© 1998 American Heart Association, Inc.
Original Contributions |
Adsorption of Fibrinogen to Droplets of Liquid Hydrophobic Phases
Functionality of the Bound Protein and Biological Implications
From the Department of Pathology and Laboratory Medicine, University of Cincinnati Medical Center, Cincinnati, Ohio.
Correspondence to Gregory S. Retzinger, MD, PhD, Department of Pathology and Laboratory Medicine, University of Cincinnati Medical Center, Cincinnati OH 45267-0529.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Abstract—Fibrinogen adsorbs spontaneously from aqueous media containing that protein to droplets of liquid hydrophobic phases dispersed in those same media. Examples of such phases include mineral oils, straight-chain hydrocarbons, and various plant- and animal-derived oils. Lecithin preexisting on the surface of oil droplets reduces significantly the amount of fibrinogen that can otherwise bind to them. When bound, fibrinogen remains active in the classic sense of fibrin gelation. As a consequence, oil droplets coated with fibrinogen can participate in a host of biologically important adhesive processes in which the protein would be expected to participate. Certain polyanions, eg, heparin, pentosan polysulfate, dextran sulfate, and suramin, bind to adsorbed fibrin(ogen) and prevent thrombin-dependent adhesion of fibrinogen-coated surfaces. Thus, these polyanions can be used to prevent adhesion between fibrin(ogen)-coated oil droplets and other fibrin(ogen)-coated surfaces. Potential practical applications and biological implications of these phenomena are presented and discussed.
Key Words: fibrinogen • atherosclerosis • oils • polyanions • drug delivery
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
We have explored the operation and biological processivity of fibrin(ogen) at interfaces1 2 because we believe such exploration will yield insight relevant to a mechanistic understanding of a host of fibrin(ogen)-dependent physiological and pathophysiological processes.3 4 5 6 7 Earlier, we showed that fibrinogen adsorbs avidly from aqueous media containing the protein to microscopic hydrophobic beads dispersed in those media.1 Phospholipids or other amphiphiles preexisting on hydrophobic beads prevent the adsorption of fibrinogen and other proteins by rendering the surface of the beads hydrophilic.8 9 When bound to beads, fibrinogen remains functional in the classic sense of fibrin gelation.1 As a consequence, fibrinogen-coated beads aggregate when stirred in the presence of thrombin, and they can either be incorporated into developing fibrin clots or adhere to other fibrin-coated surfaces they contact.1 2 Importantly, heparin binds to adsorbed fibrin(ogen) and prevents adhesion between fibrin-coated surfaces.10
During the course of our investigations, we "discovered" that fibrinogen, like many other proteins, adsorbs in stable fashion from aqueous media to droplets of liquid hydrophobic phases. As is the case when fibrinogen is bound to solid polymeric beads, that protein is functional on the surface of oil droplets. Thus, like fibrinogen-coated beads, fibrinogen-coated oil droplets can be incorporated into developing fibrin clots, and they can adhere to other fibrin-coated surfaces that they contact.
In this report, we characterize the binding of fibrinogen to microscopic droplets of several liquid hydrophobic phases. We demonstrate that the bound protein can mediate adhesion of these droplets to other fibrin-coated surfaces, and this adhesion can be prevented by certain polyanions. The relevance of these findings both to the formulation of vehicles for targeted delivery of drugs and to our understanding of pathophysiological processes involving deposition of hydrophobic phases, eg, atherosclerosis, is discussed.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Reagents and Chemicals
Human fibrinogen FIB3 was from Enzyme Research Laboratories. Before use, the buffer composition of the commercial fibrinogen was changed by gel permeation chromatography using Sephadex G-25 (Pharmacia) as the matrix material and either 0.02 mol/L Tris-HCl, pH 7.40, or 0.01 mol/L Na2HPO4 and NaH2PO4, pH 7.40, containing 0.145 mol/L NaCl (PBS) as the eluent. This fibrinogen was then divided into aliquots and stored frozen at -20°C until use. Before use, a frozen aliquot of fibrinogen solution was thawed to room temperature, diluted in buffer as appropriate, and then heated to 37°C to dissolve any residual cold-precipitable fibrinogen. Fatty acid–free BSA was from ICN. For some experiments, fibrinogen was uniformly labeled using Na125I from Amersham and Iodo-Gen from Pierce. Egg yolk L-
-lecithin was from
Avanti Polar Lipids. Drakeol 6-VR and Drakeol 32, white mineral oils
containing paraffin and naphthene hydrocarbons ranging from 
18 to 36
carbon atoms, were from Penreco.
[1-14C]Dodecane of specific activity 151.7
MBq/mmol was from Sigma Chemical Co. Dodecane, olive oil, safflower
oil, soybean oil, triolein, cholesteryl oleate, squalene, squalane,
ristocetin sulfate, human thrombin (>3000 NIH U/mg), hirudin (3000
U/mg), the sodium salt of pentosan polysulfate
[Mr(ave) 3800], and the acetate salt
of the tetrapeptide Gly-Pro-Arg-Pro (GPRP) were also from Sigma. The
sodium salts of porcine mucosa heparins of average molecular weights
14 250, 3700, and 5100 were from Calbiochem. The sodium salt of
unfractionated, pharmaceutical-grade porcine mucosa heparin, 5000 U/mL,
was from Elkins-Sinn. Sodium suramin was the generous gift of the Drug
Service of the Centers for Disease Control and Prevention, Atlanta, Ga.
The sodium salt of dextran sulfate of
Mr(ave) 8000 was from Sigma, and the
sodium salt of dextran sulfate of Mr(ave)
40 000 (range, 37 000 to 43 000) was from Chemical Dynamics.
Before its use, water was first deionized, and then distilled in an
all-glass apparatus. All organic solvents were of a grade
suitable for high-performance liquid
chromatography. All other chemicals were of the highest
quality available commercially. Fresh human plasma and fresh serum were prepared from the blood of healthy donors. For plasma, blood was drawn directly into evacuated, siliconized glass tubes (Becton Dickinson) containing sodium citrate, yielding a final concentration of the anticoagulant of 0.013 mol/L. For serum, blood was drawn directly into evacuated glass tubes (Becton Dickinson) where it was left undisturbed for 2 hours while it clotted at room temperature. Cells were removed from anticoagulated blood and from clotted blood by centrifugation at 1500g for 15 minutes, and the corresponding plasma or serum was then aspirated and, unless specified otherwise, used immediately thereafter. As necessary, the fibrinogen concentration of citrated plasma samples was determined using the method of Clauss.11 Inactivation of proteinases in citrated plasma was accomplished by heating the plasma sample at 60°C for 30 minutes. With heating, a fibrin(ogen)-rich coagulum developed in the plasma. After the sample was cooled to room temperature, this coagulum was removed. Lyophilized fibrinogen was then added to the medium, yielding a fibrinogen concentration of 2.9x10-6 mol/L.
