Effects of Flow Patterns on the Localization and Expression of VE-Cadherin at Vascular Endothelial...

Research Paper

J Vasc Res 2005;42:77–89DOI: 10.1159/000083094

Effects of Flow Patterns on the Localization andExpression of VE-Cadherin at VascularEndothelial Cell Junctions: In vivo and in vitroInvestigations

Hui Miao Ying-Li Hu Yan-Ting Shiu Suli Yuan Yihua ZhaoRoland Kaunas Yingxiao Wang Gang Jin Shunichi Usami Shu Chien

Departments of Bioengineering and Medicine, and The Whitaker Institute of Biomedical Engineering,University of California, San Diego, La Jolla, Calif., USA

Received: March 15, 2004Accepted after revision: October 7, 2004Published online: January 3, 2005

Shu Chien, MD, PhDDepartment of Bioengineering, PFBH 134University of California, San Diego, 9500 Gilman DriveLa Jolla, CA 92093-0412 (USA)Tel. +1 858 534 5195, Fax +1 858 534 5453, E-Mail [email protected]

ABCFax + 41 61 306 12 34E-Mail [email protected]

© 2005 S. Karger AG, Basel1018–1172/05/0421–0077$22.00/0

Accessible online at:www.karger.com/jvr

Key WordsEndothelial cell W Flow W In vivo W Stenosis W VE-cadherin

AbstractAtherosclerosis occurs preferentially at vascular curva-ture and branch sites where the vessel walls are exposedto fluctuating shear stress and have high endothelial per-meability. Endothelial permeability is modulated by in-tercellular adhesion molecules such as VE-cadherin. Thisstudy was designed to elucidate the effects of differentflow patterns on the localization and expression of VE-cadherin in endothelial cells (ECs) both in vivo and invitro. VE-cadherin staining at EC borders was muchstronger in the descending thoracic aorta and abdominalaorta, where the pulsatile flow has a strong net forwardcomponent than in the aortic arch and the poststenoticdilatation site beyond an experimental constriction,where the flow near the wall is complex and reciprocat-ing with little net flow. With the use of flow chambers theeffects of pulsatile flow (12 B 4 dyn/cm2 at 1 Hz) and reci-procating flow (0.5 B 4 dyn/cm2 at 1 Hz) on VE-cadherinorganization in endothelial monolayers were studied invitro. VE-cadherin staining was continuous along cell

borders in static controls. Following 6 h of either pulsatileor reciprocating flow, the VE-cadherin staining at cellborders became intermittent. When the pulsatile flowwas extended to 24, 48 or 72 h the staining around thecell borders became continuous again, but the stainingwas still intermittent when the reciprocating flow wassimilarly extended. Exposure to pulsatile or reciprocat-ing flow for 6 and 24 h neither change the expressionlevel of VE-cadherin nor its distribution between mem-brane and cytosol fractions as determined by Westernblot and compared with static controls. These findingssuggest that the cell junction remodeling induced by dif-ferent flow patterns may result from a redistribution ofVE-cadherin within the cell membrane. Both the in vivoand in vitro data indicate that pulsatile and reciprocatingflow patterns have different effects on cell junctionremodeling. The lack of junction reorganization in re-gions of reciprocating flow in vivo and in vitro may pro-vide a mechanistic basis for the high permeability andthe preferential localization of atherosclerosis in regionsof the arterial stress with complex flow patterns and fluc-tuating shear stress.

Copyright © 2005 S. Karger AG, Basel

78 J Vasc Res 2005;42:77–89 Miao/Hu/Shiu/Yuan/Zhao/Kaunas/Wang/Jin/Usami/Chien

Introduction

Atherosclerotic lesions are characterized by the accu-mulation of cholesterol which is derived from plasmalipoproteins in the intercellular matrix [1]. Atherosclero-sis tends to occur in regions of curvature (such as the archof the aorta), branches and stenosis segments in thearterial tree [2–4]. The geometries in these arterial re-gions are associated with blood flow disturbances such asflow separation, recirculation, complex secondary flowpatterns, and nonuniform shear stress distributions [2–7]. Vascular endothelial cells (ECs) which form the inter-face between the circulating blood and the arterial wallare constantly exposed to hemodynamic forces, includingfluid shear stress and mechanical stretch. Shear stress,the tangential component of hemodynamic forces actingon the ECs, is an important modulator of vascular cellu-lar functions [4, 8–11] including endothelial permeability[12, 13]. There is ample evidence that endothelial perme-ability of arteries plays a significant role in the formationof atherosclerosis [1, 6, 14–17]. Endothelial cell-cell junc-tions control the intercellular permeability to plasma sol-utes and circulating cells [18], and their integrity dependson the structure and function of adhesive proteins in-cluding VE-cadherin [18–23]. It has been shown that VE-cadherin distribution and expression are regulated byshear stress in in vitro systems [24–26], but the questionof whether different flow patterns play a role in these pro-cesses has not been addressed. We investigated the roleof different flow patterns in the remodeling of cell junc-tions in both in vivo and in vitro conditions. The resultsof the present study showed that VE-cadherin staining atEC borders is much stronger in the descending thoracicaorta and abdominal aorta where simple laminar flowprevails than in the aortic arch and poststenotic dilata-tion sites where complex flow patterns and fluctuatingshear stress dominate. Our in vitro studies using a flowchamber demonstrated that VE-cadherin staining at ECborders became intermittent with numerous gaps afterexposing monolayers of bovine aortic ECs (BAECs) toeither pulsatile or reciprocating flow for 6 h. When theexposure to a pulsatile flow was extended to 24 h or lon-ger, continuous VE-cadherin staining reappeared aroundthe entire periphery of the cells but not under reciprocat-ing flow. Pulsatile or reciprocating flow neither changethe protein expression level of VE-cadherin as deter-mined by Western blotting nor its distribution betweenmembrane and cytosol fractions.

In summary, we demonstrate that flow patterns differ-entially regulate EC junctions both in vivo and in vitro.

Pulsatile flow with a strong forward component is associ-ated with well-formed VE-cadherin junctions, whereascomplex flow patterns with fluctuating shear stress is asso-ciated with poorly-formed VE-cadherin junctions. Ourresults also suggest that a redistribution of VE-cadherinwithin the cell membrane mediates the cell junctionremodeling induced by pulsatile flow. Thus, our findingsindicate that flow pattern-regulated junction remodelingmay modulate endothelial permeability and hence thepreferential localization of atherosclerosis in the regionswith complex or disturbed flow patterns.

Materials and Methods

Animals Adult male Sprague-Dawley rats weighing 350–400 g were used.

