| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© 1997 American Heart Association, Inc.
Articles |
Effect of Atherosclerosis on Transmural Convection and Arterial Ultrastructure
Implications for Local Intravascular Drug Delivery
From the Department of Physiology, University of Arizona, and B.W. Zweifach Microcirculation Laboratories, Veterans Affairs Medical Center, Tucson, (A.L.B., L.M.W.); Krannert Institute of Cardiology, Indianapolis, Indiana (I.G.P., R.W., K.M.); and Veterans Affairs Medical Center, Indianapolis, Indiana (K.M.).
Correspondence to Ann L. Baldwin, Department of Physiology, College of Medicine, University of Arizona, Tucson, AZ 85724. E-mail A160854{at}ccit.arizona.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Abstract Local infusion of agents through perforated catheters may reduce neointimal formation following vascular angioplasty. Such treatment will succeed only if the drug is retained within the arterial intima long enough to promote repair. Drugs will be dispersed throughout the wall predominantly by transmural convection instead of diffusion if the Peclet number, Pe=J(1-
f)/P, is greater than unity, where J is the
transmural fluid flow per unit surface area and f and P
are the reflection and permeability coefficients to the drug,
respectively. Although the targets of local drug delivery will be
atherosclerotic vessels, little is known about the transport properties
of these vessels. Accordingly, we evaluated the effects of
hypercholesterolemia and
atherosclerosis on J per unit pressure (hydraulic
conductance, Lp) and on ultrastructure in femoral arteries.
Measurements were made at 30, 60, and 90 mm Hg in
anesthetized New Zealand white rabbits fed a normal diet (n=6)
and after 3 weeks of lipid feeding (n=19).
Atherosclerosis was induced in six lipid-fed animals by
air desiccation of a femoral artery. Hydraulic conductance was
significantly greater in vessels from
hypercholesterolemic than from normal animals and
decreased with pressure only in hypercholesterolemic
arteries. Atherosclerosis did not augment hydraulic
conductance compared with hypercholesterolemia
alone. Electron microscopic examination demonstrated damaged
endothelium in hypercholesterolemic
arteries and both altered endothelium and less tightly
packed medial tissue, compared with controls, in atherosclerotic
vessels, at least at lower pressures. Peclet numbers for macromolecules
exceeded unity for all three groups of arteries and reached 0.3 to 0.4
for molecules as small as heparin. Thus, convection plays a dominant
role in the distribution of macromolecular agents following local
delivery and may result in their rapid transport to the adventitia in
the femoral artery.
Key Words: rabbits • artery, femoral • hydraulic conductance • hypercholesterolemia • local drug delivery
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The use of balloon angioplasty for nonsurgical recanalization of atherosclerotic arteries has been associated with subsequent remodeling and proliferation of smooth muscle cells in the intima, leading to restenosis of the artery within a few months in about 30% of cases.1 One approach to reducing the incidence of restenosis is local administration of relatively high concentrations of agents following angioplasty. Such local delivery has been carried out by intravascular administration of agents from porous catheters,2 3 bioresorbable gels,4 or stents,5 as well as by perivascular placement of matrices containing drug.6
Several studies have evaluated the intramural residence time of soluble agents following delivery, including heparin (predominately eliminated in less than 24 hours),3 methotrexate (65% cleared within first 24 hours),7 and colchicine (90% cleared within a few hours).8 The residence time of agents delivered locally appears to play a significant role in determining their bioactivity,9 and several approaches have been developed to prolong residence time. These have involved the incorporation of agents into polymeric matrices, which may be macroscopic gels4 or coatings,10 as well as microparticulates appropriate for transcatheter delivery.2
The residence time of agents delivered locally, as well as the steady-state concentrations of compounds released from local stores, may be strongly influenced by diffusive forces causing transport away from the site of delivery, either deeper into the arterial wall or back into the bloodstream, and by convective forces promoting transport deeper into the wall. Moreover, these forces will determine both the distribution of agent deposition in the vascular wall upon catheter delivery and subsequent redistribution after delivery. Several experimental studies have shown that transmural convection promotes transport of macromolecules across the arterial wall, from blood to tissue.11 12 13 14 15 16 However, other authors17 have concluded that macromolecular transport is dominated by diffusion unless the endothelium is damaged. In addition, it has been suggested that diffusion far outweighs convection in control of transmural heparin transport in the normal artery and that convective forces increase to only one-quarter the magnitude of diffusive forces following endothelial injury.18 Thus, the effect of convection on local drug delivery has not been resolved.
One way to assess the influence of convection on the effectiveness of
drug delivery is to calculate the ratio of convective to diffusive
solute transport for a given molecule. If this ratio exceeds unity,
solute transport is governed predominately by convection. The ratio of
imposed (convective) velocity of a molecule traveling through the
arterial wall to diffusive velocity is expressed by the
Peclet number,
Pe=J(1-f)/P,19 where J
is the transmural flow per unit surface area, P is the permeability
coefficient of tissue to the solute, and f is
the filtration reflection coefficient, which represents the
relative access that the solute molecules have to water transport
channels. If water and solute molecules have equal access to such
channels, that is, there is 100% coupling between solute and solvent,
then f has a value of zero. If the solute has
no access to water transport channels, then f
is equal to unity and there is no convective transport of solute.
