micropipette aspiration of cultured bovine aortic...

Micropipette Aspiration of Cultured Bovine AorticEndothelial Cells Exposed to Shear Stress

Masaaki Sato, Murina J. Levesque, and Robert M. Nerem

The mechanical properties of cultured bovine aortic endothelial cells exposed to afluid-Imposed shear stress were studied using the micropipette technique. The cells,which were attached to a Thermanox plastic substrate, were exposed to a specificsteady shear stress of either 10,30, or 85 dynes/cm2 and for a duration ranging from0.5 to 24 hours. Morphological changes In shape and orientation were observed, andfollowing each experiment, the mechanical properties were measured using the mi-cropipette aspiration technique applied to cells detached from the substrate. Fluores-cent microscopy was carried out to observe cytoskeletal F-actln filaments stainedwith rhodamlne phalloldln. During exposure to shear, the en face shape of the endo-thelial cells on the substrate became more elongated and their long axis becameoriented to the direction of flow. There was also an alteration In the F-actln filaments.These changes were dependent on both the level of shear stress and the duration ofexposure. After detachment, the cells exposed to shear maintained their deformedshape. This Is In contrast to cells In a static, no-flow environment which becamespherical in shape upon detachment. Cells exposed to shear stress demonstrated amechanical stiffness significantly greater than that of control cells, which was depen-dent on both the level of shear stress and the duration of exposure. Furthermore, Itappears that the influence of shear stress on endothelial cell mechanical stiffnessmay be related to alterations in cytoskeletal structure. (Arteriosclerosis 7:276-286,May/June 1987)

A lthough there is considerable indirect evidence thathemodynamic forces are a factor in the process of

atherogenesis, the detailed mechanisms are still poorlyunderstood and the precise role is uncertain.1'2 The accu-mulation of lipids within the intima is believed to be animportant factor in the early stages of atherosclerosis, andthe process of transendothelial macromolecule transportand the influence of hemodynamic-related events, in par-ticular shear stress, received early attention.34 More re-cently, hemodynamic forces have been found to affect theshape and orientation of endothelial cells studied both invivo and in vttro.5"10 Furthermore, the cytoskeleton (in par-ticular stress fibers), which is an important structure in thesupport of the membrane and in the maintenance of cellshape, also appears to differ in regions of differing shearstress.11"13

It thus appears that when endothelial cells are exposedto a fluid-imposed shear stress the process of adaptationor response involves not only cell orientation and elonga-tion, but also a change in the supporting, internal structure.Furthermore, this could be reflected in the mechanicalproperties of the endothelial cells, and any change in theseproperties could be important to the deformation a cell

This research was supported by the National Institutes of Healththrough Research Grant HL-26890 and by the University of Hous-ton through Biomedical Research Support Grant RR-07147.

Address for reprints: Dr. Robert M. Nerem, Biomechanlcal Lab-oratory, School of Mechanical Engineering, Georgia Institute ofTechnology, Atlanta, Georgia 30332.

Received August 11, 1986; revision accepted January 27,1987.

undergoes as part of the adaptation process. For this rea-son, it was believed that a measurement of the mechanicalproperties of an endothelial cell would be instructive. Al-though such properties would appear to be an importantcorrelate of cell structure and function, to date there are nodata available on the mechanical properties of the endo-thelial cell.

Several experimental methods for examining the me-chanical properties of a cell have been proposed.14 Theseinclude: 1) osmotic swelling in a hypotonic solution; 2)compression between two flat plates; 3) aspiration in amicropipette; 4) deflection of the surface by a rigid spheri-cal particle; 5) fluid shear on a cell tethered to a flat plate atone point; and 6) forced flow through polycarbonatesieves. Although these methods have been mainly appliedto red blood cells, other cell types have also been consid-ered.15"25

In this investigation, the mechanical properties of cul-tured bovine endothelial cells exposed to a fluid-imposedshear stress and then detached were studied using themicropipette technique. These data demonstrate an im-portant effect of shear stress on the mechanical propertiesof the endothelial cell. It is these measurements which arepresented here, and their interpretation is discussed in thecontext of cell cytoskeletal structure.

Methods

Materials

Bovine aortic endothelial cells cultured in our laboratoryon Thermanox plastic coverslips were used. Fully conflu-ent cultured endothelial cell populations from the 7th to 9th

276

by guest on May 22, 2018

http://atvb.ahajournals.org/D

ownloaded from

http://atvb.ahajournals.org/

MICROPIPETTE ASPIRATION OF ENDOTHELIAL CELLS Sato et al. 277

generation were studied. The age or passage time of thecultures varied from 2 to 9 days. The cells were exposed toa steady shear stress using a parallel plate flow chamber.The description of the flow chamber and the analysis ofthe laminar fluid flow conditions have been reported pre-viously.9