Emulsification of Liquid Hydrophobic Phases
High-pressure extrusion was used to prepare emulsions. For this
purpose, either 140 or 220 µL of a liquid hydrophobic phase was first
added to a clean, 12x75-mm glass tube. To a tube containing 140 µL
of oil was then added 3.5 mL of PBS; to a tube containing 220 µL of
oil was then added 5.0 mL of PBS. After brief agitation, an oil-water
mixture was passed 5 times under high pressure (15 000 psi) through
the aperture of an automated homogenizer
(EmulsiFlex-20,000-B3).
Isolation of Droplets of Liquid Hydrophobic Phases
By centrifugation, droplets were separated from
the aqueous medium in which they were prepared. After the droplet-free
medium was aspirated, droplets were washed 3 times with 2.0 mL of fresh
buffer each time. For all droplets other than those containing
lecithin, the relative centrifugal force and duration of the initial
centrifugation were 1500g and 20 minutes,
respectively. Subsequently, washes of lecithin-free droplets were
performed by centrifugation at 1500g for 5.0
minutes. To isolate droplets that had been emulsified in the presence
of lecithin, the entire emulsion was first transferred to a clean,
round-bottom, glass tube. The emulsion was then centrifuged at
9400g for 60 minutes. After the droplet-free medium was
removed, droplets were washed 3 times with 2.0 mL of fresh buffer each
time. Each of these washes involved centrifugation at
9400g for 20 minutes. After their isolation and wash,
droplets were redispersed as necessary to an apparent absorbance of 1.0
at 500 nm with a cuvette of 1.0-cm-path length. The medium for this
purpose was 0.02 mol/L Tris-HCl, pH 7.40, containing 1.0 mg/mL
BSA.1
Visualization of Droplets of Liquid Hydrophobic Phases
Light microscopy was used to visualize both monodisperse
droplets and aggregates of droplets. For this purpose, a small volume
of an emulsion was placed onto a microscope slide and then overlaid
with a coverslip. Permanent records of microscopic views of
droplets were obtained by photomicroscopy.
Sizing of Droplets of Liquid Hydrophobic Phases
For 1 set of experiments, the size of droplets of various oils
was determined using a laser diffraction particle-size analyzer
(LS 230, Coulter). Refractive indices of 1.47 (olive oil) and 1.33
(water) were used when fitting the light-scattering data to the
instrument's preprogrammed sizing algorithm.
Fibrinogen Binding Studies
We assessed the binding of 125I-fibrinogen
from buffer to emulsified droplets of several liquid hydrophobic
phases. For this purpose, 1.0 mL of PBS containing 2.80 mg of
125I-fibrinogen was added to 4.0 mL of PBS
containing 176 µL of emulsified oil droplets. After
centrifugation of the dispersion, the radioactivity
associated with the droplet-free reaction medium was measured. When
droplets contained lecithin, the medium was cleared by
centrifugation at 9400g for 1.0 to 6.0
hours. When droplets did not contain lecithin, the medium was cleared
by centrifugation at 1500g for 20 minutes.
The difference between the total radioactivity of a sample and that
remaining in the medium after separation of the droplets yielded the
radioactivity, hence fibrinogen, bound to the oil droplets.
The binding of fibrinogen from a citrated plasmalike medium to olive oil droplets was also assessed. We used for these studies citrated plasma that had been supplemented with various amounts of 125I-fibrinogen. The fibrinogen concentration of the citrated virgin plasma was 7.1x10-6 mol/L. Three milliliters of 125I-fibrinogen–supplemented plasma was added to an equivalent volume of PBS containing 132 µL of freshly prepared olive oil droplets. After 30 minutes, this dispersion was mixed with an equivalent volume of aqueous sucrose, 78% (wt/vol), and the droplets were then floated by centrifugation for 3 hours at 7000g. The resulting "cream" layer was washed twice with 10 mL of aqueous sucrose and centrifuged for 1.0 hour at 7000g. The radioactivity associated with the washed cream layer was then measured. Using the known concentration of endogenous fibrinogen in the plasma, the specific activity of the 125I-fibrinogen supplementing that medium, and the radioactivity associated with the oil droplets, we determined the fibrinogen bound to the olive oil droplets.
The time dependence of the association of 125I-fibrinogen with oil droplets dispersed in either buffer or heat-treated, citrated plasma was assessed as follows. Two milliliters of oil droplets that had been coated with 125I-fibrinogen was dispersed, with continuous stirring, into 20 mL of 1 of the 2 aqueous phases. After various lengths of time, 1.0-mL aliquots of the stirred dispersion were removed, and the oil and aqueous phases of these aliquots were separated by centrifugation. Subsequently, the radioactivities associated with the oil and aqueous phases were determined.
Aggregation of Fibrin-Coated Oil Droplets and Dissociation of
Aggregates of Fibrin-Coated Oil Droplets
When stirred in the presence of thrombin, fibrinogen-coated oil
droplets aggregate, a consequence of interparticle fibrin
formation.1 Thus, thrombin-inducible aggregation
of such droplets is a convenient measure of the functionality of bound
fibrinogen. Several methods were used to monitor both aggregation of
fibrin-coated oil droplets and dissociation of aggregates of
fibrin-coated droplets. One method involved simply applying 200 µL of
an emulsion containing fibrinogen-coated oil droplets to a smooth
surface and then mixing into this emulsion an amount of thrombin, 0.5
NIH U in 20 µL, with a wooden spatula. With stirring, droplets coated
with a dense layer of fibrinogen aggregate within seconds after adding
the enzyme, a process obvious to the naked eye. With the addition of
certain substances, aggregates of fibrin-coated particles dissociate,
yielding monodisperse particles.1 10 This
dissociation was monitored visually. Permanent records of these
phenomena were obtained photographically.