Although measurements of hydraulic conductance, Lp (J per unit transmural pressure), have previously been made on healthy arteries,20 21 22 23 24 none have been performed on vessels from animals subjected to a high-lipid diet, with or without the presence of atherosclerotic lesions. Because the arteries to be treated to prevent restenosis will inevitably be atherosclerotic, this study was designed to determine whether this condition alters Lp. Accordingly, Lp of femoral arteries from rabbits fed normal or hypercholesterolemic diets was measured to enable calculation of Peclet numbers for large and small molecules. Lesions had been induced in some of the femoral arteries from hypercholesterolemic rabbits. The femoral artery was chosen for these studies because it is a site of predilection for obliterating vascular lesions and has been extensively studied with respect to atherosclerosis as well as local drug delivery.25 26 27
Control and hypercholesterolemic arteries, with and without lesions, were examined by light and electron microscopy for the purposes of obtaining structural information for determination of Peclet numbers and of relating ultrastructural changes to alterations in hydraulic conductance.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Induction of Focal Atherosclerosis
Experiments were performed on 18 male New Zealand White rabbits (4.0 to 5.0 kg) anesthetized with ketamine (100 mg/kg) and xylazine (1.5 mg/kg IM). Focal femoral atherosclerosis was induced unilaterally as described previously.27 In brief, 1 cm below the inguinal ligament, a 3-cm cutdown over the femoral artery was made, and the artery was exposed. Lidocaine solution (2%) was dripped on the artery to prevent spasm. A 1.0- to 1.5-cm-long segment of the femoral artery, devoid of side branches, was isolated, and flow was temporally interrupted with silk ligatures. The arterial segment was then cannulated with a 27-gauge needle, vented distally, and flushed with saline to remove blood. Air was passed through the arterial segment for 8 minutes at 80 mm Hg pressure to achieve endothelial desiccation, after which blood flow was reestablished. An intramuscular injection of Cefadyl (1 g) was routinely given. The injured segment was marked by a metal clip overlying the muscle layer, and the wound was closed. The rabbit was then placed on a 2% cholesterol/5% peanut oil diet (150 g/d) for 3 weeks.
An angiogram was taken 2 weeks after commencement of the diet to confirm vessel patency and consequent suitability for bilateral Lp measurements. Hydraulic conductance measurements were made on all noninstrumented femoral arteries to evaluate the effects of hypercholesterolemia without atherosclerosis. Measurements of hydraulic conductance were performed, as described below, 3 weeks after commencement of the high-lipid diet. Control arteries (n=6) were studied in rabbits fed a standard diet.
The Atherosclerotic Rabbit as a Model for Human
Angioplasty
Previous studies28 29 have demonstrated
that experimental atherosclerosis can be
consistently induced in the New Zealand White rabbit by
endothelial injury followed by a high
cholesterol (2% cholesterol/6% peanut oil)
diet. Our technique, originally described by LeVeen et
al,28 results in a mixed fibrocellular lesion.
Although the diet induces very high serum levels of
cholesterol (1146±238 mg%) and the
atheromatous lesions differ from advanced human
atherosclerosis in that necrosis, fibrosis, and
calcification are absent, there are many attractive features of this
model. These include the rapid and reliable development of stable
atherosclerotic lesions that can be assessed precisely by angiography,
with acute and chronic postangioplasty complications similar to those
documented after human angioplasty.
Surgical Procedure
The rabbits were anesthetized with pentobarbital sodium
(30 mg/kg IV). The surgical procedure was described in detail in a
previous publication.20 Briefly, the femoral
artery was exposed, and any branches were ligated close to the origin.
Heparin (1000 IU) was administered intravenously. During
the whole operative procedure, physiologically
buffered PBS, pH 7.4, warmed to 37°C and containing 4%BSA, was
applied continually to the adventitial surface of the artery to prevent
drying.
A length of femoral artery, from the inguinal ligament to the saphenous bifurcation (about 2.5 to 3.0 cm), was cannulated. The perfusate, which entered the vessel from a reservoir through the proximal cannula, was identical to the suffusate except that it contained 0.03% trypan blue dye. Dye was included so that leaks could quickly be identified. Normally, no adventitial staining was observed. Both proximal and distal cannulae were connected to pressure transducers, with output recorded using a Gould chart recorder. The inlet reservoir was positioned 60 cm above the cannulated femoral artery, and the flow rate was adjusted, using a screw clamp, to flush the blood from the segment. After the absence of leaks was established, the muscles surrounding the femoral artery were retracted to form a cavity, which was filled with the solution that had been used to keep the vessel moist.
Measurement of Hydraulic Conductance
Arteries were initially preconditioned to reduce the hysteresis
effects resulting from viscoelasticity (ie, the tendency for the vessel
to return to a diameter greater than its initial value after
pressurization). This was achieved by lowering the proximal
arterial pressure to 25 mm Hg, then raising it to
100 mm Hg, three times while occluding the outlet
cannula.
Measurements of transmural fluid flux and external radius were then made at 30, 60, and 90 mm Hg. To measure fluid flux, the upstream reservoir was connected, via a stopcock, to a 2-m length of nylon capillary tubing (0.06 cm inner diameter), a portion of which was horizontally positioned adjacent to a meter rule. The downstream end of the tubing was connected to the inlet cannula of the femoral artery. An air bubble was injected into the capillary tubing, the bubble position with reference to the meter rule was noted, and the pressure was raised to the chosen value. The position of the bubble was monitored 1, 2, 4, 6, 8, 10, 15, 20, and 30 minutes after raising the pressure. The external diameter and length of the cannulated segment of artery were measured using a mechanical caliper (accurate to 0.1 mm) initially and 6 and 30 minutes after pressurization. There was very little tapering of the segment; nevertheless, the diameter was always measured at the midpoint of the vessel length. All the tubing had been previously coated with Sigmacote to prevent sticking of the bubble.20
After the initial distension of the vessel, the air bubble was
confirmed to move at a constant rate, representative of
the rate of fluid filtration through the artery wall. The transmural
fluid flux was calculated for each pressure from the steady-state
velocity, v, of the air-bubble in the capillary tube using the
equation
 |
 |
Light and Electron Microscopic Observations
After measurements of Lp had been
performed, two control arteries, five
hypercholesterolemic arteries, and three
atherosclerotic arteries were processed for light and electron
microscopy. Five additional control femoral arteries, one additional
hypercholesterolemic vessel, and six more
atherosclerotic vessels were included in the light microscopic study.
Some of these additional vessels were also examined by electron
microscopy. In total, the following vessels were fixed and sectioned
for electron microscopy: three control vessels (one at 30 mm
Hg and two at 60 mm Hg), six
hypercholesterolemic vessels (two at 30 mm
Hg, two at 60 mm Hg, and two at 90 mm Hg), and
eight atherosclerotic vessels (four at 30 mm Hg, three at
60 mm Hg, and one at 90 mm Hg). All the vessels
were perfusion-fixed, in situ, at room temperature for 2 hours at
pressures ranging from 30 to 90 mm Hg with phosphate-buffered
Karnovsky's fixative (pH 7.4) containing enough dextran 40 to match
the colloid osmotic pressure of plasma (25 mm Hg). Fixative
was also applied to the outside of the vessels. Each vessel was then
excised and cut into annular segments about 2 mm in length. The
segments were washed in 0.15 M phosphate buffer, postfixed in 1%
OsO4 for 2 hours, dehydrated in alcohols,
embedded in Spurrs resin, and sectioned for light and electron
microscopy. Sections from each vessel were examined by light microscopy
to determine medial thickness (average of eight measurements equally
spaced circumferentially around annulus), outer diameter (circumference
of external elastic lamella/
), and intimal thickness (intimal
area/circumference of internal elastic lamella). All measurements were
made using a stylus and Bioquant software. Ultrathin sections were
examined by electron microscopy to qualitatively assess the structural
integrity of the endothelium.
