Macromolecular Transport across Arterial andVenous Endothelium in RatsStudies with Evans Blue-Albumin and

Horseradish Peroxidase

Pao-Tien Chuang, Hwai-Jong Cheng, Shing-Jong Lin,Kung-Ming Jan, Mary M.L Lee, and Shu Chien

Atherosclerotic lesions are characterized by llpid Infiltration in regions wtth highrates of endothellal cell turnover. The present investigation was designed to elucidatethe route of macromolecular transport across vascular endothelium. The aorta andvena cava of male Sprague-Dawley rats were perfusion-flxed after the intravenousin|ection of Evans-blue albumin (EBA) or horseradish peroxidase (HRP). Fluores-cence microscopic examination of en face preparation of the aorta stained withhematoxylln allowed the Identification of endothellal cells that underwent mitosis,together with the localization and quantification of fluorescent spots for EBA leakage.The HRP specimens were subjected to histochemicai treatment, and HRP leakagewas seen as brown spots under the light microscope. Sliver nitrate stain was addedIn both EBA and HRP studies to outline cell boundaries and to visualize stigmata,stomata, and dead cells. In the aorta, almost every dividing cell showed Junctionalleakage to albumin and HRP, with clustering of leaky spots around the branchorifices. Time-dependent studies showed gradual increases In the diameter andnumber of these heterogeneously sized leaky spots, which finally fused to sizescorresponding to the "blue areas" for EBA or "brown areas" for HRP. Compared wttharteries, veins had fewer mltotlc cells, but more dead cells and diffuse dye-stainingareas, Indicating a more rapid transport of macromolecules. The leaky spots In theartery were associated mainly with mltotic cells, dead cells, and stigmata, whereasthose In the vein occurred primarily at regions with dead cells. These results suggestthat the preferential association of the enhanced transport of macromolecules withmttosls In the arterial as compared to venous endothelium and the differentialbehavior In transmural transport between arteries and veins may form the basis forthe predilection of atherosclerosis in arteries.(Arteriosclerosis 10:188-197, March/April 1990)

A therosclerosis is a disease of large and medium-sized arteries. It occurs at specific sites of predilec-

tion corresponding to segmental zones of low shear withflow reversal.1 The basic lesion is an atheroma, whichconsists of the deposition of lipids, the formation of foamcells, and the proliferation of smooth muscle cells in theartery wall. Some of the important questions in athero-genesis are 1) What are the pathways for the transport oflow density lipoprotein (LDL) across the endothelium?and 2) What is the relation to hemodynamics? It has beenshown that the blue areas stained by Evans blue aJbumin-(EBA) after its intravascular administration might be as-sociated with sites of: 1) low shear with flow reversal,

From the Institute of BlomedicaJ Sciences, Academla Slnlca,and National Taiwan University College of Medicine, Taipei,Taiwan, R.O.C.

This work was supported by research Grant NSC76-0412-B001-06 from the National Science Council, R.O.C.

Shu Chlen is now at the Department of AMES-Bioenglneering,University of California, San Diego, La Jolla, California.

Address for reprints: Dr. Shu Chien, Department of AMES-Bioengineering, R-012, UCSD, La Jolla, CA 92093.

Received January 9, 1989; revision accepted September 22,1989.

2) high rate of cell turnover,2-3 and 3) enhanced LDLaccumulation.'4 In these studies, however, the circulationtime for EBA was too long to allow an identification of theinitial sites of macromolecular entry. Previous investiga-tion with horseradish peroxidase (HRP) has revealedbrown spots with an LDL concentration 47 times greaterthan the noncolored regions after a short circulation time,but the nature of those leaky spots has not beenelucidated.5 More recently, it has been shown that almostall mitotic cells in aortic endothelium are leaky to EBA,6

but the data were not analyzed in terms of their regionaldistribution, the process of lateral diffusion, or hemody-namic influences. To address the questions of: 1) theinitial sites of macromolecular entry, 2) the process oflateral diffusion in the media of vessel wall, and 3) thehemodynamic influences on atherogenesis, we usedEBA and HRP as tracers to study the macromoleculartransport across arterial and venous endothelia in normalrats. Such studies may also provide some insight into thetransport processes of a wide spectrum of macromole-cules (including LDL) across the vascular endothelia.78

The results have generated new information on endothe-lial cell mitosis and permeability in terms of: 1) time-

188

MACROMOLECULAR TRANSPORT ACROSS VASCULAR ENDOTHELIUM Chuang et al. 189

dependent changes, 2) comparison between branchingand nonbranching areas, and 3) comparison betweenarteries and veins.

