with tremendous numbers of vesicles. (Fig. 6-21).
Other pictures also show large numbers of these vesicles
everywhere in the endothelial surfaces, except at the intercell-
ular junction. Although the intercellular junction has many
vesicles at the cell-to-cell interface, it is essentially
devoid of vesicles at either the luminal or basal surface.
These vesicles can also be seen in the smooth muscle layer
underneath. The extremely high vesicle density in the
artery endothelium has significant implications in the
transport of macromolecules across the arterial endothelium,
as discussed by Dr. Schwartz in his excellent overview.

DR. DEWEY: Is it possible that the low temperatures

and the vacuum in which these tissues were prepared could
produce artifacts that appear as vesicles?

DR. CHIEN: We can apply the same technique to other

kinds of tissue, for example, red blood cells and many other
cells and you would not see this number of vesicles. Besides,
one also sees this high vesicle density in the en face
picture that Dr. Schwartz showed, where low temperature
freezing was not performed.

DR. NEREM: You show albumin data and you mention that,
if you put these data on the basis of surface area, the
numbers for pressures of 100 millimeters of mercury come
closer together.

DR. CHIEN: That is correct.

DR. NEREM: Do the data then become statistically the

same? In other words, is the statistical significance of
any difference then absent?

DR. CHIEN: Yes, it disappears.

DR. COLTON: All the slides you showed, were they of

surface attached vesicles? Secondly, did I detect an orderliness
in the array, that is uniform spacing around each row of
vesicles?
346 CHAPTER 6

Fig. 6-21: Freeze fracture of rat mesenteric artery showing

plasmalemmal vesicles in the endothelium. (43,000x) (Courtesy
of Drs. Maia and Nicholae Simonescu.) (Reduced 30% for purposes
of reproduction.)
TRANSPORT OF PROTEIN AND LIPID 347

DR. CHIEN: In the freeze fracture picture, these are

surface attached vesicles. In the transmission E.M. sectioned
normal to the surface, one sees both free and attached
vesicles. As to the arrangement of vesicles, the density is
so high that it's really hard to picture any particular
pattern. It is probably random.

DR. COLTON: Is it possible to freeze fracture through

the cell and see the distribution of vesicles that way?

DR. CHIEN: The freeze fracture will split through the

hydrophobic layer of the membrane near the surface. It is
more difficult to get to the inside.

DR. STONE: The data that you showed when you distended
the vessel to lOOmm Hg pressure, are they in vivo or in
vitro?

DR. CHIEN: These are in vitro studies using the method

of Caro and Nerem (18). The common carotid artery of the
dog was isolated and mounted on an in vitro apparatus and
incubated at 37 degrees C for 15 minutes with the l25I-albumin.

DR. STONE: Both you and Dr. Schwartz are saying that
the vesicular movement through the central core of the
endothelial cell may be impeded by the filaments in the
nucleus. Is this correct?

DR. CHIEN: Correct, yes.

DR. STONE: Then what you are inferring is that the

cytoplasmic flow is around this area. Is that a correct
assumption?

DR. CHIEN: Yes, you can see horseradish peroxidase in

vesicles in the peripheral zone just outside this area
(i.e., toward the intercellular
Horseradish Peroxidase junction), where marker-filled
in Vesicles vesicles are present through
the whole thickness of the
endothelial cells. Then you also see the horseradish
peroxidase accumulating in the subendothelial cell layer,
where there is a resistance to macromolecular diffusion. It
is possible that the macromolecules traversing the peripheral
zone via vesicles to reach the subendothelial space can then
return by vesicles into the endothelial cell or by diffusion
back into the intercellular junction.
348 CHAPTER 6

DR. CARO: This is where timing becomes very important.

DR. CHIEN: That is correct.

DR. WEINBAUM: The other thing that I wanted to emphasize

here, which I think is of interest to the fluid mechanics
people, is related to the effect
Interaction of Shearing of shearing stresses and their
Stresses with Membrane interaction with the membrane
Surface surface. I was very pleased to
see the flacid appearance of the
upper endothelial surface in the scanning E.M., taken at
magnifications of two to three thousand. If the numerous
enfoldings in the plasmalemma are not artifacts it is obvious
that the membrane will deform extremely readily to whatever
fluid mechanics is going on above it. The whole surface
seems to be rippled, but these dimensions of these ripples
are large compared to the 700 angstrom vesicles. These
ripples have dimensions of several thousand angstoms or more
and are present over the entire surface both at zero pressure
and 100 millimeters pressure.

DR. SCHWARTZ: In one of the lovely pictures you showed

there were no vesicles along the junction. I wonder if you
would comment on whether all junctions have vesicles, and
what indeed determines the distribution of vesicles along
the junctions? Why also should there be vesicles on junctions.

DR. CHIEN: We have not done a sufficiently large

number of specimens to be really sure as to what the answer
is. From what we have seen so far I wou __ d say that on the
external surface of the endothelium, the junction area is
devoid of vesicles as demonstrated by the freeze fracture
pictures. But if you go deeper, then the junction area has
a lot of vesicles, which is shown by the transmission E.M.

DR: BJORKERUD: I have two questions. One, do you

know anything about the distribution of intermembraneous
particles? Second, could the marginal zone on the intercellular
junctions devoid of vesicles correspond to the so-called
muffling zone of growing cells? Do you have any idea about
this?

DR. CHIEN: These specimens were all taken from adult

animals. So as you pointed out, they were probably not
growing except for repair. So I don't think really that the
membraneous foldings represent an extra surface area which
is available for the arteries to form new cells.
TRANSPORT OF PROTEIN AND LIPID 349

They have this extra surface area so you can redistribute

the surface area for different regions of the artery. In
this way, you can achieve deformation and flow inside endothelial
cells, As to the intra-membrane distribution of vesicles,
this is one of the things we want to study, but this requires
a large number of countings and I cannot give you a definitive
figure at this moment. We hope to get that in the near
future.

DR. COLTON: In respect to Colin Schwartz's question,

we have been doing some E.M. lately and have seen there
phenomena of vesicles attached to junctions and they seem to
appear whenever there is significant overlap of two cells
such that the junction is almost horizontal. It is as if one
cell is over the other. The cells apparently just don't
know. Their normal top and bottom behaves as it would
whether or not there was another cell under it. On the
other hand, when the junction is more verticle, that is •••

DR. SCHWARTZ: End to end.

DR. COLTON: Right. There is not much projected area,

not nearly as much, and in those cases we don't seem to get
quite the same super density, in other words it is a short
junction. When cells grow so that there is a significant
overlap that seems to be when you get the higher circulation.

DR. STONE: Is there a relationship between vesicles

and systemic arterial pressure?

DR. CHIEN: We have not done enough specimens to allow

us to give you an answer that is supported by statistics,
but the first impression is that there is not, but we have
to do more work.

DR. DEWEY: I emphasize that we are trying to achieve some

kind of constructive synthesis which will not only identify
where we are today in understanding the Dynamics of Arterial
Flow, but perhaps more importantly we are trying to suggest
how, as engineers, physiologists, lipid chemists and people
who work more than one side of the fence, we may achieve even
more accelerated progress in the future.

This is not an easy task because the demands of the inter-

disciplinary field are very great and they certainly tax our
abilities to evaluate the information that is relevant to
arterial disease as we design our own programs and try to
further this understanding.
350 CHAPTER 6

I personally feel a substantial amount of optimism about this

process for two reasons. First of all, I find that my
engineering colleagues around the table are showing a great
willingness to tackle real problems. By real problems I mean
getting into the "nitty gritties" of the mechanics of the
endothelial cell; to deal with non-linear processes in arteries
which include chemical phenomena, as well as strictly mechanical
phenomena; to look at fluid flow situations which we know we
cannot understand in detail and not let that deter us and deflect
us into problems that are irrevelant. By the same token I
think the physiologists have an appreciation for the type of
engineering approach to deal with conceptual models of the types
that we have proposed, or in fact they are proposing, is also,
I believe, a very healthy sign. I speak specifically of
Dr. Smith's suggestions regarding hypoxia; it is this type of
sophistication with respect to analyzing physiologic data that,
first of all, makes me encouraged for work with physiology
counterparts and, secondly, suggests that the interaction
is going to be a very fruitful one.

BIBLIOGRAPHY

1. Bratzler, R.L., Chisolm, G.M., Colton, C.K., Smith, K.A.,

Zilversmith, D.B., Lees, R.S.: The distribution of
labeled albumin across the rabbit thoracic aorta
in vivo. Circ. Res., 40:182, 1977.

2. Bratzler, R.L., Chisolm, G.M., Colton, C.K., Smith, K.A.,

and Lees, R.S.: The distribution of labeled low-
density lipoproteins across the rabbit thoracic
aorta in vivo. Atherosclerosis. In press.

3. Bratzler, R.L., Chisolm, G.M., Colton, C.K., Smith, K.A.:

Theoretical models for transport of low-density
lipoproteins in the arterial wall. Atherosclerosis:
Metabolic, Morphologic and Clinical Aspects.
G.W. Manning and M. Daria Haust, Eds., Plenum Press,
943-951, 1977.

4. Gimbroni, M.A., Jr.: Culture of Vascular Endothelium and

Atherosclerosis.ln: Int'] Cell Biology, (B.R. Brinkley
and K.R. Porter, Eds.) Rockefeller U. Press,
649-658, 1977.

5. Cornforth, J.W., Hunter, G.D., Popjack, G.: The biosynthesis

of cholesterol from acetate. Arch. of Biochem. &
Biophys. 42: 2, 1953.
TRANSPORT OF PROTEIN AND LIPID 351

6. Weinbaum, S., and Caro, C.G.: Macromolecule transport model

for arterial wall and endothelium based on ultra-
structural specialization observed in electron micro-
scopic studies. J. Fluid. Mech. 74, 611-643, 1976.

7. Keatinge, W.R.: Mechanical response of arteries to stimula-

tion of sympathetic nerves. J. Physiol. (London)
185: 701-705, 1966.

8. Fry, D.L.: Atherogenesis: Initiating Factors, CIBA

Symposium, New Series, No. 12., Associated Scientifc
Publishers, Amsterdam, 1973.

9. Siflinger, A., Parker, K. and Caro, C.G.: Uptake of 1251_

albumin by the endophelial surface of the isolated
dog common carotid artery: effect of certain physical
factors and metabolic inhibitors. Cardio. Res.
9:478-489, 1975.

10. Guyton, A.C., Granger, H.J., Taylor, A.E.: Interstitial

fluid pressure. Physio. Rev., Vol. 51, 527-563,
1971. and
Wiederheilm, C.A., Dynamics of transcapilary fluid. J. Gen
Physio., Vol 52, 29-63, 1968.

11. Gerrity, R.G. and Cliff, W.J.: The aortic tunic intima
in young and aging rats. Exp. Mol. Pathol.,
16:382-402, 1973.

12. Simionescu, N., Simionescu, M. and Palade, G.E.:

Permeability of muscle capillaries to small hemepeptides
Evidence for the existence of patent transendothelial
channels. J. Cell. BioI. 64: 586-607, 1975.

13. Schwartz, C.J. and Gerrity, R.G.: Arterial endothelial

structure and function with particular reference to
permeability. Atherosc. Rev. (A.M. Grotto, Ed.)
Raven Press, N.Y., 1978.

14. Bell, F.: Adamson, I.L., and Schwartz, C.J.: Aortic

endothelial permeability of albumin: Focal and
regional patterns of uptake and transmural distribu-
tion of l3lI-albumin in the young pig. Exp. Molec.
~, 20, 57-68, 1974.
352 CHAPTE R 6

15. Gerrity, R.G., Richardson, M., Somer, J.B., Bell, F.P.,

Schwartz, C.J.: Endothelial cell morphology in
areas of in vivo Evans blue uptake in the aorta of
young pigs. II. Ultrastructure of the intima in
areas of differing permeability to proteins.
Am. J. Pathol., 89: 313-334, 1977.

16. Ross, R.: Platelet mitogenic factor. The Smooth Muscle

of the Artery. Wolf, S., Werthessen, N. (Eds.) 1975.

17. Caro, C.G.: Transport of material between blood and

wall in arteries. Atherogenesis: Initiating Factors.
Ciba Foundation Symposium 12 (New Series) pp. 127-
164, Associated Scientific Publishers, Amsterdam,
1973.

18. Caro, C.G., and Nerem, R.M.: Transport of C14- 4-cholesterol

between serum and wall in the perfused dog common
carotid artery. Circ. Res. 32:187-205, 1973.
Chapter 7 HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS

DR. MEYER: I will demonstrate some structural features

of the arteries which are probably related to hemodynamic
factors and appear to be important for the localization and
further development of lesions.

The curved segment of the internal carotid artery

which, located at the base of the skull, is known as the
carotid siphon, is of particular
Carotid Siphon interest. This segment consists
of four arterial curves which
immediately follow each other (Fig. 7-lA and lB). Thus, the
blood flow is diverted in this short segment several times. The
deflection of the blood flow probably results in different
functional load upon the opposite, i.e. concave and convex
walls of the curves, and, consequently determines distinct
structural differences between the opposite arterial sectors,
which are already present at birth. In newborn children,
intimal cushions consisting of numerous fine elastic fibers
are seen at the inner walls of the curves, whereas at the
opposite convex wall only a single primary internal elastic
membrane is present (Fig. 7-2). With growth, the intimal
cushions become more prominent (Fig. 7-3 and 7-4). and in
young adults they are often as thick as the underlying
(subjacent) media. They are rather poor in elastic elements
and mainly consist of loose collagenous networks and ground
substance. At the opposite, i.e. outer (convex) wall of the
curves and in adjacent sectors numerous new elastic sheets
appear internal to the internal elastic membrane and form a
prominent (hyperplastic) subendothelial (intimal) elastic
layer (Fig. 7-3 and 7-4).

The different structural components appearing at the

opposite arterial sectors show a different affinity for
calcium and lipids. Strong
Affinity for Calcium subendothelial elastic networks
and Lipids of the outer, i.e. convex and
lateral walls of the curves early
become a predominant substrate of calification.

In the upper (fourth), sharpest curvature of the siphon,

calcific deposits have been regularly demonstrated grossly
in all children aged 1 - 16 years (1,2,3,4). They regularly
appear above the orifice of the ophthalmic artery and often
353
354 CHAPTER 7

cavernous part

Fig. 7-LA: Schematic diagram of the right carotid siphon

showing sites of predilection for early calcification (dotted
area) and indicating the petrous and cavernous portions of
the internal carotid artery. The numbers refer to the
curvatures of the siphon. D-D = dura mater. Opht - origin of
the ophthalmic artery. (From Meyer and Lind, 1972).
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 355

Fig. 7-IB: Carotid siphon of a four-year-old boy which has

been dissected free after fixation in formalin at a filling
pressure of about 90 rnrn Hg. The numbers refer to the
curvatures. (The first curvature was cut off and is not
seen in the figure). oph - cross- sectioned ophthalmic artery.
The small adherent part of the adjacent dura mater (D) indicates
the level of its perforation by the siphon. (From Meyer,
in preparation).
356 CHAPTER 7

Fig. 7-2: Small intimal cushion (C, below) mainly consisting

of fine elastic fibers is seen at the inner (concave) wall (IW)
of the fourth curvature of the carotid siphon from a newborn
child. At the opposite sector (above), e.e., at the outer wall
(OW) of the curvature only a wavy internal elastic membrane is
seen below the endothelial lining. m - media; a - adventitia.
Case 46 K/76. Cryostat section. Weigert's resorcin fuchsin.
(From Meyer, in preparation).
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 357

DC

Fig. 7-3: Diagram showing a longitudinal section of the right

carotid siphon which illustrates the location of intimal cushion
and elastic tissue elements. In childhood, the subendothelial
elastic layer is well developed along the outer walls (Oe) of
the curvatures, whereas prominent intimal cushions (arrows),
consisting mainly of loose connective tissue and sparse elastic
elements are present along the inner walls (Ie) of the curva-
tures. Inset shows a cross-section of the vessel at the site
of the arrows.
358 CHAPTER 7

Fig. 7-4: Structural differences of the opposite sections of a

carotid siphon from a l3-year-old boy. At the inner (concave)
wall (IW) of the fourth curvature (below), a prominent intimal
cushion (C) rich in elastic fibers is seen internal to the fine
primary internal membrane (arrows). At the opposite outer
(convex) wall (OW), a thick hyperplastic secondary elastic
layer is present internal to the discrete primary internal elastic
membrane (arrows) showing a fine wavy pattern. m - media,
Cryostat section. Weigert's resorcin fuchsin. Case 58 K/76.
(From Meyer, in preparation).
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 359

o ,
t

•
" ~

,
••

Fig. 7-5: Numerous, partly confluent early calcifications (black)

form a circularly arranged band above the orifice of the ophthalmic
artery (0). Scattered incrustations are also seen below the
orifice. Carotid siphon of a three-year-old girl who died after
accident. (Case 341/70). Von Kossa reaction. (Unpublished).
360 CHAPTER 7

extend around a larger part of the arterial circumference at

this level (Fig. 7-5). Later the calcifications also appear
in the subjacent, i.e. more proximal portion of the siphon.
In young adults they often encircle the intimal cushion which
regularly develops at the inner (concave) wall of the fourth
(upper) curvature. The cushion itself remains free of cal-
cific deposits for a long time (Fig. 7-6).

DR. TAYLOR: How old is this person, Dr. Meyer?

DR. MEYER: This siphon is from a 17-year-old man.

DR. TAYLOR: Seventeen. I wanted to bring that out.

It is exciting to think that this calcification occurs at
such a young age.

DR. MEYER: In the fourth decade of life, larger confluent

incrustations are also seen regularly in the dorsal wall
proximal to the fourth curvature (Fig. 7-6). Since this
wall is located at some distance from the bone, its predilection
to calcification can hardly be determined by physical parameters
of the surrounding tissues. The deposition of calcium
appears to be rather favored by hemodynamic factors arising
with deflection of the blood stream in this segment.

The lipid deposits appear later in the carotid siphon

than do the primary calcifications. With gross staining
they become visible towards the end of the second decade and
are often present in the third decade of life (5). Thus, the
lipid deposits develop in a carotid siphon, some parts of
which may already be severely affected by preceeding primary
calcifications. Nevertheless, initially the lipids do not
accumulate in the incrustated areas, but preferentially
appear in the intimal cushions of the inner (concave) wall
of the curves. In young adults the different localization
of calcific and lipid deposits is distinctly seen (Fig. 7-7).

DR. TAYLOR: How old?

DR. MEYER: This siphon is from a 35-year-old man.

At this age, calcification is often limited to the area

adjacent and proximal to the intimal cushion. Here, strong,
hyperplastic elastic sheets
Intimal Cushion become the predominant substrate
of incrustation. In contrast,
elastic sheets extending further towards the intimal cushion may
be free of calcifications, but show a diffuse, pronounced lipid
infiltration, which is also seen in the intimal cushion
itself (Fig. 7-8).
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 361

Fig. 7-6: Upper part of the medial half of the right carotid
siphon from a 17-year-old male who died in a traffic accident.
After von Kossa reaction, confluent axially arranged calcific
deposits (black) are seen above the orifice of the ophthalmic
artery (oph). The peak of the concave inner wall of the fourth
curvature (4) is free of calcifications, but tightly packed
confluent linear incrustations are seen below and in front of a
slightly raised intimal cushion (dotted line in the center of
the figure). A small intimal cushion was seen at the peak of the
third curvature (3) but it is not well seen here. Below the
peak of the fourth curvature, there is a large calcific plaque
(p) which is partly covered by grayish intimal connective tissue.
Fine calcific deposits are seen below the plaque. Gross staining
for lipid shows only a faint reddish tinge, i.e. early lipid
infiltration is present in the area of both intimal cushions
located at the peaks of the inner walls of the third and fourth
curvatures. Arrows indicate the direction of blood flow.
Case 43/72.
362 CHAPTER 7

Imml

Fig. 7-7: The inner wall of the third curvature is seen from
above in the longitudinally opened siphon from a 35-year-old man.
(Cause of death: traffic accident). Von Kossa reaction and
gross lipid staining with Fettrot 7B. Just below the height
of the curvature, the luminal layer is densely interspersed with
fine (black) partly confluent calcifications (between arrows)
which extend circularly over a large part of the circumference.
In front of the incrustations, a lipid infiltrated area (L,
grayish), red-stained in the specimen, is seen. The lipid
infiltration extends up to the intimal cushion (C, white). White
arrow indicates the direction of blood flow. (See also Fig. 7-8).
(Unpublished).
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 363

'1./

Fig. 7-8: Longitudinal cryostat section of the inner (concave)

wall of the third curvature demonstrated grossly in Fig. 7-7.
Roundish grain-like calcific deposit (C) penetrating the media
is seen to the left. In the right part of the sepment the media
(m) is superimposed by an intimal cushion (CU) consisting of two
well delineated layers. In the layer adjacent to the media fine,
grain-like calcific deposits (black) are seen. Between the
calcific deposit and the cushion the hyperplastic elastic intimal
sheets are infiltrated with lipids (L) and stained deeply red in
the section. They appear grayish in the photograph. The arrow
indicates the direction of blood flow. (Unpublished).
364 CHAPTER 7

DR. SCHWARTZ: You seem to have two components in the

cushion.