Six of these rats were used for the study of VE-cadherin expression onthe ECs of different regions of the normal aorta including aortic arch,thoracic aorta, and abdominal aorta. Six other rats were used to pro-duce the aortic stenosis in animal model. The experimental protocolwas approved by the Animal Subject Committee of University ofCalifornia, San Diego. All rats received human care in accordancewith the animal use principles of the American Physiological Society.All rats were maintained under constant environmental temperature(21 B 1°C), humidity (60 B 5%) and light/dark cycle, and had freeaccess to water and normal rat diet.

Animal Model of Aortic Stenosis Stenosis of abdominal aorta was produced by using a U-shaped

titanium clip, as described by Zand et al. [27] and Joris et al. [28].Briefly, following anesthetization with intraperitoneal ketamine(100 mg/kg body weight) and xylazine (10 mg/kg body weight), therat was laid supine, the abdominal cavity was opened, and a segmentof abdominal aorta below the renal arteries and above the iliolumbarbranches was cleared of surrounding tissue. The clip was held with apair of forceps and placed on the abdominal aorta. The extent ofclipping was controlled by placing a stopper of given size between thetwo arms of the forceps. The abdominal wound was closed and therat was returned to the cage. Thirty days after surgery, the abdominalaorta was harvested and prepared as described below.

Preparation of Aortic Tissue The rats were anesthetized as mentioned above and anticoagu-

lated with intravenous injection of heparin (100 USP U/kg). Threeminutes later, the aorta was washed free of blood by perfusing in situwith 20 ml of phosphate-buffered saline (PBS) at a pressure of120 mm Hg, with the right atrium opened to allow a free flow. Theaortic tree was then perfusion fixed for 1 h with 4.0% paraformalde-hyde in PBS at a pressure of 120 mm Hg. The fixed aorta was dis-sected free from surrounded tissues and slit opened longitudinally(fig. 1A). The aorta was pinned flat (with the intimal surface up) on apolyethylene sheet (fig. 1B) and immersed into PBS containing 4.0%paraformaldehyde for further fixation. The aortic tissue at theclipped segment was prepared by similar procedures.

Effect of Flow Patterns on VE-CadherinExpression

J Vasc Res 2005;42:77–89 79

Fig. 1. Schematic diagrams of the rat thorac-ic aorta. Showing its overall anatomy (A)and its luminal surface after having been cutopen and pinned flat (B), the labels a, b, cand d in B indicate the areas from which thepictures for VE-cadherin staining in figure 2were taken.

Brachiocephalic trunk

Left commoncarotid artery

Left subclavianartery

Line of section

Brachiocephalic trunk

Left commoncarotid artery

Left subclavianartery

Intercostalarteries

A B

cd

a b

Immunostainning of VE-Cadherin on Aortic Endothelium Following permeabilization with 0.5% Triton X-100 in PBS for

15 min, the fixed aorta was treated with 3.0% H2O2. Following wash-ing 3 times with PBS and incubation with 1.5% horse serum (to blocknonspecific sites), the specimens were incubated sequentially with ananti-VE-cadherin antibody (Santa Cruz Biotechnology), a biotiny-lated secondary antibody (Vector Laboratories Inc., Burlingame,Calif., USA), Avidin DH, and biotinylated horseradish peroxidase Hreagents (Vectastain ABC reagents; Vector Laboratories Inc.). Thespecimens were then treated with the enzyme substrate diaminoben-zidine tetrahydrochloride to develop VE-cadherin staining. The ECnuclei were stained with hematoxylin and eosin. The Häutchen pro-cedure for transferring the intact sheet of ECs onto glass slides wasperformed after staining [29]. The en face VE-cadherin staining ofECs of (a) the aorta arch and descending thoracic aorta from 6 nor-mal rats and (b) abdominal aorta from 6 rats with aortic stenosis(including sites without stenosis, the stenotic site, and the poststenot-ic dilatation site) was captured by using a Hamamatsu digital camera(Hamamatsu Photonics, Japan). The digital image files were writtento CD as TIFF files and processed in Photoshop. Only minor adjust-ments of contrast and brightness were made, which in no case alteredthe appearance of the original materials. These images were con-verted into PowerPoint for presentations in figures.

Cell Cultures and Shear Stress Experiment BAECs were isolated from bovine aorta and maintained in Dul-

becco’s modified Eagle’s medium supplemented with 10% fetalbovine serum (Omega, Tarzana, Calif., USA). All cell cultures werekept in a humidified 95% air, 5% CO2 incubator at 37°C. A flowsystem was used to impose fluid shear stress on cultured BAECs [30,31]. In brief, a 75 ! 38 mm glass slide with cultured BAECs wasmounted in a flow chamber, where a rectangular flow channel wascreated by sandwiching a silicone gasket with a central cutoutbetween the glass slide and a custom-made acrylic plate. To generatepulsatile and reciprocating flows with sinusoidal wave forms, amotor-driven syringe pump was connected to the inlet of the flowchannel to impose a sinusoidal reciprocating motion at a frequency

of 1 Hz and with an amplitude of 4 dyn/cm2 [31]. For ‘pulsatile flow’studies, BAECs were subjected to a mean shear stress of 12 dyn/cm2

superimposed on this sinusoidal flow (B4 dyn/cm2). For ‘reciprocat-ing flow’ studies, BAECs were subjected to the same sinusoidal flowof B4 dyn/cm2 with the superimposition of a very low mean shearstress of 0.5 dyn/cm2 to ensure nutrient delivery and gas exchange.The system was kept at 37°C and equilibrated with 95% humidifiedair with 5% CO2. Static control experiments were performed onBAECs maintained under identical conditions without shear stressexposure. At the end of experiments, BAECs were fixed for immu-nostaining or lysed for VE-cadherin expression/translocation assays.For the disturbed flow studies, the step flow chamber was con-structed as we previously reported [31]. Briefly, disturbed flow wascreated by a step expansion of the height of the flow channel in aparallel-plate chamber. The disturbed flow pattern is shown schemat-ically in figure 4. Position ‘a’ is the stagnation point where the shearstress is zero; ‘b’ is the center of the vortex where the shear stress ishigh, but in a direction opposite to the downstream laminar flow;‘c’ is the flow-reattachment area where the shear stress is near zerobut the shear stress gradient is high, and ‘d’ is the downstream lami-nar flow area where the shear stress is high but has no spatial gra-dient. The vortex size is defined as the distance between the foot ofthe step and the reattachment line (fig. 4A–C).