MethodsEvans Blue-Albumin Study In Arteries

EBA conjugate was prepared by mixing 0.5 g of bovineserum albumin (obtained from Sigma Chemical, StLouis, MO) and 0.1 g of Evans blue (Sigma) in 50 ml ofnormal saline.8 The EBA study was performed in arteriesof 10 young, healthy male Sprague-Dawley (SD) rats on anormal diet and weighing 250 to 350 g. After pentobar-bital anesthesia (Abbott Laboratories, North Chicago, IL;3 mg/100 g, i.p.), the femoral arteries of both sides, theright femoral vein, and the right common carotid arterywere cannulated by using PE 50 or PE 90 polyethylenetubings. EBA conjugate (5 ml) was injected through thefemoral vein over a 3-minute period and was allowed tocirculate for a total period of 10 minutes in five rats,45 minutes in four rats, and 2 hours in one rat Ten sec-onds before the termination of each circulation period,0.3 ml of heparin (5000 USP units/ml; China Chemical &Pharmaceutical, Taipei, R.O.C.) was injected through thefemoral vein to prevent coagulation, and this was fol-lowed by 0.5 ml of an overdose of pentobarbital to stopthe heart. The arterial system was perfused through thecannulated common carotid artery with 60 ml of heparin-ized saline from a reservoir of the perfusion apparatus setat a pressure of 110 mm Hg, with the femoral arteries ofboth sides serving as exits for the perfusate. Then theperfusate was changed to a 1:1 mixture of 20% formal-dehyde/4% glutaraldehyde (60 ml). Thereafter, the re-agents for silver nitrate staining were perfused in thefollowing order10: 5% glucose for 2 minutes, 0.05% silvernitrate (Sigma) for 1 minute, and a 1:1 mixture of3% CoBr (Strem Chemicals) and 1% NH4Br (Merck) for1 minute. The thoracic aorta was then removed, and theconnective tissues, including the adventitia, were care-fully peeled away with a fine forceps under a dissectingmicroscope. The aorta was cut open longitudinally alongthe ventral curvature and was sectioned Into severalsegments. Each segment was stained with Hank's hema-toxylin (Merck) for 45 seconds, was mounted onto a glassslide with glycerol gelatin (Merck), was covered with acoverslip, and was examined en face with a fluorescencemicroscope (see below). Another coverslip was carvedwith a diamond knife under the dissecting microscopeinto a grid composed of squares of 1 mm2. The grid wasthen placed on top of the coverslip, and a square waspositioned in the center of the field at 100 x magnificationunder the fluorescence microscope. A micrometer wasadded in the eyepiece to subdivide this square into 400(20x20) smaller squares (Figure 1). In this way we wereable to record the coordinates, size, and endothelialmorphology of every spot. The distribution of the spots inrelation to the branch orifices of the intercostal arterieswas plotted. Five times the diameter of the intercostalbranch orifice was used as a diameter to draw circlescentered at the branch orifice to outline the branchingregions, and the remaining luminal surface area of the

Slide

1

Cover glass 2

Mcro meter(In eyepiece)

Figure 1. Schematic representation of the method for record-ing the coordinates, size, and characteristics of endothelialmorphology of Individual Evans blue-albumin spots and thelocations of intercostal branch orifices on en face preparation ofthe rat thoracic aorta under the fluorescence microscope.

aorta outside the circles was referred to as the non-branching region. In each of these two regions, thenumber of spots and the surface areas examined weredetermined for the calculation of the number density ofspots per unit area as well as for the ratio of the numberdensities in these two regions.

