AJP - Heart Email Content Delivery
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Am J Physiol Heart Circ Physiol 286: H731-H741, 2004. First published October 2, 2003; doi:10.1152/ajpheart.00427.2003
0363-6135/04 $5.00
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
286/2/H731    most recent
00427.2003v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (9)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by DeMaio, L.
Right arrow Articles by Antonetti, D. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by DeMaio, L.
Right arrow Articles by Antonetti, D. A.

A transmural pressure gradient induces mechanical and biological adaptive responses in endothelial cells

Lucas DeMaio,1 John M. Tarbell,4 Russell C. Scaduto, Jr.,2 Thomas W. Gardner,3 and David A. Antonetti2,3

1Department of Chemical Engineering, Biomolecular Transport Dynamics Laboratory, The Pennsylvania State University, University Park 16802; Departments of 2Cellular and Molecular Physiology and 3Ophthalmology, Penn State College of Medicine, Hershey, Pennsylvania 17033; and 4Department of Biomedical Engineering, The City College of The City University of New York, New York, New York 10031

Submitted 2 May 2003 ; accepted in final form 27 September 2003


   ABSTRACT
TOP
 ABSTRACT
METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
A sudden increase in the transmural pressure gradient across endothelial monolayers reduces hydraulic conductivity (Lp), a phenomenon known as the sealing effect. To further characterize this endothelial adaptive response, we measured bovine aortic endothelial cell (BAEC) permeability to albumin and 70-kDa dextran, Lp, and the solvent-drag reflection coefficients ({sigma}) during the sealing process. The diffusional permeability coefficients for albumin (1.33 ± 0.18 x 10–6 cm/s) and dextran (0.60 ± 0.16 x 10–6 cm/s) were measured before pressure application. The effective permeabilities (measured when solvent drag contributes to solute transport) of albumin and dextran (Pealb and Pedex) were measured after the application of a 10 cmH2O pressure gradient; during the first 2 h of pressure application, Pealb, Pedex, and Lp were significantly reduced by 2.0 ± 0.3-, 2.1 ± 0.3-, and 3.7 ± 0.3-fold, respectively. Immunostaining of the tight junction (TJ) protein zonula occludens-1 (ZO-1) was significantly increased at cell-cell contacts after the application of transmural pressure. Cytochalasin D treatment significantly elevated transport but did not inhibit the adaptive response, whereas colchicine treatment had no effect on diffusive permeability but inhibited the adaptive response. Neither cytoskeletal inhibitor altered {sigma} despite significantly elevating both Lp and effective permeability. Our data suggest that BAECs actively adapt to elevated transmural pressure by mobilizing ZO-1 to intercellular junctions via microtubules. A mechanical (passive) component of the sealing effect appears to reduce the size of a small pore system that allows the transport of water but not dextran or albumin. Furthermore, the structures of the TJ determine transport rates but do not define the selectivity of the monolayer to solutes ({sigma}).

permeability; hydraulic conductivity; reflection coefficient; zonula occludens-1

THERE ARE TWO PRINCIPAL MECHANISMS of paracellular transport across the endothelium: bulk fluid movement or convection, a process driven by hydrostatic and oncotic pressure gradients, and diffusional exchange, a nonconvective process driven by concentration gradients (24, 42). These transport processes are characterized by three transport coefficients that define the barrier properties of the endothelium to water and solutes: hydraulic conductivity (Lp), diffusional permeability (Pd; i.e., when convection is zero), and the solvent-drag reflection coefficient ({sigma}). The effective permeability (Pe; measured when solvent drag contributes to solute flux) is also reported frequently in the literature. It is conventional to regard the endothelial transport coefficients as constants; in other words, any observed change in transendothelial flux is attributed to changes in pressure and/or concentration gradients across the endothelium. However, it has become increasingly evident in recent years that endothelial transport coefficients also depend on the mechanical and hormonal environment of the cell layer.

The endothelial cell layer that lines the inner surface of the vascular wall is continuously subjected to blood flow and pressure. The flow of blood generates a tangential frictional stress (i.e., wall shear stress) that directly acts on the luminal surface of endothelial cells. Several in vitro studies (11, 23, 28, 32, 40) and in vivo studies (29, 49, 50) have shown that shear stress alters transport coefficients in micro- and macrovessel model systems. Blood pressure, on the other hand, is a normal stress that is balanced by the circumferential stress of the vessel wall. Gradients of hydrostatic pressure and oncotic pressure across the wall drive fluid transport between the vessel lumen and the underlying tissue. Other investigators have observed that imposing a step increase in the transmural pressure gradient across cells in culture (6, 40, 44, 46) and intact arteries (29, 45) leads to an initial elevation, followed by a transient reduction in endothelial transport coefficients. This phenomenon, termed the "sealing effect" (44), appears to be an adaptive mechanism to protect against tissue edema in the face of elevated blood pressure. Transmural pressure gradients also regulate the myogenic response, a more recognized mechanism of edema prevention that is mediated by smooth muscle cells (SMCs). Specifically, increases in transmural pressure induce stretch-dependent SMC contraction (18) as well as transmural flow-dependent SMC contraction in response to increased interstitial flow shear stresses (12, 26); in turn, distal blood pressure and transendothelial fluxes are reduced (12, 34).

The cleft between adjacent endothelial cells is the primary pathway for the transport of water and hydrophilic solutes into the underlying tissue. Discontinuities in the tight junction (TJ) strand allow for paracellular transport of water and larger solutes, including 70-kDa dextran and albumin, which are considered in this study. Zonula occludens (ZO)-1 is a peripheral TJ protein and a member of the membrane-associated guanylate kinase (MAGuK) family (37). MAGuK proteins are defined by having PDZ and guanylate kinase-like domains, both of which are protein-protein interacting domains. ZO-1 provides a link between transmembrane TJ proteins and the actin cytoskeleton. Therefore, ZO-1 appears to play a central role in the assembly and organization of TJs (2, 3, 7) and the regulation of paracellular transport (4, 8, 30). The effect of transmural pressure gradient on the expression and localization of ZO-1, however, has not been investigated previously.

To study the role of actin in endothelial transport regulation, cytochalasin D (CytoD) has been used to disrupt microfilaments. CytoD treatment increases the number of TJ discontinuities at cell-cell contacts, which leads to increased transendothelial fluxes (25, 43). The role of microtubules in the regulation of endothelial transport is not well defined. It has been shown that the microtubule disrupting agent colchicine does not alter cerebrovascular permeability to horseradish peroxidase (HRP) in rats (39) but decreases transendothelial electrical resistance across cerebral endothelial monolayers (47). Furthermore, the roles of microtubules and microfilaments in modulating the sealing effect have not been investigated previously.

The present study was designed to investigate the dynamic barrier properties of bovine aortic endothelial cell (BAEC) monolayers after exposure to a transmural pressure gradient. We hypothesized that reduced Lp and Pe coefficients are in part a result of increased ZO-1 expression and/or localization at cell-cell contacts. Our results reveal that transmural pressure induces mechanical (passive) and biological adaptive responses in endothelial cells. The biological response appears to recruit ZO-1 to cell-cell contacts via microtubules from a presynthesized pool of ZO-1 protein, which results in the reduction of Lp and Pe. The transport properties of fixed and cytoskeletal inhibitor-treated monolayers suggest that a mechanical component reduces Lp but not Pe for 70-kDa dextran. The findings of the present study are consistent with junctional complexes dictating the magnitudes of Lp and Pe, whereas the glycocalyx determines solute selectivity ({sigma}). Our results also suggest that transendothelial water flow occurs through discontinuities in the TJ strand as well as through a small pore system that at least excludes solutes the size of 70-kDa dextran.