Immunocytochemistry After exposure to pulsatile or reciprocating flow for different

durations of time, the BAECs were fixed in 4.0% paraformaldehydein PBS for 30 min, followed by permeabilization with 0.5% TritonX-100 in PBS for 10 min, and nonspecific binding was blocked by1.5% horse serum. After washing 3 times with PBS, the specimenswere incubated with a goat anti-VE-cadherin in combination with amouse anti-ß-catenin (Santa Cruz Biotechnology). This was followedby incubation with a tetramethylrhodamine isothiocyanate-conju-gated anti-goat IgG (Sigma) in combination with a FITC-conjugatedanti-mouse IgG (Sigma). The slides were mounted with the Slow-Fade™ antifade reagent (Molecular Probes, Inc.) and staining wasobserved under a confocal microscopy system equipped with a Per-


kin Elmer Kypton-Argon laser scanning confocal imaging system(MRC-1024, Bio-Rad, Hercules, CA). A 40! oil immersion objec-tive was used to examine the specimens. By stepping the objectivethrough the depth of the specimen from the cell surface to the cellbottom, a z-series collection of ten optical sections was obtained andprojected to construct the 3D cell image in each sample. The thick-ness of each z-section was 0.5 Ìm. Rhodamine was excited at a wave-length of 568 nm and detected at 600 nm. FITC was excited at488 nm and detected at 525 nm. By stepping the objective throughthe depth of the specimen from the cell surface to the cell bottom, az-series collection of optical sections was obtained and projected toconstruct the 3D cell image. Double-stained specimens were pseudo-colored separately by a Simple PCI image analysis software of C-Imaging Systems (Compix Inc., Cranberry Township, Pa., USA).The images were transferred to a computer for further analysis; theAdobe Photoshop was used to generate RGB images to depict theVE-cadherin in red and the expressed ß-catenin in green.

Immunoblotting of VE-Cadherin and ß-Catenin To study the effects of different flow patterns on the protein level

of VE-cadherin, BAECs were lysed with a lysis buffer containing25 mM HEPES, pH 7.4, 0.5 M NaCl, 5 mM EDTA, 5 mM NaF,1 mM Na3VO4, 1 mM phenylmethylsulfonyl fluoride, 10 mg/ml leu-peptin, 2 mM ß-glycerophosphate, and 1% Triton X-100. For VE-cadherin expression, 20 Ìg of cell lysate was suspended in SDS sam-ple buffer containing 50 mM Tris-HCl, pH 6.8, 100 mM dithiothrei-tol, 2% SDS, 0.01% bromophenol blue and 10% glycerol, denaturedfor 5 min by incubating at 95°C, and separated by SDS-PAGE. Theproteins in the gel were transferred to a nitrocellulose membrane.The membrane was incubated with 3% nonfat milk in TTBS (10 mMTris-HCl, pH 7.4, 150 mM NaCl and 0.05% Tween-20) at room tem-perature for 1 h, followed by incubation with the anti-VE-cadherinantibody in TTBS containing 1% bovine serum albumin for 2 h.After washing 3 times with TTBS the bound primary antibodies weredetected by using the horseradish peroxidase-conjugated anti-goatIgG and visualization with an enhanced chemiluminescence system(Amersham, Arlington Heights, Ill., USA). To ensure equal proteinloading, the membrane containing VE-cadherin was reprobed with aß-actin monoclonal antibody (Sigma), followed by blotting withhorseradish peroxidase-conjugated anti-mouse IgG (Santa Cruz).The procedures of immunoblotting of ß-catenin are the same as thatof VE-cadherin, except that the anti-ß-catenin antibody was usedinstead of anti-VE-cadherin antibody.

VE-Cadherin Distribution in Different Cellular Compartments The distribution of VE-cadherin in different cellular compart-

ments was determined as previously described [32, 33] with somemodification. Briefly, after exposure to pulsatile or reciprocatingflow for various durations, BAECs were scraped into a lysis buffercontaining 50 mM HEPES, pH 7.4, 50 mM NaCl, 1 mM MgCl2,2 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 10 mg/ml leu-peptin, 1 mM Na3VO4, 5 mM NaF, and 1 mM dithiothreitol. The celllysates were passed through a 27-gauge needle 5 times to disrupt cellmembranes and release cytosol proteins. The lysate was ultracentri-fuged at 105 g for 1 h. After collecting the supernatant as the cytosolfraction, the pellet was resuspended in 1% Triton X-100 in the lysisbuffer, and centrifuged at 104 g for 10 min. The supernatant after thiscentrifugation was collected as the membrane fraction. The amountof VE-cadherin in each fraction was determined by immunoblottingwith a polyclonal anti-VE-cadherin antibody.

Results

Surface Expression of VE-Cadherin on AorticEndothelial Cells To assess the effects of different flow patterns on the

surface expression of VE-cadherin in vivo, we comparedthe immunostainings of VE-cadherin on rat aortic ECs atsites where laminar flow prevails (the descending thoracicaorta) with sites where fluctuating shear stress dominates(the aortic arch). The VE-cadherin staining at EC borderswas much stronger in the descending thoracic aorta(fig. 2A for ventral and fig. 2B for dorsal aspects) than inthe aortic arch (fig. 2C for superior and fig. 2D for inferioraspects). Since the ECs of these sites are exposed to thesame chemical environment the difference observed inVE-cadherin staining is most likely the result of the differ-ence in blood flow patterns between the curved (aortaarch) and straight (descending thoracic aorta) parts of theartery.

Surface Expression of VE-Cadherin on the StenoticSegment of Abdominal Aorta To investigate whether changing the simple, laminar

flow in the rat abdominal aorta to a more complex, dis-turbed flow can modulate VE-cadherin expression invivo, a local stenosis was created by using a U-shapedtitanium clip (fig. 3A). The abdominal aorta with the ste-

Fig. 2. Immunostaining of VE-cadherin of en face preparations ofdifferent sites of rat thoracic aorta. Following permeabilization of theaortic specimen with 0.5% Triton X-100 in PBS for 15 min, VE-cadherin was stained with an anti-VE-cadherin antibody. The intactsheet of endothelial cells was prepared by the Häutchen method. TheVE-cadherin staining at endothelial cell borders was much strongerin the descending thoracic aorta with laminar flow (A ventral, B dor-sal aspects) than in the aortic arch (C superior, D inferior aspects)where the disturbed flow dominates. Images shown here are repre-sentative results from thoracic aortae of 6 rats. Bar in D = 30 Ìm.Fig. 3. Effects of local stenosis on VE-cadherin distribution at celljunctions. A Sagittal view of the abdominal aorta showing the site oflocal constriction with a U-shaped clip. B Abdominal aorta pinnedflat. After 30 days of stenosis, VE-cadherin was stained with an anti-VE-cadherin antibody as described above (blood flow is from left toright). VE-cadherin was highly expressed at endothelial cell bordersin the segment of the abdominal aorta without stenosis (C: 5 mmupstream to the clip site). At the site of the constriction (D), the cellswere elongated and VE-cadherin expression at cell borders was mark-edly reduced. No detectable VE-cadherin was found at cell borderson the downstream side of the constriction, where the flow pattern iscomplex (E). Images shown here are representative results from ste-nosis of abdominal aortae of 6 rats. Bar in E = 30 Ìm.


