Ca2+ signalling and PKCα activate increased endothelial permeability by disassembly of VE—cadherin junctions

Authors


Corresponding author C. Tiruppathi: Department of Pharmacology (M/C 868), College of Medicine, The University of Illinois, 835 S. Wolcott Avenue, Chicago, IL 60612, USA. Email: tiruc@uic.edu

Abstract

  • The role of intracellular Ca2+ mobilization in the mechanism of increased endothelial permeability was studied. Human umbilical vein endothelial cells (HUVECs) were exposed to thapsigargin or thrombin at concentrations that resulted in similar increases in intracellular Ca2+ concentration ([Ca2+]i). The rise in [Ca2+]i in both cases was due to release of Ca2+ from intracellular stores and influx of extracellular Ca2+.

  • Both agents decreased endothelial cell monolayer electrical resistance (a measure of endothelial cell shape change) and increased transendothelial 125I-albumin permeability. Thapsigargin induced activation of PKCα and discontinuities in VE-cadherin junctions without formation of actin stress fibres. Thrombin also induced PKCα activation and similar alterations in VE-cadherin junctions, but in association with actin stress fibre formation.

  • Thapsigargin failed to promote phosphorylation of the 20 kDa myosin light chain (MLC20), whereas thrombin induced MLC20 phosphorylation consistent with formation of actin stress fibres.

  • Calphostin C pretreatment prevented the disruption of VE-cadherin junctions and the decrease in transendothelial electrical resistance caused by both agents. Thus, the increased [Ca2+]i elicited by thapsigargin and thrombin may activate a calphostin C-sensitive PKC pathway that signals VE-cadherin junctional disassembly and increased endothelial permeability.

  • Results suggest a critical role for Ca2+ signalling and activation of PKCα in mediating the disruption of VE-cadherin junctions, and thereby in the mechanism of increased endothelial permeability.

We have shown that the thrombin-induced increase in endothelial permeability is dependent on the generation of inositol 1,4,5-trisphosphate (InsP3) (Lum et al. 1992), rise in intracellular Ca2+ ([Ca2+]i) (Lum et al. 1989), and protein kinase C (PKC) activation (Lynch et al. 1990). The mechanism by which the rise in [Ca2+]i increases endothelial permeability involves activation of Ca2+/ calmodulin-dependent myosin light chain kinase (MLCK) (Garcia et al. 1995; Goeckeler et al. 1995; Moy et al. 1996), which promotes actin-myosin interaction by phosphorylation of 20 kDa myosin light chain (MLC20) (Garcia et al. 1995; Goeckeler et al. 1995; Moy et al. 1997). Activation of the monomeric GTPase, Rho, also contributes to MLC20 phosphorylation, and is thus involved in the mechanism of endothelial cell retraction and increased permeability (Van Nieuw et al. 1998; Vouret-Craviari et al. 1998). In addition to endothelial cell retraction, increased endothelial permeability via the paracellular pathway can result from disruption of the VE-cadherin junctional complex in endothelial cells (Rabiet et al. 1996; Corada et al. 1999). The finding that calphostin C, a protein kinase C inhibitor, prevented thrombin-induced disorganization of the VE-cadherin complex (Rabiet et al. 1996), supports a role of PKC in mediating the permeability increase by a cadherin-dependent mechanism. Despite the potential importance of this mechanism, the signalling of VE-cadherin disassembly and its role in regulating endothelial permeability remains unclear.

To address the role of Ca2+ signalling and activation of PKC in mediating the increases in endothelial permeability, we determined the responses to two agents: thapsigargin, which increases [Ca2+]i by inhibiting sarcoplasmic reticulum Ca2+-ATPase, and thrombin which also increases [Ca2+]i but by activation of the cell surface proteinase-activated receptor-1 (PAR-1) (Malik et al. 1992; Garcia et al. 1993; Lum et al. 1993; Lum & Malik, 1994; Nguyen et al. 1997; Ellis et al. 1999). Both agents not only increased [Ca2+]i but also activated Ca2+-sensitive PKC isoforms (Lum et al. 1989, 1992; Lynch et al. 1990; Tiruppathi et al. 1992a; Stevens et al. 1997; Holda et al. 1998), thus enabling us to address the role of Ca2+ signalling and activation of PKC in mediating the increase in endothelial permeability. The results show that thapsigargin and thrombin caused the translocation and activation of the Ca2+-dependent PKC isoform PKCα and increased transendothelial 125I-albumin permeability in association with disassembly of the VE-cadherin junctional complex. Inhibition of PKC activation prevented VE-cadherin disassembly suggesting an important role for PKC in the mechanism of increased endothelial permeability. The results suggest that Ca2+ signalling and PKCα activation regulate the integrity of VE-cadherin junctions, and can mediate increased endothelial permeability.