   METHODS
TOP
 ABSTRACT
 METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Materials. The following cell culture components were obtained from Sigma (St. Louis, MO): BSA (30% solution, fraction V), trypsin, minimal essential medium (MEM) (without L-glutamine), L-glutamine, HEPES, sodium bicarbonate, and fetal bovine serum (FBS). A 100x antibiotic-antimycotic mixture (penicillin G sodium, streptomycin sulfate, and amphotericin B) was obtained from GIBCO-BRL (Rockville, MD). Antibodies were obtained as follows: ZO-1 rat monoclonal antibody, clone R40–76, was provided by Dr. Bruce Stevenson (Department of Cell Biology and Anatomy, University of Alberta, Edmonton, Alberta, Canada), donkey anti-rat Cy2 antibody was obtained from Jackson Immunoresearch Laboratories (West Grove, PA), and goat anti-rat HRP-linked antibody was obtained from Amersham (Piscataway, NJ).

Cell culture. Primary BAECs were purchased from VEC Technologies (Rensselaer, NY) and grown in MEM supplemented with 10% FBS. Cells were passaged three to six times before being plated at a density of 1.25 x 105 cells/cm2 onto Transwell polycarbonate filters (0.4-µm pore size, 24.5 mm diameter) from Costar (Cambridge, MA). One hour before the cells were plated, filters were incubated at 37°C with 1 ml of 30 µg/ml fibronectin (Sigma). These cells reached confluence within 1–2 days on the filters, and experiments were performed on monolayers 3 days postseeding to allow sufficient time for the junctions between cells to develop. The day of the experiment, MEM-10% FBS cell culture media was replaced with experimental media containing 1% BSA (MEM-1% BSA) as described previously (40).

Measurement of water and solute flux. A novel apparatus was developed in our laboratory to measure water and solute flux across BAEC monolayers under near-physiological conditions (5). This apparatus, excluding the data acquisition and control computer, was housed within a Plexiglas box and kept at an ambient air temperature of 37°C. The Transwell filter containing the BAEC monolayer was sealed between a two-piece polycarbonate assembly separating the apical and basolateral compartments. The apical compartment was continuously exposed to 5% CO2-95% balance air to maintain the pH of the media at 7.4. Tygon and borosilicate glass tubing connected the basolateral compartment to a reservoir, which was lowered by 10 cm to apply a hydrostatic pressure gradient (10 cmH2O) across the monolayer. As a result, water flux was driven through the monolayer into the basolateral compartment. To measure the volumetric flow rate (Jv), a bubble was inserted into the borosilicate tubing and tracked with a photometer, which followed the air-liquid interface in real time. Bubble displacement was converted to Jv according to the following equation

(1)
where {Delta}d/{Delta}t is bubble displacement per unit time and F is a tube calibration factor (i.e., fluid volume contained in known length of tubing). Because the apical and basolateral chambers contained the same media (MEM-1% BSA), it was assumed that Lp could be calculated by the following equation

(2)
where {Delta}P is the hydrostatic pressure difference and A is the surface area of the monolayer. In the present study, Jv/A (water flow) and Lp are often used interchangeably because {Delta}P is constant; thus changes in Jv/A also reflect changes in Lp. There may have been a slight excess of protein near the luminal surface because of concentration polarization driven by volume flux, but this was estimated to have a negligible oncotic effect using a 1% BSA solution (29).

Integration of the bubble tracker with a fluorescent detection system allowed simultaneous flux measurements of water and fluorescent-conjugated solutes. Fluorescence accumulation in the basolateral chamber was measured by two silica optical fibers arranged in the standard 90° fluorometer format. One silica fiber conducted excitation light to the chamber, and the other captured emission light from the fluorescent-labeled solute. The silica fibers were interfaced to the chamber using ferrule-type fittings, which provided a watertight seal. A Uniphase 1-mW helium-argon laser (Manteca, CA) generated excitation light at 543.5 nm. The laser beam diameter was 0.8 mm and was introduced directly into the 1-mm fiber without the use of lenses. Alignment was achieved using a small X/Y fiber holder from Melles Griot (Carlsbad, CA). The emission silica fiber was coupled to a model D48 photomultiplier from C&L Instruments (Hummelstown, PA). A Windows-based software program (FluorMeasure) and model PC-DAQ control card from C&L Instruments were used for data acquisition and control of the photomultiplier.

In the present study, either 3.25 µM dextran (70 kDa)-rhodamine B or 2 µM albumin-Alexa fluor 594 (Molecular Probes; Eugene, OR) was added to the apical chamber. A trace concentration (~2 nM) of either fluorescent-labeled compound was added to the basolateral chamber so that the linear relationship between solute concentration and fluorescence intensity was maintained at low solute concentrations. In a typical experiment, fluorescence intensity was recorded every minute for 60 min with Jv = 0; afterward, a 10-cmH2O pressure gradient was applied and fluorescence was recorded every minute for an additional 240 min. To minimize measurement noise, the fluorescence intensity measurements recorded every minute were smoothed using an adaptive smoothing function (supsmooth) from Mathcad 8 (Cambridge, MA) and averaged into 5-min intervals. After the linear relationship between solute concentration and fluorescence intensity (calibration) was determined, Pd or Pe was calculated by

(3)
where {Delta}Cb/{Delta}t is the average change in basolateral solute concentration over 5 min, Vb is the fluid volume of the basolateral chamber, and Ca is the apical concentration of the solute. Changes in Vb and Ca were negligible during an experiment and Cb << Ca.

Immunocytochemistry. After most experiments in which transport was measured, the BAEC monolayer on the filter was washed twice with PBS, fixed in 1% paraformaldehyde for 10 min, permeabilized with PBS containing 0.2% Triton X-100 for 10 min, and blocked with PBS containing 10% BSA and 0.1% Triton X-100 for 1 h. Cells were incubated with anti-ZO-1 monoclonal rat antibody (1:1) for 1 h, washed with PBS containing 0.1% Triton X-100 five times, and incubated with donkey anti-rat Cy2 antibody (1:200). Filters were washed four times with PBS containing 0.1% Triton X-100, and coverslips were mounted using Aqua Poly/Mount (Polysciences; Warrington, PA). A Nikon E800 confocal microscope equipped with DIC and a PCM2000 Multi-Line Dual Laser system running Simple PCI software (C Imaging Systems) was used to capture digital images. All images within an experiment were captured in an identical fashion. ZO-1 immunoreactivity was quantified by measuring the mean gray value of all the pixels in a digital image using Optimas software (Media Cybernetics; Silver Spring, MD).