J Vasc Res 2005;42:77–89 81

2

3


Fig. 4. VE-cadherin staining in endothelial cells exposed to disturbed and laminar flows in vitro. Confluent mono-layers of BAECs were kept as controls or subjected to disturbed and laminar flows in a step flow chamber. Staining ofVE-cadherin was observed by confocal microscopy. VE-cadherin staining was continuously distributed around theentire periphery of cells under static condition (A). After a 24-hour exposure to the disturbed flow, the VE-cadherindistribution became intermittent (B), whereas 24-hour exposure to laminar flow resulted in a continuous distributionof VE-cadherin (C), as in the static condition. Images shown here are representative results from 3 separate experi-ments. Bar in A = 30 Ìm.

notic segment was harvested and stained with the anti-VE-cadherin antibody. The results showed that VE-cad-herin was highly expressed at EC borders in the abdomi-nal aorta at sites without stenosis (fig. 3C), where theblood flow is laminar. In contrast, there was little or noVE-cadherin staining in the EC borders in the poststenot-ic dilatation sites (fig. 3E), where the flow was disturbedand the wall shear stress was low and fluctuating, and alsoat the clip site (fig. 3D), where the wall shear stress wasgreatly enhanced. These results, together with thoseshown in figure 2 indicate that different flow patternsmodulate the VE-cadherin at cell junctions in vivo.

VE-Cadherin Staining in BAECs Exposed toDisturbed and Laminar Flows in vitro To mimic the poststenotic dilatation of the abdominal

aorta constriction experiments in vivo, an in vitro stepflow system was used to investigate the effects of dis-turbed and laminar flows on the localization and expres-sion of VE-cadherin. The VE-cadherin was stained byimmunocytochemistry. Under static conditions and inregions exposed to laminar flow for 24 h, VE-cadherinstaining was continuous and distributed around the entireperiphery of cells (fig. 4A, C). In the reattachment regionwhich had been exposed to 24 h of disturbed flow, how-


J Vasc Res 2005;42:77–89 83

Fig. 5. VE-cadherin and ß-catenin staining in endothelial cells exposed to pulsatile and reciprocating flows. Confluentmonolayers of BAECs were kept as controls (A, B) or subjected to different flow patterns in a flow chamber. Stainingof VE-cadherin was observed by confocal microscopy. Stainings for VE-cadherin and ß-catenin were intermit-tent after exposing to pulsatile (C, D) or reciprocating (E, F) flow for 6 h. The distribution of these junction pro-teins were continuous around the entire periphery of the cells after 24, 48 or 72 h of exposure to pulsatile flow(G, H, K, L, O, P), but not reciprocating flow (I, J, M, N, Q, R). Exposure of BAECs pre-sheared with pulsatile flow for24 h to reciprocating flow for another 24 h resulted in an intermittent pattern of adhesion junction (S, T). Imagesshown here are representative results from 3 separate experiments. Bar in T = 30 Ìm.


Fig. 6. Pulsatile and reciprocating flows did not change the expres-sion levels of VE-cadherin and ß-catenin. Confluent BAEC mono-layers were subjected to either pulsatile (P) or reciprocating (R) flowin the rectangular flow chamber in vitro or kept as static controls for6 and 24 h. Cell lysates from the various samples were separated by10% SDS-PAGE, and the proteins were transferred to a nitrocellu-lose membrane. The nitrocellulose membrane was then split into 3parts for immunoblotting with anti-VE-cadherin antibody, anti-ß-catenin and ß-actin as an endogenous control molecule (gels in A).After densitometry and normalization of the results with that of ß-actin, the relative protein amounts of the sheared samples to that ofthe static control were determined. The bar graphs in B and C repre-sent the results of three independent experiments (mean B SD) forVE-cadherin and ß-catenin, respectively. These results indicate thatexposing BAECs to pulsatile or reciprocating flows for 6 or 24 h didnot change the total expression levels of VE-cadherin (B) and ß-cate-nin (C) (ANOVA, all p 1 0.05).

0

0.2

0.4

0.6

0.8

1.0

1.2

1.4C

Rel

ativ

e -c

aten

in

amo

un

t�

0

1.2

Rel

ativ

e V

E-c

adh

erin

amo

un

t

B

0.2

0.4

0.6

0.8

1.0

A

0 6 24 6 24Shear time (h)

P R

�

VE-cadherin

�

�-catenin

�

�-actin

ever, the VE-cadherin staining became intermittent, show-ing frequent gaps (fig. 4B).

VE-Cadherin and ß-Catenin Staining in BAECsExposed to Pulsatile and Reciprocating Flows in vitro Because of the small size of the disturbed flow zone in

the step flow chamber [31], it is difficult to obtain suffi-cient amount of cells to investigate the molecular mecha-nisms by performing further experiments such as immu-noblotting. Therefore, we used the rectangular flow chan-nel without a step to investigate the effects of two differ-ent flow patterns (pulsatile and reciprocating flows) on thelocalization and expression of VE-cadherin in confluentBAEC monolayers. In these studies, we also examined thelocalization and expression of ß-catenin which binds tothe intracellular domain of VE-cadherin. Under staticconditions, VE-cadherin and ß-catenin stainings werecontinuous and distributed around the entire periphery ofthe cells (fig. 5A, B). After exposure to either pulsatile(fig. 5C, D) or reciprocating flow (fig. 5E, F) for 6 h, VE-cadherin and ß-catenin stainings at cell-cell junctionsbecame intermittent, suggesting the presence of intercel-lular gaps. Extending the application of pulsatile flow to24 (fig. 5G, H), 48 (fig. 5K, L) or 72 h (fig. 5O, P) resultedin a reappearance of the continuous stainings for VE-cad-herin and ß-catenin around the entire periphery of thecells, but their stainings remained intermittent under re-ciprocating flow extended for the same time periods(fig. 5I, J, M, N, Q, R).