Horseradish Peroxldase Study In ArteriesTen SD rats weighing 250 to 350 g were cannulated as

described above. HRP (type II, Sigma; 8 mg HRP in 1 mlof normal saline per 100 g of body weight6) was injectedthrough the femoral vein and allowed to circulate for30 seconds in four rats, 1 minute in one rat, 2 minutes infour rats, and 4 minutes in one rat. As in the EBA study,the aorta underwent perfusion fixation and silver nitratestaining. After the adventitia had been peeled off, theaorta was incubated in a mixture of 0.15% 3,3'-diaminobenzidine (DAB, Sigma) and 0.02% H2O2 in Trisbuffer (pH 7.0) at 37° C for 60 minutes. Then the aortawas cut open longitudinally and was stained with hema-toxylin; the specimen was used for en face observationunder the light microscope. In some of the HRP experi-ments, the aortic specimens were processed for electronmicroscopy. During perfusion fixation of the aorta, gluta-raldehyde was also added to the outer surface of thevessel. One hour later, the excised aortic specimen waspostfixed for 1 hour in 1% osmium tetroxide, was stainedwith 2% uranyl acetate, was dehydrated in ascendingethanol series, and was embedded in Epon. Thin sec-tions were viewed in a Joelco C100 electron microscope.

Horseradish Peroxldase Study In VeinsTen other SD rats weighing 350 g were anesthetized with

pentobarbital, and the right femoral vein was cannulated.

190 ARTERIOSCLEROSIS VOL 10, No 2, MARCH/APRIL 1990

HRP was injected through the femoral vein and was allowedto circulate for 1 minute. Perfusion fixation and silver nitratestaining were performed as in the artery study describedabove, except for the following aspects. First, the perfusionpressure was maintained at 20 mm Hg instead of110 mm Hg, and the perfusate entered the femoral vein andexited from the cut open jugular vein. Second, 0.05% silvernitrate was perfused for 5 minutes instead of 1 minute. Theinfradiaphragmatic segment of the inferior vena cava wasexcised and immersed in 2% glutarakjehyde for another30 minutes. The remaining procedures were the same as inthe artery study.

Fluorescence MicroscopyA Nikon eptfluorescence microscope (MICROPHOT-FX)

equipped with a diascopic phase-contrast attachment wasused for identifying and mapping the distribution of EBAleaky spots in en face preparations of the thoracic aorta.The EBA fluorescence was studied with a combination ofan excitation filter at 450 to 490 nm, a dichroic mirror at510 nm, and a barrier filter at 520 nm. Color slides weremade with Kodak Ektachrome P800/1600 professionalcolor-reversal film 5020 for photomicrographs.

Criteria for Identifying Mltotlc FiguresWith hematoxylin staining, different stages of cell divi-

sion in the mitotic (M) phase of the cell cycle wereidentified according to the following morphologicalcriteria.11 Cells with visible, condensed chromosomesand an intact nuclear envelope were defined as being inthe prophase. Cells with chromosomes aligned at ametaphase plate halfway between the poles and withouta nuclear envelope were defined as being in the meta-phase. Cells with separated chromatids pulled toward thepoles were defined as being in the anaphase. Cells withchromatids clustered at each pole, a contractile ringcreating a cleavage furrow, and a reformed nuclearenvelope were defined as being in the telophase.

Criteria for Identifying Dead Endothellal Cells,Stomata, and Stigmata

In en face preparations of both aorta and vena cava,silver nitrate staining produced a network of dark brownlines corresponding to endothelial cell borders. Cellsshowing intense surface silver staining were defined asthe dead cells.3 Focal black deposits on or between thecells seen by light microscopy, either as ring-shapedstructures with a clear center or as solid dots with orwithout a halo of finer granules, were identified as sto-mata and stigmata, respectively.12

ResultsEvans Blue-Albumin and HorseradishPeroxldase Studies In Arteries

As found in our previous study, nearly all mitotic cellsfound on en face preparations were associated with thepresence of EBA leaky spots. The present study showsthat the same is true for HRP. Of all the EBA leaky spots,approximately one-third were associated with dividingcells (Figures 2 and 3A) and 5% with dead cells (Fig-