Cultured endothelial cells on a coverslip with a diameterof 13 mm were positioned in the central part of the flowchamber. This flow chamber had a flow section which was220 jxm in height, 17 mm in width, and 50 mm in length.The chamber, a reservoir, and a circulation circuit werefilled with culture medium (modified Dulbecco medium(MDM), containing 25 mmol/l Hepes buffer and 20% fetalbovine serum and antibiotics). The endothelial cells in thechamber were exposed to a specific shear stress conditiondefined by the dimensions of the flow chamber and thepressure drop across the chamber. The pressure drop wasmeasured and controlled by adjusting the height of theupstream reservoir. A roller pump was used to return theoutflow from the chamber to this feeding reservoir. TheMDM in the reservoir was kept at a constant temperatureof 37 ± 0.5° C. A gas mixture of 95% air and 5% CO2 wasprovided to the MDM in the reservoir. In the 25 experi-ments reported here, the exposure time of cells to shearstress was 0.5,1,2,4,12, or 24 hours and the shear stressimposed on the cell population was either 10, 30, or 85dynes/cm2. Photographs of the cells were obtained at theend of each f tow experiment in order to observe changes inen face cell shape and orientation. These microphoto-graphs of endothelial cells were analyzed quantitativelyusing a Zeiss Videoplan image analyzer (Zeiss Incorporat-ed, New York, New York). The following morphological

parameters were determined: area, perimeter, length,width, angle of orientation, and shape index. Only oneparameter is presented in this report, the shape index,which is defined as 4 77 A/P2 where A is the en face surfacearea and P is the cell perimeter.28

Microscopy

Fluorescent microscopy was carried out: 1) on intactconfluent monolayers grown on Thermanox coverslips un-der static (i.e., no flow) conditions and after exposure toshear stress and 2) on completely detached cells frommonolayers using either trypsin or a mechanical method.The following steps were carried out at room tempera-ture.24 Actin filaments were stained with rhodamine phal-loidin (Molecular Probes, Incorporated, Eugene, Oregon)which is specific for F-actin.27 All cell samples were brieflyrinsed with phosphate-buffered saline (PBS) and thenfixed in 4% formaldehyde in PBS for 5 minutes. For de-tached cells, each step was followed by a brief centrifuga-tion. After fixation, cell samples were rinsed in PBS andwere incubated in 0.1% Triton (Sigma Chemical Company,St. Louis, Missouri) for 5 minutes. Then the cells wererinsed and incubated with 3.3 x 10~11 mol of modaminephalloidin in 200 fi\ PBS for 20 minutes for intact mono-layers and partly detached cells and for 5 minutes for de-tached cells. Cells were rinsed and mounted over glycer-ol/PBS (1M) and were observed with a Nikonphotomicroscope equipped with epifluorescence optics.Photographs were taken on Kodak TRI-X films at 400 ASA.The films were processed using the standard procedurerecommended by the film manufacturer.

Syringe

Micrometer

Monitor TV

oMicropipette

PressureTransducer

^Camera!

IaooODD

Video CassetteRecorder

Specimen Manipulator

Recorder Microscope

Figure 1. Schematic diagram of experimental setup used for micropipette technique. The left part ofthe diagram shows the pressure reservoirs and pressure measuring device. The center shows themicropipette and cell suspension on the microscopy stage. The top right shows the data acquisitionsystem.



ownloaded from


278 ARTERIOSCLEROSIS VOL 7, No 3, MAY/JUNE 1987

Mlcroplpette Experiments

After exposure of cells to a known shear stress and for aspecified length of time, the endothelial cells were de-tached from their substrate by scratching the coverslip witha sterile wood stick. The sheet of cell aggregates wasdetached further by aspiration through a 1 ml pipette. Be-fore measurement, endothelial cell suspensions werestirred for approximately 15 seconds on a test-tube stirrer.As will be seen, detached cells that had not been exposedto shear stress were spherical in appearance, while thosethat had been exposed to shear stress were elongated in

shape. The prepared cells were studied at room tempera-ture and as soon as possible after detachment, usuallywithin 6 hours. Cell shape and the measured mechanicalproperties did not show any significant change with timeduring this period.

Micropipettes with an internal diameter ranging from 2.0to 3.4 /im (2.7 ± 0.5 fim), were prepared from glass tubes(ID = 0.7 mm, OD = 1.5 mm) with the use of a micropi-pette puller (Model 700D, David Kopf Instruments, Tu-junga, California). The micropipette was filled through a0.2 /j.m membrane filter with the same MDM used for thecell culture by use of a 1 ml syringe. This fluid-filled micro-

Figure 2. Photomicrographs of rhodamine phalloidin-stained endothelial monolayers grown on Ther-manox. A. Control conditions. B. After exposure to a shear stress level of 85 dynes/cm2 for one-half hour.