Another method used to monitor both the aggregation of droplets and the dissociation of droplet aggregates involved a platelet aggregometer.1 A typical aggregation assay was performed at room temperature as follows. A 0.5-mL dispersion of droplets in 0.02 mol/L Tris-HCl, pH 7.40, containing 1.0 mg/mL BSA was added to a cylindrical, glass, sample cuvette (ID, 6 mm). The apparent absorbance of test dispersions was 1.0 at 500 nm when a cuvette of 1.0-cm-path length was used. As reference material, a dispersion of polystyrene beads of diameter 0.945±0.0064 µm (Seradyn) was used. The apparent absorbance of this reference material was 0.5 at 500 nm when a cuvette of 1.0-cm-path length was used. Once the baseline signal of the stirred (1000 rpm) test sample was established, 20 µL of an aqueous solution containing 0.5 NIH U thrombin was added to the reaction cuvette. The relative absorbance of the sample as a function of time after addition of the enzyme was then recorded. For the purpose of this report, a value of 1.0 was assigned arbitrarily to the maximal change of the aggregometry signal that occurred after the addition of 0.5 NIH U thrombin to a dispersion of fibrinogen-coated droplets of mineral oil (Drakeol 32).
Just as particle aggregation can be monitored turbidimetrically with a platelet aggregometer, so too, can dissociation of aggregates of particles be monitored with an aggregometer.1 10 Dissociation of aggregates of fibrin-coated oil droplets was assessed by using the same volumes, conditions, and instrument parameters described for aggregation assays, except that fibrinogen-coated droplets in the sample cuvette first had to be aggregated. For this purpose, 20 µL of a buffered aqueous solution of thrombin, 0.5 NIH U, was added to stirred, monodisperse, fibrinogen-coated droplets. Within 15.0 minutes after addition of the enzyme, droplets had aggregated maximally, and the resulting aggregates could be used to assess aggregate dissociation. Test reagents were added in 20 µL of water to the reaction cuvette, and the state of aggregation of the droplets was followed turbidimetrically as a function of time after the reagent was added.
Studies Involving Solution-Phase Fibrin Clots
Round-bottom, 12x75-mm glass tubes were used as reaction
vessels to explore interactions of variously coated droplets of olive
oil and [1-14C]dodecane, 95/5 vol/vol, with
solution-phase fibrin clots. Ten microliters of buffer containing 0.25
NIH U thrombin was added to 200 µL of PBS containing
1.8x10-5 mol/L fibrinogen. Solution-phase clots
formed in this fashion12 were then left
undisturbed for 60 minutes. For 1 experiment, 2.0 IU hirudin in 10 µL
of buffer was placed on each of several clots. The exposed, uppermost
surface of each clot, 1.13 cm2, was overlaid with
100 µL of buffer containing 10 µL of droplets of a particular olive
oil/[1-14C]dodecane emulsion. The contents of
the reaction vessel were then agitated for 6 seconds with a vortex
apparatus. To remove unbound oil droplets, a clot was
"washed'" twice with 1.0 mL of fresh buffer each time and gentle
agitation. After each wash, the unbound oil droplets were decanted. In
the case of clots that had been overlaid with fibrinogen-coated oil
droplets, an attempt was made to dissociate from the clots any bound
droplets by using 1 of several solutions. These solutions included
(1) 100 µL of buffer containing 0.25 IU plasmin, (2) 100 µL of
buffer containing 5.0x10-3 mol/L GPRP, and (3)
100 µL of buffer containing 100 USP U unfractionated,
pharmaceutical-grade heparin. Both untreated clots and clots that had
been treated with 1 of the test solutions were then washed another 2
times as described above using 1.0 mL of fresh buffer for each wash.
After removing from the surface those oil droplets that could be
liberated, the remaining clot-associated radioactivity was determined.
Before this radioactivity was measured, clots were dissolved by the
addition of 0.5 mL of 8.0 mol/L urea. The entire volume of solubilized
material (0.7 mL) was then added to 5.0 mL of scintillation cocktail
(Ultima Gold XR, Packard) for measurement. Each test sample was
prepared in quadruplicate.
Analysis of Data
Concentration-dependent data were paired with the corresponding
concentrations and then fit to an appropriate equation described in the
text. The best values for the parameters of an equation
were determined using the paired data and a nonlinear least-squares
regression method.13
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Fibrinogen Binds to and Stabilizes Droplets of Emulsified Liquid Hydrophobic Phases
Surfactants can be used to stabilize emulsions consisting of 2 partially miscible or immiscible phases such as mineral oil and water (Figure 1
). By occupying the interface
between the continuous and discontinuous phases, the surfactant lowers
the interfacial tension of the system, thus reducing the drive toward
self-coalescence of the individual phases. As shown on the right in
Figure 1, fibrinogen, an amphiphilic protein and as such, a surfactant,
effectively stabilizes microscopic droplets of emulsified mineral oil.
In the absence of fibrinogen or other added surfactant, the water and
oil of the emulsion separate back into discrete, continuous phases
within 20 minutes. In contrast, in the presence of fibrinogen, the oil
droplets remains monodisperse, eventually forming cream layers over a
time course ranging in oil-dependent fashion from hours to days. Liquid
hydrophobic phases that we have found can be emulsified in the presence
of fibrinogen with the resulting droplets remaining monodisperse for
various extended periods include, in addition to mineral oils, olive
oil, olive oil/cholesteryl oleate mixtures, triolein,
triolein/cholesteryl oleate mixtures, soybean oil, safflower oil,
squalene, squalane, and dodecane. The amounts of fibrinogen that bind
to various oil droplets prepared by high-pressure extrusion and then
exposed to the protein according to our standard protocol are shown in
Figure 2.
|
|
Lecithin Prevents the Binding of Fibrinogen to Droplets of Liquid
Hydrophobic Phases; Cholesteryl Oleate Does Not
Earlier, we showed that lecithin preexisting at rather high
packing density, ie, 0.014 molecule/Å2 (70
Å2 /molecule), on the surface of hydrophobic
beads prevents the binding of fibrinogen to those
beads.8 In contrast, cholesteryl oleate coated to
a similar nominal packing density does not influence the rather
significant amount of fibrinogen that otherwise binds to the surface of
the beads. We wondered whether lecithin or cholesteryl oleate, when
included in the formulation of an oil emulsion, would affect the
association of fibrinogen with oil droplets in the same way that they
affect the binding of the protein to hydrophobic beads.