Statistics
To determine whether Lp varied between
groups (control, hypercholesterolemic, and
atherosclerotic) for a given pressure, Student t tests were
performed. Variation with pressure for a given group was evaluated
using paired Student t tests. If a t test gave a
P value of <.05, we considered the null hypothesis to
be rejected.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Measurements of Hydraulic Conductance
Measurements of hydraulic conductance are shown in Table 1
. Corresponding values from the
literature, measured in healthy vessels, are shown in Table 2. Hydraulic conductance was
significantly greater in both hypercholesterolemic and
atherosclerotic vessels than in control vessels at 30 mm Hg
and exhibited the same trend at higher pressures. Hydraulic
conductances of hypercholesterolemic and
atherosclerotic vessels were not significantly different from each
other at any of the three pressures. In control arteries,
Lp did not vary significantly with pressure,
while in hypercholesterolemic and in atherosclerotic
arteries, Lp decreased with increasing pressure.
The values of Lp determined for control rabbit
femoral arteries were two to three times higher than those previously
obtained for rabbit aortas using the same technique (see references 20
and 21 in Table 2). This difference can be accounted for by the fact
that the aortic media is two to three times thicker than that of the
femoral artery. Measured dimensions of the femoral arteries are shown
in Table 3. Assuming an aortic wall
thickness of 120 µm, the product of Lp
and wall thickness, or the Darcy constant, k, will be equal to
5x10-10 cm2/s
· mm Hg-1. This is close to the k
value of 6.9x10-10
cm2/s · mm
Hg-1 calculated for control femoral arteries at
60 mm Hg (Table 4). We have
determined the Lp for rabbit aortas denuded of
endothelium to be
10.19±1.91x10-8 cm/s · mm
Hg-1 (reference 21 in Table 2), which gives a
Darcy constant of 12.2x10-10
cm2/s · mm
Hg-1, similar to the value obtained by Harrison
and Massaro (reference 22 in Table 2) for the porcine
endothelially denuded aorta. These values for denuded
aorta are similar to those obtained for the
hypercholesterolemic femoral arteries at 30 and 60
mm Hg (Table 4).
|
|
|
|
Measurements of Vessel Dimensions
As shown in Table 3
, the mean diameters of the
hypercholesterolemic and atherosclerotic vessels were
smaller than that of control vessels, but this difference was only
statistically significant for the latter. The mean medial thickness of
atherosclerotic vessels was significantly greater than that of control
and of hypercholesterolemic vessels. Intimal thickening
was eccentric and was seen only in atherosclerotic vessels.
Calculations of Peclet Numbers
In order to evaluate Peclet numbers using measured values of
Lp, compatible values for permeability and
reflection coefficients are required. Alternatively, Peclet numbers can
be expressed in terms of the Darcy constant, k, and the coefficient of
diffusivity, D, which is equal to the product of permeability
coefficient and wall thickness. By use of the parameters k
and D, the conductivity of arterial tissue to water and
solutes can be expressed independent of wall thickness. Because of the
absence of any measured values in the literature for P, D, and
f of the intact femoral artery, published
coefficients for the various single components of the
arterial wall (Table 2) were used in a model based on the
structure of the wall as revealed by our microscopic studies. The
structure of arteries has been used effectively to predict diffusion
coefficients by Penn et al.31 In this previous
study, the total resistance to HRP movement into and through the intima
was calculated from measurements of HRP concentration in the intima and
media using photomicrographs of arterial sections.
Demonstration of the arterial ultrastructure was required
to exclude layers of medial elastic tissue from the measured areas. In
addition, Fry30 used light and electron
microscopy to check that his arterial transport model took
account of the structural details of the artery. Thus, in both cases,
knowledge of the structure of the artery was essential for
determination of permeability coefficients.
In our model, the arterial wall is considered as three
components in series: endothelium (endo), internal
elastic lamina (iel), and a media consisting of fenestrated layers of
smooth muscle cells and elastic tissue embedded in a matrix ofGAGs
(gag). In the femoral artery, unlike the aorta, there is only one
elastic lamina. This structure can be seen in Figs 1, 2, a,
3, a, and 5, a. Using this model, the
macromolecular permeability coefficient of the whole artery can be
estimated from the permeability coefficients of the individual
components as follows:
 |
) were used in the above equation (unfortunately,
the permeability coefficients to the same macromolecule were not
available for both components):
 |
 |
 |
, Fry32 gave a value of
0.24x10-9 cm2/s for
Diel to albumin, but he assumed that the
internal elastic lamellar was just a thin part of a
homogeneous media. For that reason, we shall use
0.24x10-10 cm2/s for
Diel to macromolecules.
|
|
A value for Pgag was obtained from measurements
of the diffusion coefficient of serum albumin within GAGs, in
particular to HA. In cartilage interstitium, which consists largely of
a 5% to 10% HA solution, Dgag is
1.6x10-7 cm2/s (serum
albumin, reference 33). The exact GAG concentration is not
critical for our purposes, because Dgag is not
strongly dependent on concentration. For example,
Dgag for a 2.5% HA solution is
2.2x10-7 cm2/s (serum
albumin, reference 34).
 |
).
 |
 |
Thus,
 |
 |
, Fry30 gave a value of
0.137x10-7 cm2/s for the
diffusion coefficient of albumin in the media, but he did not
correct for the percentage space available for diffusion. Bratzler and
Schwartz36 obtained a value of 0.13
cm2/s in an excised rabbit aorta. Because the
aorta contracts longitudinally on excision, this could explain the
difference from our calculated value.