Immunoblotting. Upon completion of an experiment, BAEC monolayers were immediately washed with ice-cold PBS and lysed in a urea buffer [6 M urea, 0.1% Triton X-100, 10 mM Tris (pH 8.0), 1 mM dithiothreitol, 5 mM MgCl2, 5 mM EGTA, 150 mM NaCl, 10 mM NaF, 1 mM NaVO4, and 0.2 mM PMSF], which has been previously shown to optimally solubilize ZO-1 (3). Insoluble material was pelleted by centrifugation in a microfuge at 10,000 g for 10 min. Protein concentrations were determined by Bio-Rad DC protein assay, and equal protein was loaded onto 6% SDS-polyacrylamide gel. Proteins were transferred to nitrocellulose membranes (MSI; Westborough, MA), blocked with 5% milk in Tris-buffered saline-Tween, and immunoblotted using anti-ZO-1 R40–76 antibody (1:2), followed by anti-rat HRP-conjugated secondary antibody from the goat (1: 4,000). ZO-1 content was detected using 20x LumiGLO Reagent and peroxide (Cell Signaling; Beverly, MA) and quantified using Gene-Tools (SynGene) software.

Reflection coefficient calculations. The reflection coefficients for 70 kDa dextran and albumin, {sigma}dex and {sigma}alb, were calculated from the following nonlinear relation, which is based on the assumption that water and solutes share a common pathway (13)

(4)
where Z = NPe/[(exp NPe) – 1] and the Peclet number (NPe) is given by NPe = Jv/A x (1 – {sigma})/Pd. To estimate {sigma}, it was assumed that Pd was constant and equal to the average Pd value measured before the pressure gradient application (when Jv = 0). During the 4-h time course of pressure gradient application, average values of Pe and Jv/A were determined every 15 min. Then, with the use of Eq. 4, {sigma} was calculated every 15 min.

Statistical analysis. Water and solute fluxes are presented graphically as means ± SE and were analyzed for statistical significance using repeated-measures ANOVA (P < 0.05). A Tukey-Kramer multiple-comparisons test was performed when indicated. Quantification of ZO-1 immunoreactivity or total protein was analyzed for significance by unpaired Student's t-test (P < 0.05).


   RESULTS
TOP
 ABSTRACT
 METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Effect of the hydrostatic pressure gradient on transport of water, albumin, and 70-kDa dextran. The water flux (Jv/A) and albumin permeability or dextran permeability (Pealb or Pedex) after the application of a 10-cmH2O hydrostatic pressure gradient were determined over a 4-h time course (Fig. 1, A and B, respectively). In all experiments, the Pd of albumin or dextran (Pdalb or Pddex) was recorded for 1 h before the ensuing application of transmural pressure for 4 h. The averages of Pdalb and Pddex over the first hour were 1.32 ± 0.18 x 10–6 and 0.60 ± 0.16 x 10–6 cm/s, respectively. The difference between the Pdalb and Pddex values was statistically significant at all time points (P < 0.01; not shown in Fig. 1, A and B). Application of a 10-cmH2O pressure gradient immediately increased albumin and 70-kDa dextran permeability by 3.2 ± 0.4- and 4.8 ± 1.3-fold at t = 60 min relative to the average Pdalb and Pddex values, respectively. Over the next 2 h, the Pealb and Pedex were reduced by 2.0 ± 0.3- and 2.1 ± 0.3-fold, respectively. The Pealb and Pedex values at the times indicated in Fig. 1, A and B, were significantly less than the Pealb and Pedex values at t = 60 min, respectively. The Jv/A values shown in Fig. 1, A and B, were significantly reduced by 3.5 ± 0.2- and 3.9 ± 0.4-fold at t = 180 min relative to their respective peak values at t = 60 min (P < 0.001). In addition, the Pealb and Pedex values were significantly different from each other at most time points measured between t = 60 and 180 min (P < 0.05; not shown in Fig. 1, A and B). Therefore, an additional transport pathway may be quantitatively more significant for albumin flux than for dextran flux.



View larger version (17K):
[in this window]
[in a new window]
  		Fig. 1. Transmural pressure gradient induces the reduction of transendothelial water flux (Jv/A) and albumin (Alb) flux [effective permeability of Alb (Pealb); A] or 70-kDa dextran (Dex) flux [effective permeability of Dex (Pedex); B] over time. During the first 60 min, diffusional permeability coefficients of Alb and Dex (Pdalb and Pddex) were determined every 5 min in the absence of a hydrostatic pressure gradient ({Delta}P). The averages of the Pdalb and Pddex values measured during the first hour were 1.32 ± 0.18 x 10–6 and 0.60 ± 0.16 x 10–6 cm/s, respectively. At time (t) = 60 min, a 10-cmH2O pressure gradient was applied, and both water and solute fluxes were measured for an additional 240 min. All of the Jv/A values measured after 5 min of pressure application in A and B (i.e., at t >= 65 min) were significantly less than the initial values at t = 60 min (***P < 0.001). A: all of the Alb flux measurements after 1 h of pressure application (i.e., at t >= 120) were significantly less than the flux measurement at t = 60 min (***P < 0.001). B: Dex flux measurements at the indicated times were significantly less than the initial measurement taken after pressure application (at t = 60 min). *P < 0.05, **P < 0.01, or ***P < 0.001. A Tukey-Kramer multiple-comparisons test was performed for the results shown in A and B. Each point represents the mean ± SE. For Alb transport experiments, n = 8; for Dex transport experiments, n = 9.

 

Effect of fixation on diffusive and convective permeability. The diffusive flux of albumin and 70-kDa dextran across untreated monolayers or monolayers that had been fixed with 1% paraformaldehyde for 10 min is presented in Fig. 2, A and B, respectively. The average Pdalb value of control monolayers was 1.55 ± 0.20 x 10–6 cm/s during the 55-min time course; the average Pdalb value of fixed monolayers was 0.42 ± 0.07 x 10–6 cm/s (a 3.9 ± 0.4-fold reduction). The reduced flux of albumin through fixed monolayers was significant, as indicated in Fig. 2A. However, the fixative solution did not alter the diffusive flux of dextran (Fig. 2B); the average Pddex values of control and fixed monolayers were 0.24 ± 0.06 x 10–6 and 0.26 ± 0.06 x 10–6 cm/s, respectively. There was also no significant difference between the Pddex values for control monolayers and Pdalb values for fixed monolayers. These experiments suggest that the diffusive flux of dextran is passive, i.e., not involving biological activity, whereas both passive and active mechanisms contribute to the measurement of Pdalb.



View larger version (13K):
[in this window]
[in a new window]
  		Fig. 2. Transendothelial transport through bovine aortic endothelial cell (BAEC) monolayers fixed with 1% paraformaldehyde. A: Pdalb of untreated [control (Con)] and 1% paraformadlehyde-treated (Fix) monolayers. The fixative solution significantly reduced Pdalb as indicated (n = 3). Significance was determined at each time point (*P < 0.05, **P < 0.01, or ***P < 0.001). B: Pddex of untreated and 1% paraformaldehyde-treated monolayers. The fixative solution did not alter Pddex (n = 3). C: response of Jv/A and Pedex to a 10-cmH2O pressure gradient applied at t = 60 min across fixed monolayers. Fixation prevented the pressure-induced reduction in Pedex. The Jv/A values measured after 5 min of pressure application (i.e., at t >= 65 min) were significantly less than the initial value at t = 60 min (***P < 0.001). In A–C, each point represents the mean ± SE. A Tukey-Kramer multiple-comparisons test was performed at all time points.