To address this issue, we first prepared emulsions containing olive oil
and various quantities of either lecithin or cholesteryl oleate. Next,
we used laser diffraction to size droplets prepared from olive oil
alone, olive oil and 1.0 mol% lecithin, and olive oil and 2.0 mol%
cholesteryl oleate. We found that all of the measured droplets had
virtually the same size distribution, a mean diameter of
3.5±1.8 µm. We then quantified the binding of
125I-fibrinogen to the droplets. From the data of
Figure 2
and the diameter of the droplets, we determined that the
fibrinogen on droplets of virgin olive oil likely exists as a
monomolecular layer.2 14 15 Furthermore, the
packing density of the fibrinogen, 5.8x10-5
molecules/Å2 (ie, 17 300
Å2 /molecule), is consistent with the
molecules of the protein being oriented with their long axes more
normal than tangential to the interface.2 15 We
found next that no amount of cholesteryl oleate used for these
experiments (up to 4 mol% of the lipid of an emulsion) reduced the
quantity of fibrinogen that associated with droplets of otherwise
virgin olive oil (Figure 3). This result
is in keeping with both the oil solubility and the marginal
amphiphilicity of cholesterol
ester.8 16 17 18 As shown in the same figure,
however, as little as 1.0 mol% lecithin reduced to nearly zero the
amount of fibrinogen that otherwise would bind to olive oil droplets.
Indeed, from the volume of oil used for the experiment (176 µL); the
mean diameter of the spherical droplets; and the weight average
molecular weights of lecithin (760) and triolein (885, "olive
oil"), one calculates that the point of intersection of the 2
extrapolated linear regions of the figure corresponds to a nominal
molecular area for each lecithin molecule of 43
Å.2 This nominal molecular area is in good
agreement with that of lecithin maximally packed at the air-water
interface, 75 Å.2 19 The difference between
these 2 areas may be more apparent than real because, in the case of
oil droplets, the phospholipid partitions between the bulk and surface
phases of the oil.17 Taken together, these data
are eminently consistent with the proposal that fibrinogen is
excluded from the surface of the droplets when that surface is rendered
hydrophilic as a consequence of occupancy by "tightly packed"
phospholipid.8
|
) reduces the binding of fibrinogen to droplets of
otherwise virgin olive oil; cholesteryl oleate (
) does not. The
straight-line regions of data that relate to lecithin represent
an empiric fit of the data to a titration curve. The arrow indicates
lecithin content at the intersection of the extrapolated straight-line
fits. All data points represent mean±SD of duplicate
determinations.
Fibrinogen Bound to Droplets of Liquid Hydrophobic Phases Is
Functional
Fibrinogen-coated droplets of liquid hydrophobic phases can be
isolated and then redispersed in fresh aqueous medium containing either
no protein or proteins other than fibrinogen, eg, BSA (Figure 4A). The fibrinogen bound to oil droplets
is functional, as demonstrated by the macroscopic aggregation of
droplets when they are stirred in the presence of thrombin (Figure 4B).
As expected, thrombin-induced aggregation of fibrinogen-coated droplets
is inhibited by hirudin, a rapid and potent inhibitor of
this enzyme (Figure 4C). Photomicrographs of monodisperse,
fibrinogen-coated mineral oil droplets and aggregates of fibrin-coated
mineral oil droplets are shown in Figure 5.
|
|
Droplet aggregation and related phenomena can be monitored conveniently
using either a platelet aggregometer or another photometric device.
As shown in Figure 6, the apparent
absorbance of a stirred dispersion of fibrinogen-coated oil droplets
decreases rapidly after the addition of thrombin to the dispersion.
This decrease in absorbance corresponds to the aggregation of droplets,
a consequence of interparticle fibrin
dimerization.1 By inhibiting thrombin, hirudin
prevents aggregation. Thus, just as fibrin(ogen) binds to and remains
operational on solid, microscopic, hydrophobic
beads,1 so too, does fibrinogen bind to and
remain operational on droplets of liquid hydrophobic phases.
|
We assessed next whether various measures would liberate fibrin-coated
oil droplets from the surface of a solution-phase fibrin
clot.12 For this purpose, we first prepared olive
oil droplets containing 0.5 mol%
[1-14C]dodecane. These radiolabeled droplets
were coated with fibrinogen, washed, and concentrated. An aliquot of
the concentrated, fibrinogen-coated droplets was then overlaid onto the
exposed surface of a uniform, thrombin-containing, fibrin clot, and the
remaining radioactivity associated with the clot after 1 of several
treatments was determined (Figure 7). As
expected, hirudin coadministered with fibrinogen-coated oil droplets
reduced significantly the association of droplets with the clot,
indicating that adherence of the droplets to the clot is likely a
consequence of noncovalent interactions between the fibrin of the clot
and the fibrin generated on the droplets. Plasmin, GPRP, and
unfractionated, pharmaceutical-grade heparin each effectively dislodge
droplets bound to the surface of clots. The mechanism by which plasmin
liberates droplets undoubtedly involves digestion of fibrin existing
between the droplets and the clot surface, because
plasminogen does not free fibrin-coated particles from the
surface of a solution-phase clot.12
Heparin10 and GPRP,20 21 on
the other hand, likely interfere with noncovalent interactions between
fibrin on the droplets and the fibrin at the clot surface. We conclude
from these experiments that the functionality of fibrin(ogen) bound to
oil droplets is similar in all measured respects to that of
fibrin(ogen) in solution.
|
Ristocetin Flocculates Fibrinogen-Coated Oil Droplets
Ristocetin dimers flocculate fibrinogen, a consequence of
complexation of the bifunctional dimers with certain ß-turns of the
protein.22 As shown in Figure 8, dimeric ristocetin also flocculates
fibrinogen-coated oil droplets. Ristocetin dimers do not, however,
flocculate droplets coated with an irrelevant protein, ie, BSA. Thus,
as assessed using ristocetin as a structural probe, fibrinogen retains
important solution-phase features when it is bound to oil droplets.