It can be seen that the GAGs are much more permeable to macromolecules
than are the endothelium or the internal elastic
lamina. Applying the published data values to our model, we calculate
that
 |
 |
Using our calculated value for Pwall and an
Lp value of 13x10-8
cm/s · mm Hg-1 determined for
control vessels at a pressure of 60 mm Hg, a Peclet number of
33(1-f) was obtained. Therefore, the highest
possible value for Pe is 33. This assumes that there is total coupling
between convective fluid flux and macromolecular flux. There is very
little coupling between macromolecular and convective fluxes across the
endothelium.11 14 38 39
Measurements of f for serum albumin
made in skeletal muscle capillaries, which possess
endothelial cells with tight junctions, similar to
those seen in arterial endothelium, give a
value for f of 0.95. This indicates that the
rate of transendothelial albumin transport by
convection is only 5% of that of water. However, if we are considering
the effect of convective flux on molecules that have already crossed
the endothelium during the delivery process, then the
degree of coupling of solute and solvent may be different. There are no
measurements of arterial f for
macromolecules in the literature. However, Fry32
made an "educated guess" regarding coupling of LDL and water across
the internal elastic layer from the intima to the media. He estimated a
value of 0.95 for muscular arteries, such as the femoral artery. Thus,
for molecules that are lodged between the endothelium
and the internal elastic lamella after delivery, we obtain a Peclet
number of 1.65, indicating that transport is still
convection-dominated. Molecules that had penetrated both the
endothelium and the internal elastic lamella during
delivery, would be more greatly affected by convection, because of
greater solute-solvent coupling, and the Peclet number would exceed
1.65.
With regard to hypercholesterolemic vessels, the value
for Lp is greater than for controls
(18x10-8 vs 13x10-8
cm/s · mm Hg-1). Also, our
electron micrographs demonstrate that the
hypercholesterolemic vessels sustained some degree of
endothelial damage (Fig 2, a) but that the media was
similar to controls. Assuming that the permeability of the internal
elastic lamella and media are similar to that of controls and using a
medial thickness of 59 µm (Table 3), then
 |
 |
 |
 |
The highest possible value for Pe is 45. With regard to lower
bounds for Pe, the value would be 2.3 if we assume that
f=0.95, as for control vessels. Therefore, for
arteries from hypercholesterolemic animals, Pe would
still be greater than unity, and thus macromolecular transport would be
convection-dominated.
With regard to atherosclerotic arteries, the value for
Lp is 16x10-8 cm/s
· mm Hg-1. Our light micrographs
showed that medial thickness for atherosclerotic vessels was about
twice that for controls (Table 3). Therefore, we will use a value of
106 µm instead of 50 µm to calculate
Pgag. Although the intima is thickened in
atherosclerotic arteries, the thickening is highly eccentric and is
composed mainly of GAGs and cellular material. Our calculations show
that the GAGs do not appear to offer much resistance to macromolecular
transport, and unless the specific molecular agent is modified to
interact with the cells or matrix, the cellular material will not
impede transport. Therefore, the thickening of the intima will not be
included in the calculation of Pe. Insertion of the new value for
medial thickness into the equation for Pwall
leads to a value of 22.8x10-8 cm/s.
Dwall is equal to
23.9x10-10 cm2/s.
Using an Lp value of
16x10-8 cm/s · mm
Hg-1, we obtain
 |
The highest possible value for Pe is 42, and the lowest is 2.1,
assuming that f remains unaltered compared to
control vessels. It is probable that f will be
reduced in atherosclerotic arteries, and, if so, the lower bound for Pe
would be higher than 2.1.
From the above estimations, it appears that transmural convection plays a greater role in the transport of macromolecules through the femoral artery wall than does diffusion and that this tendency is retained in arteries from hypercholesterolemic and atherosclerotic animals.
Because heparin (MW, 12 to 15 kd) had been considered as a drug to
reduce arterial restenosis and because values exist
in the literature for heparin for Pmedia of rat
aorta,18 which has a similar medial thickness to
the rabbit femoral artery, we determined possible values of Pe for
heparin:
 |
 |
The highest possible value for Pe is 0.4. There are no available values
of f for heparin. However, the published value
for Pendo for heparin is similar to that for
sucrose in skeletal muscle capillaries
(1.4x10-5 cm/s, reference 39), even though
sucrose has a much lower molecular weight than heparin (342 d).
Therefore, we will use the published value of
f for sucrose (0.11, reference 39) in the
equation for Pe. This approximation gives a lower bound for Pe of 0.36.
Therefore, for heparin, diffusion appears to play as great a role as
convection in arterial transport. However, convection will
have some influence on the distribution of small to medium-sized agents
throughout the arterial tissue.
Electron Microscopic Observations
Control Vessels
Control vessels showed intact endothelium, a thin
intima (up to about 2 µm thick), and closely packed,
cigar-shaped medial smooth muscle cells, which were aligned
circumferentially. The collagen fibers were arranged in groups, each
fiber within a given group having similar orientation. An electron
micrograph of a transverse section through a control aorta fixed at
30 mm Hg is shown in Fig 1.
Hypercholesterolemic Vessels
The appearance of the hypercholesterolemic vessels
was different from that of the controls in one respect; in sections
from three of the six hypercholesterolemic vessels
prepared for electron microscopy, the endothelium was
largely missing, and in sections from two further vessels, the
endothelium was vacuolated or missing in places. This
endothelial damage was also demonstrated by the blue
staining of the intimal surface of hypercholesterolemic
vessels that was seen on their excision from the animals. The trypan
blue, which was included in the vessel perfusate, stains the
intima only if the endothelium is damaged or missing.
Control vessels did not show such staining. The appearance of the media
was similar to, or slightly less dense than, that of the controls. Fig 2, a, shows an electron micrograph of a transverse section through a
hypercholesterolemic femoral artery in which the
endothelium is absent. A transverse section through the
aorta, prepared identically from the same animal, is shown for
comparison in Fig 2, b, manifesting an intact, undamaged
endothelium. Sections from the femoral artery from the
same animal consistently showed sloughed or damaged
endothelium. The same contrast between femoral and
aortic endothelia was made in two other
hypercholesterolemic rabbits.