METHODS

Materials

Human α-thrombin was purchased from Enzyme Research Laboratories, Inc. (South Bend, IN, USA). Endothelial growth medium-2 (EBM-2) was obtained from Clonetics (San Diego, CA, USA). Dulbecco's modified Eagle's medium (DMEM), Hanks' balanced salt solution (HBSS), l-glutamine, phosphate-buffered saline (PBS) and trypsin were obtained from Life Technologies, Inc. (Grand Island, NY, USA). Fetal bovine serum (FBS) was obtained from Hyclone Laboratories, Inc. (Logan, UT, USA). Thapsigargin, calyculin A and okadaic acid were purchased from Calbiochem-Novabiochem Corp. (San Diego, CA, USA). Fura-2 AM, cell permeant calcium chelating agent bis (2-aminophenoxy)ethane N,N,N',N'-tetra-acetic acid acetoxymethyl ester (BAPTA AM), Alexa 568 phalloidin, Alexa 488 and 568 goat anti-mouse and anti-rabbit IgG, ProLong Antifade kit, and 4′6-diamidino-2-phenylindole dihydrochloride (DAPI) were purchased from Molecular Probes, Inc. (Eugene, OR, USA). Anti-phosphothreonine polyclonal antibody (Ab) was from Cell Signaling Technology (Beverly, MA, USA). Anti-PKCα polyclonal Ab and protein A/G-agarose beads were obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). Monoclonal antibodies (mAbs) against VE-cadherin, β-catenin and γ-catenin were obtained from Transduction Laboratories (Lexington, KY, USA). Transwell filters were obtained from Corning Costar Corp. (Cambridge, MA, USA). Radioactive chemicals were obtained from Amersham Pharmacia Biotech Inc. (Piscataway, NJ, USA). Anti-MLC polyclonal Ab was provided by Dr D. Mehta (University of Illinois, Chicago, IL, USA). 12-O-tetradecanoylphorbol-13-acetate (TPA) and other reagents were obtained from Sigma Chemical Company (St Louis, MO, USA).

Endothelial cell culture

Primary human umbilical vein endothelial cells (HUVECs) were obtained from Vec Technologies, Inc. (Rensselaer, NY, USA). Cells were grown in EBM-2 medium supplemented with 10 % FBS. HUVECs passaged 3-7 times were used for all experiments.

Cytosolic Ca2+ ([Ca2+]i)

Cytosolic Ca2+ concentration in single endothelial cells was measured using fura-2 AM (Tiruppathi et al. 2000). Cells grown on 25 mm diameter glass coverslips were washed × 2 with Hanks' balanced salt solution (HBSS). Cells were loaded with 3 μm fura-2 AM for 1 h at 37 °C. Cells were then washed × 2 with HBSS and treated with either thrombin or thapsigargin. Cells were imaged using an Attofluor RatioVision digital fluorescence microscopy system (Atto Instruments, Rockville, MD, USA) equipped with a Zeiss Axiovert S100 inverted microscope and F-Fluar × 40, 1.3 NA oil immersion objective. Regions of interest in individual cells were marked and excited at 334 and 380 nm with emissions collected at 520 nm at 5 s intervals. At the end of each experiment, 10 μm ionomycin was used to obtain fluorescence of Ca2+-saturated fura-2 (high [Ca2+]i) and 10 mm EGTA to obtain fluorescence of free fura-2 (low [Ca2+]i). [Ca2+]i was calculated based on a dissociation constant (Kd) of 225 nm with a two point fit curve.

Transendothelial electrical resistance

The endothelial cell retractile response was measured as previously described by the authors (Tiruppathi et al. 1992b). HUVECs were seeded onto gelatin-coated gold electrodes (4.9 × 10−4 cm2) and grown to confluence. The small electrode and the larger counter electrode were connected to a phase-sensitive lock-in amplifier. A constant current of 1 μA was supplied by a 1V, 4000 Hz, AC signal connected serially to a 1 MΩ resistor between the small electrode and larger counter electrode. The voltage between small and large counter electrodes was monitored by a lock-in amplifier, stored and processed by a personal computer. The same computer controlled the output of the amplifier and switched the measurement to different electrodes during the course of an experiment. Before the experiment, confluent endothelial monolayers were incubated in DMEM containing 10 mm Hepes, pH 7.4 for 2 h, and then thapsigargin- and thrombin-induced changes in resistance of endothelial monolayers were measured. Data are presented as absolute resistance (Ω cm2) and also changes in resistance normalized to their values at time zero as described (Tiruppathi et al. 1992b; Ellis et al. 1999).

Transendothelial 125I-albumin permeability

Transendothelial permeability of 125I-albumin in HUVEC monolayers was determined using Costar transwell units (Lum et al. 1992, 1993). This system measures transendothelial flux (luminal to abluminal) of tracer macromolecules in the absence of hydrostatic and oncotic pressure gradients. The system consists of luminal and abluminal compartments separated by a polycarbonate filter (0.4 μm pore sizes; 6.5 mm diameter). The luminal (upper) side of filters was coated with gelatin. HUVECs were seeded (105 cells per filter) and grown for 2-3 days to attain confluence. Endothelial monolayers were washed and incubated for 2 h with DMEM containing 0.1 % FBS and 10 mm Hepes, pH 7.4 (low-serum media) prior to experiments. Both luminal and abluminal compartments contained 5 mg ml−1 bovine serum albumin (BSA) in low-serum media at volumes of 0.1 and 0.6 ml, respectively. Tracer 125I-albumin (5 × 106 c.p.m. ml−1) was added to the upper compartment and 0.05 ml samples were collected from the lower compartment at 10 min intervals for 60 min for determination of transendothelial clearance rate of 125I-albumin (Lum et al. 1992, 1993).