|
Fibrinogen Is Stable on Droplets of Liquid Hydrophobic
Phases
Fibrinogen dissociates rather slowly from oil droplets dispersed
in buffer: only 4% of the radioactivity bound initially to
saturation on droplets of either mineral oil (Drakeol 32) or olive oil
is liberated from the droplets when they are incubated in buffer for 48
hours. In the case of olive oil, this holds true even when the aqueous
phase is heat-treated, fibrinogen-supplemented plasma: 95% of the
radioactivity remains associated with the droplets after 48 hours. Such
results are consistent with the notion that the binding of
fibrinogen to hydrophobic surfaces is essentially irreversible, a
phenomenon due at least in part to interfacial "denaturation" of
the protein.1 8 14 15 The protein is not so
denatured as to lose function, however. As shown in Figure 9, the thrombin-inducible aggregation of
fibrinogen-coated olive oil droplets changes little, if any, after
incubation for 24 hours in citrated plasma.
|
Fibrinogen Adsorbs From Plasma to Droplets of Liquid Hydrophobic
Phases and Is Functional
Fibrinogen adsorbs rapidly and in relatively large quantity from
blood to solid, hydrophobic surfaces in contact with that
medium.4 23 For this reason, we assessed next
whether fibrinogen would adsorb from citrated plasma to oil droplets.
Such an assessment seemed apropos, because biologically relevant lipid
particles in vivo, ie, lipoproteins, are bathed continuously in a
fibrinogen-rich medium, and fibrinogen appears to contribute to the
initiation, development, and growth of atherosclerotic
plaques.3 24 25
As shown in Figure 10, olive oil
droplets exposed to fresh, citrated plasma but not droplets exposed to
fresh serum aggregate in the presence of thrombin, indicating that
fibrinogen adsorbs from plasma to the droplets and is functional. As
shown in Figure 11, the rate and extent
of aggregation of plasma-exposed oil droplets depend in turn on the
concentration of fibrinogen in the plasma. Using
125I-fibrinogen as a tracer, we found that the
quantity of fibrinogen adsorbed to droplets incubated in plasma
containing 7.1x10-6 mol/L (241 mg%) fibrinogen
was 0.40 mg/mL oil, and the quantity of fibrinogen bound to droplets
from the plasma containing 13.2x10-6 mol/L (450
mg%) fibrinogen was 0.87 mg/mL oil. Thus, while plasma proteins, in
addition to fibrinogen, must contribute to the final layer of protein
adsorbed from plasma to oil droplets, the fibrinogen that does adsorb
is functional, and its interfacial concentration increases with
increasing plasma concentration of the protein.
|
|
Heparins and Other Polyanions Prevent the Mutual Adhesion of
Fibrin-Coated Surfaces
Earlier, we demonstrated that heparins and related polyanions, in
the absence of any cofactor, bind with high affinity to fibrin(ogen)
adsorbed to solid, polymeric beads.10 As
consequences of this binding, the mutual adhesion of fibrinogen-coated
beads that otherwise occurs in the presence of thrombin can be
prevented, and preexisting aggregates of fibrin-coated beads can be
dissociated. Discovery of these phenomena led us to suggest that the
binding of heparin to adsorbed fibrin(ogen) might serve some general,
cofactor-independent, "antiadhesive"
function.10
Believing,10 as do
others,26 27 28 29 that the direct interaction of
heparin with fibrin(ogen) contributes in vivo to the anticoagulant
activity of the mucopolysaccharide, we tested whether heparins
and other polyanions prevent thrombin-inducible aggregation of
fibrinogen-coated droplets of mineral oil. For these studies, we used
unfractionated heparin, low-molecular-weight heparins, pentosan
polysulfate, suramin, and dextran sulfates. As shown in Figure 12, all of the materials tested reduced
in a dose-dependent fashion the maximal rate of aggregation of
droplets. For all but dextran sulfate of
Mr(ave) 40 000, the reduction in rate
as a function of polyanion concentration obeys well the relationship
Vobs=Vmax/(1+ P/Kd),
where Vmax is the maximal rate of
aggregation in the absence of polyanion,
Vobs is the observed rate of aggregation in
the presence of polyanion, P is the molarity of the
polyanion, and Kd is the equilibrium
dissociation constant of the complex. The polyanions fall into 4
distinct groups. Unfractionated heparin
[Mr(ave) 14 250] is by far the most
potent inhibitor of the series, followed by, in decreasing
order of potency, the group consisting of the sulfated
polysaccharides of low molecular weight [ie,
Mr(ave)'s between 4000 and 8000], suramin
[Mr=1429], and high-molecular-weight
dextran sulfate [Mr(ave) 40 000]. For
the heparins and dextran sulfates, these results are the same as those
obtained using fibrin(ogen)-coated, solid, polymeric
beads.10 The apparent dependence on the length of
the dextran sulfate polymer compared with that of heparin and the
relative impotency of suramin with respect to unfractionated heparin
probably indicate some fundamental structural requirement for complex
formation.
|
, heparin,
Mr(ave)=5100,
Kd=16.21±6.55 nmol/L; 
, dextran sulfate,
Mr(ave)=8000,
Kd=23.78±3.47 nmol/L; 
, heparin,
Mr(ave)=3700,
Kd=24.42±1.18 nmol/L; 
, suramin,
Mr=1429,
Kd=10 040±1668 nmol/L; and 
, dextran
sulfate, Mr(ave)=40 000. See Reference 10
and text for details. |
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
We have demonstrated that fibrinogen adsorbs spontaneously from aqueous media containing that protein to droplets of liquid hydrophobic phases dispersed in those same media. Lecithin preexisting on the surface of oil droplets reduces significantly the amount of fibrinogen that can otherwise bind to the droplets. When bound, fibrinogen is stable and remains active in the classic sense of fibrin gelation. As a consequence, oil droplets coated with fibrinogen can participate in a host of biologically important adhesive processes in which the protein would otherwise be expected to participate. Certain polyanions appear to bind to adsorbed fibrin(ogen) and prevent thrombin-dependent adhesion of fibrin-coated surfaces. What follows is a discussion of potential practical applications and theoretical implications of these findings.
