Fibrinogen induces endothelial cell permeability

Abstract

Many cardiovascular and cerebrovascular disorders are accompanied by an increased blood content of fibrinogen (Fg), a high molecular weight plasma adhesion protein. Fg is a biomarker of inflammation and its degradation products have been associated with microvascular leakage. We tested the hypothesis that at pathologically high levels, Fg increases endothelial cell (EC) permeability through extracellular signal regulated kinase (ERK) signaling and by inducing F-actin formation. In cultured ECs, Fg binding to intercellular adhesion molecule-1 and to α₅β₁ integrin, caused phosphorylation of ERK. Subsequently, F-actin formation increased and coincided with formation of gaps between ECs, which corresponded with increased permeability of ECs to albumin. Our data suggest that formation of F-actin and gaps may be the mechanism for increased albumin leakage through the EC monolayer. The present study indicates that elevated un-degraded Fg may be a factor causing microvascular permeability that typically accompanies cardiovascular and cerebrovascular disorders.

Introduction

One manifestation of inflammation is an increased microvascular permeability that results in a net loss of blood plasma components into the tissue and causes edema. Albumin, the most abundant plasma protein, helps maintain oncotic pressure and protects endothelial barrier integrity by interacting with the glycocalyx [1, 2]. Thus, changes in albumin transport through the endothelial cell (EC) barrier may cause significant tissue damage and induce inflammatory responses [3].

Fibrinogen (Fg) is another plasma protein and a biomarker of inflammation [4, 5]. When elevated, it identifies individuals who have a high risk for cardiovascular disorders [4, 6]. In a meta-analysis of six cohort studies, Ernst and Resch [7] estimated that individuals in the upper tertile of Fg concentration had a 2.3-fold greater risk of subsequent cardiovascular disease than individuals in the lower tertile. These data were recently confirmed by a more comprehensive meta-analysis of 18 studies [4]. Increased plasma Fg concentration typically accompanies development of diseases such as hypertension [8, 9], diabetes [10], and stroke [11], which involve inflammatory processes. In patients with essential hypertension, for example, plasma Fg levels are higher (3.8 ± 0.1 mg/ml) than in normotensive controls (2.8 ± 0.1 mg/ml) [8]. We found that Fg concentration was greater in spontaneously hypertensive rats (3.7 ± 0.3 mg/ml) than in age-matched normotensive controls (2.4 ± 0.2 mg/ml) [9]. Fg is synthesized and assembled in hepatocytes and fibroblasts, and when secreted into the circulation, has a plasma half-life that ranges from 3 days to 4 days [12]. Synthesis of Fg involves other inflammatory mediators such as interleukin (IL)-6 and IL-1 [13, 14], which like Fg, are associated with elevation of blood pressure [15–17] and development of hypertension [15, 18, 19], respectively.

Digestion of Fg by plasmin in vivo is quite rare [19]. However, Fg degradation products generated during fibrinolysis are involved in tissue inflammation associated with various diseases [20, 21]. Increased vascular permeability has been found as a result of thrombin and Fg interaction [22]. Under normal conditions, thrombin is rapidly (t_1/2 = 5 min) inactivated and removed from circulation by serine protease inhibitors such as antithrombin III, or activated protein C [3]. Thrombin itself, has been implicated in increasing EC permeability [23, 24]. However, the role of undegraded Fg in increased vascular permeability typically accompanying these diseases is not well understood.

Of the known Fg receptors [25], only intracellular adhesion molecule-1 (ICAM-1) [26] and α₅β₁ and α_vβ₃ integrins are found on the surface of microvascular ECs [27]. We recently showed that Fg binding to ICAM-1 causes constriction of arterioles in vitro and in vivo [28]. In isolated perfused microvessels, a higher than normal concentration of intact Fg (~2 mg/ml), in addition to binding to ECs, binds to vascular smooth muscle cells [28]. These results indicate that the Fg penetrates the Endothelial cell layer and suggest that an increased Fg content can increase EC permeability. In the present study, we tested the hypothesis that at pathologically high levels (4 mg/ml), Fg increases EC leakage to albumin through extracellular signal-regulated kinase (ERK) signaling and filamentous actin (F-actin) formation.