ownloaded from



pipette was connected to a pressure control line and wasfixed to the stage of the microscope (Model BHA-P, Olym-pus, Lake Success, New York). A schematic diagram ofthe experimental system used is shown in Figure 1. Thepressure control system was modeled after that of Chien etal.23 This system is composed of three different reservoirs.One is a damping chamber which is connected to the mi-cropipette and partially filled with water. This chamber willbe used in future dynamic tests to examine viscoelasticproperties. The line between the micropipette and thechamber was filled with MDM. The other two reservoirs areadjustable, and a difference between these two heights

can produce a negative pressure at the tip of the micropi-pette. One of these reservoirs was open to the atmosphereand can be controlled with an accuracy of better than 10pirn by the use of a micrometer. A part of the line betweenthe micropipette and the chamber was connected to apressure transducer (Model P23Db, Gould, Incorporated,Cleveland, Ohio), and the pressure level of the micropi-pette system was recorded on a multichannel recorder(Gould). The transducer was calibrated with a water ma-nometer before each experiment.

The endothelial cell suspension was loaded into the cellchamber (2 mm in height and 10 mm in width) which was

Figure 2 continued. C. After exposure for 4 hours. D. After exposure for 12 hours. Bar = 50 jim. ArrowIndicates the flow direction.



ownloaded from



held by a manipulator. The suspended cells were ob-served through a long working distance objective lens(x 20) under the microscope as shown in Figure 2. The tipof the micropipette was made to approach the surface of aspherical endothelial cell by manipulating both the cellchamber and the stage of the microscope. When a nega-tive pressure, Ap, was applied to the tip of the micropipette,the cell was drawn toward and into the pipette. To deter-mine the zero pressure level, the height of the reservoirwas changed by a slight adjustment of the micrometer soas to just maintain the cell in its initial position. First, a

negative pressure of about 2 mm H2O was set by using themicrometer, and then a portion of the endothelial cell wasaspirated into the micropipette.

In preliminary experiments, the aspirated part of the cellcontinued to deform after the application of pressure, butan almost steady state was observed to occur within 8-10minutes. Therefore, in our experiments the negative pres-sure was maintained for 10 minutes at each setting andthen increased in a stepwise fashion several times, and aphotograph of the aspirated portion of the cell was taken ateach setting by a camera connected to the microscope. In

Figure 3. Photomicrographs of rhodamine phalloidin-stained endothelial monolayers grown on Ther-manox and exposed for 24 hours to fluid shear stresses of (A) 30 and (B) 85 dynes/cm2. Bar = 50 fim.Arrow indicates flow direction.



ownloaded from



this way, the aspirated length (L) of the membrane in themicropipette was measured at several different negativepressures during the loading process. After the aspiratedlength of the cell membrane had attained a length twice theradius of the micropipette, the negative pressure was de-creased and the cell was unloaded.

In a few experiments, the recovery of the deformed por-tion of the cell membrane was measured during this un-loading process. When the pressure was returned to zero,it was determined whether the aspirated portion of the cellhad fully recovered. If the cell remained deformed at zeropressure, then the data were discarded. The image of thecell shape and the deformation process was also observedby a TV camera (Model ITC-47, Ikegami Tsushinki Com-pany, Limited, Maywood, New Jersey) and was recordedon a video tape recorder (Model NV-8050, Panasonic In-dustrial Company, Secaucus, New Jersey). After each ex-periment, the aspirated length of the cell was measured ona TV monitor screen by replaying the recorded tape or fromphotographs.

The data in the form of L versus Ap represents, in effect,a stress-strain relationship. Measurements of L versus Aphave been carried out previously by other investigators forboth red cells and white cells. The data are presented herein the form of a mechanical stiffness parameter, K =

R x Ap/(UR), where R x Ap is the tension being exertedand L/R is the non-dimensional strain imposed on the cell(L is the aspirated length in the micropipette and R is themicropipette radius). With the use of a Mine II microcom-puter, linear regression analysis was performed on individ-ual experiments using a least-squares curve fit. In all ex-periments, it was possible to assess linearity over the fullrange of pressures.

Results

The changes in en face endothelial cell shape due toexposure to shear stress for a specified time have beenshown previously.9 At the control condition of time zero,the cell shape is polygonal, and after exposure to shearstresses, the cells elongate and become oriented in thedirection of flow.

F-actin microfilament bundles in confluent monolayersgrown on a Thermanox substrate are shown in Figure 2Afor static conditions and in Figures 2B, 2C, 2D, 3A, and 3Bfor shear stress exposures. Examples of F-actin microfila-ments after a short exposure time of one-half hour for ashear stress level of 85 dynes/cm2 is presented in Figure2B. With such a short exposure time, a change in cellularmorphology was not observed with phase contrast optics.However, it was observed that the microfilament system

Figure 4. Example of cellular deformation after detachment for (A) control conditions, and after exposure for 24 hours to shear stresslevels of (B) 10, (C) 30, and (D) 85 dynes/cm2. Bar = 10 /xm.



ownloaded from



was affected. First, the dense peripheral band mostly dis-appeared and an increase in stress fibers was seen. Thestress fibers seemed to have a more organized appear-ance.