Because so many pathological lesions, eg, tumors, granulomas, bacterial abscesses, atherosclerotic plaques—indeed, virtually all sites of inflammation—have a thrombotic component,30 31 32 it would seem worthwhile generally to develop vehicles to deliver therapeutic drugs or imaging agents to sites of fibrin deposition. Fibrinogen-coated oil droplets might provide such a vehicle, especially for the delivery of water-insoluble molecules. Although the use of oil droplets for the delivery of hydrophobic molecules is not new, the site-specific targeting of such droplets has been problematic.33 Functional fibrinogen adsorbed "irreversibly" to drug-loaded oil droplets might serve to focus those droplets to pathological sites that express thrombin activity, thereby locally increasing drug concentration while globally limiting nonspecific effects of therapy.
Fibrinogen-coated oil droplets might also be formulated as adjuvants for vaccines. Earlier, we proposed that fibrinogen adsorbs from plasma to the surface of mineral oil droplets of Freund's and related adjuvants and acts synergistically with amphiphilic lipids expressed on the surface of those droplets to elicit acute inflammation, granuloma formation, immune adjuvancy, and hemorrhage.5 34 Because at that time it was our intention to expose the surface of oil droplets as the site of biological activity of adjuvant lipids, we substituted solid, microscopic, hydrophobic beads for the oil droplets of conventional adjuvant emulsions.35 Using the bead system, we demonstrated conclusively that fibrinogen binds avidly and with high affinity to beads coated with adjuvant lipids and that the bound protein acts synergistically with the lipid microenvironment to produce the biological effects traditionally associated with adjuvant oil emulsions.5 34 Since then, many adjuvant preparations that include oil and 1 or another amphiphilic lipid have been formulated for successful vaccine use.36 We suggest that precoating adjuvant oil droplets with fibrinogen might optimize the immune-enhancing potential of the droplets, perhaps by either facilitating or promoting the interactions of the particles with macrophages.
Biomedical materials scientists have long appreciated that fibrinogen, of all the plasma proteins, adsorbs rapidly and preferentially to hydrophobic, polymeric materials in contact with blood.4 23 As a consequence of this adsorption, blood clots are nucleated on the surface of the material often leading, in the case of circulatory prosthetics, to the development of an occlusive thrombus. Thus, 1 aim of biomedical materials researchers is formulation of polymers that do not bind fibrinogen. (Importantly, a successful means that prevents the adsorption of fibrinogen and other proteins to a surface involves coating that surface with phospholipids.37 ) Another aim of biomedical materials researchers is identification of the region(s) of fibrinogen that actually contacts hydrophobic, polymeric surfaces.15 38 Attempts at this second aim have met with limited success, both because the binding of fibrinogen to hydrophobic surfaces is essentially irreversible and because the protein and its remnants are difficult to elute from hydrophobic surfaces.14 We propose that droplets of liquid hydrophobic phases may help determine the identity of the "surface recognition site(s)" of fibrinogen. Because droplets of the oils used here are readily soluble in organic phases, there is no need to elute the protein or an adherent remnant of the protein from the droplets. It is enough to simply dissolve the underlying hydrophobic "scaffold," leaving as residual only proteinaceous material.
Perhaps 1 of the more important concepts to derive from this study is that heparin and related polyanions prevent adhesion between fibrin-coated surfaces, modeled here by droplets of liquid hydrophobic phases. The sensitivity and specificity of the phenomenon to unfractionated heparin support the notion that this antiadhesive property exists by design,10 begging further investigation of the phenomenon. As concerns practical application, we suspect that these polyanions might be used advantageously to prevent or even reverse a host of deleterious, fibrin(ogen)-mediated, adhesive events, particularly those occurring at sites of inflammation.6 7 30 31 32 39 40
For quite some time, biomedical scientists studying atherosclerosis have focused on the role(s) of lipids in that disorder. More recently, investigators have begun to seriously consider mechanistic roles for fibrinogen in atherogenesis,41 42 43 in part because fibrin(ogen) is a ubiquitous and significant component of advanced lesions of atherosclerosis, ie, plaques.3 24 25 40 44 45 46 We ourselves have proposed that the affinity of fibrinogen for extracellular deposits of atheromatous lipids contributes to the morbid and mortal thrombotic consequences of atherosclerosis8 : because clots are nucleated by adsorbed fibrinogen, fibrinogen-coated lipid surfaces should be predisposed to thrombosis. However, fibrinogen is a significant component of even the earliest detectable precursor of the atherosclerotic plaque, the fatty streak.42 What role, if any, could fibrinogen play in the earliest stages of plaque development?
Blood, a fibrinogen-rich medium, is replete with lipid-laden lipoproteins that percolate through the walls of blood vessels and into the tissues. Popular theories hold that lipoproteins, particularly LDLs, are somehow retained within arterial walls, and this retention accounts for the accumulation there of lipids that will eventually constitute the plaque.47 48 Given the results of our studies, those of others,46 49 and current opinion regarding plaque initiation, development, and growth, it is reasonable to propose that the localization and accumulation in vivo of lipoproteins—or, for that matter, any lipid particle—will be mediated at least in part by fibrinogen adsorbed to those particles. This adsorption in turn will be dictated by the expression of hydrophobic domains on the surface of the particles. Thus, localization and accumulation of lipid particles within the vasculature need not involve any specific interaction between some particle-resident amphiphile, eg, apolipoprotein(a),50 and fibrin(ogen): for fibrinogen-coated particles to accumulate, there need only be focal production of thrombin, such as occurs at any site of inflammation.32
Many approaches could be used to assess the validity of this proposal. One approach might involve measuring directly the binding of fibrinogen to various lipoproteins and, if fibrinogen binds to them, its functionality once bound (G.S.R. et al, unpublished data, 1997). Another approach might involve correlating plasma fibrinogen concentration with some lipid-dependent aspect of plaque growth. Still another approach might involve use of anticoagulants as a means to prevent or limit plaque development. All of these biological approaches have merit, and all would likely yield valuable mechanistic insight relevant to both atherosclerosis and its therapy. However, physicochemical approaches of the sort presented here should also provide valuable information pertinent to a mechanistic understanding of atherosclerosis, because the adhesive potential conferred to a lipid particle by adsorbed fibrinogen makes the presence of that protein on the particle singularly important.
