Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Reversible protein phosphorylation, mediated by protein kinases and protein phosphatases (PPs), is an important mechanism used to regulate cellular function. At least four types of serine/threonine PPs are present in mammalian tissue: PPs type 1 (PP 1), type 2A (PP 2A), type 2B, and type 2C (Cohen, 1989; Hunter, 1995; Shenolikar and Nairn, 1991; Wera and Hemmings, 1995). The availability of cell membrane permeant inhibitors like okadaic acid (Tachibana et al., 1981) and cantharidin (Honkanen, 1993) has facilitated the study of the functional role of PP 1 and PP 2A. Okadaic acid is a polyether fatty acid first isolated from black sponges (Halichondria okadai). Cantharidin is another natural toxicant produced by as many as 1500 different species of blister beetles (Honkanen, 1993). Using these inhibitors, PPs have gained interest in recent years as potentially important regulators of cellular function.

PPs exist in endothelial cells, and their inhibition alters the phosphorylation state of regulatory proteins. For instance, in bovine pulmonary artery endothelial cells, inhibition of PP leads to phosphorylation of the regulatory light chains of myosin (MLC₂₀) (Verin et al., 1995). Recent data indicate, however, that thrombin can enhance MLC₂₀ phosphorylation by inhibition of PP activity (Shasby et al., 1997). These findings suggest an important regulatory role of PP in endothelial cells. Receptor signals may act through stimulation or reduction of PP activity. The potential coupling mechanisms between receptors and intracellular PP are as yet unknown. Thrombin or histamine increase Ca²⁺ in human umbilical vein endothelial cells and increase MLC₂₀ phosphorylation by activating Ca²⁺/calmodulin-dependent myosin light chain kinase (MLCK) (Garcia et al., 1995; Shasby et al., 1997). Others assumed that thrombin and histamine first elevate cytosolic Ca²⁺ and that this Ca²⁺ rise then triggers the activity of MLCK.

We tested the hypotheses that cantharidin can increase the phosphorylation state of proteins in endothelial cells, can inhibit PP activity in endothelial cells, and can alter endothelial barrier function.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Macromolecule Permeability

Endothelial cells from bovine aorta were isolated as previously described for porcine aortic endothelial cells (Spahr and Piper, 1990). Briefly, bovine endothelial cells were prepared by gentle mechanical scraping of the intima of the descendent part of bovine aortae. Harvests of endothelial cells were plated at a density of 10⁶ cells per 100-mm plastic Petri dish. The cells were cultured at 37°C in medium 199 with Earle's salt supplemented with 100 IU/ml penicillin G, 100 µg/ml streptomycin, and 20% (v/v) newborn calf serum. Confluent cultures of primary endothelial cell were trypsinized in PBS composed of 137 mM NaCl, 2.7 mM KCl, 1.5 mM KH₂PO₄, and 8 mM Na₂HPO₄, at pH 7.4, supplemented with 0.05% (w/v) trypsin, and 0.02% (w/v) EDTA. Cells were seeded at a density of 7 × 10⁴ cells/cm² on 24-mm round polycarbonate filters (pore size, 0.4 µm) for determination of albumin permeability. Experiments were performed with confluent monolayers 4 days after seeding.

We studied the permeability of the endothelial cell monolayer in a two-compartment system separated by a filter membrane (Noll et al., 1996). Both compartments contained as basal medium modified Tyrode's solution (150 mM NaCl, 2.7 mM KCl, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 1 mM CaCl₂, and 30 mM HEPES, pH 7.4, 37°C) supplemented with 5% (v/v) newborn calf serum. There was no hydrostatic pressure gradient between both compartments. The "luminal" compartment containing the monolayer had a volume of 2.5 ml, and the "abluminal" had a volume of 13 ml. The fluid in the abluminal compartment was constantly stirred. Trypan blue-labeled albumin (60 µM) was added to the luminal compartment. The appearance of labeled albumin in the abluminal compartment was continuously monitored by pumping the liquid through a spectrophotometer (Specord 10; Zeiss, Jena, Germany). Increases in the concentration of labeled albumin were detected with a time delay of less than 15 s. The concentration of labeled albumin in the luminal compartment was determined every 10 min of incubation. It did not change significantly in the time frame of the experiments. The albumin flux across the monolayer was determined as described previously (Siflinger-Birnboim et al., 1987; Garcia et al., 1996; Hempel et al., 1996).