In additional experiments, we first presheared BAECswith pulsatile flow for 24 h and then applied reciprocatingflow for another 24 h. The exposure of these preshearedcells to reciprocating flow disrupted the cell junctionagain and resulted in the intermittent patterns of VE-cad-herin and ß-catenin stainings (fig. 5S, T). The results indi-cate that different flow patterns led to different distribu-tions of the junction proteins.

Fluid Shear Stress neither Changes the Protein Levelof VE-Cadherin nor Its Distribution between Cytosoland Membrane Fractions Since our histochemical studies showed time-depen-

dent alterations in VE-cadherin staining at EC bordersfollowing pulsatile and reciprocating flows, we examinedby Western blots the total protein amount of VE-cadherinin BAEC monolayers after exposure to pulsatile and reci-procating flows for 6 and 24 h. The results indicate thatthe exposure of BAECs to either pulsatile or reciprocatingflow did not affect the total expression levels of VE-cad-herin and ß-catenin at both 6 and 24 h (fig. 6A–C). Fur-


J Vasc Res 2005;42:77–89 85

thermore, our fractionation studies showed that VE-cad-herin levels in the membrane and cytosol fractions werenot significantly changed by the pulsatile and reciprocat-ing flows (fig. 7).

Discussion

The early atherosclerotic lesions in humans and ani-mals typically develop in the vicinity of branch points,areas of major curvature such as the aortic arch, and aor-tic stenosis [2–4, 15]. These lesion-prone sites have anincreased endothelial permeability to plasma proteins,cholesterol and horseradish peroxidase [6, 14, 16, 17, 34,35]. A correlation has been established between theregional permeability of low-density lipoprotein and thefocal distribution of atherosclerosis in the same arterialsegment after cholesterol feeding [1, 14, 15, 35]. Low-den-sity lipoprotein tends to deposit in larger arteries atbranch sites, bifurcations and bends, suggesting a role offlow dynamics on the modulation of endothelial perme-ability. Pulsed ultrasonic Doppler velocimetry and micro-cinematographic studies on animal aorta have demon-strated the existence of extensive secondary flow andregions of flow separation and recirculation at these sites[7, 36]. All of these findings are in support of the conceptthat an increase of arterial EC permeability plays animportant role in the initiation and progression of athero-sclerosis. Endothelial permeability is controlled by spe-cific intercellular adhesion molecules including VE-cad-herin [18, 21–23]. In this study on normal rats, we presentthe first evidence that the expression of VE-cadherin atcell-cell junctions in atherosclerosis-susceptible sites (aor-ta arch) is much weaker than that in lesion-resistant sites(descending thoracic aorta). These results suggest thatflow patterns regulate the junctional expression of VE-cadherin, which would in turn modulate endothelial per-meability. In addition to VE-caherin, ECs express N-cad-herin [37]. In contrast to the VE-cadherin, immunostain-ing of N-cadherin did not show a localized distribution atcell-cell borders in the endothelium of both the aorta archand the descending thoracic aorta (data not show). Thisfinding is consistent with other reports on the expressionsof cadherins on ECs [37].

To further assess the effects of different flow patternson VE-cadherin localization, we used an arterial stenosismodel in vivo and flow systems in vitro. Creating a steno-sis in the aorta or a large artery by means of a constrictingband or ligature [27, 28, 38–41] produces high shear at thestenosis throat, flow separation with recirculation around

Fig. 7. Pulsatile and reciprocating flows did not change the distribu-tion of VE-cadherin in cytosol and membrane fractions. ConfluentBAEC monolayers were subjected to either pulsatile (P) or recipro-cating (R) flow for 6 and 24 h in the rectangular flow chamber in vitroor kept as static controls. The cell lysates were fractionated into mem-brane (A) and cytosol (B) fractions. Different fractions were sepa-rated by 10% SDS-PAGE, and the proteins were transferred to anitrocellulose membrane. The nitrocellulose membrane was thensplit into 2 parts for immunoblotting with anti-VE-cadherin anti-body (upper panels of gels in A and B) and with ß-actin as an endoge-nous control molecule (lower panels of gels in A and B). After densi-tometry and normalization of the results with that of ß-actin, the rela-tive amount of the proteins in the sheared samples to that of the staticcontrol was determined. Bar graphs represent results of three inde-pendent experiments (mean B SD) for VE-cadherin in the mem-brane and cytosol fractions. These results indicate that the VE-cad-herin levels in the membrane and cytosol fractions were not signifi-cantly changed by 6 or 24 h of pulsatile and reciprocating flows(ANOVA, all p 1 0.05).

0

1.4R

elat

ive

VE

-cad

her

in a

mo

un

t

B

�

�-actin

P

�

VE-cadherin

0.2

0.4

0.6

0.8

1.0

1.2


RCytosol

0

1.2

Rel

ativ

e V

E-c

adh

erin

am

ou

nt

�

�-actin

A

P

�

VE-cadherin


RMembrane

0.2

0.4

0.6

0.8

1.0


the poststenotic jet and turbulence in the region of jetbreak up [27, 42, 43]. We examined the cell morphologyby staining the cell border with silver nitrate and found itto be consistent with that reported in other studies [27, 40,41, 44], i.e. ECs in the area of high shear stress are elon-gated and those in the areas immediately downstream ofthe stenosis are polygonal (data not shown). Immuno-staining in figure 3 shows that VE-cadherin staining in theEC borders is very weak in the poststenotic dilatation sites(with disturbed flow) and the clip-constricted site (withvery high shear stress), as compared with regions exposedto normal pulsatile flow (fig. 3). Using the U-shaped metalclip and intraluminal hemispherical plug to create flowdisturbances in the rat aorta, Zand et al. [38, 45] haveshown that lipid deposition occurs at regions of low shearstress, stagnation and recirculation. The present resultstogether with the leaky junction hypothesis [46–49] andthe findings of Zand et al. suggest that the decrease of VE-cadherin at cell-cell borders might contribute to an in-crease in endothelial permeability and the observed lipiddeposition in the disturbed flow area. It is interesting thatthe VE-cadherin staining in the stenotic region with veryhigh shear stress above physiological range is also weak(fig. 3D) as in the poststenotic region (fig. 3E), suggestingthat this area with excessively high shear stress may alsohave an elevated transendothelial permeability.