Materials and methods

Materials and reagents

Human Fg (FIB-3, depleted of plasminogen, von-Willebrand factor, and fibronectin) was purchased from Enzyme Research Laboratories (South Bend, IN). Function-blocking monoclonal antibodies against rat α₅ integrin (clone 5H10-27), β₁ integrin (clone Ha2/5), and β₃ (clone 2C9.G2) [29] were purchased from BD biosciences (San Diego, CA). A function-blocking monoclonal antibody to rat ICAM-1 was purchased from Chemicon (Temecula, CA). Specific inhibitors of the Mitogen-activated protein kinase known as ERK, kinase-specific inhibitors PD98059 (2′-amino-3′-methoxyflavone) and U0126 (1,4-diamino-2,3-dicyano-1,4-bis[2-aminophenylthio] butadiene) were purchased from Calbiochem (La Jolla, CA). Monoclonal antibody to phosphorylated ERK1/2 (p44/42) was purchased from Cell Signaling Technology (Beverly, MA). Human thrombin (3,250 U/mg), which cleaves both fibrinopeptide A and fibrinopeptide B, was purchased from Chrono-Log (Havertown, PA). Bovine serum albumin (BSA), 1-palmitoyl-sn-glycero-3-phosphocholine (LPC), and huridin, a thrombin activity inhibitor were purchased from Sigma (St. Louis, MO). BSA conjugated with Alexa Flour-555 dye (BSA-555), Alexa Flour-594 Phalloidin (300 U), 2′,7′-bis (2-carboxyethyl)-5-(and-6)-carboxyfluorescein, acetoxymethyl ester (BCECF, AM), and a secondary antibody conjugated with Texas Red dye, were purchased from Molecular Probes (Eugene, OR).

Endothelial cell culture

Primary rat cardiac microvascular endothelial cells (MVEC) were obtained from VEC Technologies (New York, NY). Cells were grown in a complete media (MCDB131, VEC Technologies) at 37°C exposed to 5% CO₂ in a humidified environment until they reached the fifth passage, then collected and cryopreserved until use. The cells were used at the fifth passage for the experiments.

Endothelial cell permeability assay

Transwell permeable supports (Corning Inc., Corning, NJ) with polycarbonate membranes (Nuclepore Track-Etch, 6.5 mm in diameter, 0.4 μm pore size and pore density of 10⁸/cm²) were coated with fibronectin (Sigma) for 1 h. The membranes were seeded with MVECs and grown in MCDB131 media until they formed a complete monolayer.

Cell confluence and the presence of an intact monolayer on the membranes were confirmed in each series of experiments. Cell growth and their confluence in a separate well (not membrane) were monitored by light microscopy. Cells in test membranes were labeled with BCECF, AM and observed under a microscope (Carl Zeiss Axiovert-100, objective 10×) with a fluorescence filter (488 nm excitation and 516 nm emission) for absence of visible gaps in the cell monolayer. The permeability assays were done after confirming that the cells in the test wells and Transwell membranes were fully confluent and formed an intact monolayer. At the completion of experiments, the cells in the chambers were fixed with 3.7% formaldehyde for 30 min, washed twice with phosphate-buffered saline (PBS, pH 7.4), treated with anti-endothelial CD-31 antibody conjugated with fluorescein isothiocyanate (FITC, Serotech, Raleigh, NC) for 1 h and observed by a confocal microscope (Olympus FV1000, objective 20×, excitation at 495 nm, emission at 515 nm). Uniform distribution of the fluorescence suggested the presence of an intact cell monolayer.

Unlabeled BSA was added to each well to maintain its concentration similar to a normal plasma concentration of albumin (440 μM) and to maintain the activity coefficient of Fg close to that in blood [30]. To inhibit the effect of thrombin’s activity, hirudin (0.1 U/ml) was added to each well in all experiments. The surface levels of solutions in the luminal (200 μl) and abluminal (600 μl) compartments of the Transwells were the same. The experimental set up is similar to that used by Cooper et al. [31], but the presence of albumin-mediated osmotic pressure in the present study, makes it closer to in vivo conditions.