MLC20 phosphorylation

Phosphorylation of 20 kDa myosin light chain (MLC20) was determined by urea PAGE (Garcia et al. 1995). Endothelial cells grown to confluence on 100 mm dishes were washed × 2 with DMEM containing 0.1 % FBS and 10 mm Hepes, pH 7.4, followed by incubation for 2 h at 37 oC. Cells were then stimulated with either thapsigargin or thrombin. Following treatment for 2 min, reaction was stopped by addition of ice-cold 10 % trichloroacetic acid and 10 mm dithiothreitol. Endothelial cells were scraped off and centrifuged at 3000 r.p.m. in Eppendorf tubes for 5 min. Cell pellets were washed × 3 with diethyl ether and suspended in 40 μl of 6.7 m urea sample buffer as described in Garcia et al. (1995). Extracted proteins were separated by glycerol-urea PAGE and transferred to nitrocellulose membrane. Phosphorylated and non-phosphorylated MLCs were detected by incubating the membrane with anti-MLC20 Ab followed by incubation with HRP-conjugated goat anti-rabbit IgG for visualization by enhanced chemiluminescence (Pierce Chemical Company, Rockford, IL, USA). Phosphorylated and non-phosphorylated bands of MLC on nitrocellulose membranes were localized (Garcia et al. 1995).

Distribution of actin stress fibres

HUVECs were grown to confluence on gelatin-coated glass coverslips. Cells were washed × 2 with HBSS and treated with either thapsigargin or thrombin. The cells were then washed × 2 with HBSS and fixed with 4 % paraformaldehyde (PFA) in HBSS for 20 min at 25 °C. Cells were washed × 2 with HBSS and permeabilized with 0.1 % Triton X-100 for 30 min at 25 °C. The cells were subsequently stained with Alexa 568 phalloidin and DAPI for 30 min at 25 °C, washed × 2 with HBSS, mounted with ProLong Antifade mounting medium, and viewed with the Zeiss LSM 510 confocal microscope.

PKCα immunostaining

Endothelial cells grown to confluence on glass coverslips were briefly washed, and then treated with either thapsigargin or thrombin. Cells were washed × 2 with HBSS and fixed with 4 % PFA in HBSS for 20 min at 25 °C. Cells were then washed × 3 with 0.1 m glycine in HBSS and incubated with blocking buffer (HBSS containing 5 % goat serum, 0.2 % BSA, 0.01 % sodium azide and 0.1 % Triton X-100) at 25 °C for 30 min. The blocking buffer was removed and cells were incubated at 25 °C for 2 h with 2 μg ml−1 rabbit anti-PKCα Ab in blocking buffer. After incubation, cells were washed × 3 with HBSS and incubated with blocking buffer at 25 °C for 30 min. Cells were then incubated at 25 °C for 1 h with Alexa 568 goat anti-rabbit IgG in blocking buffer. Followed by wash (× 3) with HBSS, coverslips were mounted onto a slide containing ProLong Antifade mounting medium and the images were acquired with the Zeiss LSM 510 confocal microscope.

VE-cadherin immunostaining

VE-cadherin immunostaining was also performed as described above. Cells exposed to either thapsigargin or thrombin were fixed, and incubated with monoclonal anti-VE-cadherin mAb, and then with Alexa 488 goat anti-mouse IgG. Cells were washed and mounted, and images were acquired with the Zeiss LSM 510 confocal microscope.

Phosphothreonine immunostaining

Anti-phosphothreonine Ab staining was carried out as described for PKCα immunostaining above except that phosphatase inhibitors (5 nm calyculin A and 1 μm okadaic acid) were included in buffers during fixing, permeabilisation, and anti-phosphothreonine Ab incubation. Alexa 568 goat anti-rabbit IgG was used as the secondary Ab.