Figures 2C and 2D show F-actin in response to a fluidshear stress level of 85 dynes/cm2 for 4- and 12-hour expo-sures, respectively. The 4-hour exposure (Figure 2C)shows an increased number of stress fibers aligned withthe flow field. It was also apparent that there is a wideningof intercellular spaces with few connecting microfilaments.

With the longer exposure time of 12 hours, the stress fiberswere mostly aligned with the direction of flow, but a fewintercellular spaces were still present.

Figures 3A and 3B show F-actin microfilaments in cellsafter a 24-hour exposure to shear stress levels of 30 and85 dynes/cm2, respectively. In both cases, the stress fiberswere aligned with the direction of flow. For the two lowershear stress levels (i.e., 10 and 30 dynes/cm2) intercellularspaces were often seen with short microfilaments thatseem to connect adjacent cells together. For the highest

Figure 5. Photomicrographs of rhodamine phalloidin-stained endothelial cells after exposure to85 dynes/cm2 for 24 hours during the process of being detached from their substrate using trypsin(A), and in a completely detached cell using the mechanical method (B). Bar = 50 /xm.



ownloaded from



3.0

Eo^.CO<Dc

Q.

2.0

1.0

(1.620)

(1.148)

(0.871)

(0.303)

1.0 2.0 3.0

L/RFigure 6. Effects of shear stress on the stress-strain response ofendothelial cells. R x Ap is the tension Imposed on the cell and L/Ris the nondimenslonal length of aspiration, where L is the aspirat-ed length of cell in the micropipette and R is the radius of themicropipette. The lines represent the computerized linear regres-sion using a least-squares fit analysis. The numbers in parenthe-ses are the slopes In dyne/cm and represent the stiffness param-eter, K.

shear stress level, intercellular spaces were no longer visi-ble and there was a higher concentration of microfilamentbundles at the cell periphery.

Examples of cell deformation for endothelial cells whichhave been exposed to different shear stresses are shownin the series of photomicrographs in Figure 4. In contrast tocontrol cells, the shapes of the cells exposed to shearstress are quite different. As noted earlier, cells exposed tohigh shear stress do not become spherical in shape whendetached, but retain a more elongated shape. Further-more, the degree of elongation appears to depend on thelevel of the shear stress to which the cell has been ex-posed. Figure 5A shows cells detached using trypsin froma monolayer exposed to a shear stress level of 85 dynes/cm2 for 24 hours. As may be seen, there is little alteration incellular morphology occurring during detachment and awell organized F-actin microfilament system was pre-served. In a completely detached endothelial cell obtainedwith the mechanical method (Figure 5B), the F-actin micro-filament system was not disrupted. In addition, a narrowingof the cell seemed to occur. The cells exposed to highshear stress were stiffer than cells exposed to a low shearor to a static, no-flow condition. This is evidenced by thefact that a larger suction pressure was needed to deformthe cell to the same extent. An example of this is shown inFigure 6 where a typical loading curve is shown for cellsexposed to shear stresses of 10, 30, and 85 dynes/cm2,respectively, as well as for a cell from a control, no-flowcondition.

As may be seen from Figure 6, the relationship betweenthe pressure difference and the aspirated length is ap-proximately linear during the loading process. The slope of

such a relationship is used to define the stiffness param-eter, K. Values of the mechanical stiffness parameter, K =R x Ap/(L/R), obtained from cells exposed to shear stressfor different exposure times are shown in Figure 7. Theresults for 0.5 to 1.5 hours of exposure are presented inTable 1. These values are not presented in Figure 7 be-cause only the most elongated cells were measured andthese may not be a good representation of the total popula-tion. It should be noted that the cells exposed to a shearstress of 85 dynes/cm2 for only a few hours had a value ofthe stiffness parameter almost four times higher than thatfor control conditions. This is in contrast to a shear stress of30 dynes/cm2, where cells exposed for approximately 4hours had a relatively low value of the mechanical stiffnessparameter. However, with longer exposure times, the me-chanical stiffness increased to a level approximately threetimes higher than that for control cells. This time-history ofthe mechanical stiffness parameter, K, during the re-sponse of endothelial cells to shear stress is illustrated inFigure 7. Data for cells exposed to a shear stress of 10dynes/cm2 are also included. These cells show a similar