For the permeability assays, cells were washed with PBS and the following substances dissolved in the cell growth media were added to the Transwells: Fg (0.3, 2, or 4 mg/ml), Fg (4 mg/ml) with antibody against ICAM-1(50 μM), Fg (4 mg/ml) with antibodies against α₅ and β₁, or β₃ integrins (50 μMeach), Fg (4 mg/ml) with mitogen-activated protein kinase, kinase (MEK) inhibitors (PD98059, 10, 50, or 100 μMor U0126, 50 μM), or Fg (4 mg/ml) with mouse IgG (50 μM, as a control for ICAM-1). Concentrations of Fg were chosen based on the findings that during hypertension, plasma Fg content is about 4 mg/ml compared to the normal level of~2.8 mg/ml [8, 9].

To determine the effect of the antibodies and MEK inhibitors on EC permeability to albumin, the following substances were added to other Transwells without Fg: antibodies against ICAM-1, α₅ and β₁ integrins, β₃ integrin, IgG, PD98059, or U0126 (50 μM each). Cells incubated with media alone, were used as a control group. BSA-555 (3 mg/ml) was added to each of the wells described above and experiments were performed five times in duplicate for each treatment.

Endothelial cell permeability to albumin in the presence of fibrin or Fg was studied in four experiments with two concentrations of Fg and fibrin. Fibrin was prepared by incubating thrombin (0.1 U/ml) with either 2 or 4 mg/ml Fg for 1 h at 37°C. The fibrin gel (formed from 2 mg/ml to 4 mg/ml Fg) or Fg (2 or 4 mg/ml) was added to the ECs to fully cover the surface of the cell monolayer in the wells and incubated for 30 min. In separate studies, an effect of Fg was compared to that of thrombin, which is known to induce EC permeability [23, 24, 32]. The dose (0.5 U/ml) and the treatment time (15 min) for thrombin were chosen based on the work of others [23, 33]. To determine if Fg itself can penetrate EC monolayer, in separate experiments, cells were incubated with FITC-labeled Fg (Fg-FITC) at normal (2 mg/ml) or higher (4 mg/ml) concentrations.

To study Fg-induced albumin leakage, cells were incubated with Fg in the presence of BSA-555 in humidified conditions at 37°C for 30 min. To study Fg leakage, cells were incubated with Fg-FITC for 1, 6, 12, or 24 h in the same temperature conditions. After incubation, media samples were collected from lower and upper wells. Fluorescence intensity of the samples was measured by a microplate reader (SpectraMax M2, Molecular Devices Corporation, Sunnyvale, CA) with excitation at 494 nm and emission at 518 nm for Fg-FITC, or with excitation at 555 nm and emission at 565 nm for BSA-555. Results are expressed as fluorescence intensity units (FIU). Fg-induced permeability of the ECs was also assessed by estimating relative changes in total protein content in the upper and lower wells by measuring total protein content in the upper and lower wells using Bradford’s method. The results are expressed as a ratio of total protein content in lower wells to total protein content in the corresponding upper wells.

To estimate albumin leakage, fluorescence intensity changes to concentrations of BSA-555 were measured and a relative curve was plotted. Fluorescence intensity changed linearly with fluorescently labeled protein doses and was used to calculate the amount of albumin, or Fg that leaked through the EC monolayer from the upper to the lower wells of the Transwells.