PKCα activity measurement

HUVECs grown to confluence on 100 mm dishes were washed × 2 with DMEM containing 0.1 % FBS and 10 mm Hepes, pH 7.4, and incubated for 2 h at 37 °C. Cells were then treated with either thapsigargin, thrombin or TPA, washed × 2 with ice-cold PBS containing 0.2 mm sodium orthovanadate, and lysed with 1 ml of PKC assay buffer (0.25 m sucrose, 20 mm Tris-HCl at pH 7.5, 150 mm NaCl, 1.2 mm EGTA, 20 mmβ-mercaptoethanol, 1 mm phenylmethylsulphonyl fluoride (PMSF), 20 μg ml−1 leupeptin, 20 μg ml−1 aprotinin, 1 mm sodium orthovanadate, 1 mm sodium pyrophosphate, 1 mm sodium fluoride, 0.1 % Triton X-100, 0.5 % NP-40). Lysates were collected and centrifuged for 5 min at 14 000 r.p.m. as above at 4 °C. Supernatants were pre-cleared by incubating with protein A/G-agarose beads for 30 min at 4 °C with constant shaking and then centrifuged for 5 min at 3000 r.p.m. as above at 4 °C. The resulting supernatants were incubated with 2 μg ml−1 rabbit anti-PKCα Ab overnight at 4 °C with shaking. After incubation, 20 μl of protein A/G-agarose beads were added to each tube, mixed and incubated for 2 h at 4 °C with constant shaking. This was followed by washing × 3 with PKC assay buffer and by centrifugation for 5 min at 3000 r.p.m. as above at 4 °C. After the final wash, activity of PKCα was determined by PepTag assay for non-radioactive detection of PKC (Promega Corporation, Madison, WI, USA).

Immunoprecipitation and immunoblotting of VE-cadherin and its associated proteins

To determine association of catenins to VE-cadherin, HUVECs were grown to confluence on 100 mm dishes, stimulated with either thapsigargin or thrombin, and lysed with 1 ml of lysis buffer (10 mm Tris-HCl pH 7.5, 150 mm NaCl, 1 mm EDTA, 1 mm EGTA, 0.2 mm PMSF, 20 μg ml−1 aprotinin, 0.2 mm sodium orthovanadate, 0.2 mm sodium pyrophosphate, 1 % Triton X-100 and 0.5 % NP-40). Cell lysates were precleared with protein A/G-agarose beads and incubated with 2 μg ml−1 anti-VE-cadherin mAb overnight at 4 °C. Immunoprecipitated samples were collected by incubation with 20 μl of protein A/G-agarose beads for 2 h at 4 °C with constant shaking. After three separate washes with RIPA buffer, immunoprecipitated samples were resuspended in 40 μl of 2 % SDS sample buffer, boiled for 5 min, and separated by SDS-PAGE on a 7.5 % polyacrylamide gel. After electrophoresis, proteins were transferred to nitrocellulose membranes. The membranes were blocked with wash buffer (10 mm Tris-HCl pH 7.5, 100 mm NaCl and 0.1 % Tween-20) containing 5 % non-fat dried milk for 1 h and incubated overnight with primary Abs diluted in wash buffer containing 5 % non-fat dried milk. After washing × 3 with wash buffer, membrane was incubated for 2 h at room temperature with HRP-conjugated goat anti-mouse IgG. Protein bands were detected by enhanced chemiluminescence.

Statistical analysis

Statistical comparisons were made using the two-tailed Student's t test. Values are reported as means ±s.e.m. Differences in mean values between and among groups were measured by one-way analysis of variance (ANOVA) with a Bonferroni correction. Values were considered significant at P < 0.05.

RESULTS

Effects of thapsigargin and thrombin on [Ca2+]i

Fura-2-loaded HUVECs were challenged with either 10 nm (1 NIH U ml−1) thrombin (Thr) or 1 μm thapsigargin (Tp) (Fig. 1). In the presence of extracellular Ca2+ (1.26 mm), thrombin produced an increase in [Ca2+]i (peak value 1335 ± 95 nm) followed by a gradual decline during the remainder of the recording duration (Fig. 1a). [Ca2+]i at 5 min after thrombin challenge was 153 ± 20 nm, a value 4-fold greater than the basal concentration (35-40 nm). In the absence of extracellular Ca2+, the peak thrombin-induced increase in [Ca2+]i (value of 586 ± 50 nm) was reduced and the Ca2+ transient rapidly returned to baseline (Fig. 1B). In response to 1 nm thrombin, we observed a peak increase of 450 ± 40 nm (Fig. 1a inset) in the presence of extracellular Ca2+ and a value of 310 ± 28 nm in the absence of extracellular Ca2+.

Figure 1.

Effects of thapsigargin and thrombin on [Ca2+]i in HUVECs

A and B, challenged with 10 nm thrombin (Thr); inset shows the response to 1 nm thrombin. C and D, challenged with 1 μm thapsigargin (Tp). A and C, the extracellular medium Ca2+ concentration was 1.26 mm; B and D, the extracellular medium contained no Ca2+ ([Ca2+]o= 0). Arrows indicates the time at which either thapsigargin or thrombin was added.

Addition of thapsigargin produced a peak increase in [Ca2+]i (of 431 ± 42 nm) followed by a slow decline to the plateau level. As with the thrombin response, the rise in [Ca2+]i was sustained above baseline during the measurement period (Fig. 1C). At 5 min after thapsigargin, [Ca2+]i was 181 ± 17 nm. In the absence of extracellular Ca2+, thapsigargin produced a peak increase (320 ± 31 nm), which rapidly returned to basal levels (Fig. 1D). These results indicate that both thrombin and thapsigargin released Ca2+ from intracellular stores, and activated Ca2+ influx from the extracellular milieu.