DR. MEYER: Yes, this intimal cushion consists of two

components, i.e. of two layers which are superimposed upon
each other and delineated by a limiting elastic lamella.
This structural feature is probably related to the widening
and distortion of the siphon with age. The additional
luminal layer appearing above the initial intimal cushion
may represent a structural adaptation to the changes in
blood streaming (flow) resulting from increasing tortuosity
of the siphon in adults.

DR. WOLF: Is it a cellular composition?

DR. MEYER: The intimal cushions include numerous cells

embedded in rather loosely arranged collagenous networks and
ground substance. In contrast, there are only a few cells
between the hyperplastic elastic sheets which become the
substrate of calcification.

DR. TAYLOR: How old was he?

DR. MEYER: This is the siphon from a 35-year-old man

who died in a traffic accident.

Later, the selective deposition of calcium and lipids

in different structural components of the arterial wall
becomes less prominant. The lipids, for instance, appear,
additionally, in the thickened intima which develops over
the incrustated elastic sheets. Moreover, the layer of the
hyperplastic elastic intimal sheets may simultaneously be
interspersed with fine grain-like calcific deposits and be
diffusely infiltrated with lipids. Pronounced involvement
of the carotid siphon by overlapping primary calcifications,
lipid infiltration and atherosclerotic lesions occurs even
at a relatively young age (Fig. 7-9). With the progression
of lesions the assessment of factors which determine their
localization becomes increasingly difficult.

This siphon is from a 32-year-old man who died of

myocardial infarction.

The significance of the adjacent hard tissues, i.e.,

of the bone and the dura for the development of lesions
arising in the carotid
Counter Pressure of Bone siphon have not yet been precisely
assessed. The tube of the
siphon is mostly surrounded by venous sinus, which probably acts
as a hydraulic suspension and damps the recoil arising with
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 365

Fig. 7-9: Extensive primary calcifications (black) are seen in

the medial (left) and the lateral (right) halves of the long-
itudinally opened carotid siphon from a 32-year-old man who died
of myocardial infarction. In many areas calcifications are
covered by advanced atherosclerotic plaques (white). (The areas
of prominent lipid infiltration, which were red-stained in the
specimen, are not distinctly visible in the photograph). Milli-
meter scale left. Von Kossa reaction and Fettrot 7 B. Case 847/73.
(From Meyer, 1975A).
366 CHAPTER 7

transmission of the pulse wave. In this way, the siphon is

probably sufficiently protected against the counter-pressure
of the adjacent tissues. However, with age and progressive
dilatation of the arterial tube, the distance between the
widening artery and the surrounding bones may become shorter
and the "counter-pressure" of the bone more effective.

Remarkably, pronounced calcifications seldom appear in

the petrous portion of the siphon which is embedded in an
osseous channel. In contrast, only some parts of the upper
cavernous portion of the siphon are located in the vicinity
of the osseous tissue. Therefore, the greater invovlement of
this segment by calcifications and atherosclerotic lesions
appears to be related to its more tortuous pattern and
additional hemodynamic factors arising in curved arterial
segments. I should like to demonstrate three remarkable
patterns of early lipid infiltration in the cervical part of
the carotid artery, which appear to be closely related to
the preformed structural characteristics of the involved
areas (Fig. 7-6).

In the trunk of the common carotid artery the lipids

initially appear below the orifice of the external carotid
artery, extend proximally
Lipid Infiltration in along the antero-medial wall
Cervical Part of Carotid during the second decade and
Artery form a longitudinal, triangular
stripe (deposit) regularly
seen in young adults (Fig. 7-10 and 7-11). The characteristic
shape of the lipid infiltrated area persists even in later
ages.

Microscopic examination reveals conspicuous structural

differences between the opposite, i.e. predisposed and
resistant arterial sectors. At the predisposed antero-
medial wall, a well delineated musculo-elastic intimal layer
appears internal to the internal elastic membrane in the
first decade of life, i.e. before the lipid infiltration
becomes visible, grossly or microscopically. Toward the end
of the first decade, a compact sheet of tightly packed
strong, longitudinally arranged elastic fibers develops
along the luminal aspect of the musculo-elastic layer and
becomes superimposed by a thin luminal layer of a loose
connective tissue which is rich in ground substance (Fig. 7-12
arrows). Both layers develop a moderate thickening of the
intima at the predisposed arterial wall, whereas the intima
of the opposite, resistant arterial wall remains thin in
children and young adults.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 367

Fig. 7-10: Carotid bifurcation of an l8-month-old child

dissected after fixation under pressure of 60 cm water formalin
solution. CC - common carotid artery, IC - internal carotid
artery, EC- external carotid artery. The protruding sinus area
(8) is clearly seen.
368 CHAPTER 7

Fig. 7-11: Various stages in the development of gross patterns of

lipid depositis in the common carotid artery.
A. Lipid deposits (black) initially appear below the orifice of
the external carotid artery (Ee) and extend proximally along the
length of the vessel. Lipid deposits are also seen in the
peripheral parts of the sinus (S). 20-year-old man. (Case 1243/73).
B. Deposits then become confluent along the antero-medial wall of
the common carotid artery (arrow) and extend around the central
part of the sinus. In this specimen the central sinus area (CS)
has been opened longitudinally and divided into two parts which
extend to the borders of the vessel. Note also the lipid-free
carina of the carotid bifurcation (arrow). 34-year-old man
(Case 227/73 ).
C. With more advanced lesions, lipid infiltrates the antero-
medial wall and the lesion assumes a characteristic triangular
form 31-year-old man (Case 127/72). Fettrot VII B. (From Meyer
and Noll, 1974).
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 369

Fig. 7-12: Structural differences between the dorso-lateral (A)

and antero-medial (B) walls are well seen on a cross section of
an arterial ring taken from the middle of the common carotid artery
from a 20-year-old man. Cause of death: traffic accident. In
dorso-lateral wall (A), the intima (i) is thin. It consists mainly
of a subendothelial layer formed by tightly packed elastic fibers.
No internal elastic membrane is seen. In the antero-medial wall
(B), the intima (i) is thickened and has a well-developed musculo-
elastic layer (mel) between the internal elastic membrane (iem)
and the compact sheet of strong longitudinal elastic fibers (arrows).
Cross section of these fibers are seen below the thin connective
tissue layer (ctl) of the intima. No lipids are present at this
level. Cryostat frozen section. Weigert's resorcin-fuchsin.
Case 1145/71. (From Meyer and Noll, 1974) (6).
370 CHAPTER 7

DR. CARD: On the opposite side of the wall, is there a

lack of cells?

DR. MEYER: The intima is thin at the resistant arterial

wall and accordingly it includes fewer cells than the opposite
predisposed wall.

The lipids seem first to appear selectively in the

sheet of the strong longitudinal elastic fibers which develop
along the luminal aspect of the musculo-elastic luminal
layer (Fig. 7-13). Later the lipids also appear in the cells
of the musculo-elastic layer, i.e. between the sheet of
longitudinal elastic fibers and the internal elastic membrane.

In general, the extent of the thickened and differentiated

intima seems to correspond to the triangular area which
gradually becomes the site of
Possibility of a Common lipid deposition in children and
Hemodynamic Factor. in young adults. Remarkably, the
Antero-medial Wall of granular medial calcification,
Carotid Trunk independent of intimal lipid
deposits, sometimes also appears
in the antero-medial wall of the carotid trunk and extends over
approximately the same triangular area as do the lipid
deposits (4). The appearance of both intimal and medial
lesions in the same portion of the arterial tube suggests
that a common hemodynamic factor may be responsible for
their development.

The second remarkable area is the carotid sinus, i.e.

the slightly protruding proximal portion of the internal
carotid artery just above
Carotid Sinus its origin from the carotid trunk
(Fig. 7-l4a). In children, the
lipid deposits predominantly appear in the peripheral parts of
the protruding sinus area (Fig. 7-l4a). Later dot-like or
linear lesions often conglomerate into a ring-like, slightly
raised deposit, which may partly or completely encircle the
central sinus area (Fig. 7-l4b). Later, i.e., in the third
decade of life, the entire sinus area often becomes almost
completely covered by a thickened lipid-containing intima
(Fig 7-l4C and 7-140).

As in the carotid trunk, the early development of lipid

deposits in the carotid sinus appears to be closely related to
an early thickening and differentiation of the intima during
postnatal growth (Fig. 7-15).
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 371

Fig. 7-13: Cross sections of the intimal luminal layer of the

common carotid artery showing selective deposition of lipids in
the sheet of strong elastic bundles. A. With Weigert's
resorcin-fuchsin stain, cross sections of strong elastic bundles
(arrows) are distinctly seen between a thin intimal connective
tissue layer (icl) and the intimal musculo-elastic layer (mel).
iem - internal elastic membrane. B. With Fettrot, the lipid
infiltrated, red-stained cross sections of the strong elastic
bundles (arrows) appear black. Note the fragments of the
internal elastic membrane (iem). lO-year-old boy (case 51/74).
Cause of death: traffic accident. (From Meyer and Noll, 1974).
372 CHAPTER 7

Fig. 7-14: Gross patterns of early lipid deposits in the carotid

sinus. A. Note scattered punctate lipid deposits (black) around
the periphery of the carotid sinus (S). l3-year-old boy.
(Case RM 183/74). B. A ring-like lipid deposit encircles the
central protruding sinus area. lO-year-old boy. (Case 51/74).
C. Extensive lipid deposits are present in the peripheral parts
of the sinus. A pale diffuse staining ("blush") is seen in the
central sinus area and along the antero-medial wall of the
common carotid (cc). 19-year-old women. (Case 1335/74).
D. Lipid deposits in the periphery of the sinus from a 30-year-
old man. Cause of death in all cases: traffic accidents.
Fettrot 7B. S-carotid sinus; ic - internal carotid artery; ec-
external carotid artery, cc - cornmon carotid artery. (From Meyer
and Noll, 1974).
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 373

Fig. 7-15: Cross section of the protruding lipid-free area

within the ring-like lipid deposit in the carotid sinus (A)
and of the arterial wall external to the lipid deposit at the
same level (B). A. In the sinus, the intima (i) is about one-
fourth the thickness of the media (M). Arrows point to the
sheet of longitudinal elastic fibers which extends along the
inner aspect of the musculo-elastic intimal layer. B. In the
segment immediately external to the sinus, no intimal thickening
is present. IO-year-old boy. Case 51/74. Weigert's resorcin-
fuchsin stain. (From Meyer and Noll, 1974).
374 CHAPTER 7

No difference in the microscopic structure could be found

between the intima of the peripheral parts of the sinus and
the intima of the central area. The peripheral portion is
the site of early lipid deposits, whereas the central area
initially stays free. The preferential accumulation of lipids
in the periphery of the sinus may be due to the differences
in functional load and the different elasticity of the central
and peripheral portions of the sinus.

The third conspicuous gross feature is the sharp delineation

of the lipid deposits appearing in the sinus from the lipid-
free portion of the internal
Internal Carotid Artery carotid artery located immediately
Immediately above Sinus above the sinus. Even with a
pronounced involvement, the lipid
infiltration ends abruptly above the sinus. Then, the border
between the affected and lipid-free parts of the arterial
luminal surface often appears as a straight horizontal line
(Fig. 7-16). Microscopically, the upper borders of the
lipid deposits regularly correspond to the transitional area
between the elastic and muscular arterial segments, where
the media assumes the structural pattern of muscular arteries
(Fig. (7-17).

At the same level, the thickened intima of the sinus

becomes progressively thinner. Thus, the lipid deposits are
limited to the sinus which belongs to the elastic arterial
segment and which is, moreover, covered by thickened intima.
The lipids do not appear in the immediately distal muscular
segment showing a thinner intima. It is still to be elucidated
whether the structural features of the media or the intima
or of both are responsible for different involvement of
these adjacent arterial segments. (See diagram Fig. 7-18).

The next three figures show the peculiar distribution

of lipid deposits in the cervical portion of the vertebral artery.
This portion is of special interest because it passes through
several successive ring-like osseous structures with soft
tissues in between (Figs. 7-19, 7-20, 7-21). This unique
topographic relationship suggested a way of determining
whether the localization of lipid deposits and atherosclerotic
lesions may be influenced by the physical characteristics of
the surrounding tissues, i.e. whether the arterial lesions
appear at the sites surrounded by bone or softer tissue (7).
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 375

Fig. 7-16: Advanced lipid infiltration of the carotid sinus (5).

Note sharp delineation of the lipid-infiltrated area from the
lipid-free portion of the internal carotid artery (ic); cornmon
carotid artery (cc). A: carotid artery from a 22-year-old
woman (1140/71); B: carotid artery from a 46-year-old man
(1126/71) •
 w
'"0-

Fig. 7-17: Lipid deposits in the upper part of the sinus (L, black) are located
in the transitional zone between the proximal elastic (D) and distal muscular (M)
segments and do not extend into the muscular segment (even in old age groups).
Note the pale (unstained) smooth musculature of the muscular arterial segment and
the thickened intima (I( of the sinus. Resorcin-fuchsin and Fettrot stains. 27-
year-old woman (Case 118/73). (From Meyer and Noll, 1974).

()
I
:l>
CJ
-i
m
:0
-..J
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 377

Fig. 7-18: Intimal thickening (dotted) in the carotid sinsu (S)

stops abruptly at the demarcation between the elastic (E) and
muscular (M) arterial segments (arrow). CC - common carotid
artery,IC - internal carotid artery, EC - external carotid artery.

Fig. 7-19: Schematic presentation of the course of the vertebral

artery (VA) through the costovertebral foramina. SA subclavian
artery. (From W.W. Meyer, 1964).
378 CHAPTER 7

Fig. 7-20: Rhythmical arrangement of lipid deposits in the

vertebral artery from a 54-year-old woman. The red stained
deposits appear black in the photograph. The right vertebral
artery was hypoplastic (left). Scarlet red staining. (From
W.W. Meyer and Naujokat, 1964).
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 379

Fig. 7-21: Rhythmical arrangement of lipid deposits and athero-

sclerotic lesions in the vertebral artery from a 78-year-old man.
The lesions (black) predominantly appear in the ectatic segments
localized between the cos to transversal foramina (Scarlet red
staining). (From W.W. Meyer and Naujokat, 1964).
380 CHAPTER 7

But let us first look at the topographic situation. In

the cos to-transversal foramina of the cervical vertebrae,
the artery is surrounded by bone. However, the arterial
wall does not directly touch the osseous ring. A well
developed venous plexus surrounds the arterial tube and may
probably reduce the recoil of the pulse wave considerably.
The spinal nerve and spinal ganglion are located immediately
behind the artery. They obliquely cross the arterial tube
and are immediately adjacent to its dorsal wall.

The lipid deposits initially appear at the dorsal

arterial wall adjacent to the spinal ganglion. They later
extend above the spinal nerves, i.e. into the segment of the
artery which is located between the osseous rings, i.e.,
costotransversa1 foramina.

In some cases, this results in a regular sequential

arrangement of fatty streaking. Even with a further prgression
of lesions, a sequential arrangement may be distinctly
visible as demonstrated in the vertebral artery of the 78-
year-old man (Fig. 7-21).

Age changes of the arterial tube, for instance the

ectasia, probably favors the sequential arrangement of
lipids. Arterial segments
Age Changes located between the osseous
rings become more ectatic with
age compared to segments surrounded by bone. This results in
change in caliber along the intravertebra1 segment of the
artery. On the other hand, with ectasia the distance between
the arterial wall and the surrounding bone in the costovertebral
foramina decreases and the recoil of the pulse wave may
become more effective. Both factors could be significant
for the localization of lipid deposits and their sequential
arrangement.

DR. WERTHESSEN: Where are the nerves that measure the

oxygen?

DR. MEYER: The wall of the sinus is penetrated by

numerous nerves, and in some areas the media appears interrupted
by small bundles of nerves. It could not be excluded that
this structural peculiarity favors lipid deposition in the
intima of the sinus.

There is, in fact, a proliferation of the intimal cells

preceeding the lipid deposition at young age.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 381

This is true of the carotid sinus as well as of the predisposed

area of the carotid trunk. The proliferation of the intimal
cells and formation of a differentiated, thickened intima
may be interpreted as a structural adaptation to the increasing
hemodynamic load.

DR. WERTHESSEN: I was just interested in whether or

not the nerve endings which measure the blood, are in the
lipid area.

DR. MEYER: This is a very interesting problem, but as

far as I know there is no exact information on the question.

DR. TAYLOR: As far as I know, no one has studied

carotid sinus reflexes in arteriosclerotic carotid sinuses.

DR. CHIEN: It is not clear to me whether the red

stained lipid infiltration area can be free and behind the
cervical spine? Are these areas surrounded by bone or next
to the nerve?

DR. MEYER: The lipid deposits appear first at the

dorsal wall of the vertebral artery adjacent to the spinal
ganglion.

DR. LEE: In regard to your calcium deposits, which are

very interesting, have you looked at any other vessels other
than the carotid arteries? Children have such calcium
deposits. Do you know the vitamin D content of human milk
or is vitamin D supplementation the culprit?

DR. MEYER: The predilection of the common and internal iliac

arteries for calcifications may be due to the higher blood
volume flow to which these arteries are exposed during fetal
development as a part of the placental circuit. They show
an accelerated growth, and wear and tear changes may occur
in these arteries already before birth. In Germany, calcifications
of the iliac arteries and of the carotid siphon have been
demonstrated grossly in all children aged one year or older.
An overdosage of Vitamin D may occur in Germany and could be
a pathogenetic factor.

DR. LEE: We have studied Vitamin D and Fred Kommeror

studied Vitamin D, and concluded that it induces widespread
arterial calcification.
382 CHAPTER 7

DR. CARO: I would like to congratulate Dr. Meyer on

those marvelous pictures. I have seldom seen anything so
fine. I wondered if you or anyone has made any studies of
the vertebral arteries under normal pressure. Does the
artery bulge in between the areas where it is suported?

DR. MEYER: In the older individuals the segments

located between the osseous rings appear ectatic. No measurements
of the internal diameters have been made.

DR. CHIEN: I think the decisive factor responsible for

lesions of the siphon is its coiled shape. The siphon, at
least in childhood, is surrounded by a venous plexus.

DR. MEYER: Yes, the siphon seems to be well protected

against the osseous tissue by venous plexus at least in
younger individuals. However, with age, ectasia of the
siphon occurs. Then, the distance between the arterial wall
and the surrounding osseous tissues may become significantly
shorter and the recoil may become a factor influencing the
development of lesions. (2,4,8,9).

DR. TAYLOR: It makes that terrible siphonic turn too,

Dr. Meyer. I don't think bone is as good a cushion as liver,
spleen. kidney, or brain tissues. These are superb soft
external arterial cushions.

DR. CARO: Dr. Meyer tells us that there is ectasia of

the vertebrals. Now if I can go back to Dr. Dewey's presentation,
I think that we have two findings that have a lot in common.
We are seeing lipid deposition occurring in two regions of
ectasia. In one of these the deposition may stop where the
flow reattaches, or at least accelerates and Dr. Meyer
showed us that the lesion stopped where I would guess that
the flow channel narrows again.

DR. SCHWARTZ: A brief question. I was impressed with

Dr. Meyer's beautiful demonstration of the close affinity of
elastic tissue for lipid. The question I ask of the engineers
is simply this: Can they conceive of any mechanism where
abnormal fluid mechanical stresses would in some way modify
the structure and composition of elastin, and thus change
its binding capacity for lipoproteins and lipid?
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 383

DR. COLTON: Has a detailed examination been made of the

conformation of those macromolecules? Is the conformation
dependent upon anatomical position or extent of deformation
of the tissue? Perhaps the sites for interaction with lipids
are exposed to a greater extent in some regions than in others.

DR. SCHWARTZ: We still have to ask that set of questions.

DR. SMITH: Elastin is a very non-polar protein and it

appears that the non-polar aminoacids are grouped together
to form hydrophobic regions; Keller and Mandl isolated large,
non-polar polypeptides from hydro1ysates of elastin (10).
Elastin is resistant to acid and to alkali in aqueous media, but
is rapidy hydrolysed in the presence of organic solvents.
Robert (11) has suggested that organic solvents and lipids may
react with hydrophobic regions, rupturing the hydrophobic inter-
actions and leading to a partial "unfolding" of the molecule,
which then becomes more susceptible to hydrolytic or protease
attachment. He also suggests that "stretched" elastin may be
more permeable to lipid; this may be largely speculative,
but the idea that stretching facilitates lipid infiltration,
which in turn facilitates proteolysis, fits very well with
the histological picture.

DR. MEYER: I would like to ask the engineers whether

the different structure of the concave and convex walls of
the carotid siphon could be explained in terms of hemodynamics.
I discussed this question with a co11eaque privately and
learned that centrifugal forces are probably negligible in
the curved arterial segments and cannot be responsible for
pronounced hyperplasia of the intimal elastic sheets at the
outer walls of curves. What other physical forces may be
involved which could explain the structural differences in
curved arterial segments?