F-actin formation assay

Formation of F-actin in cultured ECs was studied according to the method described earlier [34, 35]. Briefly, MVECs were grown until confluent in 8-well coverglass plates coated with fibronectin. The cells were washed with PBS and incubated with media containing one of the following: Fg (2 or 4 mg/ml), Fg (4 mg/ml) with antibody against ICAM-1, Fg (4 mg/ml) with antibodies against α₅ and β₁, or β₃ integrins (50 μM each), or Fg (4 mg/ml) with MEK inhibitors (PD98059 or U0126, 50 μM each) at 37°C for 30 min. Cells incubated with media alone were used as a control. The well contents were aspirated, and the cells were incubated with Alexa Flour-594 Phalloidin (10 U) and LPC (100 μg/ml, dissolved in 3.7% formaldehyde) for 30 min at 4°Cin the dark. After incubation, the cells were washed four times with PBS. Digital images of intracellular F-actin were taken with the confocal microscope (Olympus FV1000, objective 60×) using a HeNe-G laser (596 nm) to excite the dye, while emission was observed above 620 nm. The images were compared for the: presence of nonactin staining areas in the cell monolayer (gaps), formation of individual stress fibers, and presence of actin foci [34, 35].

In separate experiments, Fg-induced formation of F-actin in ECs was compared to F-actin formation that was induced by thrombin. The cells were treated with either 4 mg/ml Fg for 30 min, 0.5 U/ml thrombin for 15 min, or medium alone (30 min). After incubation, the cells were washed with PBS and digital images of the intracellular F-actin were obtained by the confocal microscope using a 100× objective. This higher magnification allowed better detection of individual stress fibers.

Fg-induced formation of F-actin fibers (total fluorescence intensity) was assessed for each well by analyzing the total fluorescence intensity in four random fields with image analysis software (Image-Pro Plus, Media Cybernetics). Four experiments were done to study Fg-induced F-actin formation and three experiments were done to study Fg- and thrombin-induced F-actin formation. Experiments were done in duplicate (two wells per experimental group).

Fg-induced ERK phosphorylation assay

Fg-induced phosphorylation of ERK was studied using a method described previously [36]. Confluent MVECs were serum starved for 16 h and incubated with one of the following: Fg (2 or 4 mg/ml), Fg (4 mg/ml) with antibody against ICAM-1 (50 μM), Fg (4 mg/ml) with antibodies against α₅ and β₁, or β₃ integrins (50 μM each), or Fg (4 mg/ml) with MEK inhibitors PD98059 or U0126 (50 μM each) at 37°C for 30 min. Cells incubated with low serum media were used as a control. Other cells were incubated with one of the following: antibodies against ICAM-1, α₅ and β₁, or β₃ integrins (50 μMeach), PD98059, or U0126 (50 μMeach) at 37°C for 30 min. After incubation, the cells were washed twice with PBS and fixed in 3.7% paraformaldehyde before permeabilization with LPC (100 μg/ml). Blocking was done for 30 min, in 1% fetal calf serum (FCS) and PBS at room temperature. After three washes with PBS, cells were immunostained for 2 h at room temperature with polyclonal rabbit phospho-ERK-1/2 (1:300). After three to four washes with PBS, cells were stained for 1 h in the dark at room temperature with a secondary antibody conjugated with Texas Red fluorescent dye. Antibody diluent was 1% FCS in PBS [37].

After treatment, cells were washed four times with PBS. Digital images of the presence of phosphorylated ERK were obtained by the confocal microscope (objective 100×) using a HeNe-G laser (596 nm) to excite the dye, while emission was observed above 620 nm. For each series of experiments, the microscope settings were optimized for the brightest images and were kept unaltered during the analysis. Fg-induced ERK phosphorylation was assessed by measuring the total fluorescence intensity of four random fields in each well with Image-Pro Plus. Total fluorescence intensity in each image was divided by the number of cells in the image. The results are presented as an average for the four random fields and expressed as FIU/cell. Three experiments (in duplicate) were done for this protocol and data were averaged for each experimental group.

Statistics

All data are expressed as mean ± SEM. The experimental groups were compared by one-way ANOVA. If ANOVA indicated a significant difference (P<0.05), Tukey’s multiple comparison test was used to compare group means. Differences were considered statistically significant if P<0.05.