Experimental Protocol

The basal medium used in incubations was a modified Tyrode's solution (composition as described in Macromolecule Permeability). Macromolecule permeability of the endothelial monolayer, transferred to the incubation chamber, was determined after an initial equilibration period of 20 min. Then, the basal albumin permeability of each monolayer-filter system was determined for another 20 min of incubation. Agents were added as indicated, and the response of albumin permeability was recorded for a further 120 min. The solvent alone (control) did not affect permeability.

Biochemical Studies

Cell Culture. We obtained bovine aortic endothelial cells by collagenase digestion as described (Knapp et al., 1997). Cells grown in a humidified incubator under an atmosphere of 7.5% CO₂/21% O₂ at 37°C reached confluence within 5 to 7 days forming monolayers with a "cobblestone" appearance. In addition, we confirmed the nature as endothelial cells by positive staining with an antibody to factor VIII-related antigen (data not shown, see Knapp et al., 1997). We only used primary endothelial cells or cells of the first passage for the experiments.

Measurement of Cytosolic Calcium ([Ca]_i). Primary bovine aortic endothelial cells were seeded onto 15-mm-diameter coverslips and used 6 to 7 days later at subconfluent density. Cells were washed twice with physiological saline solution containing: 145 mM NaCl, 5.6 mM KCl, 1 mM MgSO₄, 2 mM CaCl₂, 20 mM HEPES, 10 mM glucose, pH 7.4, and then incubated with cell-permeant indo-1-acetoxymethylester (5 µM) and 0.5% (v/v) of the nonionic surfactant pluronic F-127 [20% in dimethyl sulfoxide (DMSO); Molecular Probes, Eugene, OR] for 60 min at 25°C. At the end of the loading period, the coverslips were washed twice with physiological saline solution and maintained an additional 60 min to allow for complete de-esterfication of the indicator at room temperature. Cells were then placed in a perfusion chamber on the stage of a modified inverted microscope (Diaphot 200; Nikon, Tokyo, Japan) and superfused with prewarmed physiological saline solution (0.8 ml/min). All experiments were performed at 25°C. [Ca]_i was recorded from a field of approximately 20 cells using a dual-emission microfluorescence system (PTI, Princeton, NJ) as described before (Knapp et al., 1998). The ratio of the two emission wavelengths (405 nm and 495 nm) was used as an index of intracellular free calcium ion concentration. Data acquisition and processing were supported by software (Felix Version 1.1, PTI) for intracellular calcium measurement.

Phosphorylation Experiments

Labeling of Bovine Aortic Endothelial Cells. Bovine aortic endothelial cells were harvested by trypsin (0.05%)/EDTA (0.02%) solution and washed twice in phosphate-free solution (Na-HEPES buffer) consisting of : 132.0 mM NaCl, 4.8 mM KCl, 1.2 mM MgSO₄, 10 mM glucose, 10 mM HEPES, and 2.5 mM sodium pyruvate, pH 7.4, at 37°C. The cells were sedimented by centrifugation (366g, 4°C) for 10 min. Bovine aortic endothelial cells were resuspended in 10 ml of Na-HEPES buffer and incubated with 5 mCi of ³²P-labeled orthophosphate for 2 h at 37°C.

Protein Phosphorylation. We added adenosine deaminase (20 U/ml) to avoid interference from endogenous adenosine upon treatment. The drug solution (100 µl) was preincubated at 37°C before mixing with the final cell suspension of bovine aortic endothelial cells (100 µl) and kept at 37°C. Reaction was stopped after 30 min by addition of 100 µl of SDS stop solution consisting of 62.5 mM tris(hydroxymethyl) aminomethane, 10% SDS (w/v), 10% glycerol (v/v), 0.6% DL-dithiothreitol (w/v), and a trace of bromphenol blue, pH adjusted to 6.8.