 

To determine the effect of fixation on the sealing of BAEC monolayers, the transport of water and 70-kDa dextran was measured for fixed monolayers after the application of a 10-cmH2O pressure gradient (Fig. 2C). The fixation eliminated Pedex sealing and partially attenuated Jv/A sealing (Jv/A was reduced 2.3 ± 0.3-fold after 2 h of pressure gradient application). Also, the Pedex values were significantly higher through fixed monolayers than through control monolayers (most time points P < 0.01; not indicated in Fig. 2C), suggesting that the sealing effect is not purely physical in origin. Furthermore, Jv/A was significantly reduced compared with the initial Jv/A measurement at t = 60 min (P < 0.001; as indicated in Fig. 2C), whereas Pedex remained approximately constant. Thus a pathway for transendothelial water flow that excludes 70-kDa dextran significantly contributes to the sealing of Lp, at least during the initial 30 min of pressure application.

Effect of the hydrostatic pressure gradient on immunoreactivity of ZO-1. In light of the pressure gradient-induced sealing effect described above and the evidence that sealing is not entirely physical in nature, we examined the distribution of ZO-1 in monolayers after application of the pressure gradient by immunofluorescence confocal microscopy. In monolayers exposed to a 0-cmH2O pressure gradient, ZO-1 was predominantly localized at the BAEC borders in a discontinuous fashion (Fig. 3A). After a hydrostatic pressure gradient of 10 cmH2O was applied for 4 h, the intensity of ZO-1 immunoreactivity increased around the borders, and there were fewer discontinuities in the ZO-1 staining pattern along the cell borders (Fig. 3B). Figure 3C summarizes the quantification of ZO-1 immunoreactivity levels in 24 paired digital images of 15 paired monolayers that were exposed or unexposed to a 10-cmH2O pressure gradient. The pressure gradient ({Delta}P) increased ZO-1 immunoreactivity levels 1.4 ± 0.1-fold (P < 0.001) relative to control monolayers unexposed to pressure. In addition, the effect of the pressure gradient on the actin cytoskeleton was determined by staining with rhodamine isothiocyanate-phalloidin. No change in the actin immunoreactivity was observed by confocal microscopy after 4 h of pressure compared with controls (data not shown).



View larger version (29K):
[in this window]
[in a new window]
  		Fig. 3. Transmural pressure gradient increases zonnula occludens (ZO)-1 immunoreactivity at cell-cell contacts. A: immunoreactivity of ZO-1 in BAEC monolayers exposed to a 0-cmH2O pressure differential for 4 h was generally continuous around cell boundaries. B: when monolayers were exposed to a pressure differential of 10 cmH2O for 4 h, the intensity of ZO-1 immunoreactivity appeared significantly increased around the cell boundaries compared with A. Bar = 20 µm. C: after 4 h, total ZO-1 intensity increased 1.4 ± 0.1-fold in monolayers exposed to a 10-cmH2O pressure gradient ({Delta}P) relative to monolayers exposed to 0-cmH2O pressure gradient (Con) (N = 15 paired monolayers, n = 24 paired digital images). ***P < 0.001.

 

Effect of transmural pressure on total ZO-1 protein content. ZO-1 protein content in BAEC monolayers subjected to 0-(control) or 10-cmH2O differential pressure ({Delta}P) for 4 h was determined by Western blot (Fig. 4A). Application of the hydrostatic pressure gradient did not significantly alter total ZO-1 content of the monolayers (Fig. 4B); thus transmural pressure may stimulate the mobilization of ZO-1 from a presynthesized pool.



View larger version (40K):
[in this window]
[in a new window]
  		Fig. 4. Transmural pressure gradient does not change the total content of ZO-1 protein. A: BAEC monolayers unexposed to transmural pressure (Con) and monolayers exposed to a 10-cmH2O hydrostatic pressure gradient for 4 h ({Delta}P) were subjected to SDS-PAGE, transferred to nitrocellulose membranes, and immunoblotted with anti-ZO-1 R40–76 (monoclonal) antibody. B: there was no significant difference in total ZO-1 protein expression (P = 0.28; n = 3) between monolayers exposed to the hydrostatic pressure gradient for 4 h ({Delta}P) vs. Con.

 

Effect of CytoD and colchicine on ZO-1 immunoreactivity in the absence of a pressure gradient. ZO-1 immunoreactivity in a monolayer unexposed to the cytoskeletal inhibitors is shown in Fig. 5A. The ZO-1 immunoreactivity in monolayers treated with 10 µM CytoD (Fig. 5B) appeared more discontinuous and less intense at cell-cell contacts compared with control monolayers (Fig. 5A). On the other hand, the staining pattern of ZO-1 in monolayers treated with 1 µM colchicine (Fig. 5C) appeared identical to untreated monolayers (Fig. 5A). These results are consistent with the CytoD-induced increase in Pddex observed in Fig. 6 and the constant Pddex values measured after colchicine addition (see Fig. 8).



View larger version (61K):
[in this window]
[in a new window]
  		Fig. 5. Effect of cytochalasin D (CytoD) and colchicine on ZO-1 immunore-activity in unpressurized monolayers. A: ZO-1 reactivity in a monolayer exposed to a 0-cmH2O pressure gradient for 2 h. B: ZO-1 reactivity in a monolayer treated with 10 µM CytoD during the last hour of a 2-h exposure to a 0-cmH2O pressure gradient. C: ZO-1 reactivity in a monolayer treated with 1 µM colchicine during the last hour of a 2-h exposure to a 0-cmH2O pressure gradient. Bar = 20 µm.

 


View larger version (21K):
[in this window]
[in a new window]
  		Fig. 6. Effect of CytoD on transmural transport. CytoD treatment increased water and 70-kDa Dex transport through BAEC monolayers, but the sealing effect was still observed. Pddex values were measured during the first 60 min. At t = 60 min, 10 µM CytoD was added and Pddex was measured for an additional 60 min. At t = 120 min, a hydrostatic pressure gradient of 10 cmH2O was applied, and both Jv/A and Pedex were determined for an additional 240 min. The Jv/A values measured after 5 min of pressure application (i.e., at t >= 65 min) were significantly less than the initial value at t = 60 min (***P < 0.001). The Pedex values were significantly reduced after 90 min of pressure application (i.e., a significance of P < 0.05 was observed at t >= 210 min relative to the initial value at t = 120 min). *P < 0.05, **P < 0.01, or ***P < 0.001 (by Tukey-Kramer multiple-comparisons test). Each point represents the mean ± SE (n = 5).

 


View larger version (19K):
[in this window]
[in a new window]
  		Fig. 8. Effect of colchicine on transmural transport. Colchicine treatment eliminated the sealing effect for 70-kDa Dex transport. BAEC Pddex (Jv = 0) was measured during the first 60 min. At t = 60 min, 1 µM colchicine was added, and Pddex was measured for an additional 60 min. At t = 120 min, a hydrostatic pressure gradient of 10 cmH2O was applied, and both Jv/A and Pedex were determined for an additional 240 min. An abbreviated sealing effect for water was observed; Jv/A values were significantly reduced compared with the initial value at t = 120 min, as indicated (*P < 0.05 by Tukey-Kramer multiple-comparisons test). Each point represents the mean ± SE (n = 3).