Results

Fg induced a dose-dependent increase in the permeability of the EC layer to albumin as determined by the fluorescence intensity of albumin that leaked through ECs into the lower chambers of the Transwells (Fig. 1A). The permeability to albumin that was induced by 4 mg/ml of Fg was decreased significantly by function-blocking antibodies to ICAM-1, α₅ and β₁ integrins (Fig. 1A), but not by IgG (Table 1). The MEK inhibitors, PD98059 or U0126 also significantly decreased Fg-induced permeability (Fig. 1A). Inhibition of MEK activity by a higher dose of PD98059 did not decrease Fg (4 mg/ml)-induced EC permeability, further and a lower dose of PD98059 was not effective (Table 1). In the absence of Fg, treatment of ECs with antibodies against ICAM-1, α₅ and β₁ integrins, or MEK inhibitors (PD98059 or U0126) had no effect on EC albumin leakage (Table 2).

Fig. 1.

Fg-induced albumin leakage through the endothelial cell monolayer. (A) Fluorescence intensity of bovine serum albumin conjugated with Alexa Flour-555 (BSA-555) in lower chamber of Transwells; (B) Total protein content in lower chamber of Transwells relative to the total protein content in the respective upper chamber. Antibodies to ICAM-1, α₅ and β₁ integrins, and MEK inhibitors, PD98059 or U0126, were used at concentration of 50 μM each. *P<0.05 versus control, ^†P<0.05 versus lower dose of Fg. Number of experiments n = 5 for all groups. Clearly defined α-, β-, and γ-chains of Fg before (left) and after the cell treatment (right) shown by the Coomassie-stained SDS-PAGE analysis (reducing conditions) confirms the purity of the Fg (insert)

Table 1.

Effect of PD98059 or IgG on Fibrinogen (Fg)-induced albumin leakage through the endothelial cell monolayer

Control	Fg (4 mg/ml)
		PD98059 10 μM	PD98059 100 μM	IgG 50μM
52 ± 5	337 ± 28*	321 ± 22*	60 ± 7	340 ± 26*

Control group had no Fg, PD98059, or IgG present

Values are mean ± SEM FIU

^*P<0.05 versus control

Number of experiments n = 4

Table 2.

Effect of antibodies against ICAM-1, α₅ and β₁ integrins, or the MEK inhibitors, PD98059 and U0126 on albumin leakage through the endothelial cell monolayer

Control	Anti-ICAM-1 50 μM	Anti-α5 and β1 50μM	PD98059 50μM	U0126 50μM
52 ± 5	58 ± 3	67 ± 6	55 ± 4	52 ± 4

Control group had no anti-ICAM-1, anti-α₅ and -β₁ antibodies, PD98059, or U0126 present

Values are mean ± SEM FIU

Number of experiments n = 4

Fg-induced permeability of the ECs to albumin, assessed by measuring changes in the total protein content in the upper and lower chambers of the Transwells (Fig. 1B), was similar to the changes in fluorescence of the samples (Fig. 1A).

In separate experiments, Fg dose-dependently increased albumin leakage (Table 3). In the same assays, albumin leakage was not altered by fibrin (Table 3). Additional experiments were done to compare effects of the high concentration of Fg (4 mg/ml) to thrombin (0.5 U/ml), which is known to increase EC layer permeability [23, 24]. Fg and thrombin both increased albumin leakage to the same extent (535 ± 29 and 580 ± 68%, respectively; n = 4) compared to the control.

Table 3.

Fibrin- or fibrinogen (Fg)-induced albumin leakage through the endothelial cell monolayer

Concentration (mg/ml)	Fibrin	Fg
2	91 ± 18%	218 ± 42%*
4	103 ± 16%	586 ± 75%*

Data are presented as a percent to a control group treated with medium alone

^*P<0.05 versus fibrin-treated cell

Number of experiments n = 4

Leakage of fluorescently labeled Fg through the EC monolayer increased with time and the dose of Fg (Fig. 2). The fluorescence intensity of 3 mg/ml of BSA-555 that was loaded into the upper wells was 38,716 ± 1,989 FIU. After 30 min, in response to Fg (4 mg/ml) treatment, the fluorescence intensity of albumin that leaked into the lower wells was 337 ± 28 FIU (Fig. 1A). From this, we estimated that <1% of the initial albumin leaked through the ECs. Knowing the fluorescence intensity (8,962 ± 52 FIU) of 0.14 mg/ml Fg that was loaded into the upper wells, and the fluorescence intensity of Fg that leaked into the lower wells during first hour, we estimated that only about 0.6% of the protein went through the EC layer.