SDS-Polyacrylamide Gel Electrophoresis and Autoradiography. Samples were heat treated (95°C), and aliquots of 100 µl corresponding to 20 to 30 µg of protein were applied to each lane. Gels were run and dried, and incorporated radioactivity was quantitated by PhosphorImager (Molecular Dynamics, Krefeld, Germany) using ImageQuant software as described (Knapp et al., 1998).

Immunological Identification of the Regulatory Light Chains of Myosin

³²P-Labeled endothelial cells were incubated with cantharidin, okadaic acid, or an appropriate amount of DMSO as control for 30 min at 37°C. The reaction was stopped, and samples were subjected to electrophoresis. Separated proteins were transferred to nitrocellulose membranes and were incubated with monoclonal antimyosin (light chains, 20 kDa). Proteins binding the antibody were visualized using alkaline phosphatase-conjugated goat anti-mouse IgM and color reagents as reported before (Knapp et al., 1998).

Immunological Identification of PPs Type 1 and Type 2A

The immunological identification of PP 1 and PP 2A in endothelial cells was performed according to the method described recently (Knapp et al., 1998).

Northern Blotting

Tissues were homogenized using a microdismembrator (B. Braun Biotech Intl., Melsungen, Germany) in TriStar-Reagent (AGS, Heidelberg, Germany). Total RNA was extracted as published (Knapp et al., 1998). First-strand cDNA was synthesized from 1 µg of total RNA from rat heart as described (Knapp et al., 1998). Primers based on published rat cDNA sequences for PP 1 (Barker et al., 1990), PP 1 (Barker et al., 1994), PP 1 (Sasaki et al., 1990), PP 2A (Stone et al., 1988), and PP 2A (Hemmings et al., 1988) were used to generate subtype-specific probes by reverse transcription-polymerase chain reaction (RT-PCR).

The primers for PP 1 were 5'-ATGCTGGGGGGGGGTCAC-3' and 5'-CCTTTATTCAAGAGACCAGATGGG-3', for PP 1 5'-CACTGTAAAACCATCCAGCCATTTTG-3' and 5'-TGACAAAATGTCCCACTGACCAGC-3', for PP 1 5'-ACCCGTCCATTCAGA AAGC-3' and 5'-CAAGCTCGCATTTAATAAGTCTG-3', for PP 2A 5'-CCTCTTGTCATCAACAGCCGT G-3' and 5'-GCAGGAAGAACCCACAAAGTG-3' 5'-GGGATCTGTCTT GGCATTAAACC-3' and for PP 2A and 5'-CACCAAATAGGATGCAAGCAC TG-3'.

All PCR reactions were carried out as described recently (Knapp et al., 1998). Each reaction was subjected to 35 cycles of denaturation (1 min, 94°C), annealing (2 min, 56°C for PP 1, 56°C for PP 1, 54°C for PP 1, 52°C for PP 2A, and 62°C for PP 2A), and extension (2 min, 72°C). To prove identity, these PCR products from rat cardiac RNA were cycle sequenced and used in Northern blots, which were performed as reported (Knapp et al., 1998). Blots were exposed to PhosphorImager screens and visualized using a PhosphorImager (Molecular Dynamics, Krefeld, Germany).

Phosphatase Activity

Preparation of Homogenates. The preparation of homogenates of endothelial cells was performed according to the method described recently (Knapp et al., 1998). The supernatants after centrifugation are termed homogenate throughout this work. Aliquots of homogenates were used for determination of phosphatase activity.

