Portal de Periódicos da CAPES

Protein phosphatase 2B inhibition promotes the secretion of von Willebrand factor from endothelial cells

2009; Elsevier BV; Volume: 7; Issue: 6 Linguagem: Inglês

10.1111/j.1538-7836.2009.03355.x

1538-7933

Leticia Nolasco, Francisca C. Gushiken, NA Turner, Tanvir Khatlani, Subhashree Pradhan, Jing Dong, Joel L. Moake, K. Vinod Vijayan,

Platelet Disorders and Treatments

Background: Secretion of Weibel–Palade body (WPB) contents is regulated, in part, by the phosphorylation of proteins that constitute the endothelial exocytotic machinery. In comparison to protein kinases, a role for protein phosphatases in regulating endothelial exocytosis is undefined. Objective and method: In this study, we investigated the role of protein phosphatase 2B (PP2B) in the process of endothelial exocytosis using pharmacological and gene knockdown approaches. Results: We show that inhibition of protein phosphatase 2B (PP2B) activity by cyclosporine A (CsA), tacrolimus or a cell-permeable PP2B autoinhibitory peptide promotes the secretion of ultralarge von Willebrand factor (ULVWF) from human umbilical vein endothelial cells (HUVECs) in the absence of any other endothelial cell-stimulating agent. PP2B inhibitor-induced secretion and anchorage of ULVWF strings from HUVECs mediate platelet tethering. In support of a role for PP2B in von Willebrand factor (VWF) secretion, the catalytic subunit of PP2B interacts with the vesicle trafficking protein, Munc18c. Serine phosphorylation of Munc18c, which promotes granule exocytosis in other secretory cells, is increased in CsA-treated HUVECs, suggesting that this process may be involved in CsA-mediated WPB exocytosis. Furthermore, the plasma VWF antigen level is also enhanced in CsA-treated mice, and small interfering RNA-mediated knockdown of the α and β isoforms of the PP2B-A subunit in HUVECs enhanced VWF secretion. Conclusions: These observations suggest that CsA promotes VWF release, in part by inhibition of PP2B activity, and are compatible with the clinically observed association of CsA treatment and increased plasma VWF levels in humans. Cyclosporine A (CsA), an immunosuppressive agent commonly used in organ transplantation, is also a potent inhibitor of protein phosphatase 2B (PP2B) activity [1]. A subset of patients treated chronically with CsA develop a type of thrombotic microangiopathy that clinically resembles thrombotic thrombocytopenic purpura or hemolytic–uremic syndrome [2]. CsA-treated patients also develop increased plasma levels of von Willebrand factor (VWF) [3]. Ultralarge VWF (ULVWF) multimers are produced and stored in the Weibel–Palade bodies (WPBs) of endothelial cells. ULVWF forms are released as long, hyperadhesive strings anchored to endothelial cells in response to different agonists (including histamine, thrombin, vasopressin, epinephrine, several cytokines, and Shiga toxins). Perfused platelets instantaneously attach to these ULVWF multimeric strings [4, 5]. ULVWF–platelet strings are subsequently cleaved by the plasma metalloprotease ADAMTS-13 (a disintegrin and metalloprotease with thrombospondin motifs) into smaller and less adhesive plasma VWF multimers. The secretion of ULVWF is followed by the fusion of WPBs to the endothelial cell plasma membrane. Various signaling pathways can be involved. Thrombin and histamine increase intracellular Ca2+ levels, whereas vasopressin and epinephrine increase intracellular cAMP levels [6, 7]. Through the agonist-specific effectors calcium/calmodulin (Ca2+/CaM), in thrombin–histamine signaling pathways, or protein kinase A (PKA) in vasopressin–epinephrine pathways, a small GTPase (Ral) is activated and promotes endothelial cell VWF secretion [6, 7]. A guanine exchange factor for Ral, the Ral GDP dissociation stimulator, has been reported to regulate Ral-dependent exocytosis of WPBs in response to agonists that elevate either intracellular Ca2+or intracellular cAMP [8]. In contrast to Ral, another GTPase, Rab3D, may play a negative stimulatory role in VWF exocytosis, as overexpression of Rab3D inhibits VWF secretion [9]. Recently, selective release of VWF but not tissue-type plasminogen activator in response to histamine was shown to be