Since it is difficult to quantitatively define the detailedcharacteristics of the hemodynamic environment in vivo,we studied the problem further in the in vitro systemswhere cultured ECs can be exposed to well-defined flowconditions [25, 30, 31, 50]. Two experimental approacheswere used in this study. First, the effects of disturbed andlaminar flow on the VE-cadherin reorganization werestudied concurrently in different regions of a step flowchamber. Under static condition, VE-cadherin was con-tinuously distributed along the cell-cell borders of con-fluent BAECs. After exposing the confluent monolayer toflow in the step flow chamber for 24 h, we found that gapswere formed between cells and that VE-cadherin stainingbecame intermittent in regions of disturbed flow, but notin the regions of the same monolayer exposed to laminarflow. It has been demonstrated by Phelps and DePaola[51] by using a similar step flow chamber that endothelialpermeability to macromolecules (dextran, mol wt 70,000)was significantly higher in regions of flow disturbancethan in the adjacent regions of the same monolayerexposed to laminar shear flow. The agreement of the VE-cadherin reorganization in the disturbed flow region ofour study and the increased endothelial permeability inthe same region found by Phelps and DePaola [51] sug-

gests a possible relationship between the reorganization ofVE-cadherin at cell junctions and endothelial permeabili-ty. The second approach for our in vitro study was basedon our hypothesis that the flow reversal (reciprocatingflow) in the regions of disturbed flow in vivo contributesto the disruption of VE-cadherin continuity at the cell-cellborders in those regions. Therefore, we compared theeffects of pulsatile and reciprocating flows on VE-cadhe-rin expression. Again, under static condition VE-cadherinwas continuously distributed along the cell-cell borders ofconfluent BAECs. After exposing the monolayer to eitherpulsatile or reciprocating flow for 6 h, gaps were formedbetween cells. The continued application of pulsatile flowfor more than 24 h, however, led to the reappearance ofVE-cadherin staining along the entire periphery of thecells. The observation of VE-cadherin continuity inBAEC monolayers exposed to sustained pulsatile shearstress is consistent with other reports on ECs exposed tosustained laminar shear stress [25, 26]. In contrast to theresults on pulsatile flow, we found that the continuedapplication of reciprocating flow for 24 h or longer did notcause the reappearance of continuous VE-cadherin stain-ing along the cell periphery. The result on the applicationof 24 h of reciprocating flow following 24 h of preshearingwith pulsatile flow indicates that the reciprocating flowcould induce VE-cadherin redistribution after its initialorganization by the preshearing, and hence its action ismore than simply preventing the BAECs from movingVE-cadherins to the junctions. The mechanisms regulat-ing cell junction organization in ECs are still not wellunderstood. Several possibilities, such as changes of theexpression level, phosphorylation status, and distributionpattern of junction proteins, have been proposed [25, 52,53]. The findings of the present study indicate that theshear-induced change in VE-cadherin staining at cell bor-ders is not likely the result of a change in the proteinexpression level (fig. 6), its redistribution from cytosol tomembrane (fig. 7), nor a change in its phosphorylationstatus (data not shown). Studies in other groups also showthat the expression levels of VE-cadherin and ß-cateninare unchanged in ECs exposed to shear stress [24, 26].

One potential explanation for this change in the inten-sity of VE-cadherin staining at cell-cell borders under var-ious conditions of shearing is that the VE-cadherin orga-nized and concentrated at cell junctions under static con-dition was disrupted by shear stress, whether pulsatile orreciprocating, to become more intermittently distributedin the membrane. Treatment of ECs with Ca2+ chelators,which interfere with homotypic interaction of cadherintransmers, causes VE-cadherin to be diffusely distributed


J Vasc Res 2005;42:77–89 87

in the cell membrane, and Ca2+ repletion by removal ofthe chelator causes a reappearance of continuous stainingof VE-cadherin at cell-cell junctions [54]. During leuko-cyte transmigration across the endothelial layer, a VE-cadherin/green fluorescent protein fusion construct wasobserved to diffuse out of endothelial junction and resultin the formation of gaps, which rapidly resealed within5 min after the completion of leukocyte transmigration,and the VE-cadherin/green fluorescent protein fusionconstruct became continuously distributed at cell junc-tions again [55]. It has been reported that C-cadherin canmigrate in the cell membrane to form focal patches of cad-herin staining at the basal cell surface in contact with theC-cadherins coated on coverslips. In contrast, cadherinstaining remains largely diffuse on the cell surface whenplated onto poly-L-lysine [56]. These modulations of ad-hesion functions of cadherin are independent of its ex-pression level and complex formation with catenins, butinvolve the redistribution of cadherin on cell surface [56].In addition, it has been demonstrated that the gap junc-tion protein connexin-43 has junctional and extrajunc-tional distributions in the cell membrane and that theextrajunctional connexin-43 accumulates at sites of cell-cell contact by lateral migration in the plane of the mem-brane bilayer as soon as the adjacent cell membranes con-tact each other [57, 58]. All these findings are in supportof the notion that junction proteins can undergo redistri-bution in the plane of the membrane.

Pathophysiological processes such as inflammation,wound healing and angiogenesis cause a dissociation andrearrangement of endothelial cell-cell junctions [59]. Sim-ilarly, experiments using flow devices have shown thatlaminar shear stress modulates cytoskeleton reorganiza-tion [60] and VE-cadherin redistribution [24–26]. Inthese situations, changes in molecular interactions be-tween cell-cell junction proteins are required to regulatecell migration, proliferation and transendothelial perme-ability. Considering the results of other studies [25, 26],the findings of the present investigation on the applica-tion of pulsatile shear stress to BAECs for 624 h can bebest explained by the relocalization of VE-cadherins,which had become discontinuous in the junctions by 6hour-shearing, into cell junctions. In contrast to pulsatileflow, the application of reciprocating shear stress toBAECs for more than 24 h did not lead to a relocalizationor reorganization of VE-cadherin. We have also foundthat the migration speed of BAECs in a monolayerincreased after 6 h of exposure to either pulsatile or thereciprocating flow, and that this increase in migrationspeed was sustained up to 24 h under reciprocating flow

but not pulsatile flow (data not shown). The parallelismbetween the observations on VE-cadherin redistributionand cell motility in response to different flow patterns sug-gests that these two types of cellular responses may berelated, e.g. through a restriction of cell movement by cell-cell junctions and/or a retardation of the formation of cell-cell junctions due to reduced contact time between neigh-boring cells as a result of cell movement.

In summary, both in vivo and in vitro data suggest thatpulsatile flow with a strong forward component and com-plex flow without a significant forward component differ-entially regulate the cell junction remodeling. This flowpattern-regulated junction remodeling may provide anexplanation for the high endothelial permeability and thepreferential localization of atherosclerosis in regionswhere flow patterns are complex, especially where theshear stress exhibits time-fluctuations without a signifi-cant net direction.