Fig. 2.

Fg leakage through the endothelial cell monolayer. Content of FITC-conjugated Fg in lower chamber of Transwells detected by fluorescence intensity measurement. *P<0.05 versus 2 mg/ml Fg, ^†P<0.05 versus previous time. Number of experiments n = 4

Fg caused a dose-dependent increase in formation of F-actin (Fig. 3) and enhanced gap formation in the EC monolayer (Fig. 3A). Formation of F-actin and gaps (cell free spaces) in the monolayer induced by the highest dose of Fg (4 mg/ml) were decreased significantly by function-blocking antibodies to ICAM-1, α₅ and β₁ integrins, or by the MEK inhibitors (PD98059 and U0126). The presence of the anti-β₃ integrin antibody did not change Fg (4 mg/ml)-induced F-actin formation compared to control (112 ± 11% of control). Formation of F-actin in cells that were treated with function-blocking antibodies against ICAM-1, α₅ and β₁, or β₃ integrins, or with the MEK inhibitors was not different from that in the control group (Table 4). Treatment of ECs with Fg (4 mg/ml) caused a slightly greater formation of F-actin compared to that induced by thrombin (0.5 U/ml for 15 min) (Fig. 4).

Fig. 3.

Fg-induced F-actin formation in cultured endothelial cells. (A) Examples of images of Fg-induced F-actin formation, (B) Total fluorescence intensity changes of F-actin staining in various groups of treatment. Antibodies to ICAM-1, α₅ and β₁ integrins, and MEK inhibitors, PD98059 or U0126, were used at concentration of 50 μM each. *P<0.05 versus control. ^†P<0.05 versus 2 mg/ml of Fg. ^#P<0.05 versus 4 mg/ml Fg alone. The arrows indicate the spaces in the endothelial cell monolayer. Each data point represents the results of duplicate determinations from four separate experiments (n = 4)

Table 4.

Effect of antibodies against ICAM-1, α₅ and β₁ integrins, or the MEK inhibitors, PD98059 and U0126 on F-actin formation in the endothelial cells

Control	Anti-ICAM-1 50 μM	Anti-α5 and β1 50μM	Anti-β3 50μM	PD98059 50μM	U0126 50μM
37 ± 2	38 ± 4	35 ± 5	40 ± 1	42 ± 2	37 ± 2

Control group had no anti-ICAM-1, anti-α₅ and -β₁, anti-β₃ antibodies, PD98059, or U0126 present

Values are mean ± SEM FIU

Number of experiments n = 4

Fig. 4.

Comparison of Fg- and thrombin-induced F-actin formation in cultured endothelial cells. (A) Examples of images; (B) Fluorescence intensity of labeled F-actin associated with Fg- and thrombin-induced F-actin formation in ECs. *P<0.05 versus control. ^†P<0.05 versus 0.5 U/ml thrombin. Each data point represents the results of duplicate determinations from three separate experiments (n = 3)

Fg induced a distinct dose-dependent increase of ERK phosphorylation that was significantly decreased by the presence of MEK inhibitors, PD98059 or U0126 (Fig. 5). Treatment of ECs with PD98059 (0.44 ± 0.03 FIU/cell, n = 4) or U0126 (0.40 ± 0.02 FIU/cell, n = 4), did not alter ERK phosphorylation in ECs compared to that in the control group (0.50 ± 0.05 FIU/cell, n = 4).

Fig. 5.