DR. DEWEY: In a vessel with an extremely sharp radius

of curvature there is a difference between the stress in the
wall at a given point of pressure on the inside versus the
outside. It may be that this difference in stress particularly
in the carotid bifurcation alters the curve.

DR. COLTON: I would like to reiterate the point raised

by Dr. Schwartz that the binding capacity of the various
macromolecules in the arterial wall, such as collagen, elastin
or the glycosaminog1ycans may be dependent on the stress to
which they are exposed and the resulting deformation they
undergo.
384 CHAPTER 7

Change in the applied stress could open sites which might

be closed at a lower degree of stress, consequently the
effect of local variations in stress might be to change the
capacity of the wall to hold the lipid that passes through
it, and not necessarily to alter the endothelial permeability
to lipoproteins.

DR. NEREM: There is no direct evidence for the involvement

of fluid mechanics in the initiation of atherosclerotic lesions.
Someone may wish to take exception with such a denial of
direct evidence but in fact it is indirect evidence upon
which we "hang our hats," base our NIH grants, and write our
papers for the variety of meetings where we go around and
talk to each other.

Fortunately, there is at least some indirect evidence,

and what I will comment on briefly is what I consider the
primary indirect evidence,
Pattern of Disease i.e., the pattern of disease.
Fig. 7-22 is the classical drawing
out of Spain's 1966 article in Scientific American (12)
illustrating that basically there is a pattern to the disease,
I think we all accept this as fact even though we may not
agree on the pattern. However, it appears that it is the
abdominal region of the aorta, the carotids, and the coronaries
which are the positions in the vasculature having the highest
predilection for developing atherosclerosis. If we look at
the coronary system in detail, and even if one accepts that
there may be involvement of most of the extramural coronary
vessels, there still are regions which seem to have a higher
predilection than others. These are primarily regions of
branching, and in some cases of sharp curvature, and these
are certainly regions of interest from a fluid mechanic
viewpoint.

Unfortunately, we do not have enough detail on the

pattern of the disease. For example, I am not sure that
anyone at this point can really state whether at a bifurcation
initiation of the disease occurs on the flow divider or on
the outer wall. Now this may be something that we will want
to talk about further. Certainly, there were some very
interesting results presented by Dr. Meyer that would suggest
a certain type of region as being the favorite. Now I am
extremely reluctant to use terms like 'low shear' and 'high
shear' because those of us who have been in this area of
research for a long time realize that low shear or high
shear may have absolutely nothing to do with it.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 385

ANTERIOR CEREBRAL- - ----+f.f1L-'!I[

MIDDLE CEREBRAL.---\~~'UJ+-_CIRCLE OF
POSTERIOR CEREBRAL WILLIS
BASILAR
VERTEBRAL------~~
COMMON CAROTID ---7"~1

INN OM I NAT E - - -"'7"----7"7""'---"",..,

AORTA---+-~--T

CORONARY ARTERI ES ===t=#===++-4'\

RENAL--~~;+~~~

ABDOM IN AL AORTA --+-fHt--t-t--"'~-.

COM MON ILIAC - - ---+...,H-If-f-I'----/
I NTERAL ILIAC - - ---HiHH'-+--ffl:'

FE MORAL----+-- -

POPLITEAL.--+-i~

Fig. 7-22: Illustration of vascular sites having greatest

predilection for the development of atherosclerosis (from
Spain, 1966).
386 CHAPTER 7

But at least there are regions which one can characterize in

this fashion as long as one realizes that the underlying
reasons may be totally different and not shear dependent.
In this context, some of Dr. Meyer's data certainly suggest
that the preferential regions are ones where maximum stresses
would not be expected to occur and where, if anything, the
opposite is true, i.e., they are suspected of being low
shear regions.

Recently there have been at least two groups that I

know of who have been trying to determine in some detail
and quantitatively where lesions initiate. One of these
efforts is at NIH. Dr. Fry is not here today, and I am not
sure how far along this effort is; however, an effort is
being made to quantitate data on the distribution of fatty
streaks and lesions in the aorta from subjects that I believe
were obtained from the Baltimore County morgue. The other
approach involves cholesterol-fed animal experiments and is
the work at the University of Western Ontario by Fred Cornhill
(13) and Margo Roach (14). Here they have used a rabbit as
a model. They have looked at branch vessels eminating from
the aorta, and using a polar coordinate system, they have
mapped the pattern. This is illustrated in Fig. 7-23. The
lesions here represent staining with Sudan III so they are
sudanaphillic lesions. The solid line here is the outline
of the celiac artery orifice and the dashed line is the
outline of the staining as measured for that particular
case. Combining all these results, one obtains a pattern in
terms of a polar coordinate system--it is nothing more than
a quantitative way of pinpointing the orifical pattern--for
particular location in an animal, and from this there is a
def£nite preference for the distal side of the orifice in
terms of flow direction. Drs. Cornhill and Roach have
looked at a number of arteries branching off the aorta
including the coronaries, the intercostals, and the renal
arteries. Although there is some difference, it is the
distal side that is favored in this particular experiment.
Note, however, that this is not normal human atheroma. This
is a feeding experiment and the distal side in terms of
stresses would be expected to be a high shear stress side
although again I would add the disclaimer that I earlier
voiced.

DR. CARO: Can I ask, were they short or long duration

experiments?
 :I:
m
s::
o
o
-<
z
l>
s::
o
1·05 o
COeliAC
o
Z
--i
:IJ
E OJ
E C
--i
~ 1' 0 o
o« Z
~
--i
... o
l>
o --i
:I:
m
~ 0 ·05 :IJ
o oen
...Z
~
o
r
m
:IJ
oen
o t.. • • • • .... ,+ •••••••• en
o 90 180 270 360
DEGREES fROM POLAR COORDINATES

(a ) (b)
Fig. 7-23: (a) Polar coordinate representation of aortic lesion (dashed line) around the
coeliac orifice (solid line); (b) rectangular coordinate drawing from (a) with the dotted
lines showing the SEM of 5 measurements. Note that the lesion is entirely distal (from w
Cornhill and Roach, 1974). ex>
......
388 CHAPTER 7

DR. NEREM: I think these are eight-week experiments.

The rabbits were on egg yolk with rabbit pellet diets. I
mention this because I think we need more detailed information
of this type and because I believe that this pattern of the
disease is the chief, if not the only, indirect evidence for
fluid mechanics--hemodynamics--being a primary or any kind of
a signficant factor in the disease process. I thus would urge
people, who are in a position to obtain this kind of data, to
help us in developing this type of quantitative approach and in
developing data bases which look very carefully at differences
between animals and between spontaneous versus induced atheroma
so that ultimately we may be able to answer this question in a
very definitive way. This is one series of questions that we
should be able to answer if we are able to organize our efforts.

In talking about fluid forces being involved, we should

recognize that we are talking about either the pressure or the
frictional force on the wall, i.e., the wall shear stress, the
frictional force or shear stress very much influenced and
determined by the detailed flow characteristics that are
associated with the particular vascular flow situation, the
local geometry and wall elastic properties and so forth. Before
going on to talk about the coronary system, I would like to say
just a word about these forces. First, if we talk about
pressure, the main reason for our interests is that hypertension
is the one risk factor which seems to have some relevance to
hemodynamics being involved. However, we are in a rather poor
state in terms of understanding what that pressure influence is,
i.e., what the role of hypertension is in terms of risk factor
information. For example, is it a mean pressure diastolic/
systolic pressure, or a pulse pressure effect? Are we
talking about vessel distension, are we talking about an
increased driving force across the wall in terms of the
transport of materials or are we talking about distortion of
the vascular geometry--the latter certainly has some appeal
to it. As I mentioned earlier, with an increased mean
pressure there is also an increased pulse pressure, and with
an increased pulse pressure, one would expect considerable
alteration in the flow characteristics, particularly in the
shear stress levels. A higher pulse pressure generally
would mean a higher peak velocity, but also a change in the
velocity wave form so as to give rise to higher shear stresses.
Thus, what exactly is the role of pressure in this whole
problem?

In considering the role of shear stress, Dr. Dewey gave

us some background as to where we think we are in terms of
understanding the general magnitude of these stresses.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 389

However, I would like to make a few additional points with

the next few figures. Fig. 7- 24 presents some calculations
by Pedley (15) of aortic wall shear stress. These are based
on the measured ascending aorta velocity waveform which is
presented. Although you can measure velocity profiles with
such instruments as hot-film anemometers and pulsed ultrasonic
doppler units, to determine a wall shear stress from a
velocity profile is a highly risky business. One approach is
to take the basic waveform and use that as the input to a
theoretical calculation. There are a number of ways to go
about this and such calculations only represent reasonable
estimates. Pedley, in taking the wave form here which was
measured by Nerem, Seed, and Wood (16) at Imperial College,
arrived at these estimates for shear rate.

There are a couple of things to note; first, the peak

shear rates independent of whether you are talking about
forward flow or reverse flow, reach a level of 4000 inverse
seconds. In terms of shear stress, that would be something
over 100 dynes per square centimeter. The interesting thing
is that there is at least as large a stress during the
period of reverse flow as there is during peak systole.
This is because as the flow reverses very steep velocity
gradients at the wall result so that the reverse flow may
have very high stresses associated with it. A second important
point is that the peak shear stresses associated with such
pulsatile flows are in order of magnitude larger than the
mean stress associated with the basic mean flow through the
vessel. Thus, when we are trying to understand the pattern
of a disease and the hemodynamic involvement, I think one
has to be looking at more than just mean flow characteristics.

At NIH, Dr. Robert Lutz and his colleagues (17) have

been carrying out some experiments in a plastic model of the
aorta, i.e., an actual cast of
Measurements of Shear Stress the aorta of a dog and some of the DOG
by a Plastic Model of Dog's major branches. The measurements
Aorta presented in Fig. 7-25 were carried
out with an electrochemical
technique and for steady flow. Without going into a lot of
detail, on the flow divider--Lutz measured a peak shear rate
of 700 inverse seconds which would be about 30 dynes per
square centimeter. This is the peak stress in the region of
the flow divider and this for a steady flow. If one allows
for a pulsatile flow through that same type of geometry, one
could again expect that the stress associated with the
pulsatile flow might be as much as an order of magnitude
greater than what Lutz measured.
390 CHAPTER 7

4000
Shear
rate(s" )
100
2000
U(t)
(ems')
50 0 -3 t(s) 0 -4

t(~ 0-1 0-3 0-4

0
-2000

-50
-4000

(a ) ( b)

Fig. 7-24: Calculation of aortic wall shear rates using velocity

waveform in (a) time history of wall shear rate shown in (b) for
different values of x', the ratio of the distance from the aortic
valves to the distance flow travels in one pulse (from Pedley,
1974).
 HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 391

FEMORAL SECTION
VENTRAL
TOP VIE\.J

Electrodes

700 700

600 I\
~ Inner \.JaIl 600
I ,
I , VI
I , ~

500 500 Cb
I ,
.,
,-..
U
I ,, II>

,
Q)
VI I :;::;,
400 I 400 II>
I ~

.... ro
Q)

,-"'""',,,,
I
ro
c:: 300 300 ......
Vl
ro
\...
ro ,, n
Q)
.r:.
VI
200 , 200

100 Outer \.JaIl 100

0
17 18 19 20 21 22 23 24 25 26 27

Electrode Position, CM

Fig. 7-25: Shear stress distribution pattern on ventral side of

model femoral artery section. Electrode positions are indicated
in the schematic drawing of the artery at the top of the figure.
Electrodes were placed on the inner wall, midline, and outer
wall of the femoral branch (from Lutz et al, 1974).
392 CHAPTER 7

Thus, instead of 30 dynes per square centimeter we are

talking more on the order of several hundred dynes per
square centimeter as a maximum shear stress. From either of
these two situations, one thus can conclude that a shear
stress on the order of 100 to 200 dynes per square centimeter
is not unrealistic in terms of the arterial system.

One other perspective on this shear stress problem is

given by the experiments presented in Fig. 7-26 and 7-27.
These were carried out in Aachen, Germany by Dr. Niranjan
Talkuder (18) when he was there as a doctoral student. The
model is that of a bifurcation, and pulsatile flow experiments
were carried out with Reynolds numbers on the order of a
couple of hundred, not unlike that in the coronary system.
Dr. Talukder went to great lengths to look at the effect of
flow division and some of these results are illustrated in
Fig. 7-26. Dr. Dewey in his presentation certainly pointed
out the importance of this, and this brings us back to some
of the comments in this meeting from researchers working on
regulatory mechanisms. These regulatory mechanisms are the
ones which will determine how the flow divides which ultimately
will decide what the actual stresses are that exist in the
arterial system and what the detailed flow characteristics
will be. It is very easy for someone like myself to say,
"Don't tell me about the system; I am interested in local
details," but in fact those local details are determined by
system characteristics.

In Fig. 7-27 a measurement of shear stress is presented

for a point which turns out to be in a separated flow region.
At the bifurcation, 30% of the flow is going one way and 70%
the other way, and a rather large separated region on the
inside wall results at point A. The measured stress over a
time of one cycle is presented here. It peaks at a little
over 30 dynes per square centimeter for this case. I think
the important thing is that the peak stress is again on the
order of a factor of ten larger than the mean stress that
one would calculate for the flow going through that branch.
The mean stress, tAm by the way, is very close to what you
would calculate for a parabolic Poiseuille profile, but tAP
the rms stress level is a factor of 2-4 higher and the peak
stress is a factor of ten larger--even though you are in a
recirculating separated flow region, a region which one
might be tempted to characterize as being a dead water or
stagnant region.
 J:
m
s:
o
o
-<
z
}>
s:
()
a. (')
.... 1.0 ~__ -.Q- -0 Station 0 o
"- Z
c: -I
.... ::0
0-- --~- Cll
II) '-r"1 - - C
ID -I
- 6 Station A
.....
'-I... " " -- -- -
o
( f) " ........... (0 2/0 1= 50/50) Z
-I
-.....
0
o
A ... }>
~ 0.5 (). 01].-- -I
J:
(f) -Z' .........
.................
~ -- m
~ Station A ::0
-- ------0
0
oen
c ~ - 0 0 ; Reo (02 /0 ,= 30/70) (')
0 r
m
II) ::0
c o
Q) Ol~ en
E en
-
"0 0
I
C 0 200 400 600 800 1000
0
Z Reynolds Number, Reo
Fig. 7-26: Steady flow shear stresses normalized by corresponding Poiseuille flow values
measured as a function of Reynolds number in a model bifurcation. (Talukder, 1974).

W
'0
W
394 CHAPTER 7

30 Reo=200
C\I
E 02:01 =30:70
u
"-
11'1

c: 20
Q)

>-
"0

l
10
~

0 '-+---+--+--+----+_ TAm

-t
Fig. 7-27: Pulsatile flow wall shear stresses as measured at
position A in a model bifurcation;\ T AI is root mean square value,
TAM is mean value, and TAP is Poiseuille shear stress value
(Talukder, 1974).
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 395

Secondary flows--this corkscrewing type of motion is certainly

very much associated with the asymmetries that we see through
vessel curvature and branching. Flow separation--in our
measurements we have not been able to detect flow separation,
but that is probably only an indication of the crude level
at which we are operating. I really think that flow separation
is possible, particularly on the outside walls. The pulsatile
nature of the flow--although we don't see turbulence in the
coronary system, we do observe oscillations which are of an
extremely interesting nature.

Just to talk a minute about the velocity asymmetries in

the coronary system, (Fig. 7-28 and 7-29) present our hot-
film measurements of the point
Velocity Asymmetries in velocity at various positions in
Coronary System the extramural coronary vessels
of the horse and with our probe HORSE
moving across the lumen of the vessel from the outside wall to
the inside wall. Fig. 7-28 shows a diagram of the left
common coronary bifurcation into the LAD and the left circumflex
in the horse. X marks the spot where we positioned the
probe. The vessels are curving over the heart, so that the
plane you are looking at is really a plane that is curving.
Thus, as we corne in with our probe, we are moving from the
outside wall, in terms of the plane of curvature, toward the
inside wall. This corresponds to going from your right to
your left in terms of these profiles. These are for various
times through the cardiac cycle -- one-fourth of the way,
one-half of the way, three-fourths and full stop through the
cardiac cycle. As you can see, there are only a very limited
number of data points. The vessel is maybe eight millimeters
in diameter, and the probe is on the order of a millimeter
in diameter. Thus, we are only able to get maybe half a
dozen points across the lumen of the vessel. And yet consistently
in the left common coronary, we get an asymmetry which is
essentially favoring the outside wall, i.e., the flow is
pushed toward the outside wall.

This is totally opposite to what one sees in the aortic

arch, where the flow is if anything pushed toward the inside
wall. However, the aortic arch is a totally different fluid
dynamic situation. Viscous effects are minimal there, being
confined to thin boundary layer regions. In the coronary
system, the flow is much more fully developed and for a
fully viscous flow in a curved pipe you would expect the
peak velocity to be pushed toward the outside wall. So
this type of profile is obtained from the left common coronary
artery and is consistent with what we know about viscous
flow in a pipe.
396 CHAPTER 7

It is in fact anything but that, particularly in a pulsatile

flow situation, and one gets considerable variation in
stress and considerable peaking, certainly during what would
be here representative of the systolic portion of the cycle.
This is presented only to provide some perspective on this
whole problem of how the shear stresses are acting in the
arterial system.

Obviously we need to obtain much better information on

what the stress levels are in the vascular system. Of
course, when we talk about pressure--and again I am interested
in pressure because hypertension is a risk factor which I
can't ignore in trying to put basic fluid mechanics and
transport studies together with the etiology of the disease--
but when we talk about pressure, it may not really be a
fluid mechanic or a hemodynamic effect at all. It conceivably
could be a hemostatic effect, in terms of the kind of stress
loading that Dr. Kenyon talked about. However, whatever it
is, we need better information on these stress levels and
this is something on which we should be able to make considerable
progress.

I would like to say a word about coronary flow characteristics.

Some of the characteristics of interest are asymmetry in
velocity profiles, secondary flows,
Coronary Flow flow separation and the oscillatory
Characteristics or pulsatile nature of the flow.
The view point I am corning from
is that you start out by saying I understand fully developed
laminar or Poiseuille flow and the parabolic profile--a
classical type of flow situation--and then say that one
never sees this in the larger vessels of interest to a group
like this. The lists above are the kinds of phenomena which
Dr. Dewey talked about and which really alter the flow from
the nice well behaved situation and make it interesting. I
just want to comment briefly on some of these relative to
the coronary system. Velocity profiles, the ones we have
measured are admittedly crude, but at least give us some
ideas of the type of skewing present in the coronary system.
Turbulence--I suppose we could spend a day talking about
turbulence in the arterial system, although I have seen in
the aorta velocity waveforms that look very much like turbulence,
I have never seen anything that looked like turbulence in
the coronary system. Distal to a highly stenosed region,
one presumably could have a turbulent jet, but in terms of
the normal vasculature--the normal coronary system--I don't
believe turbulence will be found.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 397

Velocity (cm/sec) Velocity (cm/sec)

wt: "'/2 • WI:".
20 • 20

•10 •10 •• •

- 0.4 0 0.4 - 04 o 0.4

Distance From {, (cm) Distance From t (cm) wt
Velocity (cm /sec) Velocity (cm/sec) "
RII ~ 241 tI ~ 36

WI :3"./2 WI: 2".

20 20

.10
••• • --/;~ortO
01'--
-.. ... em
~ u.
• U , Common ~ Circumfle•
,\46'"
Lt Anl.rior
- 04 0 0 ,4 04 0 OA Descendi"!!
Distance From { (cm) Distance From <t. (cm)

Fig. 7-28: Centerline velocity waveform, ECG, and velocity pro-

files for the indicated times and location of measurement in the
left common coronary artery of an anesthetized horseo Measurements
were performed in the plane of curvature of the artery.
398 CHAPTER 7

Velocity (cm/sec) Veloc ity (cm/sec)

wt: ,
w t : , /2
30 30 •
• • 20
•
20
••
. 10 10

•
-0.2 0 0.2 ."..". 2 .".
T
Distance From <t.. (cm)
..
Rtl 6 586
wt
a 6 3.9
Velocity (cm/sec)
u1 t : 2".
30

20
• •
••10
•
-0.2 o 0 .2
Distance From t (em)

Fig. 7-29: Centerline velocity waveforms, ECG, and velocity

profiles for the indicated times and location of measurement
in the circumflex coronary artery of an anesthetized horse.
Measurements were performed perpendicular to the plane of the
bifurcation.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 399

It also suggests that the highest shearing stress should be

at the outside wall and not the inside wall, again in contradiction
to what we think is the case for the aortic arch.

Now this is proximal to the bifurcation point. If you

move distal to the bifurcation point, the picture does not
change that much. As seen in Fig. 7-29 here again we have
the bifurcation into the LAD and the left circumflex. Our
probe is not now positioned perpendicular to the plane of
bifurcation but coming in as much as possible at an angle so
we can observe the effects of the flow being divided at the
bifurcation. Ideally you would want to position the probe
in the plane of the bifurcation, but we are only able to
come in at about a 45 degree angle. The velocity profiles
we have obtained do not suggest higher velocities being
toward the flow divider side, but rather suggest that the
higher velocities are away from the flow divider. Remember,
we are not in the plane of bifurcation. We are not orthognal
to that plane, but moving at about 45 degrees.