2.0

1.5

Eo*̂

COCDc>,

•D

CD

ECO

0.COCOCD

c

St

1.0

0.5

1.5

1.0

0.5

1.5

1.0

0.5• $

.1,

I-OH

0 }

10 dynes/cm2

i30 dynes/cm2

85 dynes/cm2

O*

I-OH

2 4 6 8 10 20

Time of Exposure ~ hrs

40

Figure 7. Effect of exposure time on the stiffness parameter forthe three shear stress levels studied. Each point represents mean± SD of 3 to 6 measurements, and the first data points to the leftrepresent the control values. (*p < 0.05; Student's /test, 2-tailed)



ownloaded from



o

CD

c

1 8

1.2

0.6

iamo>c

OT

Control

3085

0.4 0.6 0.8 1.0

Shape Index

Figure 8. Graphical representation of en face shape Index (4 nA/P2) and stiffness parameter for control cells and for cells ex-posed to different shear stress level from 4 to 24 hours. The solidline represents least-squares curve fit. Intercept = 2.69, slope =-2.74, r = -0 .9 , and p < 0.001.

trend to those at 30 dynes/cm2, having a lower value of themechanical stiffness parameter for times of 4 hours or less,but evidencing an increase for longer times. After 24 hoursof exposure to shear stress, the increase in the stiffnessparameter was statistically significant as compared to con-trol values for all three shear stress levels employed (p <0.05, two-tailed Student's Mest).

Discussion

Endotheiial cells in vivo are aligned with the direction offlow and have an elongated tear drop shape.5"7 Thesephenomena (i.e., cell orientation and elongation) have alsobeen reported by others for in vitro cultured endotheiialcells exposed to fluid shear stress.8'a Although much hasbeen learned about the effect of fluid mechanics on endo-theiial elongation and orientation, little is known about oth-er properties of endotheiial cells. However, since a cell'smechanical properties may be related to the process of itselongation and orientation, this investigation has focusedon endotheiial cell mechanical properties. In their prelimi-nary studes, Dewey et al.B did show that certain endotheiialcell functions, including fluid endocytosis, cytoskeletal as-sembly, and nonthrombogenic surface properties, are sen-Table 1. Stiffness Parameters of Cultured BovineAortic Endotrwllal Cells after V* to 1V» Hours ofExposure to Shear Stress

Shearstress

(dynes/cm2)

0*103085

Stiffnessparameter(dyne/cm)

0.334±0.1080.874 ±0.2700.931 ±0.2701.360 ±0.480

N

11658

p value

<0.03<0.01<0.002

Values are means ± SD; p values are from Student's 2-tailed rteat.

'Mean of all controls.N = the number of observations.

sitive to shear stress. Frangos et al.29 demonstrated en-hanced PGI2 production in endotheiial cells exposed toshear stress. These results taken together suggest thatfluid mechanical forces can directly influence endotheiialcell structure and function.

The experiments presented here were conducted underlaminar, steady flow which is in contrast to in vivo condi-tions where pulsatile flow exists. However, it is recognizedthat other types of flow (e.g., pulsatile, oscillatory, turbu-lent) may affect endotheiial cell properties differently. Thecomplexity of the flow fields would make it difficult to as-sess the effect of shear stress abne since other ftow com-ponents may indeed interfere with endotheiial cell biology.As a first step, it thus is appropriate to evaluate the effect ofshear stress on endotheiial cell mechanical properties us-ing well defined steady flow conditions.

The results presented suggested that exposure to shearstress exerted such an alteration in the microfilament net-work that even after detachment, the elongation of theendotheiial cell was preserved. This is in contrast to cellsunder a no-flow condition which, upon detachment, be-came spherical. In the latter case, there was some evi-dence that a submembranous cortical layer was formed.30

For this latter case, the analysis of the mechanical proper-ties of these detached spherical endotheiial cells was car-ried out using a membrane type of equation as originallyproposed by Evans31 and Chien et al.23 In this investiga-tion, however, the shear-exposed cells did not becomespherical upon detachment. Thus, their behavior could notbe analyzed using this previous approach.30 As a first stepin the analysis of these data, a mechanical stiffness pa-rameter, K, was used. This parameter represents the slopeof the variation in the tension, R x Ap, with the nondimen-sional length of aspiration, L/R. It is very similar to theapproach used in previous erythrocyte, leukocyte, andplatelet micropipette studies.32"34

Along with the change in endotheiial cell shape, it hasbeen observed that endotheiial cells become stiffer in re-sponse to shear stress. The effect of shear stress onshape as a function of time was reported previously.9 Thechanges with time in cell mechanical properties appear tobe due to certain structural changes in the cytoskeletonwhich are a response to shear and are discussed below.The relationship between endotheiial cell shape and themechanical stiffness parameter is demonstrated quantita-tively In Figure 8 where a correlation between the en facecell shape index and the stiffness parameter is presented.