 

Effect of CytoD on transendothelial transport and ZO-1 immunoreactivity during pressure gradient application. To determine whether actin filaments participate in the sealing response of endothelial cells, transport properties of CytoD-treated monolayers were measured in real time after pressure application (Fig. 6). The average Pddex was 0.56 ± 0.17 x 10–6 cm/s before the addition of CytoD at t = 60 min. After the addition of 10 µM CytoD, the average Pddex value between t = 60 and 115 min was 2.15 ± 0.30 x 10–6 cm/s (a 4.5 ± 0.7-fold increase relative to the Pddex values measured before the CytoD addition). Application of a 10-cmH2O pressure gradient across CytoD-treated monolayers immediately increased dextran permeability to its maximum value of 9.41 ± 1.96 x 10–6 cm/s at t = 120 min. Afterward, the sealing effect significantly reduced both Pedex and Jv/A values compared with the initial measurements at t = 120, as indicated in Fig. 6. After 4 h of pressure application, Pedex and Jv/A were reduced by 2.8 ± 1.0- and 1.7 ± 0.2-fold. The pressure-induced reduction in transport properties in the presence of CytoD indicates that the sealing process is not dependent on an intact microfilament network. However, disruption of the microfilament network significantly increases Lp, Pe, and the number of discontinuities in the TJ strand.

The effect of CytoD on ZO-1 immunoreactivity after pressure gradient application was investigated. After a 4-h application of transmural pressure to monolayers that were not treated with CytoD, the intensity of ZO-1 reactivity was increased at cell-cell contacts (Fig. 7B) relative to monolayers unexposed to transmural pressure over the same time course (Fig. 7A). For the results shown in Fig. 7C, CytoD was added to the monolayer 1 h before pressure exposure; afterward, a 10-cmH2O pressure gradient was applied across the monolayer for 4 h. The ZO-1 reactivity levels in Fig. 7C also increased relative to monolayers unexposed to pressure (Fig. 7A); however, the ZO-1 reactivity appeared significantly more discontinuous around the cell borders compared with pressurized monolayers unexposed to CytoD (Fig. 7B). Again, it appears that intact microfilaments are not required to shuttle ZO-1 to the junction but are required to assemble the protein in a continuous fashion at the cell border.



View larger version (38K):
[in this window]
[in a new window]
  		Fig. 7. Effect of CytoD on ZO-1 immunoreactivity after pressure gradient application. CytoD caused discontinuous ZO-1 immunoreactivity but did not inhibit the pressure gradient-induced increase in ZO-1 immunoreactivity at cell-cell contacts. A: ZO-1 reactivity in a monolayer exposed to a 0-cmH2O pressure gradient for 6 h. B: ZO-1 reactivity in a monolayer initially exposed to 0-cmH2O differential pressure for 2 h, followed by a 4-h application of 10-cmH2O differential pressure. C: ZO-1 reactivity in monolayer treated with 10 µM CytoD during the last hour of a 2-h exposure to a 0-cmH2O pressure gradient; afterward, a 10-cmH2O gradient was applied across the CytoD-treated monolayer for an additional 4 h. Bar = 20 µm.

 

Effect of colchicine on transendothelial transport and ZO-1 immunoreactivity during pressure application. To investigate the role of microtubules in the sealing response, the transport properties of colchicine-treated monolayers were measured in real time after pressure application (Fig. 8). The average Pddex was 0.25 ± 0.11 x 10–6 cm/s before the addition of colchicine at t = 60 min. The addition of 1 µM colchicine did not significantly alter the Pddex values measured before the application of transmural pressure (average Pddex = 0.27 ± 0.06 x 10–6 cm/s). Application of a 10-cmH2O pressure gradient across colchicine-treated monolayers at t = 120 min immediately increased dextran permeability to 2.25 ± 0.49 x 10–6 cm/s (a 9.1 ± 1.5-fold increase relative to the average Pddex value). Microtubule disruption completely eliminated the reduction in Pedex, but the reduction in Jv/A was still significant after the pressure application, as indicated in Fig. 8. After 180 min, Pedex and Jv/A increased. The Pedex values measured during the last hour of pressure application were significantly larger than the initial Pedex measurement at t = 120 min (P < 0.01; not shown in Fig. 9). Thus the reduction in Lp during the first 30 min of pressure gradient application is microtubule independent; afterward, an intact microtubule network is required to maintain sealing of the monolayers.



View larger version (44K):
[in this window]
[in a new window]
  		Fig. 9. Effect of colchicine on ZO-1 immunoreactivity after pressure gradient application. Colchicine altered ZO-1 immunoreactivity at cell-cell contacts after pressure application for 4 h. A: ZO-1 reactivity in a monolayer exposed to a 0-cmH2O pressure gradient for 6 h. B: ZO-1 reactivity in a monolayer initially exposed to 0-cmH2O differential pressure for 2 h, followed by a 4-h application of 10-cmH2O differential pressure. C: ZO-1 reactivity in a monolayer treated with 1 µM colchicine during the last hour of a 2-h exposure to a 0-cmH2O pressure gradient; afterward, a 10-cmH2O gradient was applied across the colchicine-treated monolayer for an additional 4 h. Bar = 20 µm.

 

The effect of colchicine on ZO-1 immunoreactivity after pressure gradient application was also investigated. As observed previously, a 4-h application of transmural pressure to monolayers that were not treated with colchicine increased the intensity of ZO-1 reactivity at cell-cell contacts (Fig. 9B) relative to monolayers unexposed to transmural pressure over the same time course (Fig. 9A). For the results in Fig. 9C, colchicine was added to the monolayer 1 h before pressure exposure; afterward, a 10-cmH2O pressure gradient was applied across the monolayer for 4 h. The intensity of ZO-1 reactivity shown in Fig. 9C was not increased compared with monolayers unexposed to pressure (Fig. 9A), and the distribution of ZO-1 reactivity appeared irregular around the border of most cells. These results also suggest that microtubules mediate the translocation of ZO-1 to the cell borders during the sealing process. It should also be noted that after colchicine treatment, a small fraction of the cells were washed away during the immunostaining protocol; however, the monolayers were confluent, as seen by phase-contrast microscopy immediately after the pressure application.

Effect of CytoD and colchicine on {sigma}. To estimate {sigma}(t) values during the 4-h time course of pressure gradient application, average values of Pe(t) and Jv/A(t) were determined every 15 min. It was also assumed that Pd was constant and equal to the average Pd value measured before pressure gradient application (when Jv/A = 0). Then, with the use of Eq. 4, {sigma}(t) was calculated every 15 min. To solve for {sigma} in cytoskeletal inhibitor-treated monolayers, the Pddex used in Eq. 4 was set equal to the value measured after CytoD or colchicine addition. Because the Pddex magnitudes after CytoD addition steadily increased over the next hour (see Fig. 6 between t = 60 min and 120 min), the average Pddex value between t = 105 and 115 min (2.64 ± 0.28 x 10–6 cm/s) was assumed to be equal to the steady-state value. This approximation is reasonable because Pd values that underestimate the actual steady-state value by 25% typically only cause 10% errors in the calculations of {sigma}(t) after CytoD addition. Colchicine treatment, on the other hand, did not alter Pddex over a 1-h time course (see Fig. 8); therefore, the Pddex value used in Eq. 4 was set equal to the average Pddex value measured between t = 60 min and 115 min (0.27 ± 0.06 x 10–6 cm/s).