Furthermore, although you have the flow divider which

would be expected to produce high velocities near that
inside wall, at the same time you have a curving plane of
bifurcation which is producing its own secondary flow effect.
The net result is that the peak stress point quite possibly
has been moved away from the flow divider to a slightly
different position. There has been a rotation due to the
combination of these two secondary flows of curvature effects.

I mentioned the fact that we see an interesting oscillation

in the coronary system. Fig. 7-30 presents measurements in
the left common coronary artery
Oscillation in the of the horse, both with a hot-film
Coronary System placed at the center line of the
vessel as well as with an electro-
magnetic flow meter cuff. You can observe low frequency
large amplitude oscillations which are present both on the
hot-film and on the flow meter waveforms. We also see these
oscillations in pressure waveforms, although the amplitude
is somewhat suppressed there. However, if you take the
pressure waveforms and try to compute a pressure gradient,
then again one sees rather large amplitude oscillations
which, of course, would have to be present in as much as the
flow is pressure driven. As we move the probe across the
lumen of the vessel, we see no change in the phase of these
oscillations. Thus, the entire flow seems to be sloshing
with this five to ten Hertze frequency.
400 CHAPTER 7

em/sec
30 Hot-Film
Centerline
Velocity

ml/m in
1000 EM
Flow Rate

ECG
Lead I

I Second

Fig. 7-30: Simultaneously measured centerline velocity waveform

and electromagnetic flow meter waveform together with ECG for left
anterior descending coronary artery in an anesthetized horse.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 401

We believe that this is a reflection type phenomenon, some

type of resonance phenomenon, and to test this, we have
devloped a computer model which includes a number of non-
linear characteristics of the system, e.g. the wave speed
dependence on pressure and some of the elastic characteristics
of the wall. Fig. 7-31 shows a comparison between this
standard computer model and a flow measurement from a horse, HORSE
the former using pressure measurements from the horse as
input to the computer program. The solid line is the experimental
data -- note again the oscillations, and the dashed line is
the computed waveform. Agreement is certainly not especially
good. However, we hope to be able to modify the program
systematically and parametrically to do better. One of the
things I need to point out is that with this computer program,
which in the diastolic region actually does a reasonable job
of predicting the oscillation characteristics, we can essentially
do a parametric investigation on the effect of size on this
typed phenomenon. What we find is that as one goes to
smaller and smaller animals, the frequency of the oscillations
increases slightly and the amplitude of the oscillations
decreases considerably. There is some evidence that oscillations
like these are present in dogs. Dali Patel at NIH has seen DOG
oscillations in some animals. Heinz Pieper in our Physiology
Department at Ohio State, has seen oscillations in a number
of animals. Some of Gregg's classical waveforms also show
oscillations of this type. In the horse we see these oscillations
almost all of the time, while in smaller animals the amplitude
is less, as we predict with our computer model and the
observations are not as consistent. I think when you get
down to smaller animals as opposed to the horse, the presence
of such oscillations is very sensitive to the exact wave
speed associated with the coronary vessel or system.

DR. COX: Do you see these same kinds of oscillations

in the anterior descending coronary artery or only in the
left circumflex?

DR. NEREM: You see them also in the LAD and in the
circumflex.

DR. COX: What about the right artery?

DR. NEREM: We have not been over on the right side.

DR. COX: We see them always in the left circumflex

coronary, but very rarely in the anterior descending coronary.
402 CHAPTER 7

70

60

50

u
w 40 \
I \
,
~
E 30
u \
-
>-
u
0
Qj
20 \
\/
I
>
10
- In Vivo Horse
- - - Standard Computer Model
0

-10

0 0.2 0.4 0 .6 0.8 1.0

Time, Percent of Cardiac Cycle
Fig. 7-31: Comparison of standard computer model calculation and
in vivo flow velocity measurement from a horse at a position 15 cm
distal to the left coronary ostium.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 403

I am just wondering to what extent these oscillations may be

related to compression by the cardiac muscle since there is
a very large difference in the distribution of flow in the
circumflex versus the anterior descending artery.

DR. NEREM: Our model allows for flow into branch

vessels and we have parametrically varied that effect. It
does not seem to have a major influence on the predicted
oscillation characteristics. That is all I can say in response
to your question.

DR. MANSFIELD: In man the right coronary is usually,

not always--but usually, the dominant vessel as far as flow
reception is concerned. The flows you are talking about on
the left side are mainly diastolic flows. In the right
coronary vessels there is relatively low pressure, and
systolic. So where you find your oscillations will depend
on the size of the vessel, the line of flow and whether they
are systolic or diastolic.

DR. STONE: May I add to that. One difficulty with

what you just said, which is true, is that the flow distribution
from the right to the left side would be under the same
physical principles as the left load. The flow penetration
to the myocardium through that right artery will still be a
diastolic component.

DR. MANSFIELD: The number in man is much smaller, it

is only about 20% versus 80% of the flow in man going through
the right coronary.

DR. STONE: I understand that, but what I am saying is

that the flow distributed from the right coronary artery to
the left ventricle to the posterior septal wall and the
posterior ventricular wall, will be exposed to the same
physical force as anything coming through the left side, so
the flow penetration from right to left will still be diastolic.

DR. NEREM: Even though we do have turbulence, there

are waveforms in the coronary system which can be very
complicated and the flow that goes with them can be equally
complicated. We have not investigated at all yet how certain
pharmacologic interventions would affect such oscillatory
activity. That is an area that Lowell Stone and I have
talked about and think it would be well worth pursuing. If
in fact as Dr. Dewey says the coronaries are where the
action is in terms of ultimate pay-off, there is a lot of
work for fluid mechanics researchers to do in terms of under-
standing better what is going on in that coronary vasculature.
404 CHAPTER 7

Let us approach how fluid mechanics might interact with

the endothelium. Fig. 7-32 is just one illustration of the
atherogenic process and I do
Two Types of Roles for not want us to feel constrained by
Hemodynamics or Fluid this particular series of events,
Dynamics in Disease but just to raise what I think is
Process a very important question and one
in which I know Dr. Werthessen
is interested. As you follow through with some disease process,
starting with a normal intact endothelium and proceeding,
we essentially can talk about two different types of roles
for hemodynamics or fluid mechanics. One is in terms of
primary endothelial damage. This could be a very traumatic
insult to the system or it might be a different type of
primary damage in the sense that, when we are talking about
a disease that has a 40 year time course, primary damage
presumably can be very subtle. However, at least one possibility
is some way to have hemodynamic factors involved in the
primary damage. The other possibility is that somehow
hemodynamics gets involved in the wall's ability to respond
to whatever insult has been provided it, e.g. that could be
through some sort of platelet adhesion process such as I
will talk about in a minute. But I do want to raise the
question -- knowing we are not in a position to answer it
today -- that hemodynamics, in general, even if it does play
a role, and as I mentioned earlier the evidence is indirect,
still could be either primary in nature or it could be a
response type or mediating role. Now, let us talk briefly
about some potential mechanisms through which hemodynamics
could enter into the process.

We have all been talking about yield stresses, about a

magic number of 400 dynes per square centimeter; in addition,
Dr. Mansfield had some very interesting results from his
studies on the influence of flow on the endothelium. One
view is that in some way the flow irreversibly alters
endothelial cells, that there is a primary physical insult
to the endothelium that comes from the flow. The stress
levels I mentioned earlier don't make this totally impossible.
However, I would have to admit that my own feeling is that,
for the kind of disease process about which we are talking,
I would not vote for that one.

Beyond actual physical trauma, there is the whole area

of mass transport.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 405

Intact Endothelium

!
Primary Endothelial Damage

physical trauma
biochemical alterations
hypoxia

+
Platelet Adhesion and Release Reactions

I
~ l
•
Single Injury,
Multiple Injury
Short Term

Endothelial Repair
+
Atherosclerosis

Fig. 7-32: Illustration of possible steps in atherogenic

process.
406 CHAPTER 7

There are a lot of materials moving in and out of the wall

and certainly fluid mechanics
Wall Permeability a can, in some way, playa role
Function of Wall Shear in accelerating or decelerating
Stress certain transport processes.
Now much of the emphasis has been
on the transport of cholesterol and lipoproteins. This
is true of the in vitro work, in which Dr. Caro and myself
(19) and also Dr. Carew (20) have been involved -- also some
of the work of Dr. Schwartz. We do not know the roles these
molecules play in the disease process, but it is still of
interest to study the transport problem. One of the things
that has been determined is that there is a dependence of
the transendothelial transport rate on the shear stress
applied to the endothelium. If you take all the available
data and try to correlate it into a single picture, you
obtain a pattern such as illustrated in Fig. 7-33 where the
wall permeability, a characteristic velocity for the wall
transport process in centimeters per second, is presented as
a function of wall shear stress. There are two points I
want to bring out here.

First, if you take the low shear stress data, i.e. the
work by Dr. Caro and myself and some of the later work,
there is a very weak shear stress dependence. If you talk
about the higher shear stress levels, the first piece of
work in this area was Dr. Carew's Ph.D. dissertation. We
have subsequently done some experiments at high shear stress
levels (21) and in this high shear stress region, there is a
much steeper dependence. This suggests, in fact, that there
may be more than one mechanism that is playing a role in
this shear stress dependence of wall permeability. So we
need to bear that in mind, that whether these shear dependent
transport processes are intrinsic to the disease process is
really open for question at this point.

Now the other point I want to make is that the transition

from one type of shear stress dependence to another type of
shear stress dependence occurs
Pulsatile Shifts from around 50 dynes per square
Low to High Shear Stress centimeter. Based on my values
for the level of shear stress
in the vascular system, it would appear that the mean shear
stress corresponds to the low shear stress dependence region.
However, as the flow pulses, at least in the aorta and
certain regions of the coronary system, the shear stress
moves up into the high shear stress dependence region.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 407

U
10- 6
<l.l
(/)
.........
E
u

0
u
.........
·E 10- 7

10-8L----L__L-~-LLU~__~__~~~~LL______~
1 10 10 2

Wall Shear Stress dynes /cm 2

Fig. 7-33: Illustration of shear stress dependence of wall

permeability (m/C o ) to blood-arterial wall macromolecule
transport; m, wall transport rate, and Co' blood concentration.
408 CHAPTER 7

So you are moving back and forth from a low shear stress to
a high shear stress dependence, and if there are two different
mechanisms involved here, then you are moving from one type
of shear stress influence to another type of shear stress
influence and back again as the flow pulses. Some of the
experiments we have carried out -- as well as those by
others in this area too -- suggest that the arterial wall is
in fact very sensitive to flow pulsations. To quickly
sketch an experiment that we completed approximately nine
months ago (22), we were interested in the transport of
serum cholesterol into the arterial wall for pulsatile flow
conditions. This was an in vitro experiment using excised
carotid arteries. l4C cholesterol was introduced into the
serum and we used that tagged molecule to measure the transendo-
thelial transport rate -- not looking at wall concentration
gradients, as Dr. Colton quite correctly has suggested we
should start doing, but just looking at a total wall uptake.
It is purely an oscillating flow. There is no mean flow at
all. In that oscillating flow situation, we find that the
rate of wall uptake of this tagged cholesterol molecule is
dependent on both the frequency of pulsation and the amplitude
of pulsation. In fact, however, the data all correlate with
peak shear stress, i.e., one can collapse the frequency
dependence and the amplitude dependence into a dependence on
peak shear stress --actually peak shear rate -- which gives
results which are quite consistent with the steady flow
experiments summarized earlier. Now as you can see in Fig. 7-34
there is considerable scatter in the data, and yet there is
still an unmistakable trend that the wall permeability is a
function of the wall shear rate. As may be seen, this is
true both for cholesterol and albumin. There is a definite
shear dependence of wall uptake even in this case where
there is absolutely no mean flow at all and where we have
oscillations on the order of 1 to 4 Hertz. This is a relatively
low frequency, but on the same order as some of those oscillations
I showed earlier in the coronary system.

So much for transport phenomena; now let us talk about

some other possible mechanisms through which hemodynamics
might playa role.

Actually the next one I'll mention is a transport

process also. I don't know whether Dr. Dewey and I can agree
on this one, or may be we
Hypoxia--Diffusion would agree and that agreement
of Oxygen through would be in disagreement with the
Blood to the Wall investigators who are working
specifically in this field.
 HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 409

70~--~1----~1~--~1----~1~--~1----~1~----~--~

60 I- -
• 131 I - AI bumtn Uptake

50 I- o 14C -4 Cholesterol Uptake -

U
<I>
II>
"-
•
B 40 f- o -
l
o 30
(D

• o -
I- o o
~
·E
o

0 o
••• o
o -
20 I-
o o •
dlo.
DO
~O en -
101-
•
o
0
•
0 •
I I 1 I I I I
00 20 40 60 80 100 120 140 160

Fig. 7-34: Wall permeability (m/C o ) for the transport of

l4C-4-cholesterol and 13lI-albumin in serum perfused carotid
arteries as a function of peak wall shear rates; m, wall
transport rate, and Co' serum concentration (from Mack, 1975).
410 CHAPTER 7

However, there is a sizeable effort out at U.S.C., as part of

their atheroscleorsis project, to look at hypoxia, particularly
in terms of hemodynamics playing a role. In other words, it
is their contention that the rate limiting step in the
transport of oxygen to tissue is not the movement of oxygen
through the tissue, but the diffusion of oxygen through
blood to the wall. Back (23) has made estimates of the
resistance on the plasma side of the interface versus the
wall side of the interface, and has come up with a ratio on
the order of a factor of two. I think it is a very preliminary
estimate, and I will say that the few oxygen tension measurements
through the wall with which they have been able to compare
their calculations are in agreement. I do not think we can,
at least on the evidence we now have, discount the possibility
of fluid mechanics influencing oxygen transport and playing
a role in hypoxia such as Dr. Smith suggested. Here I am
suggesting, however, that in addition to looking at how far
a location in the wall is removed from the endothelium, we
also need to look at transport characteristics in the plasma,
particularly where there are separated flow regions. Although
some separated flow regions would be very highly disturbed
and one would expect very good mixing, I am not sure that
this is true of all separated flow regions, for example the
outside wall in the region of a bifurcation. So, I think
hypoxia is a possible mechanism through which fluid mechanics
could exert an influence and which cannot be discounted.

Platelet deposition is another mechanism through which

fluid mechanics can playa role. We don't know a lot about
this either, but this is
Platelet Deposition: certainly the area where some of
A Mechanism through which the earlier work of ten or fifteen
Fluid Mechanics Plays a years ago on hemodynamics as a
Role factor was carried out. The
hypothesis is that flow somehow
influences the deposition and adhesion of platelets to the
wall and that when the wall has been insulted, it is not a
question of where the insult occurs, but where hemodynamics
will preferentially allow platelets to interact with the
wall and which thus determines the pattern of the disease.
Many of you know of the work of Perry Blackshear's group at
the University of Minnesota, and based on an admittedly
specific model as to how platelets get to the wall, they
predict that platelets will deposit on the wall based on a
parameter which involves the wall shear rate (24). This
parameter involves the shear rate in the denominator.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 411

Thus, low shear regions correspond to a large value of this

parameter, and on this basis they hypothesize that low shear
regions are the ones where you will find the disease developing
because that is where platelets will preferentially deposit
and interact.

DR. CARD: There is a requirement for a very substantial

filtering.

DR. NEREM: Yes, it includes filtering of fluid through

the wall. That is correct. There is one other possibility
that I will mention. Markle
Aortic Histamine Synthesis and Hollis at Penn State (25)
Alters Wall Permeability have been doing some work in
which they have been looking at
shear stress and its influence on albumin uptake. They also
have been looking at the effect of shear stress on aortic
wall histamine synthesis and they find a relationship that
we are totally in the dark about at this point that suggests
a different way for fluid mechanics to play a role. One
could hypothesize, purely for the sake of argument, that one
way wall permeability gets changed with wall shear stress is
that the shear stress induces aortic histamine synthesis
which alters the wall permeability. In other words, it is
this biochemical step which is influencing wall permeability.
I am not here to either support or detract from any such an
hypothesis, but only to suggest that there are a lot of
things that we don't know anything about in terms of a
hemodynamic influence and that is a definite possibility.
This also raises questions about other kinds of influences
on the biochemistry of the artery that might be present.
How does the wall respond to stresses that are imposed on
it? Is it just a mechanical response or is there something
else? Some of Dr. Smith's comments -- certainly in terms of
collagen synthesis -- would suggest that the wall does
respond in a chemical or metabolic way. So the question, or
at least another question, is through which of these mechanisms,
or through whichever other mechanisms we have not identified
yet, does hemodynamics playa role in the disease process?

I now come back to the following question: In terms of

the coronary system, where does atherosclerosis first develop?
For example, in a region of a
Where does Atherosclerosis bifurcation, is it on the flow
First Develop? divider? If so, is that because
of high shear stresses that are
denuding the endothelium?
 412 CHAPTER 7

Is it because of increased synthesis of histamine? Or, in

fact, does it start first on the outside of the wall? If it
is on the outside wall, it makes a big difference as to
whether this outside wall is slightly upstream or slightly
downstream. This is because there is possibly a separated
flow region and this has certain implications in terms of
the various types of mechanisms I have talked about. If it
is slightly downstream and in fact is in the region where
this separated flow is now re-attaching, then this is where
high shear stresses would be expected to occur. So there
are some basic questions about the pattern of the disease
that do relate to a mechanistic point of view.

The same question can be raised about regions of sharp

curvature in the coronary system. Of course in comparing
the coronary system with the aortic arch it must be remembered
that we expect in the left common coronary artery, based on
HORSE our horse experiments, that the higher shear stress will be
on the outside wall, and yet in the aortic arch we expect
the higher shear stress to be on the inside wall.

But where in fact does atherogenesis initiate and, of

course, what mechanisms then seem viable? And then finally
I bring us back to the earlier illustration of the pattern
of the disease. What are we talking about when we attempt to
explain this pattern? People like myself would like to
believe that we are talking about local flow details; however,
as I mentioned earlier those local flow details could very
well be determined by regulatory mechanisms. If we consider
the major left side coronary branch, how much of the flow
goes the circumflex route versus the LAD route? The answer
depends very much on peripheral resistance and how that flow
is going to divide. Of course, maybe it is not a hemodynamic
problem; maybe it is a hemostatic problem and it is just the
basic pressure loading on the system with stress concentration
in regions of certain types of geometry and where certain
types of tethering are important. I guess all I can do is
add to the list of questions and hopefully stimulate some
discussion.

DR. WERTHESSEN: In your Fig. 7-34 you showed the

predisposition of various areas of the vessel to become
atherosclerotic. Haimovici did
Haimovici's Translocation the classic experiment of
Experiment switching sections of the dogs'
aorta around (26). Then he put
the dog on atherogenic diet.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 413

The visceral section which is predisposed in the normal animal,

remained predisposed even when in the chest. I have a hunch
that with the improved technology available today, you could
do some really beautiful experiments.

It is also true that there is a drastic difference in

the biochemical aptitude, if I can use that term, along the
length of the aorta. For example, in both swine and bovine
aortas, there is a 10% difference in concentration of cholesterol
between the aorta and the bifurcation. There is also a wide
range of change along the aorta in its biosynthetic capacities.
Those differences could provide fluid dynamicists today with
today's surgical technology, the means to do some critical
experiments. You can take out a patch and put it somewhere
else and see how it misbehaves or rearranges its metabolism
to fit its altered location.

DR. SMITH: I am interested in Dr. Werthessen's description

of Haimovici's translocation experiment, because working
constantly with human aorta one is enormously impressed by
the different type of lesion that is developing in the lower
abdominal aorta compared with the thoracic and upper abdominal
aorta. It is certainly a more fatty lesion, it is very
different in character, so I think that this translocation
experiment is extremely interesting and I do think that you
are right and that this could very usefully be studied
further. Now this raises a whole lot of problems.

One of the things we looked at some years ago was the

relative concentrations of intact LDL in different segments
of the aorta. These had very
Translocation Experiment clear distribution. If we took
Raises Many Problems the upper abdominal as 100% the
thoracic segment was somewhere
around 150% and the abdominal segment about 95%. So, in fact,
the retention of lipoprotein was lowest in the abdominal
segment and highest in the thoracic segment (27). One
tends on the whole to get large fibrous plaques in the
thoracic segment and these very insidious fatty type plaques
in the lower segment.