The increase in cell stiffness may parallel the reorgani-zation of the cytoskeleton and, in particular, the microfila-ment system. In an independent study, we have investigat-ed the Importance of two components of the cytoskeleton,(e.g., microfilaments and microtubules) in relation to endo-theiial cell mechanical properties (unpublished results).Endotheiial monolayers under no-flow conditions were in-cubated with cytochalasin B and colchicine to disrupt themicrofilament or microtubular system respectively. Wefound that microfilaments contributed more to endotheiialcell mechanical properties than microtubules. This sug-gests that, in the presence of shear stress, the microfila-ments may make the greater contribution to endotheiialcell mechanical properties.



ownloaded from



Even exposure to shear stress for short times had adramatic effect on the F-actin microfilament system. Thiswas also true for the stiffness parameter of some cells(Table 1). With regard to the latter, it was felt that theseresults may not be representative of the actual shearstress effect for at least two reasons. First, it was the mostdeformed cells which were selected for measurement.Second, it was observed that most of the cells selected forshort exposure times did not become spherical after de-tachment, but acquired a rather potato-type appearancewith the formation of a cortical cytoskeletal layer. It is sus-pected that this cortical layer may have contributed to theincrease in cell stiffness, which may be accounted for bythe increase in stress fibers shown in Figure 2. This isconsistent with our previous work on passage effect wherewe found that cells from older monolayers were stiffer, butalso appeared to have a higher concentration of stressfibers.30 Franke et al.35 have shown that for human vascu-lar endothelium in vitro, fluid shear stress acts directly onthe stress fiber system without affecting cellular shape andorientation. They exposed cell monolayers to a shearstress level of 2 dynes/cm2 for 3 hours and observed thatthe quantity and intensity of stress fibers had dramaticallyIncreased. They also observed, as shown here, that therewas a preferential alignment, or direction, to the newlyformed stress fibers.

The results obtained for F-actin microfilaments are gen-erally consistent with the results reported by others,38"37

although others have not looked at the effect of both shearstress level and exposure time on stress fiber reorganiza-tion in as detailed a fashion. White et al.38 looked at theorientation of the endothelial cytoskeleton using culturedbovine aortic endothelial cells under both static, no-flow,and shear stress conditions in vitro. Cultured cells understatic, no-flow conditions exhibited a random stress fiberorientation. However, after exposure to 8 dynes/cm2 for 72hours, these cells became aligned with the flow direction,and their cytoskeletal structure underwent a dramatic reor-ganization; stress fibers appeared much more prominentand the cells showed an orientation with the major cell axisand with the direction of flow. The organization of the extra-cellular matrix beneath the endothelial monolayer also ap-peared to be influenced by shear stress. Similar resultswere reported by Wechezak et al.37 for fibronectin and F-actin redistribution in cultured endothelial cells exposed toshear stress.

In addition to the above, we have demonstrated that thedense peripheral band is also influenced by shear stress.For static endothelial cell monolayers, it has been suggest-ed that the dense peripheral band may be contractile andhelp maintain cell shape and yet at the same time beintimately associated with the ability of the cell to spreadafter wounding.27 This dual function is thought to reflect thedemands of the endothelial cell environment. Our resultsare perhaps more compatible with the in vivo condition,and it would seem that endothelial cells in vivo must re-main closely appositioned to each other in order to functionas a selective barrier. Figure 3 shows that higher shearmay contribute not only to an increase in stress fiber densi-ty, but also to a tightening of the monolayer. For example atthe highest shear stress, individual cells are not distin-

guishable, whereas at the lower shear stresses, not onlyare cells distinguishable but also there are intercellularspaces devoid of stress fibers. This suggests a possibleinfluence of shear stress on intercellular junctions.

The occurrence and distribution of stress fibers also hasbeen examined in vivo using mouse and rat endothelialcells In perfusion-fixed large arteries and veins.12 As anexample, in the lower thoracic aorta, the stress fibers werestrongly aligned with the flow, while in the inferior venacava there were very few stress fibers with no preferredorientation. Since the latter might be expected to havelower shear stresses than the former, these results alsosuggest that local hemodynamic factors may influenceboth the abundance and orientation of such contractile,cytoskeletal elements. The same results were obtained forvertebrate tissues by Wong et al.13 They concluded thatthe shearing force of blood may be a determinant of endo-thelial cell stress fiber elaboration and orientation.

The present results also may have implications for trans-endothelial transport, a subject which has been of interestin the study of the influence of wall shear stress on theblood-arterial wall transport of macromolecules.3'M'M Inthese earlier experiments, the accumulation of materials inthe wall was measured, but the precise pathway throughthe endothelium and into the intima was not known. Morerecently Davies et al.40 examined the effect of shear stresson endothelial cell pinocytosis using cultured bovine aorticendothelial cells. The intracellular horseradish peroxidaseaccumulation in confluent endothelial monolayers ex-posed to 0 ,1 , or 8 dynes/cm2 shear stress was measured.These results demonstrated that the rate of fluid endocyto-sis was enhanced in response to a step-change in shearstress.

Although somewhat speculative, it might be expectedthat at an early stage in atherogenesis the cytoskeleton ofendothelial cells could play an important role in determin-ing the permeability of endothelium to macromolecules,e.g., via transendothelial and junctional transport. Thus,cells in high shear regions, in response to their mechanicalstress environment, would have a more highly developedcytoskeletal structure which could inhibit both transendo-thelial and junctional macromolecule transport. The resultwould be that low shear regions, with their higher influx ofmacromolecules like LDL, would have a higher predilec-tion for the initiation of lesions.