Fg-induced ERK phosphorylation in rat cardiac microvascular ECs. (A) Examples of images of Fg-induced ERK phosphorylation (images 2 and 3), and its inhibition in the presence of MEK inhibitors PD98059, and U0126 (images 4 and 5 respectively). Image 1 represents an ERK phosphorylation in cells treated with medium alone (control). (B) Fluorescence intensity per cell associated with Fg-induced ERK phosphorylation. *P<0.05 versus control, ^†P<0.05 versus 2 mg/ml Fg, ^#P<0.05 versus 4 mg/ml Fg alone. n = 4

Discussion

The present study shows that a higher than normal concentration Fg can enhance albumin leakage through an EC monolayer (Fig. 1). To our knowledge, this effect of Fg has not been reported previously. The doses that we used are based on the results of a previous study with hypertensive rats, which had a plasma Fg content of about 4 mg/ml compared to ~2 mg/ml in normotensive controls [8, 9]. Although it was suggested that Fg and fibrin digestion products increase vascular permeability [22], it was not clear if intact Fg or fibrin can increase microvascular leakage. As in a previous study [27], we used undegraded Fg and we did not find Fg to be degraded after incubation with ECs for 30 min (Fig. 1, insert). Fg increased EC permeability to albumin in the absence of plasminogen or thrombin activity. In contrast, fibrin did not alter albumin leakage (Table 3).

An increase in Fg concentration, by itself, caused Fg leakage through the EC layer, which increased with time (Fig. 2). However, Fg leakage was slower than Fg-induced albumin leakage. An increased content of Fg may enhance albumin leakage by increasing the formation of F-actin (Fig. 3), which may cause stiffening of the cells, their retraction, and widening of inter-endothelial junctions (IEJ) [3, 23, 33, 34, 35, 38].

Comparison of effects of Fg (4 mg/ml) on EC layer permeability to albumin with those of thrombin (0.5 U/ml), a well-known inflammatory agent, show that permeability to albumin was increased to the same extent. Furthermore, Fg induced more F-actin formation than thrombin (Fig. 4), which is known to increase F-actin formation [23]. These results suggest that an increased concentration of Fg in the blood stream could have a role in increasing microvascular permeability as thrombin does.

Our data shows, that the characteristics of Fg leakage were different from those of albumin. Albumin constantly crosses the EC layer, and alterations in EC properties and the pressure gradient across the vessel wall, change the normal rate of albumin transport [3]. Fg, on the other hand, may penetrate the EC layer when its concentration is increased, but it takes a much longer time to cross the EC layer (Fig. 2) than albumin. This may be explained by the greater size of Fg compared to the albumin. However, penetration of the EC layer by Fg may be a mechanism of fibrin deposition in the extracellular matrix during various pathologies that are accompanied by an increased content of Fg (e.g., hypertension).

Interaction of Fg with microvascular ECs occurs through binding to its receptors (ICAM-1 and integrins) on the surface of ECs. Among integrins that act as Fg receptors [25], α₅β₁ and α_vβ₃ are found on ECs [27]. In addition to acting at the apical and basal surfaces, only α₅β₁ integrin was found at the endothelial cell-to-cell contact border [39]. α_vβ₃ integrin was not found in intercellular contact regions [39]. Therefore, in the present study, in addition to ICAM-1, a possible role of endothelial α₅β₁ integrin [40] was studied. Significant decreases of Fg-induced albumin leakage, and F-actin formation in the presence of function-blocking antibodies against ICAM-1 and α₅β₁ integrin, suggest that endothelial permeability could be mediated by Fg binding to its endothelial surface receptors. Fg binding to α₅β₁ integrin may occur through its two RDG sequences [40]. An effect of Fg-to-α₅β₁ integrin binding on EC permeability in the present study confirms results by Qiao et al. [33] that RGD peptide increases endothelial hydraulic conductivity. The present observation that blocking β₃ integrin function did not affect Fg-induced F-actin formation in ECs, suggests that α_Vβ₃ integrin has little role in this process. This β₃ integrin appears to be mainly involved in focal adhesion and was not found at the intercellular contract regions [39] were Fg may have its primary effect.

Treatment of EC with anti-α₅β₁ integrin antibody has been reported to increase endothelial monolayer permeability [39, 41]. However, the antibodies used in these studies were either human placenta α₅β₁ goat antiserum [39], or polyclonal antibody with no indication of an ability to specifically block α₅β₁ integrin function [41]. Therefore, it is possible that these antibodies, upon binding, activate α₅β₁ integrin, or other integrins that may not even be Fg receptors. The antibodies against α₅ and β₁ integrins used in the present study, were function-blocking antibodies [29]. Therefore, by blocking the integrin function, no increase in EC permeability to anti-α₅β₁ integrin was observed.