Thapsigargin and thrombin decrease transendothelial monolayer electrical resistance and increase 125I-albumin permeability

Human umbilical vein endothelial cells were incubated in serum-free medium for 2 h prior to exposure to either thapsigargin or thrombin. Addition of 10 nm thrombin caused a 40-50 % decrease and addition of 1 μm thapsigargin produced a 30-40 % decrease in endothelial cell monolayer electrical resistance (Fig. 2a). Endothelial resistance returned to normal values within 2 h after thrombin, whereas thapsigargin produced a delayed recovery (within 4-6 h; Fig. 2a). Both 10 nm thrombin and 1 μm thapsigargin caused 2- to 3-fold increases in transendothelial 125I-albumin clearance rate (Fig. 2B). Preincubation of HUVEC monolayer with the Ca2+ chelator BAPTA (30 μm) for 30 min, prevented the thapsigargin-induced increase in transendothelial 125I-albumin clearance and significantly reduced the thrombin-induced increase in transendothelial 125I-albumin clearance by 65 % (Fig. 2B; P < 0.05).

Figure 2.

Thapsigargin and thrombin induce a decrease in HUVEC monolayer electrical resistance and an increase in transendothelial 125I-albumin flux

In A, HUVECs grown to confluence on gold electrodes (as described in Methods) were used to measure changes in transendothelial electrical resistance. Experiments were repeated 4-6 times with similar results. Data from representative experiments are shown. Arrows indicate the time at which either 10 nm thrombin or 1 μm thapsigargin was added. In B, transendothelial clearance of 125I-albumin was determined in confluent HUVEC monolayers. Experimental details are described in Methods. The 125I-albumin clearance rates were determined after treatment with either 10 nm thrombin or 1 μm thapsigargin. BAPTA AM (30 μm) was incubated with HUVEC monolayers for 30 min at 37 oC, washed × 1, and then exposed to either thrombin or thapsigargin. Experiments were repeated four times in triplicate; values are shown as means ±s.e.m. Different from control group, **P < 0.001. Different from thrombin- or thapsigargin- treated group, †P < 0.05.

Differential effects of thapsigargin and thrombin on MLC20 phosphorylation and actin stress fibres

Exposure of HUVECs to thrombin resulted in phosphorylation of MLC20, whereas thapsigargin had no effect on this response (Fig. 3a). In control cells, few actin stress fibres were seen (Fig. 3B), whereas thrombin caused the formation of actin stress fibres (Fig. 3B). In contrast, exposure to thapsigargin for periods up to 30 min did not alter actin stress fibre distribution (Fig. 3B).

Figure 3.

Effects of thapsigargin

A, effects of thapsigargin and thrombin on MLC20 phosphorylation in HUVECs. Confluent HUVEC monolayers were stimulated with either 1 or 10 nm thrombin or 1 μm thapsigargin for 2 min. Total cell lysates were extracted for myosin light chains (MLCs) and then were separated by urea-glycerol PAGE. Other details are described in Methods. The experiment was repeated three times; data in A are from one representative experiment. B, thapsigargin-induced increase in [Ca2+]i fails to induce actin stress fibre formation. HUVECs grown on gelatin-coated glass coverslips were treated with either 1 nm thrombin for 15 min or 1 μm thapsigargin for 30 min. The cells were washed, fixed and stained with Alexa 568 phalloidin to detect actin polymerization and DAPI to visualize the nucleus as described in Methods. Results are representative of three experiments. Scale bars, all 10 μm.

Thapsigargin and thrombin induce translocation and activation of PKCα

Thapsigargin resulted in translocation of PKCα from the cytosol to plasma membrane (Fig. 4a; arrows). Thrombin also increased the plasma membrane translocation of PKCα (Fig. 4a). These changes were coupled to PKCα activation (Fig. 4B). PKCα activity increased 2- to 3-fold after exposure of HUVECs to thrombin and thapsigargin (Fig. 4B); results were comparable with the effects of positive control PKC-activating agent TPA (Fig. 4B).

Figure 4.

Effects of thapsigargin

A, thapsigargin promotes translocation of the Ca2+-dependent PKCα isoform. HUVECs grown on gelatin-coated glass coverslips were stimulated with either medium alone (control), 10 nm thrombin for 10 min or 1 μm thapsigargin for 30 min. After treatment, cells were washed, fixed and stained with an anti-PKCα Ab, and then viewed by confocal microscopy. Other details are described in Methods. The experiment was repeated three times with similar results. Data from a representative experiment are shown. Arrows indicate PKCα translocation. Scale bars, all 10 μm. B, thapsigargin increases PKCα isoform activity in HUVECs. HUVECs were stimulated with either 10 nm thrombin for 10 min, 1 μm thapsigargin for 15 or 30 min, 10 nM TPA for 15 min, or 100 nm TPA for 15 min. Other details are described in Methods. The experiment was repeated four times in duplicate; values are shown as means ±s.e.m.*P < 0.05; **P < 0.001.