1

80-

7 0 - |

«H

50-

EBA

h

ENC

Y

4 0 -

30 -

20-

Telophas*

Anaphasa

Metaphaw

Non-civking Dtvtdng Dead cotecells cats

Figure 2. Distribution profile erf the association of leaky spots withnondividing cells, dividing cells, and dead cells. About one-third ofthe leaky spots examined on en face preparations under themicroscope were associated with dividing (mitotic) cells, and 5%,with dead cells. More than one-half of the spots were associatedwith nondMcflng ceils. EBA=Evans blue-albumin.

ure 3B). The remainder (more than one-half) of the spotswere associated with ordinary, nondividing cells (Fig-ure 3C); in about half of them a dark silver nitrate stainingband exhibiting the morphology of stigmata12 could beseen either along the periphery or in the center of the cell(Figures 4A and 4B). Sometimes combinations of morethan one morphological characteristic, e.g., dividing celland stigmata (Figures 4C and 4D), and dividing cell anddead cell (Figure 4E), could be found.

Image analysis of the EBA fluorescent spots with animage processor system (Model IP 512, Image Technol-ogy, Woburn, MA) revealed a graded concentration dis-tribution profile, with the peak located at the center of thespot (Figure 5). The average diameter of the EBA fluo-rescent spots increased as the EBA circulation time wasprolonged (Figure 6A). The average number of EBAspots per unit area also increased with time (Figure 6B).For HRP, the rate of increase in spot size was much faster(Figure 7). There was a considerable degree of hetero-geneity for spot size with each dye circulation time(Figure 8). The heterogeneity was much greater in theHRP study than with EBA. Small spots and diffuse HRPstaining regions coexisted in the same sample of en faceobservations after an HRP circulation time as short as30 seconds. As time increased, the spot size increasedrapidly, and finally the spots fused together, formingdiffuse HRP staining of wide areas of the aorta at about4 minutes after its intravenous injection.

MACROMOLECULAR TRANSPORT ACROSS VASCULAR ENDOTHEUUM Chuang et al. 191

B

Figure 3. En face preparation of rat thoracic aorta. Silver nitrateand hematoxyiin staining. Bars=20/*m. A. A dividing cell In theanaphase with condensed chromosomes (arrows) In the centerof a fluorescent Evans blue-albumin (EBA) spot (bright area).x550 B. A dead cell with intense surface silver staining (arrow)associated with a horseradish peroxidase brown spot x360C. Ordinary, nondMdlng cells associated with a fluorescent EBAspot (bright area). x550

When the spatial distribution of EBA fluorescent spotswas determined in relation to intercostal branch orifices,the spots were found to exist in higher density aroundbranch orifices. Quantitative calculation showed an aver-age of two times more spots per unit area in the branch-ing than in the non-branching area, but no preference ofspot distribution was found with respect to the upstreamor downstream side of brach orifices (Table 1). There wasa tendency for the spots to distribute spirally along thelongitudinal direction of the aorta, curving around thebranch orifices (Figure 9).

Extensive electron microscopic examinations of thinsections of aortic specimens in HRP experiments yieldedthe finding of one endothelial cell in mitosis (Figure 10),as revealed by the nuclear pattern of this cell. Thejunction to the right of this dividing cell shows abnormalwidening and is filled with the electron-dense HRP reac-tion product (Figures 10B and 10C), thus providing directexperimental support of the cell turnover-leaky junctionhypothesis.13 (This picture was reproduced with permis-sion from a Proceedings paper which we reported at ameeting.14) Due to the rare occurrence of mitotjc cells(1/104) compounded by the difficulty of obtaining themitotic nucleus and leaky junction on the same thinsection, it is estimated that more than one thousand gridsneed be examined to find one such picture.