Separation of Phosphatases. The separation was performed as described (Neumann et al., 1995; Knapp et al., 1998). Homogenates were centrifuged for 30 min at 27,000g (4°C). To the supernatant, 351 g/liter solid ammonium sulfate was added and kept at 4°C for 30 min under stirring. The suspension was sedimented at 3000g for 20 min (4°C). The resulting pellet was resuspended in buffer A, and five volumes of ethanol were added. The sample was stirred for 30 min (4°C). The suspension was centrifuged for 30 min at 27,000g (4°C). The resulting pellet was extracted with buffer A and sedimented at 27,000g for 20 min (4°C). The supernatant was kept. The pellet was re-extracted with buffer A and sedimented again at 27,000g for 20 min (4°C). Supernatants were combined and dialyzed against a 10-fold volume of buffer A containing 10% glycerol. The dialyzed sample was applied to a column containing heparin-Sepharose equilibrated in buffer A. Fractions of 3 ml were collected in the flow through (peak 1, PP 2A) and in a linear gradient from 0 to 0.5 M NaCl (peak 2, PP 1) in buffer A. Aliquots of fractions obtained were used for determination of phosphatase activity.

Phosphatase Assay. Phosphatase activity was determined as described previously (Neumann et al., 1993; Knapp et al., 1998) using [³²P]phosphorylase a as substrate. The reaction was started by adding aliquots of homogenates or aliquots of peak fractions. Reaction was stopped by addition of 50% trichloroacetic acid. Precipitated protein was sedimented by centrifugation, and the supernatant was counted in a liquid scintillation counter.

Protein Determination

Protein was measured according to the method of Bradford (1976).

Chemicals

Medium 199, amphotericin B, benzamidine, leupeptin, phenylmethylsulfonyl fluoride, and cantharidin were purchased from Sigma; L-glutamine and penicillin/streptomycin solution were purchased from Serva (Heidelberg, Germany) and Life Technologies, Inc. (Gaithersburg, MD), respectively. Gentamicin and trypsine/EDTA solution were obtained from Boehringer Mannheim (Mannheim, Germany). Fetal calf serum was obtained from Eurobio (Raunheim, Germany) and Boehringer Mannheim. Indo 1/AM and pluronic were purchased from Molecular Probes (Leiden, the Netherlands). Antibodies against PP 1 (rabbit polyclonal IgG, lot 12641) and PP 2A (rabbit polyclonal IgG, lot 13949) were obtained from BIOMOL Research Laboratories (Plymouth Meeting, PA). The antibody directed against the regulatory light chains of myosin (mouse monoclonal antimyosin light chains, 20 kDa; clone MY-21) and the antibody to factor VIII-related antigen were obtained from Sigma. All other chemicals used were of analytical or best commercially grade available.

Statistics

Results are expressed as mean ± S.E.M. Significance was estimated by Student's t test for paired and unpaired observations as appropriate. A P value <.05 was considered to be significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effects of Cantharidin on Albumin Flux. To assess the effect of cantharidin on endothelial function, we studied the permeability of albumin in bovine endothelial monolayers. In the present study, albumin permeability was continuously monitored by determining the flux of labeled albumin across endothelial monolayers. Under control conditions, mean albumin flux set at 100% in Fig. 1 was 3.9 ± 0.4 × 10¹³ mol/(s × cm²) corresponding to permeability coefficient of 6.5 ± 0.6 × 10⁶ cm/s. Albumin flux of untreated aortic endothelial monolayers was constant during the entire period of observation.

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Effects of cantharidin (5 µM, ; 10 µM, ; 25 µM, ; 100 µM, ) on albumin flux of aortic endothelial monolayers. Values represent mean ± S.E.M. of six separate experiments. Abscissa, time (min); ordinate: albumin flux as % of control (Ctr) (= 100%). The asterisks denote the first significant differences versus 0 min.