mediated by phospholipase D1 [10]. Collectively, these signaling events could result in the reorganization of the cytoskeleton, particularly actin and microtubules, in order to achieve WPB redistribution and fusion to the endothelial cell plasma membrane. Microtubular 'motors', including dynein–dynactin, contribute to the contractile forces that facilitate the clustering of WPBs and secretion of their contents [11]. In other secretory cells (e.g. mast cells, neurons, and pancreatic islet cells), vesicle trafficking proteins of the soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) superfamily are also involved in granule exocytosis. Endothelial cells contain SNARE proteins, and there is evidence that SNARE and SNARE-associated proteins may control the exocytosis of WPBs and secretion of their contents [12, 13]. Reversible post-translational modifications, including S-nitrosylation of N-ethylmaleimide-sensitive factor [12] or phosphorylation of tyrosine and/or serine/threonine residues on vesicle trafficking proteins, are capable of regulating granule exocytosis [14]. The phosphorylation status of any trafficking protein is determined by the action of both protein kinases and protein phosphatases, and alterations in the activity of either type of enzyme can alter exocytosis. A role for PKA [11] or protein kinase Cδ [15] in endothelial cell VWF exocytosis has been investigated. In contrast, any contribution of serine/threonine phosphatases to the regulation of WPB exocytosis and secretion of their contents is unknown. PP2B (or calcineurin) is a serine/threonine phosphatase that exists as a heterodimer composed of a 58–64-kDa catalytic subunit (A subunit; PP2B-A) and a 19-kDa Ca2+ regulatory subunit (B subunit; PP2B-B). PP2B-A has three isoforms (α, β, and γ); the α and β isoforms are ubiquitous, whereas the γ isoform is testis-specific. PP2B is a Ca2+/CaM-activated enzyme involved in the regulation of a variety of cellular functions [16]. PP2B-A contains an autoinhibitory (AI) domain and the binding sites for PP2B-B and CaM. In the absence of CaM, the AI domain binds to the active cleft of PP2B and inhibits the phosphatase activity [17]. Peptides mimicking the AI domain are potent and specific inhibitors of PP2B activity [18]. The immunosuppressants CsA and tacrolimus (FK506) are also inhibitors of PP2B. In order for this inhibition to occur, CsA must first form a complex with the resident cytoplasmic protein, cyclophilin, and tacrolimus must complex with FK506-binding protein 12 (FKBP12) [1]. This phosphatase inhibitory activity could provide, in part, a mechanism for the CsA-associated thrombotic microangiopathies found in a small subset of organ transplant patients. As PP2B is a serine/threonine phosphatase, we hypothesized that inhibition of PP2B activity may induce ULVWF secretion (at least in part) by increased serine phosphorylation of a vesicle trafficking protein that promotes WPB exocytosis and ULVWF secretion. We evaluated this possibility in vitro, using human umbilical vein endothelial cells (HUVECs), and in vivo in mice. HUVECs were obtained as previously described [5]. Briefly, umbilical cord veins were washed in phosphate buffer, subjected to collagenase digestion, and seeded onto six-well or 35-mm culture dishes coated with 1% gelatin. HUVECs were maintained in medium 199 supplemented with 20% fetal bovine serum, 3% penicillin–streptomycin–neomycin, and 0.2 mm l-glutamine, and grown to confluency (5–7 days). Most of the cells used in these studies were either primary or used after one passage. The presence of mRNA for PP2B-Aα and PP2B-Aβ was determined using SYBR-Green-based real-time PCR. Total RNA was isolated from HUVECs with an RNeasy Mini kit (Qiagen, Valencia, CA, USA), and cDNA was synthesized from 2 μg of total RNA using the Superscript III First RT-PCR system (Invitrogen, Carlsbad, CA, USA). PCR amplification of the cDNA samples was performed using 7 μL of 2 × Power SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA, USA), 1 pmol of primers, and 13 μL of cDNA preparation, to a final volume of 20 μL. The following primers were used: 5′-ACGCCAACCTTAACTCCATCA-3′ (forward) and 5′-TGCTGTCCGTGCCGTTAGT-3′ (reverse) for PP2B-Aα; 