Another question. All this beautiful work on vesicular

transport and yet nobody, as far as I can see, knows what
the condition of the endothelium in the middle aged human
is. Do we, in fact, actually have a beautiful endothelium
as has been shown in the experimental animals?
414 CHAPTER 7

Or, by middle age have we got an endothelium that is pretty

fatty so that there is passage of lipoprotein between cells
as well as vesicular transport? I don't know how this
question can be answered but it seems to me that it is
extremely relevant when one is thinking of the endothelial
barrier. And going back again to the transposed aorta it
seems to me that it is extraordinarily difficult to isolate
the structure from atherosclerosis. I suppose that these
were adult dogs that Haimovici was using; if you did this to
a newborn puppy and let the animal grow up with this patch
and after it was fully grown fed it cholesterol, I wonder
what you would get then? This brings me around to Dr.
Meyer's work. Dr. Nerem and I tried to do a joint exercise
to see if we could decide where the lesions developed at the
bifurcation. So far the results have been so complicated
that we have not been able to arrive at an answer.

DR. CARD: What bifurcation was this?

DR. SMITH: The iliac bifurcation. The wall is different

around the iliacs, and you have thin patches and thick
patches. Are these really innate structural differences or
are they early atheroma? How do they influence atheroma
development? This is extremely relevant at branching
points where the whole architecture of the underlying vessel
is changed. I find it extremely difficult to know whether
an intimal cushion is atherosclerosis or architecture.
Quite apart from the fluid dynamics, we do not even know
what the actual structure itself is doing?

DR. MANSFIELD: A couple of points about the translocation

experiments. It is a very interesting way of going at it.
You have significantly altered the segment that you are
talking about and when you have denervated it because you
have removed it from its initial nervous connection, you
have also removed at least the common pathways in the vasa
vasorum in that area when you are translocating, for example,
from the abdominal aorta to the thoracic aorta. So I think
you have to be very careful in terms of your interpretation
of what you find. It is not at all infrequent as you may
see, for example, clinically to take an artery from one area
and move it to another. For example, the hypogastric artery
moved up and used as a bypass for the renal artery, the flow
going to each of these then becomes different and it is a
denervated segment and it does tend to get quite atherosclerotic
and dilated with time, but it does not tend to obliterate
its lumen, although it may increase its size three and four
fold.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 415

Why this should occur under certain circumstances we really

cannot explain. But part of it is probably due to the fact
that we have changed the base line and we have denervated a
segment. These are important considerations when we are
doing that type of experiment.

DR. WERTHESSEN: Do these transplants ever become

innervated? Have you been able to check? The reason I
asked is that we have data that indicates that denervating
the artery alters lipid metabolism (28).

DR. MANSFIELD: I would not be surprised. Reinnervation

has not been thoroughly studied. You can demonstrate,
however, that the vessels can respond locally to drugs.

DR. MALLIANI: We should indicate that atherogenesis is

likely to be the result of very many different processes,
different in the various species and in the various types of
patients. In others words, a kind of common end point for
different pathways. In addition, it is obvious that in the
course of decades of life the mUltiple factors may variously
act at different times.

DR. BJORKERUD: Dr. Smith suggested possible presence of

injured endothelium and/or changed endothelium in the artery
segments of human arteries obtained at surgery and kept
alive in tissue culture medium for a short time. We stained
the specimens by a technique which is rather similar to the
one used in the tissue cultures to indicate decreased viability,
the so-called dye exclusion technique. We tried a lot of
stains for this purpose and found the stereo isomer Trypan
Blue, Evans blue to be most suitable. Fig. 7-35, shows a
segment from a human femoral artery cut open; the blue spots
are endothelial cells that cannot exclude the stain, i.e.
they are, to some extent, injured. Interestingly, there are
fewer injured cells around the orifice of a branch which is
difficult to reconcile, I think, with current hemodynamic
concepts. The explanation could be that the branches constitute
a source of new endothelium for areas in the major vessel
subject to severe strain. Another thing which is interesting
and should be emphasized is that more visible (i.e. stainable)
lipid is present in areas with intact endothelium. However,
this is histochemically stained lipid. As measured with
chemical methods there is much more lipid in areas with
injured endothelium. I think one should keep this paradox
in mind when interpreting histological slides, because
strong lipid staining does not necessarily mean there is
more lipid there.
416 CHAPTER 7

Fig. 7-35: View of inner surface of an area of a segment of

the femoral artery from a 36-year-old man. The leg was removed
for clinical reasons and the artery subjected to dye exclusion
test with uncomplexed Evans blue immediately after excision.
The black spots (arrows) are nuclei of injured endothelial
cells which take up the stain. Uninjured endothelium does not
stain and appears white. In the upper left corner is the
opening of a small branch. Ca. 40 x.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 417

I want to make a comment on what Dr. Nerem said. Is

endothelial injury really something of importance or is it a
secondary rather unimportant
Mechanical Factors may be factor as compared to classical
as Important as Lipoprotein atherogenic factors such as
Abnormalities in hypercholesterolemia? If
Atherogenesis endothelial integrity changes
from the most integrated endothelium
to the most disintegrated endothelium which we find in the
normal non-manipulated rabbit, we find an increase of lipoprotein
transfer -- as evaluated from transfer of labelled cholesterol
ester, of about 4 times. If we denude endothelium, i.e. we
create a situation with minimal endothelial integrity. The
transfer is about 30 times higher. These figures should be
compared with the ca. 10 times increase of lipoprotein
transfer which occurs when altering the serum cholesterol
level in the rabbit from ca. 50 mg% to 400 mg% (29). This
might indicate that the state of the endothelium as influenced
by mechanical factors may be as important as lipoprotein
abnormalities in atherogenesis.

DR. COLTON: What was the label of cholesterol you

used?

DR. BJORKERUD: It was tritiated cholesterol.

DR. COLTON: How was that incorporated into the lipoprotein?

DR. BJORKERUD: We labeled the lipoproteins with the

technique described by Whereat and Staple (30). This technique
has been shown not to denature the lipoproteins.

DR. COLTON: Did you determine the extent to which you

had labeled cholesterol ester and free cholesterol?

DR. BJORKERUD: Yes. The radioactivity was rather

evenly distributed between the d- and the combined B- and pre -B
fractions at all time intervals tested during 24 hours after
the injection.

DR. CARO: It would seem entirely analogous to what Dr. Fry

and we find, something like a six to ten fold increase in
flux when the vessel is denuded of endothelial cells. On
the other hand, I do not think this helps to answer the
question of what happens in an experiment of twenty years
duration.

DR. TAYLOR: I will review two phenomena of arteriosclerosis

and arterial repair which, in the main, defy teleological
exp lana tions •
418 CHAPTER 7

1. The immunity of healed intimal arterial scars to fatty

atherosclerotic deposits baffles me; these intimal scars
must have first healed in
Immunity of Healed animals with normally low serum
Arterial Scars to lipid levels. My co-workers and
Atherosclerosis I, did such a study in 1963 (31);
this study was done in monkeys.
Review of the literature, in depth, revealed that a similar
phenomenon had been observed in rabbits in 1929 (32).
Interestingly after damaged arterial scars had been permitted
to heal for 3 months and animals were then made hypercholesterol-
emic, those portions of arteries with no healed scars were
highly susceptible to atherosclerosis whereas the scars that
had healed while serum lipid levels were normal were immune
to lipid deposits (atherosclerosis).

2. In 1960 Economou et al., (33) produced aortic injury in

dogs by employing several methods. Two methods resulted in
central and/or external medial
Restoration of Functional damage with no injury to the inner
Vascular Tissue media, internal elastic membrane,
subendothelial space and endothelium
Fig. 7-36 injection of Urokon (70% acetizoate) into central
media, and Fig. 7-37 excision of the outer 1/2 to 2/3 of
the aortic media. In both of these types of aortic injury
there was no reparative reaction and aneurysms present at
the time of surgery, still persisted 30 weeks later, Fig. 7-38.
Two different forms of injury resulted in damage to the
subendothelium and inner media Fig. 7-39 transmural freezing
of the entire aortic wall, (33) and Fig. 7-40 - surgical
removal (employing a cork borer) of the endothelium, subendothelium
and internal elastic membrane and inner media. Within 6
weeks the aneurysms produced by transmural freezing Fig. 7-39
or excision of the inner media, subendotehlium and endothelium
Fig. 7-40 were completely healed by the proliferation of new
vascular tissue. In a third study we removed about 1/2 of
the outer media then tramsmurally killed all medial cells
of 1/2 of the aneurysmal lesion by hypothermal injury Fig. 7-41
and diagrammatically shown in Fig. 7-39 and illustrated in
the photomicrograph Fig. 7-42 Only the transmurally frozen
half of the lesion had a reparative intimal scar and half
its aneurysmal dilatation had been corrected Figs. 7-41 & 42.
Our group concluded that in order for completely restorative
arterial repair to occur the subendothelial space and internal
elastic membrane must be stimulated by injury; we feel that
there are multipotential cells in the subendothelial space
which respond to injury by rapid proliferation and formation
of new elastic, collagen and smooth muscle cells to restore
an adequate quantity of functional vascular tissue.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 419

A
c
~
U)JJJljJ,j)JI",~;;j;i-~"II!))J/'

Fig. 7-36: Diagrammatic illustration of method of intramural

injection of 70 per cent acetrizoate. A. A 45 degree curve in
the needle makes it easier to maintain the orifice of the needle
in an intramural position. B. The contrast material dissects
laterally and separates the elastic lamellae. a. Adventitia;
b. media; c. internal elastic membrane; d. endothelium. C.
Aneurysmal dilatation observed as early as 3 weeks after intra-
mural injection of the contrast material. D. Microscopically
there is destruction of the central portion of the media. There
is no apparent injury to the internal elastic membrane and sub-
endothelial layer. No subendothelial reparative scar was noted
as long as 30 weeks after intramural injection of Urokon
(acetrizoate).
420 CHAPTER 7

Fig. 7-37: Diagrammatic presentation of formation of aneurysm

by stripping of outer 70 per cent of the medi.a. A. With gentle
strokes of the scalpel the desired depth of di.ssection is
reached. Band C. at the depth of the incision the media could
be peeled along a cleavage plane. An elipse of the dissected
media is exised, allowing full development of the aneurysm.
D. With removal of more than 70 per cent of the media, acute
rupture of the resultant aneurysm occurred. E. Aneurysmal
formation showing stretching of internal elastic membrane.
Even though 70 per cent of the media had been removed without
injury to the subjendothelial space and internal elastic
membrane, there was not reparative response.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 421

Fig. 7-38: Persistent aortic aneurysms in dogs abdominal

aorta 30 weeks after surgical removal of 70 per cent of outer
media and adventitia. Aorta has been opened longitudinally
and shows intimal surface.
422 CHAPTER 7

JJJ1!!)j}J»)j)jJ)jJj)))}j))))))jJl ~
IJl))U!/JJ)JJJJ1JJ1U)})j})jJJJ}j11l .

Fig. 7-39: Diagrammatic illustrations of effect of transmural

freezing, (A) on arterial tissue. Immediately after freezing
aneurysm develops at site of freezing (b) and persists for
several weeks. C. Illustrates straightening and disruption
of elastic lamellae and dissolution of killed cells in area
frozen. During the first 2 weeks there was not proliferation
of multipotential subendothelial cells (C); however, during
the third week these cells proliferated abundantly and
essentially filled the aneurysmal defect (D). The multi-
potential cells differentiated and formed an essentially new
vascular wall at about the sixth week (E), after 6 weeks mature
cells and new elastic and collagen fibers have restored the
artery to its original caliber and repaired the aneurysm.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 423

~ P:%~
~-----::::--~.::.

o
~~:~~
~ -~
-
-:::::;- -- ~
-.....---::"- -

Fig. 7-40: Diagrammatic illustration of response of aorta to

stripping of endothelium and subjacent elastic lamella. A
shallow circular cut is made by gently rotating a 1/2 inch
diameter cork borer (A). Endothelium, subendothelium. and
internal elastic membrane are then dissected away leaving
shallow circular defect (B). As shown in C at about 3 weeks a
thick tuft of subendothelial cells and a thin layer of endothelial
cells cover the defect. At about 6 weeks the intimal scar
covering the small area of injury has differentiated into mature
vascular scar tissue with elastic and collagen fibers (D).
424 CHAPTER 7

Fig. 7-41: A and B, transmural freezing of one-half of aneurysm

which had been produced by excision of 70 per cent of outer
media. The other half was left undisturbed. C. Whereas there
was no reparative process in the portion of the aneurysm which
was left undisturbed, there was brisk proliferation of sub-
endothelial cells in the portion of the aneurysm that had been
transmurally frozen. D. The multipotential subendothelial cells
have differentiated and formed an essentially new arterial wall
in a thickened intima at the site of transmural freezing.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 425

Fig. 7-42: Photomicrograph of Weigert's elastic tissue stain of

lesion produced by excision of outer media. Cut ends of outer
elastic lamellae are visible on the lower surface of the aorta
near each lateral margin of the photograph. After excision of
the outer media two-thirds of the aorta on the right was
trasmurally frozen; 2 months later this right-hand portion of
the aneurysm showed a thick intimal scar containing a rich
elastic network. The aneurysmal dilatation on this side of
the aneurysm was markedly corrected by this intimal scar.
 426 CHAPTER 7

DR. SCHWARTZ: Dr. Bjorkerud has shown something very

important, I think. The hypercholesterolemic effects on
endothelium are probably
Hypercholesterolemic terribly important; they occur
Effects on Endothelium at very low levels before any
significant hypercholesterolemia
has in fact occurred. We, like yourself, have demonstrated
very marked increases in the influx of protein into the arterial
PIG wall in hypercholesterolemia in the pig. Stefanovitch and Gore
RABBIT have done this in the rabbit with iodinated albumin (35).
The influence of very minor elevations in the serum cholesterol
level in a matter of days or weeks produced a spectrum of
changes in endothelial ultrastructure. I think this is a
point which is perhaps not appreciated, and it appears that
hypercholesterolemia is producing in the human an enhanced
level of permeability, part of which may be accounted for on
the basis of endothelial injury.

DR. LEE: In this connection we have done a study on

thymidine incorporation and mitosis in the endothelial cell
RABBIT of the rabbit after feeding cholesterol for one day and
three days. After three days of cholesterol feeding, the
labeling indices of endothelial cells doubled from 1/2 of
1% to 1 or 2%. This indicates that certain rapid changes
are taking place in the arterial wall including endothelial
cells. The increased mitosis and labeling indices indicated
that some endothelial cells are dying and being replac~d at
a faster rate. As I mentioned earlier, after the dead
endothelial cells are washed away into the bloodstream and
before the denuded vascular surface is covered again,
noxious blood constituents would have an easier access to
subendothelial spaces through these denuded areas.

DR. SMITH: I want to comment on these very short term

effects; I have already referred to a paper by Day and
Proudlock (36) showing greatly increased cholesterol esterification
just three days after feeding cholesterol. In those three
RABBIT days, the rabbit's cholesterol went from something like 70
mg% to something over 400 mg%. And when you get this extraordinary
change in environment, I wonder whether one can relate it to
the sort of thing that might happen to you and me. I cannot
imagine any situation by which I could put my cholesterol up
seven fold in three days.

DR. LEE: I have to correct my earlier statement. I

said that we have carried out thymidine incorporation and
mitosis studies in rabbits.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 427

This statement is not correct; the animal we used for those

studies was young swine and it was not rabbits. The serum RABBIT
cholesterol level of control swine was around 90 mg% and it
increased to approximately 150 mg% after three days of PIG
cholesterol feeding.

DR. SCHWARTZ: Do the ultrastructural changes also

occur without any marked increase in the circulating cholesterol?

DR. LEE: Right. In the early stages of atherosclerosis

in these swine, the pre-proliferative phase of atherosclerosis,
we have observed certain ultrastructural changes including
more frequent dead or dying cells and necrotic debris in the
media of the aorta. These findings suggest that some active
metabolic changes are taking place in the arterial wall even
in the very early stages.

DR. BJORKERUD: As related to what Dr. Schwartz said

about the relationship between endothelium integrity and r-
lipoproteins, HDL; we had a series of rabbits which were RABBIT
kept at a constant serum cholesterol level for about a year.
We dosed the cholesterol individually as guided by the
individual serum cholesterol levels in each specific animal.
We had obtained levels which were rather constant which is
similar to the situation in the human being. Interestingly
enough, in the animals below about 350 mg% of cholesterol in
serum the integrity of the endothelium was largely unchanged.
Above a level of ca. 350 mg% the integrity of the endothelium
seemed to decrease.

DR. CARO: Margot Roach at Houston, at the Cardiovascular

Fluid Mechanics course last November, reported on some
volunteers who were starved, as I recall it, for several
days and then ate several eggs and bacon and it seems their
serum cholesterol rose from a hundred to many hundreds mg%.
I think she mentioned the number 800 mg/100 ml.

DR. DEWEY: There are at least two very disturbing

pieces of information that have come out of this conference
and that have been subliminally
What is the Relationship plaguing me for several years. I
between Animal Experiments would like to pose this as a
and Human Conditions? question. What is the relationship
between animal experiments and
human conditions? Dr. Mansfield very subtly touched on the
differences in the endothelial cell, from site to site
within man or differences between one animal species and
another.
42B CHAPTER 7

And I find this very disturbing, if we expect the endothelial

cell to be an important element in the whole process that we
are looking at.

The second disturbing fact is the difficulty which

exists in trying to infer natural disease-producing mechanisms
from intervention trials in
Inference of Natural animals. If you take data from
Disease-Producing experimental animals--coarctation
Mechanisms for Trials experiments for example--one sees
in Animals the flow dividers will be spared
in certain circumstances in test
animals such as dogs' can one then apply that knowledge to the
human circulation? So I pose this as an open question:
What should we do? Should we try to duplicate human physiology
or do experiments we feel relevant, in a very strict sense,
to the human condition? Or should we attempt to understand
animals and animal conditions where we have more access to
details of the biochemistry and the physiology and the
uptake for example, for the various labeled materials in the
arterial wall? It seems to me that this is a very real
problem.

DR. NEREM: I will respond in two ways, Dr. Dewey.

First, I think many of us have felt for a long time that
talk about high shear and ~ow
High Shear and Low Shear shear was almost irrelevant in
the sense that one had to speak
in the context of a particular position in the vascular system.
What is high shear one place, may be low shear some place else.
Because whatever way shear is entering into the process, it
makes a difference in terms of how it relates to the wall
biochemistry and that is critical,
Shear & Relation to the balance between the various
Biochemistry steps involved. One situation may
require a different level of
shear than another--assuming that shear is a factor. There
is a more general problem. I feel we need basic data independent
of whether they are related to the disease or not. I don't
know whether people working on vesicular transport are
really working on the right thing, but I think that is a
process that we need to study in order to understand this
whole area of the dynamics of arterial flow.

DR. WEINBAUM: What is your intuition at this time,

Dr. Schwartz, as to the relationship of your short term
studies to either the formation of fatty lesions, or to the
development of early lesions?
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 429

I think we could eventually understand your experiments, but

it is the gap between short term experiments and the actual
disease process that is the big mystery.

DR. SCHWARTZ: I quite frankly can't answer that. We

have to be able to answer numerous biologically-oriented
questions that relate to the different components of the
process, including vesicular transport. mechanisms of endothelial
injury, modifications to endothelial structure and so on.
Intuitively one has the feeling that transport of certain
macromolecules into the vessel wall may be very important,
and that certain hemodynamic factors may modify transport
rates and/or fixation of these molecules in the arterial
wall. Even to cope with the problem of getting macromolecules
in or out of a vessel is a significant issue and to extrapolate
further to the details of atherogenesis is a tall order. We
do know that certain areas with arterioles as delineated by
the uptake of Evans Blue, have abnormal transport rates, and
a wide variety of other differences. We also know from Dr.
Fry's data that lesions occur preferentially at these sites.
These sites differ metabolically. using a wide variety of
labeled precursors. and in their response to insulin.
Additionally, in early hypercholesterolemia there is an
accelerated rate of accumulation of lipids in blue areas
relative to white areas. While I believe that blue areas are
pre-lesion sites, the sequence of events leading to the
established atheromatous lesion is complex and not fully
understood.

DR. CARO: We must appreciate that for material entering

the artery wall there are a number of resistances arranged
in series. and we lack the information at present to assess
their significance. Thus. whether an area is 'blue' or
'white.' in relation to Evans Blue labeled albumin. may be
of no significance in a transport process with a time scale
of thirty or forty years. Nonetheless, I think we should
continue to study the basic biology. for example endothelial
cell pinocytosis, because it may have great relevance to
some other process.

I would also like to ask what work you have done with
low molecular weight materials because, apart from our own
work with acetate which is reported in the CIBA Foundation
Symposium (37), I know of no other with small molecule
materials. We found a negligible shear stress dependence
and that the transport was not controlled by the diffusion
boundary layer.
430 CHAPTER 7

DR. SCHWARTZ: Yes, but only using the Evans Blue areas
for in vitro metabolic studies.

DR. WERTHESSEN: I would like to comment on an implication

in Dr. Fry's earlier remarks. The risk factor in hypertension
is, as I see the epidemiological data, an expression clinically
of the underlying disease and not necessarily of the amount
of atherosclerosis. I think this is a point we often forget.