Horseradish Peroxidase Study In VeinsHRP staining in veins was rather diffuse in comparison

with that in arteries. After a 1 -minute dye circulation time,the HRP staining area in veins was similar to that inarteries with a dye circulation time of 4 minutes (Fig-ure 7). Dead cells (Figure 11) constantly appeared in thediffuse tracer staining regions. Mitotic cells were occa-sionally seen, but no association of HRP staining withthese cells could be identified in the vein.

DiscussionAn analysis of the various pathways for transendothe-

lial transport of macromolecules has led to the hypothesisthat the endothelial cells that are involved in cell turnoverhave poorly formed or leaky junctions.13 As described inour previous studies, 99% of mitotic cells are associatedwith leaky spots for EBA,8 and 80% of mitotic cells areassociated with leaky spots for Lucifer yellow-labeledLDL.18 However, mitotic cells account for only 23% of thetotal EBA leaky spots6 and 42% of the LDL leaky spots15;thus, macromolecular leaky spots are also found aroundnonmitotic cells. According to the cell turnover-leakyjunction hypothesis, one possible explanation for thefinding of spots not associated with mitotic cells is that theintercellular junctions become disrupted before the meta-

192 ARTERIOSCLEROSIS VOL 10, No 2, MARCH/APRIL 1990

Figure 4. En face preparation of rat thoracic aorta. Silver nitrateand hematoxylin staining. Bars=20 nm. A. Dark silver nitratestaining band (arrow) In the center of the ceil In a fluorescentEvans blue-albumin (EBA) spot (bright area). x550 B. Dark silvernitrate staining spots (arrows) in the periphery of the cells in ahorseradish peroxldase (HRP) brown spot. x360 C. Appearanceof a dividing endothelial cell in the metaphase with condensedchromosomes aligned at a metaphase plate (arrow), and astigmata (black dot, short arrow) within the same fluorescenceEBA spot (bright area). x550 D. Appearance of a dividingendothelial cell In the metaphase (arrow), and stigmata (blackdots, short arrows) within the same HRP brown spot x360E. Coexistence of a dividing cell in the metaphase (arrow) anda dead cell with Intense sliver staining (short arrow) in the samefluorescent EBA spot (bright area). x550

phase and need time to reform completely after the latetelophase.14 Another possibility is that EBA leakage alsooccurs in association with endothelial cell death andduring denuding or nondenuding desquamatJon.

The increase in the number of EBA leaky spots as dyecirculation time is prolonged (Figure 6B) indicates that thejunctions of different cells become leaky to macromoleculesat various times. That is, because endothelial ceOs are indifferent stages of the cefl cycle, there would be a progres-

sive increase in the number of mitotic ceQs whose junctionsbecome leaky to EBA. The increase in average spot size asdye circulation time increases (Figure 6A) reflects the contin-uous influx of EBA and its lateral diffusion process in themedia At 2-hour dye circulation time, diffuse dye-stainingregions are seen compatible to the so-called blue area318

The heterogeneousty sized spots at each dye circulation timecan also be explained by the asynchrony of cell mitosis withattended junction leakiness and the continuous influx and

MACROMOLECULAR TRANSPORT ACROSS VASCULAR ENDOTHELIUM Chuang et al. 193

Figure 5. Image analysis of an Evans blue-albumin fluorescentspot (micrograph of Rgure 3A) with the IRIS system. Note thegraded concentration distribution profile with a peak located atthe center of the spot.

lateral diffusion of the macromolecules. The HRP studydemonstrated the same trend as that observed in the EBAstudy, but the lateral diffusion in the subendothelial spacewas more rapid. The spreading velocity of HRP was approx-

EBA FLUORESCENT SPOTS

UJ

a «

I

— EC

O N > a o 3 0 4 o o o e o r o » g o n o t i o t »

DYE CIRCULATION TIME (min.)

Figure 6. The average diameter of the fluorescent spots (A)and average spot number per unit area (B) measured on en facepreparations of rat thoracic aorta at various dye circulation timesafter intravenous Injection of Evans blue-albumin (EBA). Curvesshow gradually Increasing spot diameter and number densitywith time. Arrow shows the diameter of the endothelial cell (EC)for comparison. Dots represent the mean values; bars arestandard errors.