Atherosclerotic Vessels
In areas of atherosclerotic vessels distant from the actual
lesion, the tissue appearance was similar to that of
hypercholesterolemic arteries that had not undergone
air dessication, with an incomplete endothelium, thin
intima, and closely packed and highly organized medial tissue.
Endothelial cells were generally present overlying
lesions but were not in perfect condition. Often, on excision, these
vessels demonstrated relatively little intimal trypan blue staining
over the lesion but showed intense staining of the surrounding area.
Examples of the intimas of transverse sections through lesions in
vessels fixed at 30, 60, and 90 mm Hg are shown in Figs 3, a,
4, a, and
5, a, respectively. The corresponding
medias are shown in Figs 3, b, 4, b, and 5, b.
|
In lesions, the endothelial cells were vacuolated
(arrows, Figs 3, a, 4, a, and 5, a), with some gaps between cells
(arrowhead, Fig 3, a), and occasional sparse cytoplasm suggesting cell
rupture and release of cytoplasmic contents (large arrow, Fig 3, a).
The intima reached up to 30 µm (Fig 4, a), with significant
eccentricity as well as variability noted between vessels. Many of the
cells within the intima had the appearance of smooth muscle cells,
which may have migrated from the media through fenestrae in the
internal elastic lamina (Fig 3, a). These cells were characterized by
an elongated shape and an abundance of cytoplasmic filaments and
plasmalemmal vesicles. Other intimal cells were present
that resembled "modified smooth muscle cells," characterized by
their ovoid shape and abundant rough-surfaced endoplasmic reticulum and
prominent Golgi apparatus.40 An
example of two such cells can be seen in Fig 4, a, just beneath the
endothelial cell, indicated by an arrow. A large
fraction of the intimal cells of all lesions examined by electron
microscopy contained large fat globules (Figs 3, a, 4, a, and 5, a).
Lipid-filled intimal cells have previously been noted in the femoral
arteries of rabbits intermittently fed a high-cholesterol
diet for 3 to 14 weeks.41 Fat globules were often
seen within smooth muscle cells in the adventitia, close to vasa
vasorum, suggesting that the processes responsible for the deposition
of blood-derived lipids within the intimal cells are also operating
across the walls of the vasa vasorum. Occasionally, fat globules were
found within smooth muscle cells within the central regions of the
media (see Fig 3, b), but this was a rare occurrence. The intima of the
artery fixed at 90 mm Hg (Fig 5, a) appeared to be more
densely packed (ie, less extracellular space per unit volume of tissue)
than the vessels fixed at 30 and 60 mm Hg (Figs 3, a and 4,
a, respectively).
|
|
The appearance of the media within lesions was variable. In many
cases the extracellular space per unit volume of tissue appeared to
larger than that seen in controls and in nonlesion regions of
hypercholesterolemic vessels (see Figs 3, b, and 4, b).
In these specimens the smooth muscle cells exhibited many pronounced
pseudopodia, in contrast to the smooth outlines displayed by the smooth
muscle cells in control preparations (Fig 1). In the lesion from a
vessel fixed at 90 mm Hg, the media resembled that of control
vessels (Fig 5, b). The smooth muscle cells within the media were quite
densely packed and showed few pseudopodia.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
In summary, this study has shown that delivery of drugs may be significantly influenced by convective fluid flux through the artery wall, from the lumen to the adventitia. Convection-dependent washout of agents from the intima and inner media is expected to occur. This finding should be taken into account when designing drug treatments to reduce arterial restenosis. In this atherosclerosis model of the rabbit femoral artery, diet-induced hypercholesterolemia for a period of 3 weeks caused an increase in hydraulic conductance of the femoral artery, regardless of whether lesions were initiated by mechanical damage of the endothelium. It might be expected that the occurrence of even early atherosclerosis in human vessels would favor increased convectional drug washout.
Our conclusion that macromolecular transport is affected by convective fluid flux is in disagreement with the conclusion of Penn et al.17 Those authors measured solute (HRP) transport and obtained values for convective velocity by nonlinear parameter estimation. The optimal values of the medial parameters were those for which the model output provided the best least-squares fit to the experimental data. The authors assumed in their model that transendothelial transport is dominated by diffusion. They stated that their estimations of arterial transport properties had very poor precision regarding animals with undamaged endothelium. However, their estimation of P was close to other reported permeability coefficients to macromolecules. For this reason, we used their estimation of P in our calculations of Pe. In fact, it is their value for Lp that is in disagreement with other published data. In particular, their value of 10 cm/s · mm Hg-1 for the rat aorta is three orders of magnitude smaller than the value that we measured in the femoral artery. This deviation could explain their very low values of Pe. The optimum experiment, that is, the measurement of macromolecular transport properties and Lp of the same vessel, remains to be performed.
Convective washout, and hence failure to maintain a high concentration
of drug in the arterial intima, is consistent with
the lack of success achieved so far by certain intramural treatment
regimens targeting restenosed arteries, such as for administration of
heparin by porous balloon catheter.27 The
limitation of convective washout needs to be addressed in order to
exploit the advantage of focal administration of drugs to an
atherosclerotic site. One possibility is to alter the physicochemical
characteristics of drugs that affect their f
or P values. For example, an agent could be selected that rapidly binds
specifically to intimal smooth muscle cells and prevents their
proliferation. Delivery of the agent could also be via liposomes, which
would remain entrapped within the intima.42
Another interesting possibility is raised by recent studies that suggest that intimal atherosclerotic lesions may be triggered by occlusion or thrombosis of the adventitial vasa vasorum.43 44 It is thought that the resulting tissue hypoxia in the media may cause migration of the medial smooth muscle cells toward the intima. In addition, mRNA for platelet derived growth factor may be increased in the hypoxic endothelial cells of the vasa vasorum, and the platelet-derived growth factor may cause smooth muscle cell proliferation. If this is the case, then it would be advantageous for luminally administered drugs to be redistributed to the adventitia by convective flux, since the adventitia would be the targeted site of need. It is much easier to deliver drugs intraluminally and rely on convection to rapidly propel them to the adventitia than it is to administer drugs directly into the adventitia. Although drugs can be delivered locally from polymers implanted in the adventitia, surgical placement of drug-eluting matrices is impractical for most clinical applications. Thus, it may be possible to exploit the role of convection in transporting agents from the lumen to the adventitia in preventing development of atherosclerotic plaques.