DR. FRY: The principle increases in risk are an increased

incidence of coronary occlusion and stroke. Your point is
well taken. So far as I know, careful measurements o'f the
amount of atherosclerosis in these two circulations as a
function of chronically elevated blood pressure have not
been made. Myocardial infarction correlates with the presence
of coronary atherosclerosis and presumably with the extent
or amount of coronary disease. However, the relationship of
the extent of atherosclerosis in the cerebral vascular
system to the incidence of either hypertension or stroke is
less clear. Cerebral hemorrhage secondary to medial degeneration
of the small intracerebral vessels apparently accounts for a
large amount of the increased risk of stroke in hypertension.
The relationship of this disease process to atherosclerosis
is not obvious.

DR. SCHWARTZ: There is a fairly consistent relationship

between the elevation of arterial pressure and the amount of
atheroma in the coronary arteries. The relationship of hyper-
tension to atheroma in the cerebral vessels is nowhere near
so clear cut (38,39). One can obtain a correlation coefficient,
but for the single individual it does not mean much.

DR. FRY: I did not mean to imply that fatty streaks

always evolve into atheromatous plaques. I presented data
from our studies in experimental atherosclerosis that showed
that the sites at which atheromata occur frequently correspond
to sites that were originally occupied by fatty streaks.
This correlation is particularly convincing at the bifurcation
of the left circumflex and anterior descending coronary
arteries, on the ventral aspect of the lower abdominal
aorta, and in the internal iliac arteries just above the
deep femoral orifices.

DR. GREENE: One means of investigating the importance

of the fluid mechanical parameter to atherogenesis is to
change the nature of the fluid.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 431

This methodology has not been discussed thus far at the

conference, but if fluid mechanics is important, then let us
alter the fluid's constitutive equation so as to change its
response to some of the short term phenomena such as turbulence,
separation, and reattachment. Any variation in physiological
or other effects in response to fluid manipulation would
offer a degree of confirmation to this hypothesis.

The methodology used in the present experiments has

been to change fluid behavior in a manner which primarily
affects the short time processes
Change of Fluid Behavior occurring in the flowing
fluid. Addition of an elastic
component to the normal fluid response yields a fluid which
is termed viscoelastic. This response is quite separate and
distinct from the viscoelasticity we have talked about
concerning the arterial wall, or the erthyrocyte itself.
Thus, under appropriate conditions we are able to talk about
blood as a continuum with some component of induced elastic
response. Viscoelastic response is brought about primarily
by adding small amounts of water soluble polymers of very
high molecular weight. This causes the fluid to exhibit a
certain amount of drag or friction reduction under some
circumstances, but in all cases to follow a constitutive
equation that involves a certain amount of elasticity. From
a spring-dashpot model approach, we have added the response
of a Hookean spring to the fluid. One limiting criterion,
however, is that the amount of elasticity added to the blood
response is so very small, and the memory of the fluid is so
terribly faint, that one only notices this effect for very
short time processes which occur in the arterial system, in
vitro, or wherever. Fortunately, viscoelastic effects are
observed at low concentrations of soluble polymer addition,
so that any gross effects of adding a foreign material to
the fluid are minimized. Polymer levels in the neighborhood
of 50 to 75 parts per million on a weight basis are typical.

Several previous studies have dealt with the effects of

polymer addition on blood turbulence. In one study, the
mean frequency of disturbances distal to a contrived occlusion
in the descending aorta of dogs was measured with and without
polymer addition.

Results showed that with a small amount of the viscoelastic

agent added to the blood, the frequency of the turbulence
generated distally drops significantly, in the neighborhood
of 60 or 70%. If one relates the frequency of these disturbances
to the flux of momentum, the fluid shear stress at the wall
distal to the occlusion is diminished.
432 CHAPTER}

A second study has involved the problem of blood hemolysis

in roller pumps. Perfusionists have been working for quite
some time with the problem of
Hemolysis During Life hemolysis during life support
Support with extracorporeal pumps, and
have found problems with
mechanical trauma to erythrocytes related to long term
support. Roller pumps manufactured by Pemco and Travenol
were tested to determine the effects of polymer addition on
the level of hemolysis in calf blood. Interestingly, it was
found that hemolysis levels decreased by 30-60% when certain
polymers were added prior to pumping.

To understand this phenomenon we must review the design

of this pump along with the fluid mechanical problems it
poses. Fluid is pumped through a tube by squeezing it between
a roller and a wall. The rollers are adjustable so that
one can change the minimum clearance between the tube walls
and, as most perfusionists point out, one does not run at
quite complete occlusion of the tube walls. The reason for
this is that many red blood cells would be crushed between
the opposing walls of the tube. Instead one tends to run
these pumps so that there is a slight unoccluded area,
perhaps the thickness of a few sheets of paper, between the
opposing walls of the tube which nearly eliminates crushing
red blood cells in the process of perfusion. This presents
a very interesting fluid mechanics situation as shown in
Fig. 7- 43. Viewing flow from right to left, one has the
possibility during motion of the roller, of a rather giant
axial pressure gradient between forward and reverse sides of
this moving roller, allowing for considerable fluid motion
and a great deal of regurgitation of fluid from the high
pressure side to the low pressure side. It has been hypothesized
that high rates of fluid acceleration, large local shear
stresses, and high frequency flow disturbances in this
region are responsible for much of the observed hemolysis.

To confirm this hypothesis, sucrose solutions with

effective viscosities about the same as blood and containing
small quantities of tiny wood particles were circulated
through the roller pumps. High speed cinematograph was
utilized, with the aid of a Fairchild, Model HS-40l Motion
Analysis Camera, to photograph the roller region at about
1500 frames/sec. Tracing the paths of the small particles
in the course of their motion from frame to frame allowed
estimation of regurgitation velocities as high as 10 meters/sec.
along with generally turbulent conditions.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 433

MOVING
ROLLER

TUBING

Fig. 7-43: Roller Pump Schematic.

 434 CHAPTER 7

Hence, much of the hemolysis that occurs in roller pumps may

be caused by the extreme trauma of these red cells being
regurgitated under extreme conditions of turbulence; certainly
this would imply that high shear stresses must be exerted on
the wall of the erythrocyte. With the addition of polymer
to the sucrose system we have also looked at the possibility
of reducing a certain amount of that turbulence, reducing
the short-time response as a semi-elastic material for very
short-time processes. Results after polymer addition (100 PPM)
showed maximum regurgitation velocities of only 6 meters/sec.
but more importantly, eddies appeared considerably larger
and less chaotic than previously. A typical frame is shown
in Fig. 7-44.

We have also completed an in vivo experiement using

RABBIT rabbits on high cholesterol diets. The experiment, carried
out by Drs. Mostardi and Nokes
Fluid Mechanic Effects at the University of Akron,
involved a four month duration
test with a series of rabbits fed high cholesterol diets
such that their cholesterol levels ran upwards of a thousand
mg%. If fluid mechanics are important and short time processes
are important, as Dr. Nerem has questioned, then let us
again do something about changing the fluid response in vivo
for short time processes. This was done as before by maintaining
a circulating polymer level of about 60 PPM in the vascular
systems of half of the test animals. The animals were
sacrificed and examined at the conclusion of the experiment.
The pieces of tissue shown in Fig. 7-45 are descending
thoracic aorta from a matched pair of rabbits. The one on
the left has received the diet for approximately four months.
The one on the right was fed the same diet but was also
given periodic injections of a viscoelastic polymer which
altered the blood response to a very short term process,
meaning that many of the turbulent phenomena, disturbances,
separation, occurring at the highest frequencies were partially
suppressed. I cannot explain the process physiologically;
we have to date made no detailed studies of the tissue other
than these gross ones but certainly the fatty streaking, for
the no-polymer rabbit aorta on the left is in marked contrast
to the rabbit whose blood was doctored, so to speak, with a
small amount of viscoelastic polymer.

DR. CARO: Is that a stained section?

DR. GREENE: No, it is not. Next, Fig. 7-46 shows

comparative sections in the abdominal aorta of test rabbits;
the differences in plaque formation are not quite as apparent
here.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 435

Fig. 7-44: Typical high speed cinematography of area behind

moving roller showing wood particles in sucrose solutions.
436 CHAPTER 7

Fig. 7-45: Samples of descending thoracic aorta from no-

polymer rabbit (left) versus polymer rabbit (right).
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 437

Fig. 7-46: Samples of abdominal aorta from no-polymer

rabbit (left) versus polymer rabbit (right).
438 CHAPTER 7

It is the section again on the left which was taken from

the control rabbit on the high fat diet. The rabbit on the
right received the high fat diet as well as periodic injections
of the polymer solution. Again, on a percentage involvement
basis, there seems to be a significant difference between
the involvement of the rabbit aorta on the right versus the
involvement of the aorta on the left.

In summary, I want to point out that these studies have

all been designed to involve a new dimension in experiments
in blood flow. That dimension is to change the rheological
nature of the fluid, and, if fluid mechanics are important,
to see what effect this has on the course of a vascular
disease or a rheological event. There are many additional
studies which could be performed in this manner, for example,
some of the oscillating wall experiments that Dr. Nerem has
done would be very interestingly repeated with additions of
a small amount of polymer, so that the fluid was made viscoelastic.
We might see some significant changes in wall permeability
or uptake rate of albumin.

VOICE: Did you see any difference in growth rates

between those two?

DR. GREENE: I believe that while the external factors

such as growth and weight increase, as best as I can recall,
were reasonably similar, there may have been some slight
hemolytic effects with the polymer. Nevertheless, note that
I am not suggesting that we use polymer addition as a therapeutic
device at this time, I am primarily stressing the fluid
mechanical effects.

DR. WOLF: How frequently did you have to inject the

polymer over a four month period?

DR. GREENE: Roughly two and half times a week. We had

run some previous experiments with tritiated polymer and it
appears to have a half-life of about 36 hours in vivo.

DR. SMITH: Were the serum cholesterols the same?

DR. GREENE: Yes.

DR. COLTON: What polymer did you use?

HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 439

DR. GREENE: This happened to be a polyacrylamide

(Separan AP-30), a polymer made by Dow Chemical Co. It has
a molecular weight of about two to three million. Of course
it degrades with time, I suspect, both biologically and
because of mechanical trauma in vivo.

DR. KENYON: How do you characterize the influence of

the polymer on blood?

DR. GREENE: The apparent viscosities of the blood

versus the blood with small amounts of polymer are nearly
indistinguishable. There is a very slight increase in blood
viscosity as one might suspect if one adds a small amount of
polymer solution to the blood but only by a couple of per
cent. The biggest chang_ appears to be in the addition of
the viscoelastic factor 'which can be characterized rheometrically
by the first normal stress difference. I have no really
good data from making total rheological measurements on
blood yet; it depends strongly on what instruments you use.
We have found in equivalent "viscosity" sucrose solutions,
that adding sufficient polymer to induce a just normal stress
difference of about five dynes per square centimeter generally
gives significant differences in in vivo or in vitro effects
for a given study.

DR. CHIEN: What was the shear rate at which you measured
the blood viscosity?

DR. GREENE: It went from one reciprocal second or less

to well over a thousand, but we tended to be interested
primarily in the values obtained at higher shear rates since
these are probably more appropriate in vivo for the larger
arteries. Now whether this is the optimum shear rate, I
don't know.

DR. CHIEN: The reason why I asked that is because the

polymer induced aggregation and the increase in viscosity
probably can be seen only at very low shear rates.

DR. GREENE: We have carried out some ESR measurements

and microscopic observations on treated blood and we have
found that even though many of these polymers are used for
particle flocculation (and this scares one off immediately
in terms of any kind of in vivo study) one finds that for
circulating polymer concentrations of less than 150 parts
per million these properties seem to be absent. So I don't
think RBC aggregation is important here.
 440 CHAPTER 7

DR. DEWEY: One frightening feature that you have

pointed out to us in these results is the decrease in the
high frequency components of what apparently is turbulent
flow. At least disturbed flow. I would propose that perhaps
this might be the way of altering the fluid mechanic stress
at the wall, that could be potentially significant from the
experimental point of view. There have been a number of
studies. Dr. Schwartz has done some, for example, on coarctation
of the aorta. Dr. Fry and many people here have performed
experiments in which they have found significant changes in
the staining charactersitics of the endothelial cells distal
to a coarctation and one would correlate the regions of
damage and change with regions in which you would expect
very high wall pressure fluctuations and wall shear fluctuations.

There are relatively straightforward methods by which

one could obtain the power spectral density of the fluctuations
themselves. By that I mean that if you look at the spectral
content and the energy contained at the various frequencies,
one finds a rather characteristic type of turbulent spectrum
in these coarctation groups. It is easily characterized for
ordinary blood or for example in a Newtonian fluid. I would
expect on the basis of your results that the turbulent power
spectral density would change dramatically with the addition
of the polymer and if it is the high frequency content, and
as Dr. Weinbaum and others have suggested, which is as
important if not more important than the actual magnitude of
the shear, this might be a possible way on an experimental
basis to change this specific parameter without changing
other gross fluid properties. In this way one might be able
to see differences in staining or the uptake in which we are
able to vary one specific hemodynamic parameter without
influencing gross properties in flow.

DR. NEREM: I accept what you say, Dr. Dewey, in terms

of that particular type of situation; but I would be somewhat
surprised if those were present in the rabbit aorta under
similar conditions. However, in terms of your story, Dr. Greene,
are you sure that what you have done is to alter fluid
dynamics? Is it possible at this point that this polymer
has had a chemical effect on the endothelial cells?

DR. GREENE: I have no conclusive answer to that question

RAT at this time. In one set of early experiments with rats using
tritiated polymer, we did try to find out whether there was
any build up of polymer in or on various organs or the
arterial wall itself; we found no specific activity in those
regions.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 441

DR. NEREM: I was thinking of the cells in some way

changing shape or whatever.

DR. GREENE: I think what Dr. Weinbaum was discussing,

the possible alteration of the motion of endothelial cells,
would have some bearing on the result, but we have found no
changes in shape of erythrocytes.

DR. WERTHESSEN: I do not know if you can do this but

as an experimentalist may I suggest the use of a spontaneously
developing lesion such as in the White Carneau pigeon?

DR. GREENE: We now have about 30 pigeons under study PIGEON

and are currently working on this.

DR. WERTHESSEN: Good, good. Because if you can inhibit

the development of lesions, then I think you have something.

DR. GREENE: That is precisely where we are proceeding.

DR. STONE: When you are increasing viscoelastic properties,
does this tend to increase the viscosity?

DR. GREENE: Not significantly, that is a separate and

distinct property. We are not making the fluid thicker
although I mentioned that inadvertently one increased the
viscosity slightly by polymer addition. We have run in
vitro controls where we have significantly increased only
the viscosity of the blood just by adding a viscous agent,
but have not found corresponding reductions in say, hemolysis.

DR. STONE: Would this be similar to a hyperproteinemia?

DR. GREENE: What do you mean?

DR. STONE: Where you have increased protein content in

the blood itself.

DR. GREENE: Does that affect the apparent viscosity of

blood?

DR. STONE: Yes.

DR. GREENE: Molecular weight, what would it be?

DR. CARO: Fibrinogen, principally.

DR. GREENE: We have found no significant viscoelastic

effects for molecules under a couple of million.
442 CHAPTER 7

DR. STONE: They build them now in the M.A.A. Radio

Isotope field, these are relatively large macroalbumin
particles, you can get them quite long.

DR. GREENE: A much more appropriate set of experiments

would be to try to build something that one could be sure
would be fairly compatible •..

DR. CHIEN: The largest dextran has a molecular weight

of approximately two million.

DR. GREENE: 1 have not seen any data on that.

DR. CHIEN: This is the blue dextran manufactured by

the Pharmacia Laboratories in Sweden.

DR. STONE: What is the actual count of Siren D, the

hydrogenated size of those two million molecular weight
particles?

DR. GREENE: It varies quite a bit depending on what

fluid you put it in and the stresses present during flow.

DR. SMITH: Have you looked at the interaction between

this material and any of the plasma constituents? 1 was
thinking that if LDL in fact stuck itself on to this material
it could make such a large molecule that it would not filter.

DR. GREENE: We have not considered that possibility.

DR. WERTHESSEN: You have an excuse for getting an

ultracentrifuge.

VOICE: Does it change the behavior of the platelets,

first question. Second question: Did you investigate the
serum lipoprotein profile in the control and experimental
animals with regard to, for instance, the presence of HDL?

DR. GREENE: I would guess that the answer to all of

those is no, we have not investigated such factors, which is
the reason why 1 call these extremely preliminary experimental
findings.

DR. CHIEN: 1 am glad that Dr. Greene brought up the

problem of blood properties in relation to the problem of
atherogenesis.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 443

Just to continue this vein of discussion, Fig. 7-47 is

taken from the work of Goldsmith (40). What it shows is
that the velocity profile is
Relation of Blood altered, if you have cells
Properties to suspended in the media. And
Atherogenesis this parabolic profile is
shown by the dotted line when
you turn your fluid so that you have 38% red blood cells
made into ghosts by removing the hemoglobin after osmotic
lysis. The reason for doing that is to allow the visualization
of the few cells that remain with the ghosts. Then you can
map the velocity profile which shows blunting in the center
and a steepening of the gradient near the wall. When a very
high cell concentration (92% cells) is used, the degree of
blunting in the middle is even more severe and the steepening
of the gradient near the wall becomes more prominent. This
will change the shear rate near the wall and hence also the
shear stress at the wall. Of course, the velocity profile
is also dependent on the flow velocity, the deformability of
the cells, and a number of other factors. So the conversion
of the mean velocity into shear stress at the wall is a
rather complicated problem in the presence of cells (Fig. 7-48).
In the presence of cells another thing that happens is that
there will be collision between the cells and this again is
taken from the work of Dr. Goldsmith (40). He tracked the
path of individual red cells with time as the cells flow
through a cylindrical tube. The abscissa shows the radial
location in the tube plotted as a fraction of the tube
radius. His observations indicate that the path of the
cells is not a straight one, i.e. it is erratic. In the
presence of neighboring cells, cell-to-cell collisions cause
sideway motions of these red cells (8 micron in diameter) in
the tube. When latex particles with the diameter of one or
two microns are used to simulate platelets, it is observed
that the path is even more erratic. The particles move
across the tube over a very wide radial distance with time.
Therefore, for the smaller particles such as platelets,
collisions with red cells cause a very large excursion in
their paths. What this means is that there is a larger
chance for the interaction of the particles with the wall.
This is true for the red cells to some degree and even more
so for the smaller particles.

Figure 7-49 is again taken from Goldsmith's work (40)

on the flow pattern beyond a partial obstruction created in a
tube by gluing a latex sphere to the side of the tube. The
behavior of particles such as red cells or latex particles
flowing past such a spherical obstruction can be observed micro-
scopically. An enlargement shows the detailed picture as the cells
streaked by at rather high velocity past the obstruction, but once
444 CHAPTER 7

0
d .'ce".
,
I
I
\
\
\
I
I
I
\
\
\
\
\
\
\
\
\
T
\
\
i 0.015
\
\
~

.,
ii" • I
0.025
-;~ \ I
\ I
\ I
0.010 \ I

\\ ~M /t \ , I

, , 0.030 \
I

"
\ I I
\

, ,,
, I
,
\
, I
. \ I
0.025 \ ,'..., ........ flow ,
,, YM , .4,.'01...... '10..
,,
0.035
'
' ...

'U.tOj
I ",'
"'---. ... '
uWll
0.0J0 0.040
Q5 05 1.0
1.0 0.5 o 0.5 1.0 R 1.0 0

Ro
Fig. 7-47: Velocity distribution obtained with tracer human red
cells in ghost cell suspensions at two cell concentrations: left,
92%, right 38%. The velocity along the longitudinal direction of
the tube, U3(R), is plotted as a function of the radial position
in the tube, R, normalized for the tube radius, RO' RO = 42 p.
Courtesy of Dr. H. L. Goldsmith.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 445

Time, seconds
1.0 ~-r--.-~r--r--~~-'-';-j35

0.8 30
25
0.6 20
0.4 15
10
0.2 5
_ ____ _ 0
0
0.2 5
R 5 10 15 20
Rc 1.0 5 10 15 20 2S 30 35 R,~
r-~--~~~~~~~~--~~5

0.8 30
o 2.0)llD
25
0.6
20
0.4 15
10
0.2
5
0 ________________ _
o
5
0.2

Time, seconds

Fig. 7-48: Paths of tracer red cells (top) and of tracer 2}lm
latex spheres (bottom) with time in a 40% ghost cell suspension
flowing through a 73 urn diameter tube (RO = 3.65)Um) at a mean
velocity of 0.030 em/sec. The sizes of the particles relative
to the tube are shown. After Yu and Goldsmith, in Platelets,
Drugs and Thrombosis (J. Hirsch, ed.), pp. 78-93. S. Karger,
Basel, 1975.
446 CHAPTER 7

of ffow

Fig. 7-49: Schematic drawing of the model of a spherical

obstruction in the circulation. The general pattern of the
streamlines in the vortex is given. The position of reattach-
ment of the flow is indicated.
B: The vortex region is enlarged to show the forward (F),
reverse (R), and stagnation (5) regions downstream of the
obstruction. The orbits of a red cell as it spiraled out from
the center are shown. Large changes in the linear velocity of
cells were noted. Open circles show particle positions at
0.12 sec intervals; dots show particle positions at 5.9 msec
intervals. From Yu and Goldsmith, Microvascular Research
6:5-31, 1973.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 447

getting into the post-obstruction area, spend a long time recircu-

lating in this area of flow separation. Thus, platelets would have
a longer time to undergo various types of chemical interactions.
When red cells go into the stagnant, low shear region, they
tend to form rouleaux and cause further reductions in flow.
Therefore, cells and other solutes spend a long time in such
a separated flow region. The reattachment of the flow is
marked by arrows in Fig. 7-49.