Acknowledgments

We thank Lilian Wei-Ling Huang for her assistance with the dataanalysis. This study was supported by USPHS research grantsHL19454, HL43026 and HL64382.


References

1 Nielsen LB, Nordestgaard BG, Stender S,Kjeldsen K: Aortic permeability to LDL as apredictor of aortic cholesterol accumulation incholesterol-fed rabbits. Arterioscler Thromb1992;12:1402–1409.

2 Gimbrone MA: Vascular endothelium, hemo-dynamic forces, and atherogenesis. Am J Pa-thol 1999;155:1–5.

3 Thubrikar MJ, Robicsek F: Pressure-inducedarterial wall stress and atherosclerosis. AnnThorac Surg 1995;59:1594–1603.

4 Glagov S, Zarins C, Giddens DP, Ku DN:Hemodynamics and atherosclerosis. Insightsand perspectives gained from studies of humanarteries. Arch Pathol Lab Med 1988;112:1018–1031.

5 Chiu JJ, Wang DL, Chien S, Skalak R, UsamiS: Effects of disturbed flow on endothelial cells.J Biomech Eng 1998;120:2–8.

6 Barakat AI, Uhthoff PA, Colton CK: Topo-graphical mapping of sites of enhanced HRPpermeability in the normal rabbit aorta. JBiomech Eng 1992;114:283–292.

7 Barakat AI, Karino T, Colton CK: Microcine-matographic studies of flow patterns in theexcised rabbit aorta and its major branches.Biorheology 1997;34:195–221.

8 Al-Mehdi AB, Zhao G, Fisher AB: ATP-inde-pendent membrane depolarization with isch-emia in the oxygen-ventilated isolated rat lung.Am J Respir Cell Mol Biol 1998;18:653–661.

9 Caro CG, Fitz-Gerald JM, Schroter RC: Athe-roma and arterial wall shear. Observation, cor-relation and proposal of a shear dependentmass transfer mechanism for atherogenesis.Proc R Soc Lond B Biol Sci 1971;177:109–159.

10 Giddens DP, Zarins CK, Glagov S: The role offluid mechanics in the localization and detec-tion of atherosclerosis. J Biomech Eng 1993;115:588–594.

11 Soler HM, Watkins MT, Albadawi H, Kadowa-ki H, Patton GM: Effects of oxygen tension andshear stress on human endothelial cell prosta-cyclin production. J Surg Res 1997;67:46–53.

12 Waters CM: Flow-induced modulation of thepermeability of endothelial cells cultured onmicrocarrier beads. J Cell Physiol 1996;168:403–411.

13 Ogunrinade O, Kameya GT, Truskey GA: Ef-fect of fluid shear stress on the permeability ofthe arterial endothelium. Ann Biomed Eng2002;30:430–446.

14 Fry DL, Herderick EE, Johnson DK: Localintimal-medial uptakes of I-125-albumin, I-125-LDL, and parenteral Evans blue dye pro-tein complex along the aortas of normocholes-terolemic minipigs as predictors of subsequenthypercholesterolemic atherogenesis. Arterio-scler Thromb 1993;13:1193–1204.

15 Herrmann RA, Malinauskas RA, Truskey GA:Characterization of sites with elevated LDLpermeability at intercostal, celiac, and iliacbranches of the normal rabbit aorta. Arterio-scler Thromb 1994;14:313–323.

16 Bell FP, Somer JB, Craig IH, Schwartz CJ: Pat-terns of aortic Evans blue uptake in vivo and invitro. Atherosclerosis 1972;16:369–375.

17 Somer JB, Schwartz CJ: Focal 3H-cholesteroluptake in the pig aorta. Atherosclerosis 1971;13:293–304.

18 Dejana E, Corada M, Lampugnani MG: Endo-thelial cell-to-cell junctions. FASEB J 1995;9:910–918.

19 Navarro P, Caveda L, Breviario F, Mando-teanu I, Lampugnani MG, Dejana E: Catenin-dependent and -independent functions of vas-cular endothelial cadherin. J Biol Chem 1995;270:30965–30972.

20 Corada M, Mariotti M, Thurston G, Smith K,Kunkel R, Brockhaus M, Lampugnani MG,Martin-Padura I, Stoppacciaro A, Ruco L,McDonald DM, Ward PA, Dejana E: Vascularendothelial-cadherin is an important determi-nant of microvascular integrity in vivo. ProcNatl Acad Sci USA 1999;96:9815–9820.

21 Dejana E, Spagnuolo R, Bazzoni G: Interendo-thelial junctions and their role in the control ofangiogenesis, vascular permeability and leuko-cyte transmigration. Thromb Haemost 2001;86:308–315.

22 Dejana E: Endothelial adherens junctions: Im-plications in the control of vascular permeabili-ty and angiogenesis. J Clin Invest 1996;98:1949–1953.

23 Lampugnani MG, Dejana E: Interendothelialjunctions: Structure, signalling and functionalroles. Curr Opin Cell Biol 1997;9:674–682.

24 Shay-Salit A, Shushy M, Wolfovitz E, YahavH, Breviario F, Dejana E, Resnick N: VEGFreceptor 2 and the adherens junction as a me-chanical transducer in vascular endothelialcells. Proc Natl Acad Sci USA 2002;99:9462–9467.

25 Noria S, Cowan DB, Gotlieb AI, Langille BL:Transient and steady-state effects of shearstress on endothelial cell adherens junctions.Circ Res 1999;85:504–514.

26 Ukropec JA, Hollinger MK, Woolkalis MJ:Regulation of VE-cadherin linkage to the cyto-skeleton in endothelial cells exposed to fluidshear stress. Exp Cell Res 2002;273:240–247.

27 Zand T, Nunnari JJ, Hoffman AH, SavilonisBJ, MacWilliams B, Majno G, Joris I: Endothe-lial adaptations in aortic stenosis. Correlationwith flow parameters. Am J Pathol 1988;133:407–418.

28 Joris I, Zand T, Majno G: Hydrodynamic inju-ry of the endothelium in acute aortic stenosis.Am J Pathol 1982;106:394–408.

29 Poole JCF, Sanders AG, Florey HW: The re-generation of aortic endothelium. J Path Bact1958;75:133–143.

30 Frangos JA, Eskin SG, McIntire LV, Ives CL:Flow effects on prostacyclin production by cul-tured human endothelial cells. Science 1985;227:1477–1479.