Cytosolic Calcium Concentrations ([Ca²⁺]_i). To determine the effect of cantharidin on bovine aortic endothelial [Ca²⁺]_i, we exposed Indo 1/AM-loaded monolayers of endothelial cells to 100 µM cantharidin for 30 min. Cantharidin did not affect [Ca²⁺]_i within 30 min (94.4% ± 1.7% of predrug value, n = 3). We obtained similar results in control cells treated with an appropriate amount of the solvent DMSO (96.3% ± 0.3% of predrug value, n = 3). However, the cells responded to 10 µM ATP at the end of the incubation period with a rapid increase in [Ca²⁺]_i (151.8% ± 30.6% of predrug value, n = 3). Furthermore, 0.1 µM bradykinin rapidly elevated free calcium levels (267.6% ± 78.4% of predrug value, n = 3), indicating that the cells respond to an adequate stimulus (Himmel et al., 1994). Thus, cantharidin did not affect [Ca²⁺]_i in bovine aortic endothelial cells.

Protein Phosphorylation in Bovine Aortic Endothelial Cells. We studied the effects of cantharidin on protein phosphorylation in ³²P-labeled intact bovine aortic endothelial cells. For comparison, we used okadaic acid, a more potent and selective inhibitor of phosphatase activity (Honkanen, 1993; Neumann et al., 1995; Knapp et al., 1998). Both cantharidin and okadaic acid increased the phosphorylation state of various endothelial proteins in a concentration-dependent manner. A representative autoradiogram for cantharidin is depicted in Fig. 2. Specifically, cantharidin and okadaic acid increased the phosphorylation state of a protein tentatively identified as the regulatory light chains of myosin (MLC₂₀) based on immunoblot analysis (Fig. 3). The quantitative effects of cantharidin and okadaic acid on MLC₂₀ phosphorylation are depicted in Fig. 4. Of note, we detected numerous additional phosphoproteins (Fig. 2) but did not address their identity in this study.

View larger version (54K): [in this window] [in a new window]   Fig. 2.   Concentration-dependent effect of cantharidin on protein phosphorylation in 32P-labeled bovine aortic endothelial cells. 32P-labeled aortic endothelial cells were incubated with cantharidin (Cant) for 30 min at 37°C. The reaction was stopped by addition of SDS stop solution, and samples were subjected to electrophoresis and autoradiography. Incorporated radioactivity was quantified as described in Materials and Methods. A representative autoradiogram is shown. On the left side, molecular mass standards are indicated (in kDa).

View larger version (57K): [in this window] [in a new window]   Fig. 3.   Immunological identification of the regulatory light chains of myosin (MLC) in 32P-labeled bovine aortic endothelial cells. 32P-labeled aortic endothelial cells were incubated with 10 µM okadaic acid (lane 3), 1 mM cantharidin (lane 2), or an appropriate amount of DMSO (lane 1) as control (Ctr) for 30 min at 37°C. The reaction was stopped, and samples were subjected to electrophoresis. Separated proteins were transferred to nitrocellulose membranes and were incubated with monoclonal antimyosin (light chains, 20 kDa). Proteins binding the antibody were visualized by use of alkaline phosphatase-conjugated goat anti-mouse IgM and color reagents. A corresponding immunoblot is depicted on the right side. Incorporated radioactivity of nitrocellulose membranes was quantitated by PhosphorImager with ImageQuant software. On the left side, molecular mass standards are indicated (in kDa).

View larger version (12K): [in this window] [in a new window]   Fig. 4.   Concentration-dependent effects of cantharidin and okadaic acid on protein phosphorylation in 32P-labeled bovine aortic endothelial cells. 32P-labeled aortic endothelial cells were incubated with cantharidin () or okadaic acid () for 30 min at 37°C. The reaction was stopped by addition of SDS stop solution, and samples were subjected to electrophoresis and autoradiography. Incorporated radioactivity was quantified as described in Materials and Methods. Symbols represent mean ± S.E. mean of at least six experiments. Abscissa, concentrations of cantharidin and okadaic acid; ordinate, phosphorylation as % of control (Ctr). (), significant differences versus control (DMSO).