280-

240-

200-

I "

cc111

120-

80-

40-

20-

HRP BROWN SPOTS

0 05 1 2 3 4

HRP CIRCULATION TIME (min.)

Figure 7. The average diameter of the horseradish peroxidase(HRP) brown spots measured on en face preparations of ratthoracic aorta at various HRP circulation times after intravenousinjection of HRP. With progression of time, the spot size in-creased rapidly, and finally the spots became confluent to formdiffuse HRP staining of the whole segment of the aorta at about4 minutes. In comparison with Evans blue-albumin (Rgure 6),the rate of increase in spot size was much greater.

imatety 50 /un/min in the first minute and 10 ^m/min be-tween 2 and 4 minutes; in contrast, the spreading velocity ofEBA was approximately 02 /im/min.

Certain sites in the circulation are more prone to thedevelopment of atherosclerosis than others. In humans,these sites include the inner curvature of a curved flow,the upstream side of the side branch, the downstreamside of the main branch, and the inner sides of thebifurcation.17 It has been suggested that this focal natureof atherogenesis is determined by hemodynamic factors,e.g., shear stress,1818 and disturbed flow20 at the vesselwall. The number and density of EBA spots was found inthe present study to be twice as high in a branchingregion as compared to a nonbranching region, suggest-ing that the concentration of macromolecules transportedinto the subendothelial space would be correspondinglyhigher in the branching region than the nonbranchingregion. If the local concentration of atherogenic mole-cules is an important factor that initiates local atheroscle-rotic changes,21 •22^3 this regional variation in macromo-lecular transport in relation to local hemodynamic patternmay provide a physical basis for the focal pattern ofatherogenesis.78 While the mitotic cells were nearly allleaky to HRP and EBA, a lower percentage (80%) of thesecells were leaky to LDL15 These results suggest that

194 ARTERIOSCLEROSIS VOL 10, No 2, MARCH/APRIL 1990

BO

IE

CRANIAL SIDE OF AORTA

«0 200

SPOT SIZE ( • » )

Figure 8. Distribution profile of the size of the fluorescent spots inthe Evans blue-albumin (EBA) study. The degree of heterogeneityof spot size became greater with increasing dye circulation time.

there is some degree of size restriction of macromolecu-lar passage through the leaky junction14 and that thereare quantitative differences in the behavior of LDL fromthe smaller macromolecules.15 Further investigations onhemodynamic influences and the interactions amongendothelium, macrophages, and platelets in these foci

0

. 0

o

o

o

o

o

o ;

o

o

o

A'10MIN.

o-

o •

o

•o-

o

&45MIN.

O

o '

o-

• 'O

. 0

CAUDAL SIDE OF AORTA

A BFigure 9. Spatial distribution of the fluorescent Evans blue-albumin (EBA) spots on the aortic surface In relation to Intercostalbranch orifices. Each dot represents a spot under the fluores-cence microscope; circles are the branch orifices A. 10 minutes ofEBA circulation time. B. 45 minutes. Note the tendency for thespots to distribute spirally along the branch orifices.

Table 1. Evans Blue-Albumin Spot Distribution In Branching and Nonbranching Regions

Dye circulation time (min)

Spot numberBranching region*

UpstreamDownstream

Nonbranching regionTotal

Surface area (lOVm2)Branching regionNonbranching regionTotal

Spot number per unit area (per 10%Branching region (A)Nonbranching region (B)TotalA/Bt

1

10

853

2432

35.27173.24208.51

im2)

0.230.140.151.6

2

10

1138

1728

30.75135.94166.69

0.360.130.172.9

3

10

2821

76795

27.68104.10131.78

1.010.640.721.6

4

10

1798

3249

26.8785.00

111.87

0.630.380.441.7

Rat number

5

10

523

1318

30.83129.86160.69

0.160.100.111.6

6

45

3519165691

22.88102.12125.00

1.530.550.732.8

7

45

211297

68

31.04135.65166.69

0.680.340.412.0

8

45

2412125983

30.47171.22201.69

0.780.340.412.3

9

45

41192292

133

29.64102.05131.69

1.380.901.011.5

10

120

219

127596

20.2486.86

107.10

1.041.700.901.2

*There were 111 spots on the upstream side and 100 spots on the downstream side.tThe means±SEM for A/B was 1.9±0.2 at 10 minutes and 2.2±0.3 at 45 minutes.