These considerations indicate the microrheologic behavior

of the particles in the blood flowing in our circulatory
system may have considerable
Relevance of Abnormalities influence on their interactions
in Blood Rheology to with the wall, particularly when
Problem of Atherogenesis there are abnormal situations in
the circulatory system. The major
determinants of the blood viscosity are plasma viscosity, cell
concentration, cell deformability and cell aggregation.
There are a number of clinical situations in which the blood
viscosity is abnormal due to alterations in one of these
factors. Clinically we have classical examples of abnormality
for each one of these factors. For example, in multiple
myeloma we have plasma viscosity elevation, in polycythemia
the cell concentration is high, in sickle cell anemia the
cell deformability is abnormal, in macroglobulinemia red
cell aggregation is enhanced. I don't know whether there is
any statistical information on the correlation between
atherosclerosis and these various kinds of disease states.
Maybe our colleagues in pathology or epidemiology have some
information showing whether there is any correlation between
anyone of these conditions and atherogenesis. I think it
would be interesting to find out if there is any relevance
of abnormalities in blood rheology to the problem of atherogenesis.

DR. WERTHESSEN: Dr. Chien, on this point of deformability

of erythocytes, do you know of Rasmussen and Allen's work?
(41). I think it would be a superb tool in these studies to
change their deformability by the addition of prostaglandins.

DR. CHIEN: Yes and they used catecholamines too.

DR. SCHWARTZ: I wish to refer back to remarks that Dr.

Nerem made about the problems of animal models and Dr. Smith
too, and really raise what I think could be an important
point in terms of the logic of looking at factors responsible
for the development of atherosclerosis. We are tempted to
make an assumption which may be quite wrong, that the progression
of disease from its initiation to its ultimate development
is a continuum. There may be important mechanisms involved
in initiation, and different mechanisms involved thereafter.
448 CHAPTER 7

DR. WOLF: Dr. Mansfield's presentation mentioned the

need to have access to early changes and Dr. Smith pointed
out the difficulty in distinguishing the difference between
early changes and different sites in the structure of the
artery for architectural changes, such as cushions and so
forth.

Dr. Schwartz just pointed out that one should not

assume that one is looking at a long term continuum process
and relevant to this I think were the observations that Ted
Gillman made at the Lindau Conference, in which he was
studying the uterine artery and pointed out that the changes,
namely the hyperplasia of smooth muscle cells in the intima
occurred in the process of enlargement of the uterine artery
with pregnancy and that these reversed after parturition
when the uterine artery diminished in size.

With repeated pregnancies there ultimately appeared

typical atherosclerotic lesions. Here then an opportunity
is offered to look at early lesions, to look at the production
and resolution of lesions, and to make distinctions between
adaptive changes in the vessel and pathologic changes, if
you will. In this connection there is an interesting paper
I want to get into the record by Ryan, Clark and Brodie
called Neurogenic and Mechanical Control of Canine Uterine
Vascular Resistance. It is in the AMERICAN JOURNAL OF
PHYSIOLOGY. (42). What they show is that there is cholinergic
vasodilator innervation of the uterine artery and adrenergic
vasoconstrictor innervation in the uterine artery. During
pregnancy the vaso constrictor activity is very severely
blunted but reappears following parturition. This is an
interesting article and pertinent to all three of these
observations cited.

DR. WEINBAUM: We have not mentioned this previously,

because the experiments were preliminary, but it is the
reason that we initially suspected that we could increase
vesicle transport through fluid mechanics or mechanical
factors. The experiments are briefly described in a short
note that Dr. Caro and I put in PHYSIOLOGICAL PROCEEDINGS.
(43). These experiments need to be carefully repeated but
their essence is the following. The diffusion velocity of a
vesicle, as best we could estimate, is of the order of a
hundred angstroms per second. Thus, we tried to design a
mechanical disturbance experiment in which we would cause a
boundary motion of the endothelial cell which would produce
velocities at the plasmalemma which were of that order.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 449

We had known from the stretch experiments of Dr. Fry that if

you introduce very small stretches on the order of four or
five per cent, that there is almost no change in transport.
We attempted to introduce disturbances of the same magnitude
as Dr. Fry's but at a frequency such that the boundary
motion would be roughly comparable to the diffusion velocity
of the vesicle. The results showed that there was a very
curious frequency response. When we increased the frequencies
to the order of about one Hertz the transport seemed to rise
sharply whereas if we decreased the frequency the transport
would return to the normal static situation at a frequency
of about a tenth of a Hertz. So there was roughly a ten fold
range over which there was a very substantial increase in
transport of roughly 60 to 100%. Again this data is preliminary.
The fascinating feature is that the increase in transport
occurs where the boundary velocity of the plasmalemma membrane
of the cell is of the same order as is our best estimate of
the diffusion velocity of the vesicles. If any experiment
has been the motivation for the work that we have done since
then, I think this is probably the key one.

DR. DEWEY: How did you make the motion?

DR. CARO: These are very difficult experiments and

many more than the 14 we have done need doing. but essentially
a dog's common carotid is carefully excised and mounted in a
rig at its in vivo length in a bath. A cam is then adjusted
which allows it to relax by three or four percent of its
length. The motion is simple harmonic and the frequency can
be adjusted.

DR. DEWEY: Longitudinal •••

DR. CARO: Yes, the extension is longitudinal. Transmural

pressure was about one cm. of water.

DR. COLTON: In these experiments what does the backside

of the artery look like where you would normally have adventitia?

DR. CARO: I have never done EMs, but to the naked eye
it is a tatty piece of adventitia.

DR. COLTON: I suggest that the total uptake depends on

the state of the backside of the artery you measure, as well
as of the luminal endothelial surface. If there is adventitia
and any vasa vasorum, the tissue will be invaginated with
lots of vessels that do not communicate well with the bath.
450 CHAPTER 7

One additional thing that could happen is that the stretching

serves to help perfuse the fluid in these vessels, thereby
increasing the uptake rate.

DR. CARO: I will accept this absolutely. On the other

hand, we do dissect off the adventitia at the end of an
experiment and it is not counted as part of the wall. I
should also mention that we did some of these experiments
with sodium acetate as the tracer and there was no effect of
stretching on its uptake.

DR. COLTON: The sodium acetate and albumin are likely

to behave quite differently. The sodium acetate may diffuse
so readily through the tissue that it is insensitive to the
detailed structure of the backside of the artery, whereas
the albumin, because it moves much more slowly through
plasma and tissue, would be quite sensitive to the anatomical
structure and extent of perfusion on the back side.

DR. CARO: But the sodium acetate will presumably come

into the vasa vasorum in the same way as the albumin will.

DR. COLTON: The agreement that it is appearing at the

endothelium would be more convincing if you had evidence of
that.

DR. CARO: I totally agree that it has got to be done

by looking, absolutely.

DR. COX: I would like to ask Dr. Caro a question about

these studies involving uptake of markers by perfused segments.
Do the markers ever appear in the surrounding fluid? I
presume there is surrounding fluid.

DR. CARO: Yes they do. We always monitor the bath.

The volume of the bath is actually very large (about 50 mI.)
compared to the volume contained in the segment. about 0.1
mI. Unfortunately there is no very good solution to the
problem of leakage. which we have all encountered. We tried
to seal off the vasa vasorum. and test by putting visible
tracer and pressurizing the vessel. Then if there is a leak
we reject the specimen. In the end I have to carry out most
of my studies at zero, or very low, transmural pressure.
The conditions are artificial but leakage is avoided.

DR. COX: Could you get around this problem by testing

your segment and selecting it from ones where there was no
vasa interna originally in your sections so you could eliminate
the problem.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 451

Do you have any information that it represents leaking

through the vasa or is it simply continuous diffusion through
the wall?

DR. CARO: This is leakage. No question at all. When

you dissect out the vessel you see the vasa on the surface,
the arteries and the veins and if you are clumsy and cut one
of these and do not tie it securely, when you mount the
segment in the rig a little jet comes out of the vessel.

DR. COLTON: Have you ever thought about coating the

backside with something like beeswax?

DR. WEINBAUM: This is in reference to the questions

that were asked previously about denuding the endothelium.
The adventitia is purposely left on for that reason. The
experiments I described were conducted over a 15 minute
period. This is the shortest time in which one could expect
albumin to diffuse in substantial quantities to the adventitial
surface. If a leak is present it shows up very easily in
the adventitia because we strip it off separately after the
experiment is over. There is never any question as to when
you have a leak because the adventitia will show readings
which are ten fold larger than normal control.

DR. COLTON: Is that data then rejected?

DR. WEINBAUM: Yes.

DR. DEWEY: I have a question for Dr. Smith. I found

it very exciting to look at the data that you presented
earlier regarding the type of
Oxygen Util~zation metabolism within the artery
and how, in fact, it can be
anaerobic and aerobic and how that may in fact be a function
of the diffusion depth to which the oxygen can go in the
artery. One of the rather interesting things I think is that
the diffusion depth of the oxygen and the rate at which the
oxygen is utilized in the metabolic process, will obviously
vary from artery to artery, certainly in terms of the diffusion
coefficient from one artery to another, that should not
change dramatically, I suggest. The utilization rate might,
however, and we do find a dramatic difference in the arterial
wall thickness from, for example, the coronary intermediate
arteries to the aorta and it would suggest to me that
perhaps this balance between aerobic and anaerobic metabolism
may change from one artery to another or, for example,
during vasoconstriction or vasodilation, particularly in the
coronary arteries.
452 CHAPTER 7

DR. SMITH: This is a very good point and I don't think

there has been any work on this at all. All these measurements
on oxygen uptake and lactate production have been made in
aorta. I would imagine that a very muscular artery like the
coronary, has a much higher metabolic rate so that you might
expect the more adequately perfused coronary artery to have
a higher oxygen utilization and conversely to produce more
lactic acid when deprived of oxygen. I don't think that
there is any data on this and I think this is a very interesting
area indeed. If it was consuming more oxygen then obviously
the critical diffusion distance would become shorter, and if
it was in a state of violent stretch, i t would again consume
more oxygen. Certainly, looking at the things that happen
to coronary arteries they seem to have a very rough life
indeed, they might need to consume more oxygen to keep
themselves in shape. I think this is a very interesting
area and does really require further investigation.

DR. CARO: I have a question for Dr. Dewey. You have

said, I think, that you have found on your arteriograms
signs of early changes in the wall in the carotid sinus. Is
the wall there significantly different in thickness than it
is in the parent vessel?

DR. DEWEY: I think the most pertinent data here are

data which Dr. Meyer has on the carotid sinus. I have not
had a chance to analyze them in detail, but I have the
impression from one of the reprints that he furnished to me
that there is a very substantial difference in the wall
properties, between the sinus area and the internal carotid
distal to the sinus itself. In other words, there are very
substantial changes in the elastic properties; and I am sure
there are physiological reasons for this in terms of chemoreceptor
control, but I cannot comment in detail on the actual physiological
differences.

DR. COLTON: Is it known at what P0 2 the smooth muscle

cells of the arterial wall shift over from aerobic to anaerobic
metabolism?

DR. SMITH: I cannot answer that. In the bovine artery

segment data that Arnqvist and Lundholm (44) presented, it
seemded to be a gradual shift. Reducing P0 2 from about 70
mm Hg (normal level for arterial blood) to 35 mm doubled
lactic production, and reducing it from 35 mm to zero,
doubled it again (Fig. 5-2).
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 453

DR. BJORKERUD: I think that is different in different

situations though. Because we constantly see injury to the
plaque tissue when the plaque is completely re-endothelialized.
Of course we cannot tell if that is due to a sudden deficiency
of oxygen or due to a sudden deficiency of nutriments. We
see tissue vary very reproducibly upon re-endothelialization
of experimental lesions in the rabbit. Maybe Dr. Mansfield
has similar experience from lesions in other species?

DR. COX: I have a question for Dr. Smith in this

regard. concerning hypoxia and cold. I have routinely
stored blood vessel samples in the icebox overnight. I go
back the next day, incubate them at 37 degrees C for a
couple of hours. Comparing refrigerated samples, which were
between 0 and 4 degrees C, and fresh blood vessels. there
was no statistically significant difference in the ability
of the smooth muscle to generate force. This is for periods
of storage from 24 to 72 hours. So it would seem to me,
from a simple point of view, that cold storage is not affecting
the contractile ability of cells. These are fairly large
vessels which are certainly more than 350 microns in wall
thickness. Did you not show data yesterday that indicated,
or suggested that storage at 4 degrees had some sort of
deleterious effect on arterial smooth muscle?

DR. SMITH: It appears to switch them over from oxidative

phosphorylation to the glycolytic pathway. This is the
Scott and Morrison data (45,46,47) and this has been pretty
amply confirmed (48,49). But this does not mean to say that
they can't contract using glycolysis.

DR. COX: I would call everyone's attention to the fact

that studies in cardiac muscle mechanics never use capillary
muscles more than something like 800 microns in diameter.
There have been a number of demonstrations that larger sized
capillary muscles studied in vitro have reduced contractile
responses.

DR. MEYER: Reminiscent of observations of Gillman on

the uterine arteries referred to by Dr. Wolf. (28) I have
further evidence of the importance of hemodynamics and
associated structural features manifest in the development
of some early lesions in the iliac arteries on the side of
the obliterated single umbilical artery.
454 CHAPTER 7

Usually there are two umbilical arteries, but in 0.72 -

1.0 per cent of newborn infants one umbilical artery is
missing (49,50). With a
Effect of a Congenital single umbilical artery a unique
Single Umbilical Artery hemodynamic situation arises
during fetal development: the
entire blood to the placenta is transported from the abdominal
aorta through the common iliac arteries of only one side of
the body (Fig. 7-50). Therefore, the common and internal
iliac arteries on the side of the SUA become exposed to a
higher hemodynamic load during fetal development and,
consequently, assume a larger caliber and thicker wall than
those on the other side of the body which do not participate
in the placental circuit.

The differences in caliber are closely associated with

different structure of the arterial wall (51). On the side
of the single umbilical artery, the common and internal
iliac arteries show the structure of elastic arteries.
Their media consists of tightly packed elastic membranes
(Fig. 7-51a). In contrast, the common and internal arteries
on the opposite side, which do not participate in placental
circuit. display the structural pattern of muscular arteries
(Fig. 7-SIb) •

Berry et al., (52) measured vessel compliance in vivo

in 18 children born with a single umbilical artery with a
non-invasive method and found significantly different compliance
values in iliac arteries on both sides of the body. These
differences are probably determined by different structural
features of iliac arteries present in children with a single
umbilical artery.

Differences in functional load and structure probably

also account for different patterns of early calcifications
often occurring in iliac arteries of infants and children
with a single umbilical artery. On the side of the closed
or already obliterated single umbilical artery the early
membrane calcifications are usually irregularly distributed
along the luminal surface (Fig. 7-52). In contrast. in the
narrower common iliac artery on the other side of the body,
calcific incrustations often assume a more regular pattern
and appear as transverse streaks. Microscopically. the
streaks represent calcified edges of membrane gaps, a typical
structural feature of muscular arteries. Hence, the calcification
pattern is in accord with the muscular structural pattern of
the narrower common iliac artery.
HEMODYNAMIC CONTR IBUTION TO ATHEROSCLEROSIS 455

NORMAL SINGLE UMBILICAL ARTERY

Fig. 7-50: Diagram of a part of the systemic arterial tree in

a fetus with two umbilical arteries (left) and in a fetus with
a single umbilical artery (right) showing different calibers of
the iliac arteries. UA - umbilical artery, IIA - internal
iliac artery, CIA - common iliac artery. UV - umbilical
vein (from Meyer and Lind, 1974).
456 CHAPTER 7

Fig. 7-51: Cross-sections of common iliac arteries of a term

newborn infant. The common iliac artery (A) on the side of the
single umbilical artery is an elastic artery with well-developed
elastic sheets in its media. In contrast, the thin-walled common
iliac artery on the opposite side (B) is a muscular artery with
a media (m) poor in elastic networks. a- adventitia. Cryostat
frozen section. Weigert's resorcin-fuchsin stain. (From Meyer
and Lind, 1974).
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 457

Fig. 7-52: Iliac arteries from a one-month-01d infant with a

single left umbilical artery. Note the large caliber of the
common iliac artery (Lci) and that of the internal iliac artery
(Lii) on the side of the obliterated single umbilical artery
(SUA). Conspicuous calcifications (black) are present in both
iliac arteries, but the pattern differs in the iliac arteries on
the side of the single umbilical artery. In the right common
iliac artery (Rci) circular calcium - free bands (arrows) are
seen. Microscopically, they correspond to the gaps in the
internal elastic membrane.
Von Kossa reaction. Millimeter scale at top left. Case 906/73.
(From Meyer and Lind, 1974).
458 CHAPTER 7

Fig. 7-53: A raised longitudinal atherosclerotic lesion

(between arrows) is seen in the enlarged right common iliac
artery on the side of the obliterated single umbilical artery
(sua) from a four-year-old boy. The red stained lesion
appears black in the photograph. Fine calcific incrustations
(c) are seen in the right internal iliac artery which is
significantly larger than the left internal iliac artery. A -
abdominal aorta, CI - common iliac artery, II - internal iliac
artery, EI - external iliac artery. Case 116/72. Von Kossa
reaction and gross fat staining with Fettrot 7B.
 :x:
m
s::
o
o
-<
z
»
s::
("')
("')
o
Z
--I
::u
OJ
c
--I
o
Z
--I
o
»
--I
:x:
m
::u
o
Ul
("')
r
m
::u
oUl
Ul

Fig. 7-54: (A) In the luminal layer of the atherosclerotic plaque from the right common
iliac artery (on the side of the obliterated single umbilical artery) numerous anisotropic
crystals and droplets are seen. B. Same section seen in non-polarized light. The deeper
intima includes numerous lipophages (B, arrows). iem - internal elastic membrane. Cryostat ~
01
frozen section, Fettrot 7B stain. X 500. -0
460 CHAPTER 7

In the enlarged common iliac arteries on the side of a

single umbilical artery established atherosclerotic lesions
have been found in two children aged 18 months and 4 years
(51). (Fig. 7-53 and 7-54). Since fatty streaks were absent
in other arteries, general metabolic factors were probably
not involved. Hence, the development of lesions may essentially
be due to local structural and hemodynamic factors. After
birth and cessation of the umbilical circulation, the enlarged
iliac arteries must accommodate to a considerably diminished
blood flow. This accommodation is associated with a remodeling
of the arterial tube, especially with an intimal thickening
which contributes to the narrowing of the lumen. The prolifera-
ting intima may develop an increased affinity to lipids (53,54)
become the site of lipid infiltration and, in this way, favor
the development of early atherosclerotic lesions.

DR. WERTHESSEN: I must confess that your Chairman

succumbed to a bit of showmanship in having Dr. Meyer prove
at the end of the conference that nature could do a better
experiment than any of us. I would like to see it repeated
in the laboratory.

DR. MANSFIELD: I want to emphasize the point that

Dr. Meyer has made so eloquently, that atherosclerotic disease
and arterial changes are surely occurring immediately after
birth, and in fact he finds evidence that they are starting at
parturition. Atherosclerosis is not just a disease of old
age; we only see gross changes at that time, rather it may
be an aspect of vascular adaptation called into play at
any age.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 461

BIBLIOGRAPHY

1. Meyer, W.W. and Lind, J.: Calcifications of the carotid

siphon--a common finding in infancy and childhood.
Arch. Dis. Childh. 47:355-363, 1972.

2. Walsh, S.Z., Meyer, W.W. and Lind, J.: The Human Fetal
and Neonatal Circulation. Function and Structure.
Charles C. Thomas, Springfield, Ill., 1974.

3. Meyer, W.W.: The mode of calcification in atherosclerotic

lesions. Internal Workshop Conference on Atherosclerosis,
London, Ontario, 1975 (In press).

4. Meyer, W.W.: Early forms of calcinosis in human arteries,

Sandorama, Basel (Switzerland) 11-16, 1975.

5. Meyer, W.W.: The interrelation between the early calci-

fications and lipid deposits in the carotid siphon
demonstrated by dual gross staining method. IX.
International Congress of Angiology, Florence, 1974.

6. Meyer, W.W. and Noll, M.: Gross patterns of early lipid

deposits in the carotid artery and their relation to
the preformed arterial structures. Artery 1:31-45,
1974.

7. Meyer, W.W. and Naujokat, B.: Uber die rhythmische

Lokalisation der atherosklerotischen Herde im
cervicalen Abschnitt der Vertebralarterie. Beitr.
~ ~ 130:24-39, 1964.