Ultrastructural evaluations provide further data confirming the basis for the observation of hydraulic conductance. In hypercholesterolemic vessels without lesions, our ultrastructural studies indicate that hydraulic conductance is increased because the endothelium is damaged or missing. In the one case in which the endothelium was intact, the Lp was close to control values. The endothelium is known to present a considerable barrier to transarterial water transport.21 26 Vessels with focal lesions may have demonstrated a high Lp for the same reason. Although the endothelium was usually in place over the lesions, it was frequently vacuolated or showed widened interendothelial junctions. In addition, the endothelium was often missing in regions adjacent to the lesion that were included in the cannulated segment.
Our finding that hypercholesterolemia alone caused endothelial denudation and damage in five of the six femoral arteries examined was surprising. Other studies have concentrated on the effect of hypercholesterolemia on the structure of the aorta and coronary arteries. Rabbits fed a 2% cholesterol and 10% corn oil diet for 2 weeks did not show endothelial damage of aortas.40 Rabbits fed 1% cholesterol for 10 weeks did show some endothelial damage of coronary arteries in the form of vacuolated cells and widened intercell clefts.45 However, the damage was not as extensive as we observed in most femoral arteries from hypercholesterolemic animals in this study. Because we did not observe endothelial damage in the aortas of our hypercholesterolemic rabbits, it is possible that the endothelium of femoral arteries is particularly fragile. In fact, previous authors have shown that the femoral artery is more susceptible than the carotid artery to atherosclerosis.46 One of the first signs of endothelial dysfunction during hypercholesterolemia is the impaired ability of arteries to undergo endothelial-dependent relaxation.47 48 49 It is likely that hypercholesterolemia reduces the synthesis and/or availability of the endothelially derived vasodilator, NO, because exposure of vessels to the NO substrate, L-arginine, reverses this effect.50 In the femoral artery, if NO production is deliberately impaired by perfusing the vessel with the NO inhibitor, NG-nitro-L-arginine methyl ester, the endothelium becomes structurally damaged.51 This does not occur if similar experiments are performed on the aorta. Thus, the endothelial damage we observed in the femoral arteries of hypercholesterolemic rabbits may be initiated by the accompanying impaired release of NO.
The less dense media observed beneath lesions may also have contributed to the higher Lp. Vessels with lesions always showed similar Lps to hypercholesterolemic arteries without lesions, even though the intima and media were thicker in the former case. It is possible that the effect of increased wall thickness on Lp was canceled out by the effect of a less dense media beneath lesions. As pressure was increased, Lps of hypercholesterolemic vessels, with and without lesions, decreased and approached control values. This finding suggests that at high pressures the endothelium becomes less important as a barrier to transarterial water flux and that the media plays a more dominant role.
The endothelium is also important because it can act as a barrier to diffusion of material, for example, a locally delivered drug, from the arterial wall into the bloodstream. Immediately after angioplasty the affected region of the artery is denuded of endothelium.52 53 54 Therefore, this would not be an optimal time to deliver drugs into the arterial wall, intraluminally, to prevent restenosis. Endothelial regeneration is usually complete within a few weeks, and the endothelial cells have completely matured within 3 months.52 This would be a more appropriate time for drug delivery, because the intact endothelium would reduce the rate of diffusion of intimally delivered drug into the bloodstream and would reduce the convective water flux from the lumen to the outer regions of the artery wall. However, this type of approach would be successful only if the drug delivery system itself did not damage the endothelium.
To summarize, we made direct measurements of Lp on normal hypercholesterolemic and atherosclerotic arteries and used these data, together with information obtained from our ultrastructural studies, to determine Peclet numbers for large and small molecules. The ultrastructural component provides a structural basis with which to interpret measurements of Lp. This study demonstrates that convective flux may significantly affect the distribution of small and large molecules within the artery wall in normal and in atherosclerotic vessels. Both the local administration and subsequent redistribution of drugs to lesions after angioplasty will be affected. These measurements of hydraulic fluid flux provide a framework for further studies to evaluate the possibility of influencing drug residence time and distribution by modifying the transport properties of the drug.
|
Selected Abbreviations and Acronyms |
|---|
|
|
Acknowledgments |
|---|
This work was supported by Eli Lilly and Company, a Grant-in-Aid from the American Heart Association, Arizona Affiliate, and by National Heart, Lung, and Blood Institute grant HL-17421.
Received March 8, 1996; accepted May 7, 1997.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
1. Nobuyoshi M, Kimura T, Nosaka H, et al. Restenosis after successful percutaneous transluminal coronary angioplasty: serial angiographic follow-up of 229 patients. J Am Coll Cardiol. 1988;12:616–623.[Abstract]
2. Wilensky RL, March KL, Hathaway DR. Direct intraarterial wall injection of microparticles via a catheter: a potential drug delivery strategy following angioplasty. Am Heart J. 1991;122(suppl 4, pt 1):1136–1140.
3. Wolinsky H, Thung SN. Use of a perforated balloon catheter to deliver concentrated heparin into the wall of the normal canine artery. J Am Coll Cardiol. 1990;15:475–481.[Abstract]
4. Bier JD, Zalesky P, Li ST. A new bioabsorbable intravascular stent: in vitro assessment of hemodynamic and morphometric characteristics. J Interven Cardiol. 1992;5:187–194.[Medline] [Order article via Infotrieve]
5.
Herrmann HC, Okada SS. Inhibition of smooth muscle
cell proliferation and experimental angioplasty restenosis by
beta-cyclodextrin tetradecasulfate. Arterioscler Thromb. 1993;13:924–931.
6.
Edelman ER, Adams DA, Karnovsky MJ. Effect of
controlled adventitial heparin delivery on smooth muscle cell
proliferation following endothelial injury. Proc
Nat Acad Sci U S A. 1990;87:3773–3777.