Thrombin, but not RGD peptide-induced intercellular gaps, were accompanied by formation of F-actin [33]. This observation coincides with our data showing the slightly lesser (but not significant) effects of α₅β₁ integrin on Fg-induced F-actin formation compared to that of ICAM-1 (Fig. 3). The greater role of ICAM-1 in Fg-induced albumin leakage and F-actin formation, may be explained by the greater Fg binding affinity to ICAM-1, than to α₅β₁ integrin [26]. In addition, an increased content of Fg upregulates expression of ICAM-1 [42]. However, to affect albumin leakage from the vasculature, Fg binding to both of its endothelial receptors (ICAM-1 and α₅β₁ integrin) may be necessary.

Mitogen-activated protein kinase, kinase (MEK) is a protein kinase that is phosphorylated and activated by Raf. Once activated, MEK phosphorylates and activates ERK [43]. PD9805 and U0126 used in the present study, are chemically unrelated MEK/ERK inhibitors with different mechanisms of action [44, 45]. While PD98059 binds to the inactive forms of MEK and prevents its activation by upstream activators such as Raf [44], U0126 binds to the activated form of MEK [45]. Our data show a significant decrease of Fg-induced albumin leakage and ERK phosphorylation in the presence of these MEK/ERK inhibitors (Figs. 1, 5). These results, in combination with the data showing decreased albumin leakage in the presence of antibodies against ICAM-1 and α₅β₁ integrin, suggest that binding of Fg to its endothelial receptors involves the MEK/ERK signaling pathway. Activation of ICAM-1 or α₅β₁ integrin, lead to ERK signaling [46, 47 respectively] and formation of F-actin through ERK signaling [48]. Furthermore, binding of Fg to endothelial ICAM-1 leads to activation of ERK1/2 signaling [49]. In the present experiments, the presence of the MEK/ERK inhibitors significantly decreased formation of F-actin in ECs in the presence of Fg (Fig. 3). Therefore, increased blood Fg concentration may increase the binding of Fg to endothelial ICAM-1 and α₅β₁ integrin, which in turn may activate ERK signaling and cause formation of F-actin in the ECs. F-actin formation may increase the rigidity of the cells, widen the IEJ, and may even form gaps between the cells [23, 33–35, 38, 50]. The enlarged IEJ gaps could be enough to allow albumin leakage, which has a diameter of approximately 3.6 nm [51] to pass though the EC layer.

In parallel, Fg-induced albumin leakage may also occur through transendothelial extravasation. Binding of Fg to ECs and activation of ERK signaling, triggers albumin extravasation by caveolae via an absorptive (receptor-mediated) or fluid-phase pathway [52]. Albumin can be easily taken up by caveolae with a radius of about 25 nm at the neck region of vesicles [53]. Several studies showed that a 60-kDa glycoprotein (gp60) is a specific albumin-binding protein, which is involved in albumin transcytosis [52, 54–56]. Thus, a gp60 serves as a specific receptor for albumin in its internalization into caveolae, migration through cells, and release (caveolae exocytosis) at their opposite side. The length of a typical trinodular structure of Fg is about 46 nm [57], while its Stokes radius is about 8.4 nm [58], which is significantly greater than that of albumin. However, despite their difference in size, albumin and Fg were both found to permeate the capillary glycocalyx at the same rate [59]. The present study shows that albumin moves across the EC layer at a higher rate than Fg, indicating that the EC layer is more of a barrier to Fg than to albumin. Further experiments are needed to determine the prevailing role of Fg-induced albumin leakage through the EC monolayer via paracellular versus transcellular pathways.

Increased blood Fg concentration is typical of many cardiovascular and cerebrovascular diseases [4, 6, 7, 11]. Almost all of these diseases are accompanied by inflammation and increased microvascular permeability. The results of the present study show that elevated levels of Fg could have a role in causing increased EC permeability during these diseases.