8. Meyer, W.W. and Lind, J.: Calcifications of iliac arteries

in newborns and infants. Arch. Dis. Childh. 47:364-
372, 1972.

9. Meyer, W.W. and Ehlers, U.: Early calcification patterns

of the iliac arteries and their relation to the
arterial structure. Z. Zellforsch. 130:378-388, 1972.

10. Keller, S. and Mandl, I.: Non-polar peptides from elastin·

Prot ides of the Biological Fluids, 22, 127 (Ed.
H. Peters) Pergamon Press, 1975.
11. Robert, L.: The macromolecular matrix of the arterial wall:
collagen, elastin, mucopolysaccharides. In: Athero-
sclerosis (Ed. R.J. Jones) p. 59 (Springer-Verlag,
1970.
462 CHAPTER 7

12. Spain, D.M: Atherosclerosis. Scientific American,

13. Cornhill, J.F. and Roach, M.R.: Quantitative method

for the evaluation of atherosclerotic lesions.
Atherosclerosis,

14. Cornhill, J.F. and Roach, M.R.: A Quantitative study of

the localization of atherosclerotic lesions in the
rabbit aorta. Atherosclerosis,

15. Pedley, T.J.: Flow in the entrance of the aorta.

Proceedings from a Specialists Meeting on Fluid
Dynamic Aspects of Arterial Disease, Ed. by
R.M. Nerem, held in Columbus, Ohio, Sept. 19-20, 1974.
(See also: Viscous boundary layers in reversing
flow. J. Fluid. Mech. Vol 74, Part 1. 59-79, 1976.

16. Nerem, R.M., Seed, W.A. and Wood, N.B.: An experimental

study of the velocity distribution and transition
to turbulence in the aorta. J. Fluid. Mech.
52:137-160, 1972.

17. Lutz, R.J., Cannon, J.N. and Munore, R.E.: Shear stress
measurements in model arteries during steady and
pulsatile flow. Proceedings from a Specialists
Meeting on Fluid Dynamic Aspects of Arterial Disease,
Ed. by R.M. Nerem, held in Columbus, Ohio,
Sept. 19-20, 1974.

18. Talkuder, N.: Untersuchung uber die Stromung in arteriellen

Verzweigungen. Dissertation, Technische Hochschule
Aachen, 1974.

19. Caro, C.G. and Nerem, R.M: Transport of l4C-4-cholesterol

between serum and wall in the perfused dog common
carotid artery. ~C~i~r~c~.~R~e~s~., 32:187-205, 1973.

20. Carew, T.E.: Mechano-Chemical Responses of Canine Aortic

Endothelium to Elevated Shear Stress In Vitro.
Thesis. Catholic University of America, Washington,
D.C., 1971.

21. Reif, T.H., Nerem, R.M. and Kulacki, F.A. An in vitro study
of transendothelial albumin transport in a steady state
pipe flow at high shear rates. J. of Fluids Engineering,
1976.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 463

22. Mack, P.J.: An In Vitro Study of 14C-4-cho1estero1

Transport Between Serum and the Arterial Wall in
the Presence of a Sinusoidally Oscillating Flow.
Master's Thesis. The Ohio State University, 1975.

23. Back, L.H.: Analysis of Oxygen Transport in the Vascular

Region of Arteries. Submitted for publication.

24. Forstrom, R.J., Voss, G.O. and Blackshear, P.L., Jr.:

Fluid dynamics of particle (platelet) deposition for
filtering walls: Relationship to atherosclerosis
ASME Journal of Fluids Engineering, 96-168, 1974.

25. Markle, R.A. and Hollis, T.M.: Regional aortic histamine

synthesis and transmural albumin uptake following
focal changes in shearing stress. Presented at the
60th Annual Meeting of the Federation of American
Societies for Experimental Biology. Anaheim, Calif.,
April 11-16, 1976.

26. Haimovici, H., and Maier, N.: Facts of aorta homograph in

canine atherosclerosis. Arch. of Surg. 89:
961-969, 1964.

27. Smith, E.B. and Slater, R.S.: Lipids and low-density

lipoproteins in intima in relation to its morphological
characteristics. In: Atherogenesis: Initiating
factors. Ciba Syrnp. No. 12 (NS) 39-52, 1973.

28. Haimovici, H., Maier, N.: Experiment in canine athero-

sclerosis in autogenous abdominal aortagraph
implanted into the jugular vein. Atherosclerosis
13:375-384, 1971.

29. Bjorkerud, S. and Bondjers, G.: Arterial repair and

atherosclerosis after mechanical injury. Part I.
Permeability and light microscopic characteristics
of endothelium in non-atherosclerotic and athero-
sclerotic lesions. Atherosclerosis 13:355-363, 1971.

30. Whereat, A.F. and Staple, E.: The preparation of lipo-

proteins labeled with radioactive cholesterol.
Arch. of Biochem. 90: 224-228, 1960.

31. Taylor, C.B., Trueheart, R.E. and Cox, G.E.: Athero-

sclerosis in Rhesus Monkeys. III. The role of
increased thickness of arterial walls in atherogenesis.
Arch. Path. 76:14-28, 1963.
464 CHAPTER 7

32. Ssolowjew, A.: Experimentalle Untersuchungen uber die

Bedeautung von lokaler Schadigung fur die
Lipoidablagerung in der Arterienwand. Z. Ges. Exp.
Med. 69:94-104, 1929.

33. Economou, S.G., Taylor, C.B., Beattie, Jr., E.J. and

Davis, Jr., C. B. : Persistent experimental aortic
aneurysms in dogs. Surgery 47:21-28, 1960.

34. Taylor, C.B., Baldwin, D. and Hass, G.M.: Localized

arteriosclerotic lesions induced in the aorta
of the juvenile rabbit by freezing. Arch. Path.
39:623-640, 1950.

35. Stefanovitch, V., and Gore, I.: Cholesterol diet and

permeability of rabbit aorta. Exp. Mo1ec. Path.
14:20-29, 1971.

36. Day, A.J. and Proud lock, J.W.: Changes in aortic

cholesterol-esterifying activity in the rabbits
fed cholesterol for 3 days. Atherosclerosis
19:253-258, 1974.

37. Caro, C.G.: Atherogenesis: Initiating factors.

Associated Scientific Publishers, Amsterdam,
Ciba Foundation Symposium 12, 1973.

38. Mitchell, J.R.A. and Schwartz, C.J.: Arterial Disease

Blackwell Scientific Pubications, Oxford, 1965.

39. Schwartz, C.J., Stenhouse, N.S., Taylor, A.E., White, T.A.:

Coronary disease severity and necropsy. Brit. Heart. J.
27:731-739, 1965.
40. Goldsmith, H.L.: Blood flow and thrombosis. Thromb.
Diath. Haemorrh. 32:35-48, 1974.
41. Alan, J.E., Rasmussen, H.: In: Prostaoglandins in cellular
biology and the inflammatory process. Phariss, B.,
and Ramwelle, P., (Eds.) Plenum Press, N.Y., 1977

42. Ryan, M.J., Clark, K.E. and Brody, M.J.: Neurogenic and
mechanical control of canine uterine vascular resistance.
Am. J. Physiol., 227(3): 547-55, 1974.

43. Caro, C.G., Lewis, C.T., Weinbaum, S.: A mechanism by which

mechanical disturbances can increase the uptake of macro-
molecules by the artery wall. Proc. of Phys., C.20,
p. 77, 1974.
HEMODYNAMIC CONTRIBUTION TO ATHEROSCLEROSIS 465

44. Arnqvist, H.J. and Lundholm, L.: Influence of oxygen

tension on the metabolism of vascular smooth muscle;
demonstration of a Pasteur effect. Atherosclerosis
25:245-253, 1976.

45. Scott, R.F., Morrison, E.S., and Kroms, M.: Effect of

cold shock on respiration and glycolysis in swine
arterial tissue. Am. J. Physiol. 219:1363-1365,
1970.

46. Scott, R.F., Morrison, E.S., and Kroms, M.: Aortic

respiration and glycolysis in the pre-proliferative
phase of diet-induced atherosclerosis in swine.
J. Atheroscler. Res. 9:5-16, 1969.

47. Morrison, E.S., Scott, R.F., Kroms, M., and Frick, J.:
Glucose degradation in normal and atherosclerotic
aortic intima-media. Atherosclerosis, 16:175-184,
1972.

48. Morrison, A.D., Berwick, L., Orci, L., and Winegrad, A.I.:
Morphology and metabolism of an aortic intima-media
preparation in which an intact endothelium is
preserved. J. Clin. Invest. 5:650-660, 1976.

49. Benirschke, K. and Bourne, G.L.: The incidence and

prognostic implication of congenital absence of
one umbilical artery. Am. J. Obst. Gyn. 79:251-254,
1960.

50. Bryan, E.M. and Kohler, H.G.: The missing umbilical

artery. I. Prospective study based on a maternity
unit. Arch. Dis. Childh., 49:844-852, 1974.

51. Meyer, W.W. and Lind, J.: Iliac arteries in children with
a single umbilical artery. Structure, calcifications,
and early atherosclerotic lesions. Arch. Dis. Childh.
49:671-679, 1974.

52. Berry, C.L., Gosling, R.G., Loagun, A.A. and Bryan, E.:
Anomalous iliac complicance in children with a
single umbilical artery. Brit. Heart. J. 38:
510-515, 1976.

53. Taylor, C.B.: The reaction of arteries to injury by

physical agents with a discussion of arterial repair
and its relationship in atheroslcerosis. In Symposium
on Atherosclerosis, p. 74, National Research Council
Publ. No. 338, Academy of Sciences, Wash. D.C., 1955.
466 CHAPTER 7

54. Hass, G.M.: The pathogenesis of human and experimental

atheroarteriosclerosis. In Cowdry's Arteriosclerosis.
Survey of the Problem 2nd ed •• p. 689. Ed. by
H.T. Blumenthal. Thomas. Springfield. Ill •• 1967
 INDEX
Acetycholine Arterial Contractility
effect on arterial contrac- autonomic influences on,
tility, 120 157-158
Aging control of, 105-179
arterial changes in, 6-12 effect of acetycholine
changes in carotid artery on, 120
with, 380-381 effect of angiotensin
changes in carotid sinus on, 149-150
with, 382 effect of vasopressin on,
endothelial changes during, 151
16 hormonal control of, 149-
sensitivity of aortic arch 152
in, 147 importance of calcium in,
subendothelial changes in, 110
22 isotonic vs. isometric,
Albumin 112
transport across arterial relation to atherogenesis,
wall, 331 120
Angiotensin Arterial Pressure
effect on aortic contrac- setting and resetting
tility, 149-150 mechanisms, 155-156
in endothelial cell, 317 Arterial Wall
interaction with circula- energy production in, 245
ting catecholamines, 318 growth factors in, 287-288
Aorta, 1-12, 120-126 metabolic activity in, 245-
effect of sheer stress on, 288
66, 92 oxyge~ tension in, 248-254
reflex role of, 170-176 Pasteur effect in, 245
Aortic Arch, 123-149 protein and lipid transport
effect of hypoxia and into, 299-350
hypercapnia on, 142 tangential stress in, 111-
effect on pulsatile 117
pressures on, 124-127 transport of LDL in 313-
hypothalamic interactions 317
with, 144 Arteries
neural control of, 144-146 anatomical characteristics
norepinephrine effect on, of, 1-12
129-131 autoregulation of, 49, 50
pathological changes in, cell renewal in, 13, 14
146-147 changes in aging, 6-12
patterns of flow in, 395 collagen content of, 108
role of in blood pressure coronary circulation in,
control, 170-176 55-57, 71, 92, 166-170
sensitivity changes with curvature of, 83
aging, 147 development of, 2
467
468 INDEX

distensibility of, 85, 97- Autonomic control of arterial

98 contractility, 117-123,
endothelial changes 157-158
during growth of, 16 of baroreceptors, 144-146
elastic, 1-12 of cardiac functions, 144-
elasticity of, 112 146
elastin content of, 108 of cardiac functions, 167-
Arteriosclerosis 168
development of, (see of carotid sinus, 144-146
atherogenesis) of cerebral circulation,
etiologic role of 159-163
thrombosis in, 87 of general circulation,
in aorta, 66 165-170
in descending aorta, 87 Baroreceptors, 123-149
in pulmonary artery, 87, 88 control of blood flow
understanding of, 55 distribution by, 149
Arteriovenous shunt, 95 effect of hypoxia and
Atherogenesis, 87, 88 hypercapnia on, 142
at arterial bifurcations, effect of pulsatile
82 pressures on, 124-127
fluid mechanics in, 384 hypothalamic interactions
hemodynamic contribution with, 144
to, 353-460 neural control of, 144-146
neoplastic theory of, 283 norepinephrine effect on,
neurohumoral effects on, 129, 131
152-153 pathological changes in,
predilection of various 146, 147
arterial sites for, 384- sensitivity of with aging,
388 147
relation of blood proper- Blood flow
ties to, 443-447 control of baroreceptors,
relation of blood rheology 149
to, 447-451 Fourrier analysis of, 155
permeability, 205 pulsatile, 55
relation to hypertension, turbulent, 61
430 velocity of, 60
relation to mechanical wave forms of, 60
factors to, 417 Calcium
relevance of shear stress affinity of arterial wall
to, 384-395 for, 353-378
Atrial receptor importance of in arterial
types of, 157 contraction, 111
Autoregulation of arteries, importance of in endothe-
49, 50 lial cell adhesion, 238
Cardiac functions
neural control of, 167-168
INDEX 469

Carotid artery, 1-12, 93 Cholesterol oxidation products

age changes in, 380-381 angiotoxic properties of,
calcification in, 366 279
circulation in, 74, 78, Collagen
94 binding of LDL by, 305
external, resonant spectra content in arteries, 108
in, 94 effect of hyperlipedemia
intimal cushions in, 353- on synthesis of, 270
364 formation of, 1-6, 90
structural features of, synthesis of, 281
353-374 Connective tissue
ultrastructure of, 342 bifurcations in, 92
Carotid sinus, 123-149 synthesis of, 6
age changes in, 382 Coronary arteries
control of blood flow circulation in, 55-57, 71,
distribution by, 149 92, 166-170
effect of hypoxia and effect of shear stress on,
hypercapnia on, 142 67, 92, 96
effect on peripheral innervation of, 167-169
resistance, 133-136 neurogenic control of, 167-
effect of pulsatile press- 169
ures on, 124-127 patterns of flow in, 395-
hypothalamic interactions 403
with, 144 resonance phenomena in, 94
neural control of, 144-146 shear stress in, 92
norepinephrine effect on, Cronweight, 96
129-131 Denervation of artery
pathological changes in, relation to arterial
146-147 metabolism, 414-415
sensitivity of with relation to atherogenesis,
aging, 147 414-415
structural features of, Edema
370-374 subendothelial, 95
vorticity in, 82 Elastin
Carotid syphon binding of LDL by, 305
structural features of, content in arteries, 108
353-374 formation of, 1-6, 90
Cathecholamines synthesis of, 281
interaction with angio- Endothelial cells
tensin, 318 contraction of, 153
Caveolae, 325 deformation of by shear
Cerebral circulation stress, 154
autonomic control of, 158- hydrodynamic effects on,
163 193-240
effect of arterial pC02 human vs. bovine, 229
on, 165 reverse transport of
effect of symathectomy on, cholesterol in, 278
165 role in transport of
macromolecules, 154, 229, 329
470 INDEX

tissue culture of, 216- Fibrinogen

220 transport across arterial
Endothelium wall, 331
aortic, critical shear Fibrous plaques, 256-262, 273,
stress in, 93 275, 288, 287
arterial vs. capillary, location of, 413
203 pathogenesis of, 252
effect of hemodynamic Gelatinous lesions, 256-262,
forces on, 203-204 272
effect of shear stress on Glycosaminoglycans
permeability of, 204, 239, binding of LDL by, 305
240, 406 Gortler instability, 97
erosion of, 95 Histamine
hemodynamic damage to, 404- effect on endothelial
405 permeability, 411-412
injury, associated with synthesis of, 411-412
critical shear, 94 Hooke's Law, 106
integrity of, 88, 89, 98 Hydroxyproline
mitosis in, 20 synthesis of, 283
permeability, 94 Hypercholesterolemia
permeability to LDL, 317 effect on endothelium,
repair of, 234-237 426-427
rheologic behavior of, relation to atherogenesis,
94, 95 277
structure of, 16 Hyperlipedemia
venous vs. arterial, 212 effect on collagen
Erythrocytes synthesis, 270
deformability of, 432-434, Hypertension
447 relation to atherogenesis,
effect of prostoglandins 430
on, 447 relation to macromolecules,
Fatty streaks, 256, 275, 283, 264
286, 428, 430 Hypoxia
distribution of, 386, 413 effect on baroreceptors,
mechanism of production, 142
277-278 in atherogenesis, 248-254,
relation to low oxygen 408-410
tension, 254 relation to fatty streaks,
Femoral, 93, 120 254
flow patterns in, 79 relation to lactate
geometry of, 55, 74 production, 262
growth and development of, relation to lysosomal
90 damage, 258
growth of, 14, 15 relation to smooth muscle
Ferritin cell proliferation, 262,
transport of, across endo- 284-286
thelium, 338-340 Iliac, 1-12, 112, 117, 453-459
innervation of, 117
INDEX 471

in vitro modeling of, 57 role in endothelial cell

mechanical properties of, attachment, 221, 229
106 Myofi1aments
metabolic activity in, 245- anchoring, 20
288 cytoplasmic, 20
Innervation development of, 6
(see autonomic control) Navier-Stokes equations, 70
Lactic acid Neurohumoral effects
accumulation in arterial on contractile proteins,
wall, 252-254 153
effect of accumulation on on arterial contractility,
1ysosomes, 254 149-152
LDL on arterial wall shear
binding of by collagen stress, 152
and elastin, 305 on vasa vasorum, 152-153
concentration in different relation to atherogenesis,
segments of aorta, 413 152-153
concentration in smooth Norepinephrine
muscle cells, 317 effect on arterial
transport of in arterial circulation, 129, 131
wall, 312-317 effect on aortic arch,
Lipids 129, 131
affinity of arterial wall effect on carotid sinus,
for, 353-378 133-136
Macromolecules Pasteur effect on arterial wall,
interaction with plasma 245-248
in intima, 272 Phonoangiography, 82
transport through arterial Plasmalemma membranes transport
wall, 196-205 through, 197-202
relation to hypertension, vesicles distribution of
264 20-23
Mechanical stress in endothelial cell,
effect on arterial smooth 342-350
muscle, 262-264 Platelet rich plasma
Metabolic activity in arterial effect of on cultured cells,
wall, 245-286 228
Mesenteric, 117, 120 Platelet
muscular, 1-12 growth stimulation factor,
obstruction in, 79 334
oxygen utilization by, 451, deposition, mechanism of,
453 410-411
permeability of, 44 Poisseui11e's Law, 61
Microfi1aments Pulmonary, 63
affinity for platelet pulsatile flow in, 76
attachment, 234, 238 repair of, 417-425, 453
in smooth muscle cells, response to cell loss, 12
237 rheology of, 47
senility of, 14, 15
472 INDEX

species differences in Strain,

structure, 15 circumferential, 48
steady vs. unsteady flow values of, 49
in, 76 Stress
stenosis of, 78 concepts of, 41
structural features of, 353 distribution of in arteries,
turbulent flow in, 1-12, 50
83 fluid shear, 55
Reynolds number stress relaxation, 32
calculation of, 70, 71 stress vs. strain, 31
Rheology Sympathetic nerves
continuum mechanics in, (see autonomic control)
50 Tethering
Shear rates, 66 importance of, 89, 93
in relation to blood Thrombus generation
viscosity, 439-442 in relation to atherogenesis,
Shear stress 205-208
calculation of, 216 Turbulence, 93
critical, 67, 89, 92, 93, Umbilical artery, 453-460
94, 95, 96 Uterine, changes during
effects on arterial pregnancy, 448
structure, 27 velocities and pressure
in aorta, 66, 92 wave forms in, 57, 64, 68,
in arterial wall, 60, 63 92
in bifurcations, 67, 68 Van der Waals forces, 200, 202
in coronary vessels, 67, 236, 322, 325
92, 96 Vasoactive substances, 105
in divided flow, 389-395 Vasomotor function
in iliac and carotid control of, 105-179
arteries, 88 (see also autonomic control)
of fluid, 55 Vasa vasorum
relation to arterial role in arterial nutrition
metabolism, 428-429 248-250, 289, 290, 325
relation to endothelial Velocity profiles, 70
permeability, 204, 239, Vertebral, 1-12
240, 406 viscoelastic properties
relation to turbulence, 67, of, 106
71 vorticity in, 79
relevance to atherogenesis, Vesicles
384-395, 411-412 motion - dynamics of, 328
Smooth muscle of artery transport - effect of
activation of, III fluid mechanics on, 448-449
culture of, 266-270, 281 model of, 320-329
proliferation of, 273-275 through endothelial cells,
Stellate ganglion 193-197, 305-317, 428
role in distribution of Weibel-Palade bodies, 20
coronary blood flow, 178 Womersley formula, 61-66