5′-TGGTGAAAGAAGGTCGAGTAGATG-3′ (forward) and 5′-TGGCAGCACCCTCATTGA-3′ (reverse) for PP2B-Aβ; and 5′-ATGGAAATCCCATCACCATCTT-3′ (forward) and 5′-CGCCCCACTTGATTTTGG-3′ (reverse) for glyceraldehyde-3-phosphate dehydrogenase. PCR cycling conditions included one cycle of initial activation at 50 °C for 2 min, one cycle of initial denaturation at 95 °C for 10 min, 40 cycles of denaturation at 95 °C for 15 min, and 40 cycles of annealing at 60 °C for 1 min. The PCR products, their dissociation curves and absolute quantity of mRNA were detected with the 7300 Real-Time PCR System (Applied Biosystems). Cells were stimulated for 15 min with either CsA, FK506 [1 and 10 nmol L−1 diluted in 0.001% dimethylsulfoxide (DMSO)] (Sigma, St Louis, MO, USA), vehicle control DMSO (0.001%), or cell-permeable peptides containing the scrambled or PP2B AI domain (1 or 2 μmol L−1) or 100 μmol L−1 histamine as positive controls. PP2B AI cell-permeable peptide was composed of the calcineurin AI domain fused to a poly-arginine-based protein transduction domain (11R), Ac-RRRRRRRRRRRGGGRMAPPRRDAMPSDA-NH2 (Calbiochem, San Diego, CA, USA). Ac-RRRRRRRRRRRGGGRMRDRPAPAMDPSA-NH2 was the scrambled peptide synthesized by Gene Script Corporation (Piscataway, NJ, USA). Any cellular debris in the supernatant was removed by centrifugation at 150 × g for 3 min. VWF antigen levels were measured in cell supernatants using a standard enzyme-linked immunosorbent assay (ELISA) technique. Samples were incubated for 1 h on microtiter plates precoated with 1 μg mL−1 rabbit anti-human VWF antibody (Ramco Laboratories, Stafford, TX, USA). After several washes, samples were incubated with 0.5 μg mL−1 of goat anti-human VWF and rabbit anti-goat horseradish peroxidase (HRP) antibodies (Bethyl Laboratories, Montgomery, TX, USA) and color-developed by the addition of 3,3′,5,5′-tetramethylbenzidine (Invitrogen). The reaction was stopped with 1 m HCl, and the absorbance was read at 450 nm. Human plasma was used as a standard in the VWF antigen ELISA. Fold differences in VWF antigen levels were determined after normalizing the absorbance values obtained for test samples to that obtained from media only. In some experiments, equal volumes of samples (cell supernatants) obtained from equivalent numbers of HUVECs treated with various agents were separated by sodium dodecylsulfate (SDS)–1% agarose gel electrophoresis, and VWF multimers were detected by immunoblotting with polyclonal anti-human VWF antibody and secondary anti-IgG HRP conjugate (Bethyl Laboratories). PP2B activities were measured in HUVEC lysate obtained after treatment with the following agents: DMSO, CsA, FK506, AI and scrambled peptides, control small interfering RNA (siRNA) and also in cells depleted of PP2B-Aα or PP2B-Aβ using the Calcineurin Cellular Assay Kit Plus (Biomol International, Plymouth Meeting, PA, USA). Free phosphate within the lysates was removed by gel filtration using desalting resin. PP2B activity in the lysate was measured by removal of phosphate from the PP2B substrate (RII phosphopeptide) provided. The released phosphate was detected by addition of Biomol Green reagent, and absorbance measurements at 620 nm were converted into released phosphate using a generated phosphate standard curve. Isolation and washing of platelets from healthy voluntary donors was performed as described previously [5]. Human subjects gave written informed consent, and the study was approved by the Baylor College of Medicine and Rice University Review Boards. The number of ULVWF strings (with adherent platelets) secreted by PP2B inhibitor-induced stimulated endothelial cells was quantified under flowing conditions (2.5 dynes cm−2) in a parallel-plate flow chamber system, as previously described [5]. HUVECs seeded onto 35-mm culture dishes were treated for 2 min with either: 100 μmol L−1 histamine (positive control); or 0.001% DMSO (solvent control); or 1, 10 nmol L−1 (CsA and FK506); or 1, 2 μmol L−1 (AI peptide) or Tyrode's buffer (basal). Washed platelets resuspended in Ca2+/Mg2+-free Tyrode's buffer were perfused over the treated HUVECs for 2 min and the number of ULVWF strings was calculated by counting the number of platelets that adhered to the ULVWF strings in 20 