Finally, the present data and that of others suggest thatthe response of an endothelial cell to shear stress includesan alteration in both its structure and its mechanical prop-erties. More detailed comparisons of structure and cellfunction between in vivo and cultured endothelial cells areneeded to apply the results of in vitro observations of me-chanical properties to the endothelial cells of living bloodvessels.

AcknowledgmentsThese experiments were carried out in the Bioengineering Re-

search Center, University of Houston, Houston, Texas.Special thanks are due to Robert Faith (University of Houston)

for the use of his Nikon fluorescent microscope and to Ira M.Herman (Tufts University) for many stimulating discussions andIdeas on endothelial cell cytoskeleton.



ownloaded from



References1. Nerem RM, Cornhill JF. The role of fluid mechanics in ather-

ogenesis. J Biomech Eng 1980;102:181-1892. Ross R. Atherosclerosis: a problem of the biology of arterial

wall cells and their interaction with blood components. Arte-riosclerosis 1981 ;1:293-311

3. Fry DL. Acute vascular endothelial changes associated withincreased blood velocity. Circ Res 1968;22:165-197

4. Caro CG, Fitz-Gerald JM, Schroter RC. Atheroma and arte-rial wall shear — observation, correlation and proposal of ashear-dependent mass-transfer mechanism for atherogene-sis. Proc R Soc Lond [Biol] 1971 ;177:109-159

5. Flaherty JT, Pierce JE, Ferrans VJ, Patel DJ, Tucker WK,Fry DL. Endothelial nuclear patterns in the canine arterialtree with particular reference to hemodynamic events. CircRes 1972;31:23-33

6. Silkworth JB, Stehbens W. The shape of endothelial cells inen face preparations of rabbit blood vessels. Angiology1975;26:474-487

7. Nerem RM, Levesque MJ, Cornhill JF. Vascular endothelialmorphology as an indicator of the pattern of blood flow. JBiomech Eng 1981;103:172-176

8. Dewey CF Jr, Bussolari SR, Gimbrone MA Jr, Davies PF.The dynamic response of vascular endothelial cells to fluidshear stress. J Biomech Eng 1981;103:177—185

9. Levesque MJ, Nerem RM. Elongation and orientation of cul-tured endothelial cells in response to shear. J Biomech Eng1985:341-347

10. Levesque MJ, Liepsch D, Moravec S, Nerem RM. Correla-tion of endothelial cell shape and wall shear stress in a sten-osed dog aorta. Arteriosclerosis 1986;6:220-229

11. Rogers KA, Kalnins VI. Comparison of the cytoskeleton inaortic endothelial cells in situ and in vitro. Laboratory Investi-gation 1983:49:650-654

12. White GE, Gimbrone MA Jr, Fujiwara K. Factors influenc-ing the expression of stress fibers in vascular endothelial cellsin situ. J Cell Biol 1983:97:414-424

13. Wong AL, Pollard TD, Herman IM. Actin filament stressfibers in vascular endothelial cells in vivo. Science 1983;219:867-869

14. Fung YC. Biomechanics-mechanical properties of living tis-sues. New York: Springer-Verlag, 1981:122-131

15. Yoneda M. Tension at the surface of sea-urchin egg: a criticalexamination of Cole's experiment. J Exp Biol 1964;41:893-906

16. Gregersen Ml, Bryant CA, Hammerle WE, Usami S, ChienS. Flow characteristics of human erythrocytes through poly-carbonate sieves. Science 1967;157:825-827

17. Hochmuth RM, Mohandas N. Uniaxial loading of the red cellmembrane. J Biomech 1972;5:501-509

18. Mela MJ. Elastic-mathematical theory of cells and mitochon-dria in swelling process. I. The membranous stresses andmodulus of elasticity of the egg cell sea urchin. Biophys J1967;7:95-110

19. Fung YCB, Tong P. Theory of the sphering of red blood cells.Biophys J 1968:8:175-198

20. Hiramoto Y. Rheological properties of sea urchin eggs. Bio-rheology 1970:6:201-234

21. Evans EA. Bending elastic modulus of red blood cell mem-brane derived from buckling instability in micropipet aspira-tion tests. Biophys J 1983:43:27-30

22. Zarda PR, Chien S, Skalak R. Elastic deformations of redblood cells. J Biomech 1977;10:211-221

23. Chien S, Sung K-LP, Skalak R, Usami S, Tozeren A. Theo-retical and experimental studies on viscoelastic properties oferythrocyte membrane. Biophys J 1978;24:463-487

24. Hochmuth RM, Buxbaum KL, Evans EA. Temperaturedependence of the viscoelastic recovery of red cell mem-brane. Biophys J 1980;29:177-182