Immunological Identification of PPs Type 1 (PP 1) and Type 2A (PP 2A) Catalytic Subunits. We subjected extracts of bovine aortic endothelial cells to gel electrophoresis and transferred separated proteins to nitrocellulose membranes. After incubating these blots with antibodies raised against the catalytic subunits of PP 1 and PP 2A (Fig. 5), we detected prominent bands at the expected molecular weight of about 36 kDa, indicating the presence of catalytic subunits of PP 1 and PP 2A/ in these cells. Next, we extended our study to the mRNA level by performing Northern blot experiments.

View larger version (40K): [in this window] [in a new window]   Fig. 5.   Immunological identification of phosphatase type 1 (PP 1) and type 2A (PP 2A). Homogenates of bovine aortic endothelial cells (lane 1), purified rabbit skeletal muscle PP 1 (lane 2), and purified rabbit skeletal muscle PP 2A (lane 3) were subjected to electrophoresis. Separated proteins were transferred to nitrocellulose and were incubated with polyclonal antiphosphatase 1 (PP 1) or antiphosphatase 2A (PP 2A) antibodies. Proteins binding antibodies were visualized using alkaline phosphatase-conjugated goat anti-rabbit IgG and color reagents. The arrows indicate prominent bands at the expected molecular mass of about 36 kDa. In the middle, molecular mass standards are indicated (in kDa).

Northern Blots. We isolated total RNA from bovine aortic endothelial cells. After separation by agarose electrophoresis and hybridization on membranes with specific DNA probes, we detected transcripts coding for PP 1, PP 1, and PP 2A at the expected sizes of about 1.8 kb, 3.2 kb, and 2.0 kb, respectively, with probes based on rat sequences (Fig. 6). We did not detect specific signals for PP 1 and PP 2A in bovine aortic endothelial cells under our experimental conditions.

View larger version (63K): [in this window] [in a new window]   Fig. 6.   Expression of different isoforms of phosphatase catalytic subunits. Northern blot analysis for PP 1, PP 1, and PP 2A. Total RNA was extracted from aortic endothelial cells. RNA was separated on 1% agarose gels (20 µg per lane) and blotted to nylon membranes, which were hybridized overnight in buffer containing the probes. Membranes were hybridized with probes specific for PP 1, PP 1, and PP 2A, washed, exposed to PhosphorImager screens, and visualized in a PhosphorImager.

PP Activity in Homogenates of Bovine Aortic Endothelial Cells. Cantharidin inhibited phosphatase activity in homogenates from bovine aortic endothelial cells with an IC₅₀ value of about 200 nM as reported before (Knapp et al., 1997). Here, we also studied okadaic acid, a structurally unrelated potent inhibitor of phosphatase activity in many tissues (Honkanen, 1993; Neumann et al., 1995; Verin et al., 1995; Knapp et al., 1998) for comparison. Okadaic acid inhibited phosphatase activity in homogenates from bovine aortic endothelial cells in a concentration-dependent fashion with an IC₅₀ value of about 3 nM (n = 3). Thus, cantharidin is about 70-fold less potent but equieffective as okadaic acid.

Effects of Cantharidin and Okadaic Acid on Separated Phosphatase Activities. We separated catalytic subunits of PP 1 and PP 2A from bovine aortic endothelial cells by column chromatography (Fig. 7). The first peak of activity corresponds most likely to PP 2A and the second peak to PP 1. Cantharidin inhibited the activities of PP 1 and PP 2A with IC₅₀ values of about 3 µM and 30 nM, respectively (Fig. 8A). Okadaic acid was equieffective asbut more potent thancantharidin, with IC₅₀ values of about 30 nM for PP 1 and 0.1 nM for PP 2A (Fig. 8B).

View larger version (12K): [in this window] [in a new window]   Fig. 7.   Separation of catalytic subunits of PP 1 and PP 2A from bovine aortic endothelial cells. The separation of catalytic subunits of PP 1 and PP 2A was performed as described in Materials and Methods. Fractions of 3 ml were collected in the flow-through (peak 1, PP 2A) and in a linear gradient from 0 to 0.5 M NaCl (peak 2, PP 1) in buffer A. Aliquots of fractions obtained were used for determination of phosphatase activity. Abscissa, fractions obtained after column chromatography; ordinate, phosphatase activity of fractions in cpm.