Disassembly of VE-cadherin junctions induced by thapsigargin is prevented by PKC inhibition

In control cells, intense anti-VE-cadherin mAb staining was evident at intercellular junctions (Fig. 5a). Addition of thrombin or thapsigargin caused marked VE-cadherin staining (Fig. 5C and E). Pretreatment with the PKC inhibitor, calphostin C (100 nm), in cells challenged with thapsigargin or thrombin, prevented the response (Fig. 5D and F). As PKCα may mediate VE-cadherin disassembly, we studied whether PKCα co-localized with VE-cadherin upon activation of HUVECs with either thrombin or thapsigargin. Results showed that both thrombin and thapsigargin induced PKCα and VE-cadherin co-localization at intercellular junctions (Fig. 6).

Figure 5.

Thapsigargin and thrombin induce disruption of VE-cadherin junctions and phosphorylate junctional proteins

HUVECs grown to confluence on gelatin-coated glass coverslips were incubated with serum-free medium, and then used for the experiments. A and B, challenged with medium alone (control); C and D, treated with 10 nm thrombin for 10 min; E and F, treated with 1 μm thapsigargin for 15 min. B, D and F, preincubated with calphostin C (100 nm) for 1 h at 37 °C prior to adding either thrombin or thapsigargin. HUVECs were exposed to light during calphostin C treatment. After fixing, cells were stained with anti-VE-cadherin mAb and anti-phosphothreonine Ab overnight at 4 oC, and then stained with Alexa 488 (green) and 568 (red) goat-anti-mouse and anti-rabbit IgG. Cells were also stained with DAPI (blue) to visualize the nucleus. Other details are described in Methods. Co-localization of phosphothreonine and VE-cadherin immunostaining was observed in yellow. The experiment was repeated three times; data shown are from a representative experiment. Arrows indicate irregularities in VE-cadherin staining. Scale bars, all 10 μm.

Figure 6.

Thapsigargin and thrombin induce translocation of PKCα to VE-cadherin junction

HUVECs grown to confluence on gelatin-coated glass coverslips were incubated with serum-free medium, and then used for the experiments. Cells were either stimulated with 10 nm thrombin for 10 min or 1 μm thapsigargin for 15 min and fixed. Cells were stained with anti-VE-cadherin mAb and anti-PKCα Ab overnight at 4 oC. The cells were then stained with Alexa 488 (green) and 568 (red) goat-anti-mouse and anti-rabbit IgG. Co-localization of PKCα and VE-cadherin immunostaining was observed in yellow. Other details are described in Methods. The experiment was repeated three times; data shown are from a representative experiment. A, control; B, thrombin; C, thapsigargin. Scale bars, all 20 μm.

We stained HUVECs with anti-phosphothreonine Ab to address the possibility that PKCα activation mediated phosphorylation of junctional proteins. Both thapsigargin and thrombin increased anti-phospho-threonine Ab staining at intercellular junctions (Fig. 5C and E), and moreover, the staining co-localized with anti-VE-cadherin mAb staining (Fig. 5C and E). Calphostin C prevented these effects (Fig. 5D and F).

We also investigated the effects of calphostin C on endothelial monolayer resistance to assess whether the changes in immunostaining described above were coupled to the loss of endothelial barrier function. Calphostin C treatment alone had no significant effect on basal transendothelial monolayer electrical resistance (Fig. 7 and Table 1); however, calphostin C (100 nm) pretreatment inhibited ≈80 % of thapsigargin-induced and ≈60 % of thrombin-induced decreases in electrical resistance (Fig. 7 and Table 1).

Figure 7.

Calphostin C prevents thapsigargin-, thrombin- and TPA-induced decreases in endothelial monolayer electrical resistance

HUVECs were incubated with or without 100 nm calphostin C in the presence of light at 37 °C for 1 h prior to treatment with thrombin (10 nm), thapsigargin (1 μm) or TPA. Maximum decrease in resistance was observed at 10, 20 and 15 min, respectively, for thrombin, thapsigargin and TPA. Results are expressed as percentage decrease in monolayer resistance. Calphostin C treatment alone did not significantly alter HUVEC monolayer resistance. Results are also presented in Table 1 in Ω cm2. Other details are described in Methods. The results are from four experiments carried out in duplicate; values are shown as means ±s.e.m.

Since calphostin C-sensitive PKCα was an important determinant of VE-cadherin disassembly and the loss of endothelial barrier function, we also addressed the direct effects of activation of PKC with TPA. Exposure of HUVECs to either 10 nm or 100 nm TPA decreased monolayer electrical resistance (Fig. 7 and Table 1) in a manner sensitive to calphostin C (Fig. 7 and Table 1). Moreover, exposure to TPA induced VE-cadherin disassembly and PKCα co-localization with VE-cadherin (cf. Fig. 8a and B). Calphostin C prevented the TPA-induced VE-cadherin disassembly and PKCα translocation (Fig. 8C and D). As with the effects of thapsigargin, TPA failed to induce the formation of actin stress fibres even though it resulted in VE-cadherin disassembly (Fig. 8D and F).