7. Muller DW, Topol EJ, Abrams GD, Gallagher KP, Ellis SG. Intramural methotrexate therapy for the prevention of neointimal thickening after balloon angioplasty. J Am Coll Cardiol. 1992;20:460–466.[Abstract]
8. Wilensky RL, Gradus-Pizlo I, March KL. Efficacy of local intramural injection of colchicine in reducing restenosis following angioplasty in the atherosclerotic rabbit model. J Am Coll Cardiol. 1995;26:1549–1557.[Abstract]
9. Riessen R, Isner JM. Prospects for site-specific delivery of pharmacologic and molecular therapies. J Am Coll Cardiol. 1994;23:1234–1244.[Abstract]
10. Fram DB, Aretz T, Azrin MA. Localized intramural drug delivery during balloon angioplasty using hydrogel-coated balloons and pressure-augmented diffusion. J Am Coll Cardiol. 1994;23:1570–1577.[Abstract]
11. Colton CK, Schneiderman G, Ramirez CA, Smith KA, Lees RS, Stemerman MB. Labeled albumin transport into the normal and de-endothelialized rabbit aorta in vivo. In: Nerem RM, Guyton AC, eds. Hemodynamics and the Arterial Wall. Houston, Tex: University of Houston; 1980;42–46.
12. Tedgui A, Lever MJ, Caro CGC. Effect of bulk flow on the transport of 1#1. I-albumin across the rabbit thoracic aorta. Proceedings of the Third International Conference on Medicine and Biology, Compiegne, France. 1982. Abstract.
13.
Tedgui A, Lever MJ. The interaction of convection and
diffusion in the transport of 131I-albumin within
the media of the rabbit thoracic aorta. Circ Res. 1985;57:856–863.
14.
Tedgui A, Lever MJ. Effect of pressure and intimal
damage on 131I-albumin and
[14C]sucrose spaces in aorta. Am J
Physiol. 1987;253:H1530–H1539.
15. Truskey GA, Colton CK, Smith KA. Quantitative analysis of protein transport in the arterial wall. In: Schwartz CJ, Werthessen NT, Wolf S, eds. Structure and Function of the Circulation, vol 3. New York, NY: Plenum Press; 1981:287–355.
16.
Wilens SL, McCluskey RT. The permeability of excised
arteries and other tissues to serum lipid. Circ Res. 1954;2:175–182.
17.
Penn MS, Saidel GM, Chisolm GM. Vascular injury by
endotoxin: changes in macromolecular transport parameters
in rat aortas in vivo. Am J Physiol. 1992;262:H1563–H1571.
18.
Lovich MA, Edelman ER. Mechanisms of transmural heparin
transport in the rat abdominal aorta following local vascular delivery.
Circ Res. 1995;77:1143–1150.
19. Curry FRE. Mechanisms and thermodynamics of transcapillary exchange. In: Renkin EM, Michel CC, eds. Handbook of Physiology, sect 2, The Cardiovascular System, vol IV, Microcirculation, part 1. Bethesda, Md: American Physiology Society; 1984:309–374.
20.
Baldwin AL, Wilson LM, Simon BR. Effect of pressure on
aortic hydraulic conductance. Arterioscler Thromb. 1992;12:163–171.
21.
Baldwin AL, Wilson LM. Endothelium
increases medial hydraulic conductance of aorta, possibly by EDRF.
Am J Physiol. 1993;264:H26–H32.
22. Harrison RB, Massaro TA. Water flux through porcine aortic tissue due to a hydrostatic pressure gradient. Atherosclerosis. 1976;24:363–367.[Medline] [Order article via Infotrieve]
23. Lever MJ, Sharifi N. The effects of transmural pressure and perfusate albumin concentration on the hydraulic conductivity of the isolated rabbit common carotid artery. J Physiol. 1987;387;68. Abstract.
24. Tarbell JM, Lever MJ, Caro CG. The effect of varying albumin concentration on the hydraulic conductivity of the rabbit common carotid artery. Microvasc Res. 1988;35:204–220.[Medline] [Order article via Infotrieve]
25. Bergstrand L. Femoral and coronary atherosclerosis in patients with hyperlipidaemia: arteriographic findings correlated to clinical and biochemical parameters. Acta Radiol Suppl. 1994;35(suppl 392):I–27.
26. Cashin-Hemphill L. Femoral and coronary angiographic trials. Am J Cardiol. 1993;71:20B–25B.[Medline] [Order article via Infotrieve]
27. Smedby O, Hogman N, Nilsson S, Olsson AG, Walldius G. Two-dimensional tortuousity of the superficial femoral artery in early atherosclerosis. J Vasc Res. 1993;30:181–191.[Medline] [Order article via Infotrieve]
28. LeVeen RF, Wolf GL, Villanueva TG. New rabbit atherosclerosis model for the investigation of transluminal angioplasty. Invest Radiol. 1982;17:470–475.[Medline] [Order article via Infotrieve]
29.
Sarembock IJ, LaVeau PJ, Sigal SL, et al. Influence of
inflation pressure and balloon size on the development of intimal
hyperplasia after balloon angioplasty. Circulation. 1989;80:1029–1040.
30. Fry DL. Effect of pressure and stirring on in vitro aortic transmural 125I-albumin transport. Am J Physiol. 1983;245:H977–H991.
31. Penn MS, Saidel GM, Chisolm GM. Relative significance of endothelium and internal elastic lamina in regulating the entry of macromolecules into arteries in vivo. Circ Res. 1994,74:74–82.
32. Fry DL. Mass transport, atherogenesis and risk. Arteriosclerosis. 1987;7:88–100.[Abstract]
33. Maroudas A. Physical chemistry of articular cartilage and the intervertebral disc. In: Sokoloff L, ed. Joints and Synovial Fluid, vol II. New York, NY: Academic Press; 1980:240–241.
34. Winlove CP, Parker KH. Diffusion of macromolecules in hyaluronate gels. I. Development of methods and preliminary results. Biorheology. 1984;21:347–362.[Medline] [Order article via Infotrieve]
35. El-Kareh AW, Brownstein SL, Secomb TW. Effect of cell arrangement and interstitial volume fraction on the diffusivity of monoclonal antibodies in tissue. Biophys J. 1993;64:1638–1646.[Medline] [Order article via Infotrieve]
36. Bratzler RL, Schwartz ER. Convective and diffusive transport of labeled albumin through rabbit aortic tissue. Proc Annu Conf Eng Med Biol. 30th, Los Angeles, CA 1977; p318.
37.
Bratzler RL, Chisolm GM, Colton CK, Smith KA,
Zilversmit DB, Lees RS. The distribution of labeled albumin
across the rabbit thoracic aorta in vivo. Circ Res. 1977;40:182–190.