View larger version (15K): [in this window] [in a new window]   Fig. 8.   Inhibition of purified phosphatase type 1 (PP 1) and 2A (PP 2A) activities. Inhibition of purified PP 1 () and PP 2A () activities from bovine aortic endothelial cells by cantharidin (A) and okadaic acid (B). Phosphorylase a phosphatase activity is expressed as % of solvent control (DMSO). The concentration of solvent was constant under all experimental conditions. Symbols represent mean ± S.E. mean of at least three experiments. Abcissae, concentrations of cantharidin or okadaic acid; ordinates, phosphorylase a phosphatase activity as % of control (Ctr). (), first significant differences versus control.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The main findings of the present study are that cantharidin can increase the phosphorylation state of proteins in bovine aortic endothelial cells, can inhibit the activity of PPs type 1 (PP 1) and type 2A (PP 2A) of endothelial cells, and can alter endothelial barrier function.

In this study, we show the presence of PP 1 and PP 2A in bovine aortic endothelial cells by several lines of evidence. First, the enzymatic activity of PP 1 and PP 2A was present in homogenates from aortic endothelial cells. We could separate the activities of the catalytic subunits of PP 1 and PP 2A from the regulatory subunits by means of ethanol precipitation and then separate PP 1 and PP 2A from each other by affinity chromatography on heparin Sepharose. This methodology is similar to previous separation procedures in rabbit liver (Erdödi et al., 1985), guinea pig cardiomyocytes (Herzig et al., 1995), guinea pig heart (Neumann et al., 1995), human heart (Neumann et al., 1997), and bovine coronary arteries and bovine smooth muscle cells (Knapp et al., 1998).

Second, we could identify the catalytic subunits of PP 1 and PP 2A in aortic endothelial cells immunologically by a method used before in bovine smooth muscle cells (Knapp et al., 1998). Others have identified catalytic subunits of PP 1 and PP 2A in bovine pulmonary endothelial cells and human umbilical vein endothelial cells by immunoblot analysis (Verin et al., 1995).

Third, we identified the expression of PP 1, PP 1, and PP 2A by Northern blot analysis using subtype-specific probes. These catalytic subunits are also present in bovine coronary arteries, isolated smooth muscle cells, and myocardium (Knapp et al., 1998). Others have shown the expression of catalytic subunits of PP 1 in human umbilical vein endothelial cells using PCR (Shasby et al., 1997). Taken together, our data demonstrate that PP 1 and PP 2A are present on mRNA and protein level in bovine aortic endothelial cells. We conclude that PP 1, PP 1, and PP 2A are ubiquitous in the cardiovascular system.

Cantharidin is a known specific inhibitor of PP 1 and PP 2A (Honkanen, 1993; Neumann et al., 1995; Knapp et al., 1997, 1998). Although structurally unrelated to other PP inhibitors, cantharidin shares functional effects with other PP inhibitors like okadaic acid (Kodama et al., 1986; Takai et al., 1987). In the past and in this present study, we have used experimental conditions where the sum of PP 1 and PP 2A activity is measured in homogenates. We confined our interest to PP 1 and PP 2A because they comprise more than 90% of PP activity in eukaryotic cells (DePaoli-Roach et al., 1994). Cantharidin inhibited PP activity in homogenates of bovine aortic endothelial cells with a similar potency as in homogenates from guinea pig cardiomyocytes, human cardiac muscle, bovine cardiac muscle, bovine coronary arteries, bovine aortic smooth muscle cells, and bovine coronary artery smooth muscle cells (Linck et al., 1996; Neumann et al., 1995; Knapp et al., 1997, 1998). In all PP measurements in homogenates, cantharidin was less potent than okadaic acid.