31 Hsu PP, Li S, Li YS, Usami S, Ratcliffe A,Wang X, Chien S: Effects of flow patterns onendothelial cell migration into a zone of me-chanical denudation. Biochem Biophys ResCommun 2001;285:751–759.

32 Miao H, Yuan SL, Wang YX, Tsygankov A,Chien S: Role of cbl in shear-activation of PI3-kinase and JNK in endothelial cells. BiochemBiophys Res Commun 2002;292:892–899.

33 Li S, Chen BPC, Azuma N, Hu YL, Wu SZ,Sumpio BE, Shyy JYJ, Chien S: Distinct rolesfor the small GTPases Cdc42 and Rho in endo-thelial responses to shear stress. J Clin Invest1999;103:1141–1150.

34 Chuang PT, Cheng HJ, Lin SJ, Jan KM, LeeMML, Chien S: Macromolecular transportacross arterial and venous endothelium in rats– studies with Evans blue-albumin and horse-radish-peroxidase. Arteriosclerosis 1990;10:188–197.

35 Chien S: Molecular and mechanical bases offocal lipid accumulation in arterial wall. ProgBiophys Mol Biol 2003;83:131–151.

36 Farthing S, Peronneau P: Flow in the thoracicaorta. Cardiovasc Res 1979;13:607–620.

37 Navarro P, Ruco L, Dejana E: Differentiallocalization of VE- and N-cadherins in humanendothelial cells: VE-cadherin competes withN-cadherin for junctional localization. J CellBiol 1998;140:1475–1484.

38 Zand T, Majno G, Nunnari JJ, Hoffman AH,Savilonis BJ, MacWilliams B, Joris I: Lipiddeposition and intimal stress and strain. Astudy in rats with aortic stenosis. Am J Pathol1991;139:101–113.

39 Gerrity RG, Naito HK: Alteration of endothe-lial cell surface morphology after experimentalaortic coarctation. Artery 1980;8:267–274.

40 Langille BL, Reidy MA, Kline RL: Injury andrepair of endothelium at sites of flow distur-bances near abdominal aortic coarctations inrabbits. Arteriosclerosis 1986;6:146–154.

41 Reidy MA, Langille BL: The effect of localblood flow patterns on endothelial cell mor-phology. Exp Mol Pathol 1980;32:276–289.

42 Hutchison KJ, Karpinski E: In vivo demon-stration of flow recirculation and turbulencedownstream of graded stenoses in canine arter-ies. J Biomech 1985;18:285–296.

43 Ahmed SA, Giddens DP: Velocity measure-ments in steady flow through axisymmetric ste-noses at moderate Reynolds numbers. J Bio-mech 1983;16:505–516.

44 Hutchison KJ: Endothelial cell morphologyaround graded stenoses of the dog commoncarotid artery. Blood Vessels 1991;28:396–406.

45 Zand T, Hoffman AH, Savilonis BJ, Under-wood JM, Nunnari JJ, Majno G, Joris I: Lipiddeposition in rat aortas with intraluminalhemispherical plug stenosis – A morphologicaland biophysical study. Am J Pathol 1999;155:85–92.


J Vasc Res 2005;42:77–89 89

46 Weinbaum S, Tzeghai G, Ganatos P, Pfeffer R,Chien S: Effect of cell turnover and leaky junc-tions on arterial macromolecular transport.Am J Physiol 1985;248:H945–H960.

47 Lin SJ, Jan KM, Weinbaum S, Chien S: Trans-endothelial transport of low-density lipopro-tein in association with cell mitosis in rat aorta.Arteriosclerosis 1989;9:230–236.

48 Huang AL, Jan KM, Chien S: Role of intercel-lular-junctions in the passage of horseradish-peroxidase across aortic endothelium. Lab In-vest 1992;67:201–209.

49 Chen YL, Jan KM, Lin HS, Chien S: Ultra-structural studies on macromolecular perme-ability in relation to endothelial cell turnover.Atherosclerosis 1995;118:89–104.

50 Helmlinger G, Geiger RV, Schreck S, NeremRM: Effects of pulsatile flow on cultured vascu-lar endothelial cell morphology. J Biomech Eng1991;113:123–131.

51 Phelps J, DePaola N: Spatial variations in en-dothelial barrier function in disturbed flows invitro. Am J Physiol Heart Circ Physiol 2000;278:H469–H476.

52 Konstantoulaki M, Kouklis P, Malik AB: Pro-tein kinase C modifications of VE-cadherin,p120, and beta-catenin contribute to endothe-lial barrier dysregulation induced by thrombin.Am J Physiol Lung Cell Mol Physiol 2003;285:L434–L442.

53 Alexander JS, Jackson SA, Chaney E, KevilCG, Haselton FR: The role of cadherin endocy-tosis in endothelial barrier regulation: Involve-ment of protein kinase C and actin-cadherininteractions. Inflammation 1998;22:419–433.

54 Gao XP, Kouklis P, Xu N, Minshall RD, San-doval R, Vogel SM, Malik AB: Reversibility ofincreased microvessel permeability in responseto VE-cadherin disassembly. Am J PhysiolLung Cell Mol Physiol 2000;279:L1218–L1225.

55 Shaw SK, Bamba PS, Perkins BN, LuscinskasFW: Real-time imaging of vascular endothe-lial-cadherin during leukocyte transmigrationacross endothelium. J Immunol 2001;167:2323–2330.

56 Yap AS, Brieher WM, Pruschy M, GumbinerBM: Lateral clustering of the adhesive ectodo-main: A fundamental determinant of cadherinfunction. Curr Biol 1997;7:308–315.

57 Musil LS, Goodenough DA: Biochemical-anal-ysis of connexin43 intracellular-transport,phosphorylation, and assembly into Gap junc-tional plaques. J Cell Biol 1991;115:1357–1374.

58 Rook MB, Dejonge B, Jongsma HJ, Masson-pevet MA: Gap junction formation and func-tional interaction between neonatal rat cardio-cytes in culture – a correlative physiologicaland ultrastructural-study. J Membrane Biol1990;118:179–192.

59 Schnittler HJ: Structural and functional as-pects of intercellular junctions in vascular en-dothelium. Basic Res Cardiol 1998;93:30–39.

60 Galbraith CG, Skalak R, Chien S: Shear stressinduces spatial reorganization of the endothe-lial cell cytoskeleton. Cell Motil Cytoskeleton1998;40:317–330.

Effects of Flow Patterns on the Localization and Expression of VE-Cadherin at Vascular Endothelial...

Documents

Transcript of Effects of Flow Patterns on the Localization and Expression of VE-Cadherin at Vascular Endothelial...