Figure 8.

TPA induces PKCα translocation and irregular and reduced VE-cadherin immunostaining at intercellular junctions

HUVECs grown to confluence on gelatin-coated glass coverslips were stimulated with medium alone (A, C and E) or 100 nm PMA for 15 min (B, D and F). Calphostin C (100 nm) was preincubated with cells as described in Fig. 5 prior to adding medium or TPA (C and D). After fixing, cells were stained with anti-VE-cadherin mAb and anti-PKCα Ab (A-D) or Alexa 568 phalloidin (E and F). Secondary antibodies and other details are described in Methods. The experiments were repeated three times; data shown are from a representative experiment. Arrows indicate the discontinuous and markedly irregular pattern of VE-cadherin staining. Scale bars, all 20 μm.

Thapsigargin and thrombin fail to disrupt VE-cadherin interaction with catenins

Since the possibility exists that thapsigargin- and thrombin-induced alterations in VE-cadherin staining may reflect disruption of cadherin-catenin interactions, we determined whether severing of the VE-cadherin- catenin complex could explain our findings. Endothelial cell lysate was immunoprecipitated with anti-VE-cadherin mAb (see Methods), and the immunoprecipitated proteins were blotted with anti-β-catenin mAb, anti-γ-catenin mAb and anti-VE-cadherin mAb. Treatment with 10 nm thrombin for 15 min or 1 μm thapsigargin for 15 or 30 min failed to dissociate VE-cadherin from its interacting catenin proteins (Fig. 9).

Figure 9.

Thapsigargin and thrombin treatment failed to alter VE-cadherin-catenin assembly

HUVECs were stimulated with either 10 nm thrombin for 15 min or 1 μm thapsigargin for 15 and 30 min and the cell lysates were immunoprecipitated with anti-VE-cadherin mAb. The immunoprecipitates were then blotted with mAbs directed against VE-cadherin, β-catenin and γ-catenin. Other details are described in Methods. The experiment was repeated three times; data shown are from a representative experiment.

DISCUSSION

Although the rise in [Ca2+]i and activation of PKC have been implicated in the mechanism of increased endothelial permeability (Lum et al. 1989; Lynch et al. 1990; Van Nieuw et al. 1998), the underlying basis of the response remains unclear. Prevention of the rise in [Ca2+]i by loading cells with the intracellular Ca2+ chelator bis(2- aminophenoxy)ethane-N,N,N',N'-tetraacetic acid acetoxymethyl ester (BAPTA AM) and inhibition of PKC activation both inhibited the thrombin-induced increase in endothelial permeability (Lum et al. 1989, 1990; Van Nieuw et al. 1998). The rise in [Ca2+]i may be important in signalling the response since Ca2+ can bind to calmodulin (CaM), and thereby induce Ca2+/CaM-dependent activation of myosin light chain kinase (MLCK) and phosphorylation of MLC20 (Garcia et al. 1995; Wysolmerski & Lagunoff et al. 1990). In addition, it is possible that Ca2+ signalling can regulate the function of VE-cadherin junctions, and thereby mediate increased endothelial permeability by interfering with VE-cadherin interactions at the level of the adherens junctional complex (Rabiet et al. 1996; Corada et al. 1999). As the role of Ca2+ signalling in mediating increased endothelial permeability may be complex, in the present study we specifically addressed the basis by which Ca2+ signalling induces VE-cadherin disassembly.

As a means of increasing [Ca2+]i, we used thapsigargin, an inhibitor of sarcoplasmic-endoplasmic reticulum Ca2+-ATPase (Holda et al. 1998). Thapsigargin is known to promote intracellular store depletion and trigger the capacitative entry of Ca2+ in cells (Stevens et al. 1997; Holda et al. 1998). We compared the results with thrombin, a mediator that increases endothelial permeability by a Ca2+-dependent mechanism (Lum et al. 1989). We showed that both thapsigargin and thrombin induced the release of Ca2+ from intracellular stores followed by Ca2+ influx. Due to the similarities in Ca2+ transients, we could reliably compare the permeability responses to both agents. We showed that both thapsigargin and thrombin increased transendothelial 125I-albumin permeability by 2- to 3-fold over baseline values and similarly decreased endothelial cell monolayer electrical resistance. Together these results can be interpreted as indicating the opening of paracellular permeability pathways. In addition thrombin induced the phosphorylation of MLC20, whereas thapsigargin had no effect on the actin-myosin contractile machinery. Thrombin also induced actin stress fibre formation consistent with the MLC20 phosphorylation, whereas thapsigargin had no effect on actin fibres. These results support the concept that the rise in intracellular Ca2+ concentration is a key mechanism mediating the increased endothelial permeability and that the response activated by Ca2+ signalling (as with thapsigargin) can occur independently of MLC20 phosphorylation and activation of the actin-myosin contractile motor.