separate fields at 200× using phase contrast microscopy. Balb/C mice (8–12 weeks) were intraperitoneally injected with either 20 mg kg−1 CsA or control vehicle DMSO, or mock injected with an equal volume of 0.9% saline (NaCl) solution under a protocol approved by the Institutional Animal Care and Use Committee at Baylor College of Medicine. Blood was collected 15 min after injection. VWF levels in the plasma were measured using a sandwich ELISA described above, with the exception that the antibodies used were rabbit anti-human VWF (Dako, Glostrup, Denmark). This antibody cross-reacts with mouse VWF, and has been previously used to detect mouse VWF antigen levels [19]. Mouse plasma was used as a standard in the VWF antigen ELISA. Fold differences in mouse VWF antigen levels were determined after normalizing the absorbance values obtained for test samples (CsA or DMSO treatment) to that obtained from control (mock NaCl treatment). A preformed mix of four independent siRNAs targeting PP2B-Aα (PPP3CA) (M-008300-02), and PP2B-Aβ (PPP3CB) (M-009704-01) were purchased from Dharmacon (Lafayette, CO, USA). A non-specific control siRNA pool with no sequence homology to any human or mouse sequence (D-001206-13-05) was used as control in this study. HUVECs were transfected with 100 nmol L−1 siRNA oligonucleotides, using DharmaFect transfecting reagent (Dharmacon) according to the manufacturer's instruction. After 48 h, the supernatant was collected, and VWF antigen levels were measured with a VWF ELISA assay kit (Ramco Laboratories). The percentage of VWF antigen was determined using an internal prediluted human VWF calibrator solution. Fold differences in VWF antigen levels were determined after normalizing the percentage of VWF antigen obtained for cells treated with PP2B-Aα or PP2B-Aβ to that obtained from control siRNA-treated cells. Silencing of PP2B-Aα and PP2B-Aβ were confirmed by RT-PCR and immunoblotting with isoform-specific antibodies (Santa Cruz Biotechnology, Santa Cruz, CA, USA). The blots were stripped and reprobed with an antibody to actin (loading control). The interaction of the catalytic subunit of PP2B (PP2B-Aα) with the vesicle trafficking protein Munc18c was determined by a PP2B pull-down assay. PPP3CA truncated at residue 347 and tagged to glutathione S-transferase (GST) in the pGEX-6P plasmid was obtained from A. Rao (Harvard Medical School, Boston, MA, USA). This cDNA lacks the AI domain, but retains the catalytic subunit and many of the PP2B-B-binding subunits that are necessary to sustain interaction with any potential substrates. GST-tagged protein was expressed in Escherichia coli after isopropyl-β-d-thiogalactopyranoside induction, and purified using glutathione beads. The GST and GST–PP2B-Aα proteins were characterized by Coomassie Blue staining. Purified GST–PP2B-Aα and GST proteins (3 μg) were precoupled with glutathione beads and mixed overnight at 4 °C with 100 μg of lysates obtained from resting HUVECs. Beads were washed three times, analyzed on a 10% SDS polyacrylamide gel electrophoresis (PAGE) gel, and immunoblotted with anti-Munc18c antibody (gift from J.E. Pressin, State University of New York at Stony Brook, NY, USA). Lysates obtained from HUVECs (750 μg) following treatment with DMSO or CsA (1 and 10 nmol L−1) for 15 min were immunoprecipitated using 3 μg of anti-phosphoserine antibody (Invitrogen) and protein A–Sepharose. Proteins were separated by 10% SDS-PAGE, transferred to nitrocellulose, probed with polyclonal antibody to Munc18c, and developed using the enhanced chemiluminescence system (Amersham Bioscience, Piscataway, NJ, USA). The signals were scanned using photoshop version 6 software (Adobe, San Jose, CA, USA), and densitometric quantification was performed using a BioRad gel documentation system (Hercules, CA, USA). Experimental conditions were compared by using paired Student's t-tests. A P-value of 0.05 was considered significant. The absolute amount of VWF secreted from HUVECs showed some variation from experiment to experiment, but there was consistency in the ratio of control to PP2B-inhibited samples. Therefore, the VWF antigens were primarily expressed as a fold difference in PP2B-inhibited samples as compared with