Cantharidin inhibited activity of PP 1 less potently than the activity of PP 2A in several tissues. This holds true for PP 1 and PP 2A purified from bovine aortic endothelial cells (present study), bovine vascular smooth muscle (Knapp et al., 1998), and guinea pig heart (Neumann et al., 1995). Moreover, cantharidin is consistently less potent than okadaic acid to inhibit purified PP 1 and PP 2A activity. For instance, cantharidin and okadaic acid inhibit the activity of PP 1 from guinea pig heart with IC₅₀ values of 2.70 µM and 120 nM and of PP 2A with IC₅₀ values of 0.13 µM and 0.7 nM, respectively.

How do PPs affect endothelial function? We studied the effect of PP inhibition by cantharidin on endothelial permeability by using a well characterized technique, namely albumin flux (Siflinger-Birnboim et al., 1987; Garcia et al., 1996; Hempel et al., 1996). Cantharidin elevated the albumin flux through aortic endothelial cell monolayers. To the best of our knowledge, this has not been reported before. This enhanced permeability is a clear indication that PP could play an important role in the barrier function of endothelial cells. It is possible that the enhanced flux is the result of phosphorylation of regulatory proteins in these cells. Others have noted before that thrombin enhances the permeability of endothelial cells (Garcia et al., 1995). The authors attributed enhanced permeability to MLC₂₀ phosphorylation and contraction of endothelial cells. Consistent with the assumption that cantharidin is able to permeate intact cell membranes, cantharidin increased MLC₂₀ phosphorylation in ³²P-labeled aortic endothelial cells. It is noteworthy that we required higher concentrations of cantharidin than of okadaic acid to elevate protein phosphorylation in these cells. These observations strongly support the view that cantharidin acts mainly, if not exclusively, by inhibition of PP 1 and PP 2A activities. Furthermore, cantharidin did not increase cytosolic Ca²⁺ in aortic endothelial cells (this study) and vascular smooth muscle cells (Knapp et al., 1998). Thus, it seems unlikely that cantharidin enhances MLC₂₀ phosphorylation by activating Ca²⁺/calmodulin-dependent myosin light chain kinase.

Interestingly, indirect evidence indicates that thrombin acts at least in part by inhibition of PP activity in endothelial cells (Shasby et al., 1997). However, a caveat is in order. Others used nonradioisotopic methods and noted that histamine, thrombin, and PP inhibitors could elevate MLC₂₀ phosphorylation (Verin et al., 1995; Shasby et al., 1997). The present work, using the more powerful method of metabolic labeling of endothelial cells with [³²P]orthophosphate, clearly indicates that numerous proteins in addition to MLC₂₀ are also phosphorylated in the presence of a PP inhibitor. Hence, the enhanced permeability in endothelial cells could involve other substrates in addition to MLC₂₀ (for review, see DePaoliRoach et al., 1994; and Lum and Malik, 1994). It has not been reported before that PP inhibition by cantharidin stimulates the permeability of bovine aortic endothelial cells. Although both PP 1 and PP 2A are known to dephosphorylate MLC₂₀ (Haeberle et al., 1985; Shirazi et al., 1994), it has been hypothesized that PP 1 is the most important PP that regulates the phosphorylation state of MLC₂₀ in endothelial cells. This conclusion was based on the observation that 100 nM okadaic acid did not affect MLC₂₀ phosphorylation (Verin et al., 1995). Likewise, we noted that at least 10 µM okadaic acid is necessary to stimulate MLC₂₀ phosphorylation (Fig. 4). Hence, our data indicate that MLC₂₀ phosphorylation in aortic endothelial cells like in pulmonary artery endothelial cells (Verin et al., 1995) probably involves mainly PP 1.

In summary, we present evidence that PP 1 and PP 2A are involved in the regulation of aortic endothelial permeability. This conclusion is based on the fact that inhibition of these PPs by cantharidin increases macromolecule permeability and at the same time increases the phosphorylation state of endothelial regulatory proteins.