The thapsigargin-induced increase in transendothelial 125I-albumin permeability was prevented by the intracellular Ca2+ chelator BAPTA, indicating that the thapsigargin-mediated response is primarily dependent on Ca2+ signalling. In contrast, BAPTA reduced the thrombin-induced increase in transendothelial 125I-albumin clearance and decrease in monolayer resistance by about 65 %. This finding is consistent with the observation that thrombin can also activate Ca2+-independent signalling pathways such as the small GTPase, Rho, which induce MLC phosphorylation and can increase endothelial permeability (Garcia et al. 1995; Van Nieuw et al. 1998).

We addressed possible alterations in VE-cadherin junctions as a basis of the increased permeability induced by Ca2+ signalling. Previous studies have shown that the loss of VE-cadherin function in adherens junctions is important in the mechanism of increased permeability (Rabiet et al. 1996; Vouret-Craviari et al. 1998; Corada et al. 1999). VE-cadherins are transmembrane adhesion proteins located in intercellular adherens junctions where they are linked in the cytoplasm to β- and γ-catenins, and in turn to α-catenin and the actin cytoskeleton (Lampugnani et al. 1995; Dejana, 1996). VE-cadherin disorganization and the resulting loss of function of adherens junctions activated by mediators such as thrombin has been proposed as the basis of increased endothelial permeability (Rabiet et al. 1996). Intravenous injection of a monoclonal antibody against VE-cadherin in mice increased vascular permeability (Corada et al. 1999; Gao et al. 2000), indicating that the VE-cadherin complex is involved in the in vivo regulation of endothelial barrier function. In the present study, we showed that both thapsigargin and thrombin induced the disassembly of VE-cadherin at cell junctions as evident by decreased endothelial monolayer resistance and the markedly irregular pattern of VE-cadherin immunostaining. However, the normal VE-cadherin-catenin interactions were not severed by either thrombin or thapsigargin in that immunoprecipitation studies showed that the cadherin-catenin complex remained intact. Thus, discontinutities in VE-cadherin immunostaining and increased permeability could not be explained by the dissociation of VE-cadherin from the catenin complex. As thapsigargin caused disruption of VE-cadherin in a manner independent of MLC phosphorylation and actin stress fibre formation, the results indicate that loss of VE-cadherin function as regulated by Ca2+ signalling is important in the mechanism of increased endothelial permeability.

Since Ca2+ signalling can activate PKC isoforms and PKC activation is known to disrupt VE-cadherin assembly in endothelial cells (Rabiet et al. 1996), we next addressed the possible role of PKC in mediating the permeability response. We measured the activity and membrane translocation of the Ca2+-dependent PKC isoform, PKCα, that is abundantly expressed in endothelial cells (Haller et al. 1996; Rahman et al. 2000). Both thapsigargin and thrombin increased PKCα activity and induced its translocation from the cytosol to intercellular junctional sites where it co-localized with VE-cadherin. The PKC inhibitor, calphostin C, prevented both thapsigargin- and thrombin-induced VE-cadherin disruption as well as the decreased endothelial cell monolayer resistance. Thus, PKC activation regulates endothelial barrier function by interfering with VE-cadherin junctional assembly.

The mechanism of the action of PKCα may involve phosphorylation of VE-cadherin serine and/or threonine residues (Jaken, 1996). We observed that both thapsigargin and thrombin induced the staining of anti-phosphothreonine antibody at intercellular junctions and its co-localization with anti-VE-cadherin staining. The results therefore suggest the phosphorylation of threonine residues in VE-cadherin and its associated junctional proteins. Thus, the mechanism of loss of VE-cadherin function may involve activation of the Ca2+-dependent PKCα isoform, and the resultant phosphorylation of components of the VE-cadherin- catenin complex.

If PKC activation is required in the mechanism of increased endothelial permeability, then direct activation of PKC should recapitulate the response observed with thapsigargin and thrombin. To address the effects of PKC activation in mediating of VE-cadherin junctional diassembly, we used the PKC activator, TPA. As with the thapsigargin response, TPA induced translocation and activation of PKCα, and disassembly of VE-cadherins at sites of cell-cell contact without affecting the actin stress fibre distribution. The TPA-activated response was also prevented by calphostin C. The results support the hypothesis that PKC activation can increase endothelial permeability independently of actin stress fibre formation by a mechanism involving VE-cadherin disassembly.

In conclusion, the present results demonstrate that loss of VE-cadherin function in cell junctions is regulated by Ca2+ signalling and PKCα activation and that signalling via these second messengers is involved in mediating increased endothelial permeability. Although permeability increasing mediators such as thrombin activate the MLC- and actin-dependent contractile apparatus in endothelial cells, it appears that the increase in endothelial permeability induced by thrombin is also critically regulated by Ca2+- and PKCα-dependent disruption of VE-cadherin function.

Acknowledgements

This work was supported by National Institutes of Health grants no. GM 58531 and HL 45638.

Ancillary