Neither CytoD nor colchicine-treatment significantly altered the BAEC {sigma} magnitudes to dextran compared with untreated monolayers (Fig. 10). The average {sigma} value during the 4-h exposure to CytoD was 0.69 ± 0.05 compared with 0.67 ± 0.08 for control (P > 0.86). The average {sigma} value during the 4-h exposure to colchicine was 0.68 ± 0.04 compared with 0.67 ± 0.08 for control (P > 0.90). These results suggest that intact TJs do not determine solute selectivity ({sigma}).



View larger version (20K):
[in this window]
[in a new window]
  		Fig. 10. Cytoskeletal inhibitors do not alter the reflection coefficient. Neither CytoD (10 µM) nor colchicine (1 µM) significantly altered the reflection coefficient magnitudes to Dex compared with Con. Each point represents the mean ± SE (n = 5 for CytoD and colchicine and n = 9 for Con).

 


   DISCUSSION
TOP
 ABSTRACT
 METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
Endothelial cells are naturally found in a mechanically active environment. Several studies have shown that endothelial cells alter their morphology, growth rate, and metabolism in response to the wall shear stress generated by blood flow (for reviews, see Refs. 14, 22, and 41). Traditionally, endothelial transport properties (Lp, Pd, and {sigma}) have been regarded as passive (constant) properties of the endothelium; however, it is now becoming clear that endothelial permeability coefficients depend on hemodynamic forces. Several investigators have shown that endothelial Lp increases in response to an increase in fluid wall shear stress using in vitro (11, 15, 19, 40) and in vivo (29, 33, 49, 50) models. In the present study, a step change in the hydrostatic pressure gradient from 0 to 10 cmH2O induced a prominent sealing effect in BAEC monolayers (Fig. 1). This phenomenon significantly reduced water, albumin, and dextran fluxes to minimum values ~2 h after application of a 10-cmH2O pressure gradient. Other investigators, using cells in culture (6, 40, 44, 46) and intact arteries (29, 45), have also observed a similar pressure-induced reduction in endothelial transport coefficients (Lp and Pe). This adaptive response or sealing effect may provide a mechanism by which the body protects against edema formation when vascular pressure increases suddenly.

A previous study (46) of the endothelial sealing phenomenon suggested that it is a physical process not requiring biological activity. This group reported that the sealing of bovine pulmonary artery endothelial monolayers in vitro was quantitatively similar to the sealing of fixed monolayers and speculated that the mechanism is a result of pressure-induced deformations of transendothelial channels. However, it should be pointed out that the postsealing Lp value in that study was more than an order of magnitude greater than the postsealing Lp values reported in the present study (41 x 10–7 vs. 3 x 10–7 cm·s–1·cmH2O–1, respectively) and Pe was not measured.

The results of the present study indicate that sealing in BAEC monolayers depends on biological mechanisms. The reduction in Lp and Pe magnitudes during the sealing effect was coincident with increased ZO-1 immunoreactivity at cell-cell contacts (Fig. 3). Fixation increased Pe values compared with control values and eliminated the pressure-induced reduction in Pe values over time (Fig. 2C). Also, the reduction in Lp across fixed BAEC monolayers was largely attenuated compared with untreated monolayers; specifically, Lp was reduced 2.3 ± 0.3-fold for fixed monolayers versus 3.7 ± 0.3-fold for untreated monolayers after 2 h of pressure application. Nonetheless, a very significant reduction in Lp (but not Pe) was observed for fixed monolayers, indicating that the sealing phenomenon may include a passive or mechanical component that reduces Lp.

A large body of evidence indicates that actin plays a significant role in controlling endothelial permeability. Various actin disrupting drugs such as cytokinins, phalloidin, and cytochalasins disturb TJ structure and permeability (9, 31, 38). We observed that CytoD induced significant discontinuity of ZO-1 immunoreactivity (Fig. 5) and dramatically elevated Lp, Pd, and Pedex (Fig. 6). Surprisingly, however, Lp and Pe were significantly reduced by 1.7 ± 0.2- and 2.8 ± 1.0-fold, respectively, after 4-h of pressure gradient application across CytoD-treated monolayers. Furthermore, CytoD treatment did not inhibit the pressure-induced increase in ZO-1 immunoreactivity at the cell-cell contacts (Fig. 7), albeit a discontinuous staining pattern was observed. These results suggest that the translocation of ZO-1 to the cell border does not depend on intact actin filaments because both Lp and Pe were significantly reduced over time and the immunoreactivity of ZO-1 was increased at the cell border. However, the maintenance of a continuous junctional complex does depend on intact actin filaments.

In contrast to microfilamentous actin, the role of microtubules in endothelial cell barrier regulation has received much less attention in the literature. Historically, microtubules and actin filaments have been viewed as separate and distinct cytoskeletal networks, but are now known to interact functionally during dynamic cellular processes (17, 27). Treatment of BAEC monolayers with colchicine, a microtubule disrupting agent, did not alter Pddex or the immunoreactivity of ZO-1 at the cell-cell contacts after 1 h (Fig. 9). The sudden application of a pressure gradient across colchicine-treated monolayers did not attenuate the Lp reduction over the next 30 min relative to untreated monolayers (Fig. 8). However, the pressure-induced reduction in Pe was eliminated and the sealing of Lp was reversed over longer time periods. These results, together with the measurements of water and solute flux across fixed monolayers, imply that the initial reduction in Lp after pressure gradient application is primarily due to a mechanical or passive sealing process; over longer time periods, a functional microtubule network is required to induce a reduction in Pe and a further reduction in Lp. It should also be noted that an increase in transmural pressure did not stimulate an increase in total ZO-1 protein. Therefore, the transmural pressure gradient may stimulate the recruitment of presynthesized TJ proteins, and possibly other molecules that induce TJ assembly, to cell-cell contacts via vesicles that travel along microtubules. Suttorp et al. (44) reported that cooling endothelial monolayers to 4°C attenuated the endothelial sealing effect, which also suggests that the phenomenon requires biological activity such as vesicular-mediated protein transport along microtubules.

The data from the present study also suggest that two pathways are available for water transport through the interendothelial cleft: the first is through discontinuities in the TJ strand, and the second through narrow pores that exclude large solutes (1, 10). While there is little doubt that discontinuities in the TJ strand provide a paracellular pathway for water and solutes, it is not clear if a separate small pore pathway contributes to the transendothelial flux of water (16). Here, Jv/A across untreated monolayers was reduced to a larger extent and over a shorter time course than the reduction in Pedex or Pealb (Fig. 1). Furthermore, colchicine treatment and fixation did not attenuate the reduction in Jv/A from occurring immediately after pressure application but did block the reduction in Pedex (Figs. 8 and 2C, respectively). These results suggest that the application of transmural pressure immediately induces a mechanical or passive adaptive response by possibly compressing narrow gaps (small pores) within the TJ strand that allow the transport of water but not 70-kDa dextran. Figure 11 shows a schematic of a two-pathway model of paracellular transport, which includes a small pore system (distance h across the narrow slit) and larger pores formed from the discontinuities in the TJ strand. We propose that an increase in the hydrostatic pressure gradient leads to a passive reduction in h even after treatment with the cytoskeletal inhibitors and paraformaldehyde. The biological component of the sealing effect, on the other hand, mobilizes ZO-1 at cell-cell contacts via an intact microtubule network, which effectively reduces the length (L) of the TJ discontinuities. Fixed and colchicine-treated monolayers would, therefore, lack a mechanism to transport TJ proteins to the cell border. In CytoD-treated monolayers, all of the dimensions (h, L, and width W) are presumably larger due to the lack of microfilament support; after pressure application, h is passively reduced and the biological response reduces L because the microtubules are relatively intact.