MACROMOLECULAR TRANSPORT ACROSS VASCULAR ENDOTHELIUM Chuang et al. 195

ljon

Figure 10. Electron micrographs of rat aortic endothelium. A. An endothelial cell undergoing celldivision, x 5000 B. was taken from another thin section of the same cell and shows a junctionwidened to a bizarre shape (arrow) on the right side of the dividing cell (x 4700). C. Anenlargement of the same junction, which is filled with horseradish peroxidase (x 15 500).Reproduced from reference 14, with the permission of Plenum Press.

196 ARTERIOSCLEROSIS VOL 10, No 2, MARCH/APRIL 1990

Figure 11. Diffuse horseradish peroxidase staining region inthe inferior vena cava. Several dead cells (arrows) are seen Inthis region. En face preparation with silver nitrate and hematox-ylln staining, x 550

may help to elucidate the mechanism leading to the finaloccurrence of atheromatic plaques in loci of macromo-lecular leakiness.24'25

One of the important features of atherosclerosis is thatit occurs in arteries but not in veins. Both types of vesselsare exposed to blood with a comparable lipid composi-tion and most other chemical constituents, but the hemo-dynamic factors (e.g., pressure, flow pattern, etc.) andthe wall properties (e.g., wall thickness, wall composition,etc.) are markedly different. Fluid mechanical consider-ations suggest that the preferential occurrence of athero-sclerosis in the thlck-walled arterial vessels may beexplained by the transport phenomena within the wall.26

Theoretical modeling shows that the thickness of a vesselis an important determinant of the hydraulic resistance,which in turn would Influence the macromolecular trans-port across the vessel wall.2526 The finding of a fastertransendothelial transport through the thin-walled veins incomparison with the thick-walled arteries is in agreementwith the theoretical prediction of a lower resistance to thetransmural diffusion of macromolecules in thin-walledvessels. This difference in transport behavior betweenarteries and veins may lead to a greater tendency formacromolecular accumulation in the arterial wall and to arelative vulnerability of arteries to atherosclerosis. Al-though such consideration ignores the details of lipidchemistry and cellular factors, which are also important inaffecting atherogenesls, it provides a possible answerand points the way to further investigations.26

AcknowledgmentsThe authors gratefully acknowledge Chen Ung Chang for her

valuable assistance In image processing, Mei Ung Chen for hertechnical assistance in specimen preparation, Yi Fang Hsleh forher assistance In manuscript preparation, and Sheldon Wein-baum for his helpful discussions.

References1. Karino T. Microscopic structure of disturbed flows in the

arterial and venous systems and Its Implication In thelocalization of vascular diseases. Int Anglo 1986;5:197-313

2. Caplan BA, Schwartz CT. Increased endottiellal cell turn-over in areas of In vivo Evans blue uptake In the pig aortaAtherosclerosis 1973; 17:401-417

3. Qerrtty RG, Richardson M, Somer JB, Bell FP, SchwartzCT. Endothellal cell morphology In areas of In vivo Evansblue uptake in the aorta of young pigs. Am J Pathol1977;89:313-334

4. Hoft HF, Qerrtty RG, Natto HK, Kusak DM. Quantitatton ofapo B in aortas of hyperchotesterolemlc swine. Lab Invest1983:48:492-504

5. Stsmerman MB, Morrel EM, Burke KR, Cotton CK, SmithKA, LAOS RS. Local variation in arterial wall permeability totow density lipoprotein in normal rabbit aorta. Arteriosclero-sis 1986;6:64-69

6. Un SJ, Jan KM, Schuessler G, Wslnbaum S, Chlen S.Enhanced macromolecular permeability of aortic endothellalcells in association with mitosis. Atherosclerosis 1988;73:223-232