25. Schmid-Schonbein GW, Sung K-LP, Tozeren H, Skalak R,Chien S. Passive mechanical properties of human leuko-cytes. Biophys J 1981:36:243-256

26. Cornhill JF, Levesque MJ, Herderick EE, et al. Quantitativestudy of the rabbit aortic endothelium using vascular casts.Atherosclerosis 1980:35:321-327

27. Wong MKK, Gotlieb Al. Endothelial cell monolayer integrity.1. Characterization of dense peripheral band of microfila-ments. Arteriosclerosis 1986:6:212-219

28. Eskin SG, Ives CL, Mclntire LV, Navarro LT. Response ofcultured endothelial cells to steady flow. Microvasc Res1984:28:87-94

29. Frangos JA, Mclntire LV, Eskin SG, Ives CL. Flow effectson prostacyclin production by cultured endothelial cells. Sci-ence 1985:227:1477-1479

30. Sato M, Levesque MJ, Nerem RM. An application of themicropipette technique to the measurement of the mechani-cal properties of cultured bovine aortic endothelial cells. JBiomech Eng 1987;109:27-34

31. Evans EA. New membrane concept applied to the analysis offluid shear and micropipette-deformed red blood cells.Biophys J 1973:13:941-954

32. Evans EA, LaCelle P. Intrinsic material properties of theerythrocyte membrane indicated by mechanical analysis ofdeformation. Blood 1975:45:29-43

33. Chien S, Schmid-Schonbein GW, Sung K-LP, SchmalzerEA, Skalak R. Viscoelastic properties of leukocytes. In: Mei-selman HJ, Lichtman MA, LaCelle PL, eds. White cell me-chanics; Basic science and clinical aspect. New York: Alan R.Liss, 1984;19-51

34. Burris SM, Smith CM II, Tukey DT, Clauson CC, White JG.Micropipette aspiration of human platelets after exposure toaggregating agents. Arteriosclerosis 1986;6:321-325

35. Franke RP, Grafe M, Schnitter H, Seiffge D, MittermayerC. Induction of human vascular endothelial stress fibers byfluid shear stress. Nature 1984;307:648-649

36. White GE, Fuijiwara K, Shefton E, Dewey CF Jr, GimbroneMA Jr. Fluid shear stress influences cell shape and cyto-skeletal organization in cultured vascular endothelium [abstr].Fed Proc 1982;41:321

37. Wechezak AR, Viggers RF, Sauvage LR. Fibronectin andF-actin redistribution in cultured endothelial cells exposed toshear stress. Lab Invest 1985:6:639-647

38. Caro CG, Nerem RM. Transport of 14C-4-cholesterol be-tween serum and wall in perfused dog common carotid artery.Circ Res 1973;3:197-210

39. Nerem RM, Mosberg AT, Schwerin WD. Transendothelialtransport of 131 I-Albumin. Biorheology 1976;13:71-77

40. Davies PF, Dewey CF Jr, Bussolari SR, Gordon EJ, Gim-brone MA Jr. Influence of hemodynamic forces on vascularendothelial function. In vitro studies of shear stress and pino-cytosis in bovine aortic endothelial cells. J Clin Invest 1984;73:1121-1129

Index Terms: atherosclerosiswall shear stress

cell mechanical properties • cytoskeleton • endothelium • hemodynamics



ownloaded from


M Sato, M J Levesque and R M NeremMicropipette aspiration of cultured bovine aortic endothelial cells exposed to shear stress.

Print ISSN: 1079-5642. Online ISSN: 1524-4636 Copyright © 1987 American Heart Association, Inc. All rights reserved.

Avenue, Dallas, TX 75231is published by the American Heart Association, 7272 GreenvilleArteriosclerosis, Thrombosis, and Vascular Biology

doi: 10.1161/01.ATV.7.3.2761987;7:276-286Arterioscler Thromb Vasc Biol.

http://atvb.ahajournals.org/content/7/3/276World Wide Web at:

The online version of this article, along with updated information and services, is located on the

http://atvb.ahajournals.org//subscriptions/

at: is onlineArteriosclerosis, Thrombosis, and Vascular Biology Information about subscribing to Subscriptions:

http://www.lww.com/reprints

Information about reprints can be found online at: Reprints:

document.Permissions and Rights Question and AnswerFurther information about this process is available in theis being requested is located, click Request Permissions in the middle column of the Web page under Services.Clearance Center, not the Editorial Office. Once the online version of the published article for which permission

can be obtained via RightsLink, a service of the CopyrightArteriosclerosis, Thrombosis, and Vascular Biology Requests for permissions to reproduce figures, tables, or portions of articles originally published inPermissions:



ownloaded from

http://atvb.ahajournals.org/content/7/3/276

http://www.ahajournals.org/site/rights/

http://www.lww.com/reprints

http://atvb.ahajournals.org//subscriptions/


micropipette aspiration of cultured bovine aortic...

Documents