38.
Tedgui A, Lever MJ. Filtration through damaged and
undamaged rabbit thoracic aorta. Am J Physiol. 1984;247:H784–H791.
39. Crone CC, Levitt DG. Capillary permeability to small solutes. In: Renkin EM, Michel CC, eds. Handbook of Physiology, sect 2, The Cardiovascular System, vol IV, Microcirculation, part 1. Bethesda, Md: American Physiological Society; 1984.
40.
Buja LM, Kita T, Goldstein JL, Watanabe Y, Brown MS.
Cellular pathology of progressive atherosclerosis in
the WHHL rabbit: an animal model of familial
hypercholesterolemia.
Arteriosclerosis. 1983;3:87–101.
41. Bocan TM, Mueller SB, Uhlendorf PD, Ferguson E, Newton RS. Dietary and mechanically induced rabbit iliac-femoral atherosclerotic lesions: a chemical and morphological evaluation. Exp Molec Pathol. 1991;54:201–277.[Medline] [Order article via Infotrieve]
42.
Wu NZ, Da D, Rudoll TL, Needham D, Whorton AR, Dewhurst
MW. Increased microvascular permeability contributes to preferential
accumulation of stealth liposomes in tumor tissue. Cancer
Res. 1993;53:3765–3770.
43.
Barker SGE, Talbert A, Cottan S, Baskerville PA, Martin
JF. Arterial intimal hyperplasia after occlusion of the
adventitial vasa vasorum in the pig. Arterioscler Thromb. 1993;13:70–77.
44. Martin JF, Booth RFG, Moncada S. Arterial wall hypoxia following thrombosis of the vasa vasorum is an initial lesion in atherosclerosis. Eur J Clin Invest. 1991;21:355–359.[Medline] [Order article via Infotrieve]
45. Wang BY, Singer AH, Tsao PS, Dexler H, Kosek J, Cooke JP. Dietary arginine prevents atherogenesis in the coronary artery of the hypercholesterolemic rabbit. J Am Coll Cardiol. 1994;23:452–458.[Abstract]
46.
Wendelhag I, Wiklund J. Atherosclerotic changes in the
femoral and carotid arteries in familial
hypercholesterolemia: ultrasonic assessment of
intima–media thickness and plaque occurrence. Arterioscler
Thromb. 1993;13:1404–1411.
47. Andrews HE, Bruckdorfer KR, Dunn RC, Jacobs M. Low-density lipoproteins inhibit endothelial-dependent relaxation in rabbit aorta. Nature. 1987;327:237–239.[Medline] [Order article via Infotrieve]
48.
Cohen RA, Zitnay KM, Haudenschild CC, Cunningham LD.
Loss of selective endothelial cell vasoactive functions
caused by hypercholesterolemia in pig
coronary arteries. Circ Res. 1988;63:903–910.
49.
Osborne JA, Siegman MJ, Sedar AW, Mooers SU, Lefer AM.
Lack of endothelial-dependent relaxation in
coronary resistance arteries of cholesterol-fed
rabbits. Am J Physiol. 1989;256:C591–C597.
50. Cooke JP, Dzau J, Creager A. Endothelial dysfunction in hypercholesterolemia is corrected by L-arginine. Basic Res Cardiol. 1991;86(suppl 2):173–181.
51. Baldwin AL, Wilson LM. Nitric oxide affects hydraulic conductance of different arteries in different ways. FASEB J. 1993;7:A491. Abstract.
52. Zollikofer CL, Salomonowitz E, Sibley R, et al. Transluminal angioplasty evaluated by electron microscopy. Intervent Radiol. 1984;153:369–374.
53.
Ohkawa M, Fujiwara N, Tanabe M, et al. Cerebral
vasospastic vessels: histological changes after
percutaneous transluminal angioplasty.
Radiology. 1996;198:179–184.
54. Chan PD, Findlay JM, Vollrath B, et al. Pharmacological and morphological effects of in vitro transluminal balloon angioplasty on normal and vasospastic canine basilar arteries. J Neurosurg. 1995;83:522–530.[Medline] [Order article via Infotrieve]
This article has been cited by other articles:
 |
|
 |
|
  Z.-D. Shi, X.-Y. Ji, H. Qazi, and J. M. Tarbell Interstitial flow promotes vascular fibroblast, myofibroblast, and smooth muscle cell motility in 3-D collagen I via upregulation of MMP-1 Am J Physiol Heart Circ Physiol, October 1, 2009; 297(4): H1225 - H1234. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. J. M. Roberts, A. C. Chapman, and M. J. Cipolla PPAR-{gamma} agonist rosiglitazone reverses increased cerebral venous hydraulic conductivity during hypertension Am J Physiol Heart Circ Physiol, October 1, 2009; 297(4): H1347 - H1353. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.-B. Michel, O. Thaunat, X. Houard, O. Meilhac, G. Caligiuri, and A. Nicoletti Topological Determinants and Consequences of Adventitial Responses to Arterial Wall Injury Arterioscler Thromb Vasc Biol, June 1, 2007; 27(6): 1259 - 1268. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. P. Alberding, A. L. Baldwin, J. K. Barton, and E. Wiley Effects of pulsation frequency and endothelial integrity on enhanced arterial transmural filtration produced by pulsatile pressure Am J Physiol Heart Circ Physiol, August 1, 2005; 289(2): H931 - H937. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Tada and J. M. Tarbell Internal elastic lamina affects the distribution of macromolecules in the arterial wall: a computational study Am J Physiol Heart Circ Physiol, August 1, 2004; 287(2): H905 - H913. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. P. Alberding, A. L. Baldwin, J. K. Barton, and E. Wiley Onset of pulsatile pressure causes transiently increased filtration through artery wall Am J Physiol Heart Circ Physiol, May 1, 2004; 286(5): H1827 - H1835. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C.-W. Hwang and E. R. Edelman Arterial Ultrastructure Influences Transport of Locally Delivered Drugs Circ. Res., April 19, 2002; 90(7): 826 - 832. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C.-W. Hwang and E. R. Edelman Arterial Ultrastructure Influences Transport of Locally Delivered Drugs Circ. Res., April 19, 2002; 90(7): 826 - 832. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||