Which factors should be expected to influence the "nonspecific" binding of fibrinogen to hydrophobic lipid particles? We have demonstrated several, including the solution-phase concentration of fibrinogen, the solution-phase concentration of species that compete with fibrinogen for the surface of particles, and amphiphiles preexisting on the surface of particles. The concentration of particles, a measure of lipid "load," will also influence the equilibrium distribution of the protein, as will measures that affect the affinity and/or capacity of individual particles for the protein. Given all of these factors, it is reasonable to presume that the disposition of lipid particles that percolate through the walls of the vasculature will depend on the dynamic equilibrium that must normally exist between bound and free forms of fibrinogen.
|
Acknowledgments |
|---|
This work was supported by funds to G.S.R. from the Department of Pathology and Laboratory Medicine, University of Cincinnati, and by a Focused Giving Award to G.S.R. from Johnson & Johnson. G.S.R. thanks Ruth Mary Retzinger for inspiration.
Received April 16, 1998; accepted June 5, 1998.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
1. Retzinger GS, McGinnis MC. A turbidimetric method for measuring fibrin formation and fibrinolysis at solid-liquid interfaces. Anal Biochem. 1990;186:169–178.[Medline] [Order article via Infotrieve]
2. Cook BC, Retzinger GS. Lipid microenvironment influences the processivity of adsorbed fibrin(ogen): enzymatic processing and adhesivity of the bound protein. J Colloid Interface Sci. 1994;162:171–181.
3. Haust MD, Wyllie JC, More RM. Atherogenesis and plasma constituents, I: demonstration of fibrin in the white plaque by the fluorescent antibody technique. Am J Pathol. 1964;44:255–267.
4. Baier RE. Initial events in interactions of blood with a foreign surface. J Biomed Mater Res. 1969;3:191–206.[Medline] [Order article via Infotrieve]
5. Retzinger GS, Meredith SC, Hunter RL, Takayama K, Kézdy FJ. Identification of the physiologically active state of the mycobacterial glycolipid trehalose 6,6'-dimycolate and the role of fibrinogen in the biologic activities of trehalose 6,6'-dimycolate monolayers. J Immunol. 1982;129:735–744.[Medline] [Order article via Infotrieve]
6. McRitchie DI, Girotti MJ, Glynn MFX, Goldberg JM, Rotstein OD. Effect of systemic fibrinogen depletion on intraabdominal abscess formation. J Lab Clin Med. 1991;118:48–55.[Medline] [Order article via Infotrieve]
7.
Tang L, Eaton JW. Fibrin(ogen) mediates acute
inflammatory responses to biomaterials. J Exp Med. 1993;178:2147–2156.
8.
Retzinger GS. Adsorption and coagulability of
fibrinogen on atheromatous lipid surfaces.
Arterioscler Thromb Vasc Biol. 1995;15:786–792.
9. O'Connor SM, Patuto SJ, Gehrke SH, Retzinger GS. Fibrinogen-dependent adherence of macrophages to block copolymer-coated polystyrene-divinylbenzene beads. Ann N Y Acad Sci. 1997;831:138–144.[Medline] [Order article via Infotrieve]
10.
Retzinger GS, Chandler LJ, Cook BC. Complexation with
heparin prevents adhesion between fibrin-coated surfaces. J
Biol Chem. 1992;267:24356–24362.
11. Clauss A. Gerinnungsphysiologische Schnellmethode zur Bestimmung des Fibrinogens. Acta Haematol. 1957;17:237–246.[Medline] [Order article via Infotrieve]
12. DeAnglis AP, Retzinger GS. Preparation and characterization of fibrinogen-coated, reversibly adhesive, lecithin/cholesterol vesicles. J Pharm Sci. 1995;84:399–403.[Medline] [Order article via Infotrieve]
13. Yamaoka K, Tanigawara Y, Nakagawa T, Uno T. A pharmacokinetic analysis program (multi) for microcomputer. J Pharmacobio-Dyn. 1981;4:879–885.[Medline] [Order article via Infotrieve]
14. Cook BC, Retzinger GS. Elution of fibrinogen and other plasma proteins from unmodified, and from lecithin-coated, polystyrene-divinylbenzene beads. J Colloid Interface Sci. 1992;153:1–12.
15. Retzinger GS, Cook BC, DeAnglis AP. The binding of fibrinogen to surfaces and the identification of two distinct surface-bound species of the protein. J Colloid Interface Sci. 1994;168:514–521.
16.
Smaby JM, Brockman HL. Regulation of cholesteryl oleate
and triolein miscibility in monolayers and bilayers. J Biol
Chem. 1987;262:8206–8212.
17. Miller KW, Small DM. The phase behavior of triolein, cholesterol, and lecithin emulsions. J Colloid Interface Sci. 1982;89:466–478.
18. Miller KW, Small DM. Triolein-cholesteryl oleate-cholesterol-lecithin emulsions: structural models of triglyceride-rich lipoproteins. Biochemistry. 1983;22:443–451.[Medline] [Order article via Infotrieve]
19. Gaines GL. Insoluble Monolayers at Liquid-Gas Interfaces. New York, NY: Interscience; 1966;257.
20. Laudano AP, Doolittle RF. Studies on synthetic peptides that bind to fibrinogen and prevent fibrin polymerization: structural requirements, number of binding sites, and species differences. Biochemistry. 1980;19:1013–1019.[Medline] [Order article via Infotrieve]
21.
Shimizu A, Schindlauer G, Ferry JD. Interaction
of the fibrinogen-binding tetrapeptide gly-pro-arg-pro with fine clots
and oligomers of -fibrin; comparison with ß-fibrin.
Biopolymers. 1988;27:775–788.[Medline]
[Order article via Infotrieve]
22.
Scott JP, Montgomery RR, Retzinger GS. Dimeric
ristocetin flocculates proteins, binds to platelets, and mediates
von Willebrand factor-dependent agglutination of platelets.
J Biol Chem. 1991;266:8149–8155.
23. Bagnall RD. Adsorption of plasma proteins on hydrophobic surfaces, II: fibrinogen and fibrinogen-containing protein mixtures. J Biomed Mater Res. 1978;12:203–217.[Medline] [Order article via Infotrieve]
24. Haust MD, Wyllie JC, More RM. Electron microscopy of fibrin in human atherosclerotic lesions: immunohistochemical and morphologic identification. Exp Mol Pathol. 1965;4:205–216.
25. Smith EB. Molecular interactions in human atherosclerotic plaques. Am J Pathol. 1977;86:665–674.[Abstract]
26. Mohri H, Ohkubo T. Fibrinogen binds to heparin: the relationship of the binding of other adhesive proteins to heparin. Arch Biochem Biophys. 1993;303:27–31.[Medline] [Order article via Infotrieve]
27.