7. Welnbaum S, Wan GB, Ganatos P, Pfeffer R, Chlen S. Onthe time dependent diffusion of macromotecules throughtransient open junctions and their subendothelial spread; 1.Snort time model for cleft exit region. J Theor Biol 1988;135:1-30

8. Wen GB, Welnbaum S, Ganatos P, Pfeffer R, Chlen S. Onthe time dependent diffusion of macromolecules throughtransient open Junctions and their subendothelial spread; 2.Long time model for Interaction between leakage sites. JTheor Biol 1988;135:219-253

9. Freedman FB, Johnson JA. Equilibrium and kinetic prop-erties of the Evans-blue albumin system. Am J Physio!1969:216:678-681

10. Zand T, Underwood JM, Nunnari JJ, Majno Q, Jorls I.Endothellum and "silver line." An electron microscopicstudy. Virchows Arch [Pathol Anat] 1982:395:133-144

11. Alberts B, Bray 0, Lewis J, Raft M, Roberts K, Watson JD.Chapter 11: Cell growth and division. In: Molecular biologyof the cell. New York: Garland Press, 1983:648-652

12. Ma|no G, Underwood JM, Zand T, Jorts I. The significanceof endothellal stomata and stigmata in the rat aorta. Vir-chows Arch pathol Anat] 1985;408:75-91

13. Welnbaum S, Tzeghal G, Pfeffer R, Ganatos P, Chlen S.Effect of cell turnover and leaky junctions on arterial macro-molecular transport Am J Phystol 1985;248:H945-H960

14. Chlen S, Un SJ, Welnbaum S, Lee MML, Jan KM. The roleof arterial endothellal cell mitosis in macromolecular perme-ability. In: Chien S, ed. Vascular endothelium In health anddisease. New York: Plenum Press, 1988:59-76

15. Un SJ, Jan KM, Welnbaum S, Chlen S. Transendothelialtransport of low density lipoprotein in association with cellmitosis in rat aorta. Arteriosclerosis 1989:9:230-236

16. Bernard AC, Gerrtty RG, Schwartz CJ. Endothellal cellmorphology in focal areas of in vivo Evans blue uptake in theyoung pig aorta. Exp Mol Pathol 1974;21:102-117

17. Texon M. Hemodynamlc basis of atherosclerosis. Washing-ton: Hemisphere, 1980:95

18. Caro CC, FHz-Gerald JM, Schroter RC. Atheroma andarterial wall shear. Observation, correlation and proposal ofa shear dependent mass transfer mechanism for atherogen-esis. Proc R Soc Lond B 1971; 177:109-159

19. Fry DL Acute vascular endothelial changes associated withincreased blood velocity gradients. Clrc Res 1968;22:165-197

20. Nerem RM, Levesque MT. The case for fluid dynamics as alocalizing factor in atherogenesls. In: SchettJer S, et al, eds.Fluid dynamics as a localizing factor for atherosclerosis.Berlin: Springer-Veriag, 1983:26-37

21. Wright HP. Endothelial mitosis around aortic branches Innormal guinea-pigs. Nature 1968;220:78-79

22. Schwartz SM, Bendltt EP. Clustering of replicating cells inaortic endothelium. Proc Natl Acad Sci USA 1976;73:651-653

23. Fry DL Mass transport, atherogenesls, and risk. Arterioscle-rosis 1987;7:88-100

MACROMOLECULAR TRANSPORT ACROSS VASCULAR ENDOTHEUUM Chuang et al. 197

24. Steinberg D. Upoproteins and atherosclerosis. A look back 26. Nerem RM, Levesque MX. Fluid mechanics In atheroscle-and a look ahead. Arteriosclerosis 1983:3:283-301 rosis. In: Skalak R, Chien S, eds. Handbook of bkwngineer-

25. Majno G, Jorls I, Zand T. Atherosclerosis: New horizons. ing. New York: McGraw-Hill Book Company, 1987:Hum Pathol 1985;16:3-5 21.1-21.22

Index Terms: albumin • atherosclerosis • endothelium • horseradish peroxidase • leaky junctions• macromolecular transport • mitosis • permeability