View larger version (20K):
[in this window]
[in a new window]
  		Fig. 11. A two-pathway model for paracellular transport (top view). In frog mesenteric capillaries, the width of the interendothelial cleft, W, is ~15–20 nm, and the distance across the slit (or small pore diameter), h, is 2 nm (1). The length, L, and frequency of the tight junction (TJ) break may be more cell specific than h or W. In Con monolayers, the passive component of the sealing effect reduces h, whereas the active component (TJ recruitment to cell borders via microtubules) reduces L. In colchicine-treated monolayers, h is reduced but L is constant, at least initially, because there is no mechanism to transport TJ proteins. In CytoD-treated monolayers, both L and W are larger due to lack of microfilament support; after pressure application, the physical and biological components of the sealing effect reduce h and L, respectively.

 

The addition of the cytoskeletal inhibitors not only allows us to examine the dependence of the adaptive response on the cytoskeleton but also provides an assessment of the structures responsible for the sieving of solutes. Through the use of combined experimental and theoretical approaches, other authors have shown in vivo that the glycocalyx (20, 21, 35, 48) serves as the primary molecular filter for plasma proteins and thus determines the magnitude of {sigma} (20, 21, 35, 48). Here, the simultaneous measurements of Lp and Pe permitted a straightforward estimation of {sigma} that was similar in magnitude (~0.7) to estimations typically reported in vivo. The addition of the CytoD did not alter the value of {sigma}, indicating that structures other than the TJ strand determine the selectivity of BAEC monolayers because the paracellular pathways are wide open (TJs are disassembled). These results strongly suggest the presence of a material within the interendothelial cleft that acts as a molecular filter. The underlying physical picture associated with this model is similar to the one described by Weinbaum and Curry (20): breaks in the TJ provide the dominant transport pathways, but selectivity to solutes is conferred by the glycocalyx within the interendothelial cleft.

It should be noted that {sigma}alb values (data not shown) were lower than the {sigma}dex values, averaging 0.53 ± 0.02 over the time course of pressure application. These results indicate that solvent drag also contributes to paracellular albumin flux in BAEC monolayers. However, transcellular transport of albumin (for a review, see Ref. 36) appears to significantly contribute to the total flux of albumin. The Pd for 70-kDa dextran was significantly lower than the Pd for albumin at all time points in Fig. 1 (P < 0.01). In separate experiments (Fig. 2A), the diffusive flux of albumin through unfixed monolayers (1.55 ± 0.20 x 10–6 cm/s) was significantly greater than the flux through fixed monolayers (0.42 ± 0.07 x 10–6 cm/s). However, paraformaldehyde treatment did not affect the diffusive flux of dextran (Fig. 2B). These results suggest that fixation blocked the active transcellular transport of albumin and that transcellular transport provides a quantitatively significant pathway for albumin through control monolayers.

We conclude that transmural pressure gradients stimulate biological and mechanical responses in BAEC monolayers. Our data support the presence of a small pore pathway for water that appears to seal passively after pressure application. After longer periods of pressure application, a significant increase in ZO-1 expression at cell-cell contacts was observed. We envision that the mechanical component of the sealing process reduces the size of the small pores within the TJ strand and the biological component reduces the length of existing TJ breaks along endothelial cell borders. Also, junctional complexes dictate the magnitudes of Lp and Pe, but selectivity to solutes is probably conferred by a glycocalyx.


   ACKNOWLEDGMENTS
 
GRANTS

This study was supported by National Institutes of Health Grants HL-57093 and EY-12021, National Aeronautics and Space Administration Grant NAG3–2746, and Juvenile Diabetes Research Foundation Grant JDRF-298212.


   FOOTNOTES
 

Address for reprint requests and other correspondence: D. A. Antonetti, The Pennsylvania State Univ., College of Medicine, 500 University Dr., Hershey, PA 17033 (E-mail: dantonetti{at}psu.edu).