Hotchkiss KA, Chesterman CN, Hogg PJ. Inhibition
of heparin activity in plasma by soluble fibrin: evidence for ternary
thrombin-fibrin-heparin complex formation. Blood. 1994;84:498–503.
28.
Odrljin TM, Shainoff JR, Lawrence SO, Simpson-Haidaris
PJ. Thrombin cleavage enhances exposure of a heparin binding domain in
the N-terminus of the fibrin ß chain. Blood. 1996;88:2050–2061.
29. Raut S, Gaffney PJ. Interaction of heparin with fibrinogen using surface plasmon resonance technology: investigation of heparin binding site on fibrinogen. Thromb Res. 1996;81:503–509.[Medline] [Order article via Infotrieve]
30. Dunn CJ, Willoughby DA. Leukocyte and macrophage migration inhibitory activities in inflammatory exudates: involvement of the coagulation system. Lymphokines. 1981;4:231–269.
31. Helin H. Macrophage procoagulant factors: mediators of inflammatory and neoplastic tissue lesions. Med Biol. 1986;64:167–176.[Medline] [Order article via Infotrieve]
32.
Marcus AJ. Thrombosis and inflammation as multicellular
process: pathophysiologic significance of transcellular
metabolism. Blood. 1990;76:1903–1907.
33. Prankerd RJ, Stella VJ. The use of oil-in-water emulsions as a vehicle for parenteral drug administration. J Parenter Sci Technol. 1990;44:139–149.[Medline] [Order article via Infotrieve]
34. Retzinger GS. Dissemination of beads coated with trehalose 6,6'-dimycolate: a possible role for coagulation in the dissemination process. Exp Mol Pathol. 1987;46:190–198.[Medline] [Order article via Infotrieve]
35.
Retzinger GS, Meredith SC, Takayama K, Hunter RL,
Kézdy FJ. The role of surface in the biological activities of
trehalose 6,6'-dimycolate: surface properties and development of a
model system. J Biol Chem. 1981;256:8208–8216.
36.
Rabinovich NR, McInnes P, Klein DL, Hall BF. Vaccine
technologies: view to the future. Science. 1994;265:1401–1404.
37. Hayward JA, Johnston DS, Chapman D. Polymeric phospholipids as new biomaterials. Ann N Y Acad Sci. 1985;446:267–281.[Medline] [Order article via Infotrieve]
38. Gebhardt K, Lauvrak V, Babaie E, Eijsink V, Lindqvist BH. Adhesive peptides selected by phage display: characterization, applications and similarities with fibrinogen. Pept Res. 1996;9:269–278.[Medline] [Order article via Infotrieve]
39. Kamitsuji H, Sakamoto S, Matsunaga T, Taira K, Kawahara S, Nakajima M. Intraglomerular deposition of fibrin/fibrinogen-related antigen in children with various renal diseases. Am J Pathol. 1988;133:61–72.[Abstract]
40.
Smith EB, Keen GA, Grant A, Stirk C. Fate of fibrinogen
in human arterial intima.
Arteriosclerosis. 1990;10:263–275.
41.
Kannel WB, Wolf PA, Castelli WP, D'Agostino RB.
Fibrinogen and risk of cardiovascular disease.
JAMA. 1987;258:1183–1186.
42. Ernst E. The role of fibrinogen as a cardiovascular risk factor. Atherosclerosis. 1993;100:1–12.[Medline] [Order article via Infotrieve]
43.
Heinrich J, Balleisen L, Schulte H, Assmann G, van de
Loo J. Fibrinogen and factor VII in the prediction of coronary
risk: results from the PROCAM study in healthy men. Arterioscler
Thromb. 1994;14:54–59.
44. Smith EB. Fibrinogen, fibrin and fibrin degradation products in relation to atherosclerosis. Clin Haematol. 1986;15:355–370.[Medline] [Order article via Infotrieve]
45.
Bini A, Fenoglio JJ, Mesa-Tejada R, Kudryk B, Kaplan
KL. Identification and distribution of fibrinogen, fibrin, and
fibrin(ogen) degradation products in
atherosclerosis: use of monoclonal antibodies.
Arteriosclerosis. 1989;9:109–121.
46.
Kunz F, Pechlaner C, Erhart R, Zweirzina W-D, Kemmler
G. Increased lipid binding to thrombi in coronary artery
disease: findings in patients without premedication in native (not
anticoagulated) test systems. Arterioscler Thromb. 1992;12:1516–1521.
47. Ross R. The pathogenesis of atherosclerosis: a perspective for the 90s. Nature. 1993;362:801–809.[Medline] [Order article via Infotrieve]
48.
Williams KJ, Tabas I. The response-to-retention
hypothesis of early atherogenesis. Arterioscler Thromb Vasc
Biol. 1995;15:551–561.
49. Smith EB, Massie IB, Alexander KM. The release of immobilized lipoprotein fraction from atherosclerotic lesions by incubation with plasmin. Atherosclerosis. 1976;25:71–84.[Medline] [Order article via Infotrieve]
50.
Loscalzo J. Lipoprotein (a): a unique risk factor for
atherothrombotic disease.
Arteriosclerosis. 1990;10:672–679.
This article has been cited by other articles:
 |
|
 |
|
  C. M. Einhaus, A. C. Retzinger, A. O. Perrotta, M. D. Dentler, A. S. Jakate, P. B. Desai, and G. S. Retzinger Fibrinogen-Coated Droplets of Olive Oil for Delivery of Docetaxel to a Fibrin(ogen)-Rich Ascites Form of a Murine Mammary Tumor Clin. Cancer Res., October 15, 2004; 10(20): 7001 - 7010. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. S. Jakate, C. M. Einhaus, A. P. DeAnglis, G. S. Retzinger, and P. B. Desai Preparation, Characterization, and Preliminary Application of Fibrinogen-Coated Olive Oil Droplets for the Targeted Delivery of Docetaxel to Solid Malignancies Cancer Res., November 1, 2003; 63(21): 7314 - 7320. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. H Reinhart Fibrinogen - marker or mediator of vascular disease? Vascular Medicine, August 1, 2003; 8(3): 211 - 216. [Abstract] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||