The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


   REFERENCES
TOP
 ABSTRACT
 METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Adamson RH and Michel CC. Pathways through the intercellular clefts of frog mesenteric capillaries. J Physiol 466: 303–327, 1993.[Abstract/Free Full Text]
  2. Anderson JM. Cell signalling: MAGUK magic. Curr Biol 6: 382–384, 1996.[CrossRef][ISI][Medline]
  3. Anderson JM, Stevenson BR, Jesaitis LA, Goodenough DA, and Mooseker MS. Characterization of ZO-1, a protein component of the tight junction from mouse liver and Madin-Darby canine kidney cells. J Cell Biol 106: 1141–1149, 1988.[Abstract/Free Full Text]
  4. Antonetti DA, Barber AJ, Hollinger LA, Wolpert EB, and Gardner TW. Vascular endothelial growth factor induces rapid phosphorylation of tight junction proteins occludin and zonula occluden 1. A potential mechanism for vascular permeability in diabetic retinopathy and tumors. J Biol Chem 274: 23463–23467, 1999.[Abstract/Free Full Text]
  5. Antonetti DA, Wolpert EB, DeMaio L, Harhaj NS, and Scaduto RC Jr. Hydrocortisone decreases retinal endothelial cell water and solute flux coincident with increased content and decreased phosphorylation of occludin. J Neurochem 80: 667–677, 2002.[CrossRef][ISI][Medline]
  6. Baetscher M and Brune K. An in vitro system for measuring endothelial permeability under hydrostatic pressure. Exp Cell Res 148: 541–547, 1983.[CrossRef][ISI][Medline]
  7. Balda MS, Anderson JM, and Matter K. The SH3 domain of the tight junction protein ZO-1 binds to a serine protein kinase that phosphorylates a region C-terminal to this domain. FEBS Lett 399: 326–332, 1996.[CrossRef][ISI][Medline]
  8. Bazzoni G and Dejana E. Pores in the sieve and channels in the wall: control of paracellular permeability by junctional proteins in endothelial cells. Microcirculation 8: 143–152, 2001.[CrossRef][ISI][Medline]
  9. Bentzel CJ, Hainau B, Edelman A, Anagnostopoulos T, and Benedetti EL. Effect of plant cytokinins on microfilaments and tight junction permeability. Nature 264: 666–668, 1976.[CrossRef][Medline]
  10. Bundgaard M. The three-dimensional organization of tight junctions in a capillary endothelium revealed by serial-section electron microscopy. J Ultrastruct Res 88: 1–17, 1984.[CrossRef][ISI][Medline]
  11. Chang YS, Yaccino JA, Lakshminarayanan S, Frangos JA, and Tarbell JM. Shear-induced increase in hydraulic conductivity in endothelial cells is mediated by a nitric oxide-dependent mechanism. Arterioscler Thromb Vasc Biol 20: 35–42, 2000.[Abstract/Free Full Text]
  12. Civelek M, Ainslie K, Garanich JS, and Tarbell JM. Smooth muscle cells contract in response to fluid flow via a Ca2+-independent signaling mechanism. J Appl Physiol 93: 1907–1917, 2002.[Abstract/Free Full Text]
  13. Curry F-RE. Mechanics and thermodynamics of transcapillary exchange. In: Handbook of Physiology. Microcirculation. Bethesda, MD: Am. Physiol. Soc., 1984, sect. 2, vol. IV, pt. 1, chapt. 8, p. 309–374.
  14. Davies PF. Flow-mediated endothelial mechanotransduction. Physiol Rev 75: 519–560, 1995.[Abstract/Free Full Text]
  15. DeMaio L, Chang YS, Gardner TW, Tarbell JM, and Antonetti DA. Shear stress regulates occludin content and phosphorylation. Am J Physiol Heart Circ Physiol 281: H105–H113, 2001.[Abstract/Free Full Text]
  16. Fu BM, Weinbaum S, Tsay RY, and Curry FE. A junction-orifice-fiber entrance layer model for capillary permeability: application to frog mesenteric capillaries. J Biomech Eng 116: 502–513, 1994.[ISI][Medline]
  17. Goode BL, Drubin DG, and Barnes G. Functional cooperation between the microtubule and actin cytoskeletons. Curr Opin Cell Biol 12: 63–71, 2000.[CrossRef][ISI][Medline]
  18. Hill MA, Zou H, Davis MJ, Potocnik SJ, and Price S. Transient increases in diameter and [Ca2+]i are not obligatory for myogenic constriction. Am J Physiol Heart Circ Physiol 278: H345–H352, 2000.[Abstract/Free Full Text]
  19. Hillsley MV and Tarbell JM. Oscillatory shear alters endothelial hydraulic conductivity and nitric oxide levels. Biochem Biophys Res Commun 293: 1466–1471, 2002.[CrossRef][ISI][Medline]
  20. Hu X, Adamson RH, Liu B, Curry FE, and Weinbaum S. Starling forces that oppose filtration after tissue oncotic pressure is increased. Am J Physiol Heart Circ Physiol 279: H1724–H1736, 2000.[Abstract/Free Full Text]
  21. Hu X and Weinbaum S. A new view of Starling's hypothesis at the microstructural level. Microvasc Res 58: 281–304, 1999.[CrossRef][ISI][Medline]
  22. Ishida T, Takahashi M, Corson MA, and Berk BC. Fluid shear stress-mediated signal transduction: how do endothelial cells transduce mechanical force into biological responses? Ann NY Acad Sci 811: 12–23, 1997.[Abstract/Free Full Text]
  23. Jo H, Dull RO, Hollis TM, and Tarbell JM. Endothelial albumin permeability is shear dependent, time dependent, and reversible. Am J Physiol Heart Circ Physiol 260: H1992–H1996, 1991.[Abstract/Free Full Text]
  24. Kedem O and Katchalsky A. Thermodynamic analysis of the permeability of biological membranes to non-electrolytes. Biochim Biophys Acta 2: 229–246, 1958.
  25. Kevil CG, Okayama N, Trocha SD, Kalogeris TJ, Coe LL, Specian RD, Davis CP, and Alexander JS. Expression of zonula occludens and adherens junctional proteins in human venous and arterial endothelial cells: role of occludin in endothelial solute barriers. Microcirculation 5: 197–210, 1998.[CrossRef][ISI][Medline]
  26. Kim MHN and Tarbell J. Control of the arteriolar myogenic response by transmural fluid filtration. Second Joint EMBS-BMES Conference: IEEE, 2002, p. 247, 2002.
  27. Klymkowsky MW. Weaving a tangled web: the interconnected cytoskeleton. Nat Cell Biol 1: E121–E123, 1999.[CrossRef][ISI][Medline]
  28. Lakshminarayanan S, Gardner TW, and Tarbell JM. Effect of shear stress on the hydraulic conductivity of cultured bovine retinal microvascular endothelial cell monolayers. Curr Eye Res 21: 944–951, 2000.[CrossRef][ISI][Medline]
  29. Lever MJ, Tarbell JM, and Caro CG. The effect of luminal flow in rabbit carotid artery on transmural fluid transport. Exp Physiol 77: 553–563, 1992.[Abstract]
  30. Madara JL. Regulation of the movement of solutes across tight junctions. Annu Rev Physiol 60: 143–159, 1998.[CrossRef][ISI][Medline]
  31. Madara JL and Pappenheimer JR. Structural basis for physiological regulation of paracellular pathways in intestinal epithelia. J Membr Biol 100: 149–164, 1987.[CrossRef][ISI][Medline]
  32. McIntire LV, Wagner JE, Papadaki M, Whitson PA, and Eskin SG. Effect of flow on gene regulation in smooth muscle cells and macromolecular transport across endothelial cell monolayers. Biol Bull 194: 394–399, 1998.[ISI][Medline]
  33. McIntire LV, Wagner JE, and Whitson PA. Effect of flow on macromolecular transport across bovine brain endothelial cell monolayers. Bioeng Con ASME BED 9: 79–80, 1995.
  34. Michel CC. Microvascular permeability, venous stasis and oedema. Int Angiol 8: 9–13, 1989.[ISI][Medline]
  35. Michel CC. Starling: the formulation of his hypothesis of microvascular fluid exchange and its significance after 100 years. Exp Physiol 82: 1–30, 1997.[ISI][Medline]
  36. Michel CC and Curry FE. Microvascular permeability. Physiol Rev 79: 703–761, 1999.[Abstract/Free Full Text]
  37. Mitic LL and Anderson JM. Molecular architecture of tight junctions. Annu Rev Physiol 60: 121–142, 1998.[CrossRef][ISI][Medline]
  38. Montesano R, Gabbiani G, Perrelet A, and Orci L. In vivo induction of tight junction proliferation in rat liver. J Cell Biol 68: 793–798, 1976.[Abstract/Free Full Text]
  39. Nag S. Role of the endothelial cytoskeleton in blood-brain-barrier permeability to protein. Acta Neuropathol (Berl) 90: 454–460, 1995.[Medline]
  40. Sill HW, Chang YS, Artman JR, Frangos JA, Hollis TM, and Tarbell JM. Shear stress increases hydraulic conductivity of cultured endothelial monolayers. Am J Physiol Heart Circ Physiol 268: H535–H543, 1995.[Abstract/Free Full Text]
  41. Stamatas GN and McIntire LV. Rapid flow-induced responses in endothelial cells. Biotechnol Prog 17: 383–402, 2001.[CrossRef][Medline]
  42. Starling EH. On the absorption of fluids from connective tissue spaces. J Physiol 19: 312–326, 1896.
  43. Stevenson BR and Begg DA. Concentration-dependent effects of cytochalasin D on tight junctions and actin filaments in MDCK epithelial cells. J Cell Sci 107: 367–375, 1994.[Abstract]
  44. Suttorp N, Hessz T, Seeger W, Wilke A, Koob R, Lutz F, and Drenckhahn D. Bacterial exotoxins and endothelial permeability for water and albumin in vitro. Am J Physiol Cell Physiol 255: C368–C376, 1988.[Abstract/Free Full Text]
  45. Tedgui A and Lever MJ. Filtration through damaged and undamaged rabbit thoracic aorta. Am J Physiol Heart Circ Physiol 247: H784–H791, 1984.[Abstract/Free Full Text]
  46. Turner MR. Flows of liquid and electrical current through monolayers of cultured bovine arterial endothelium. J Physiol 449: 1–20, 1992.<