control or mock-treated samples. Figure 1A demonstrates the presence of mRNA for PP2B-Aα and PP2B-Aβ in HUVECs. Immunoblotting of the lysates from HUVECs with isoform-specific antibodies confirmed the presence of PP2B-Aα and PP2B-Aβ proteins (Fig. 1B,C). Furthermore, the regulatory subunit of PP2B (PP2B-B) was also expressed in HUVECs (Fig. 1D). Inhibition of protein phosphatase 2B (PP2B) was associated with von Willebrand factor (VWF) secretion from human umbilical vein endothelial cells (HUVECs). (A) Expression of PP2B-Aα and PP2B-Aβ mRNA by reverse transcription polymerase chain reaction (RT-PCR). An amplification product of 714 bp specific for PP2B-Aα and one of 737 bp specific for PP2B-Aβ were identified by RT-PCR. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (496 bp) was used as a loading control. (B–D) Expression of PP2B-Aα, PP2B-Aβ and PP2B-B proteins. HUVEC lysates (40 μg) were separated by sodium dodecylsulfate polyacrylamide gel electrophoresis and immunoblotted with antibodies specific for PP2B-Aα (B), PP2B-Aβ (C), or the regulatory subunit of PP2B (PP2B-B) (D). Lysate obtained from mouse brain (brain) is also shown as a positive control. (E) Cyclosporine A (CsA), FK506 and autoinhibitory (AI) peptide inhibited PP2B activity. Data are the mean ± standard error of the mean of three to eight experiments. As compared with dimethylsulfoxide (DMSO): *P = 0.0043 (for 1 nmol L−1 CsA); #P = 0.0028 (for 10 nmol L−1 CsA); †P = 0.056 (for 1 nmol L−1 FK506); and ⋄P = 0.04 (for 10 nmol L−1 FK506). As compared with the scrambled peptide: ♦P = 0.03 and ‡P = 0.02 for 1 and 2 μmol L−1 AI peptide, respectively. (F) PP2B inhibitors promoted VWF release. HUVECs were incubated with the indicated compounds, the supernatant collected, and the released VWF was measured by enzyme-linked immunosorbent assay. Data are mean ± standard error from five to 11 experiments. As compared with media: *P < 0.001 (for 1 nmol L−1 CsA); #P = 0.004 (for 10 nmol L−1 CsA); †P = 0.05 (for 1 nmol L−1 FK506); ⋄P = 0.01 (for 10 nmol L−1 FK506); ○P = 0.40 (for 1 μmol L−1 scrambled peptide); •P = 0.069 (for 2 μmol L−1 scrambled peptide); and ♣P = 0.005 (for 100 μmol L−1 histamine). As compared with scrambled peptide: ♦P = 0.044 and ‡P = 0.01 for 1 and 2 μmol L−1 AI peptide, respectively. In order to evaluate the possible functional role of PP2B in VWF secretion from WPBs, we treated primary HUVECs in culture with PP2B inhibitors, and measured VWF antigen release using an ELISA assay. As compared with the controls (DMSO or cell-permeable scrambled peptide), exposure of HUVECs to CsA, FK506 or a cell-permeable peptide containing the PP2B AI domain [20] caused a maximum of approximately 48–56% inhibition of PP2B activity (Fig. 1E). The inhibition of PP2B activity by these agents was associated with the rapid release of VWF antigen from HUVECs in a dose-dependent manner (Fig. 1F). Because AI, but not the scrambled peptide, significantly enhanced VWF secretion, only AI peptide was used for the rest of the studies. CsA, FK506 and AI peptide were about as effective as histamine in promoting the release of VWF antigen from HUVECs. HUVECs secrete ULVWF multimers from the WPBs rapidly in response to histamine [4]. We conducted experiments to determine whether inhibition of PP2B activity also promotes the secretion of ULVWF multimeric strings from the HUVECs. As compared with normal pooled plasma (reference), HUVECs exposed either to CsA, AI peptide or FK506 secreted VWF multimers enriched in ULVWF forms (Fig. 2A). In complementary experiments, PP2B inhibitor-induced ULVWF strings were visualized using microscopy by the adherence of washed platelets under flowing conditions (approximately 2.5 dynes cm−2 shear stress) to the hyperadhesive ULVWF strings. As compared with the DMSO-treated endothelial cells, HUVECs secreted approximately 4–5-fold or approximately 2.5–3-fold more ULVWF–platelet strings in response to CsA and FK506, respectively (Fig. 2B). Moreover, HUVECS treated with AI peptide or histamine (positive control) also showed approximately 3.5-fold more ULVWF strings than cells in Tyrode's buffer (basal conditions) (Fig. 2B). These studies suggest that inhibition of PP2B provokes the secretion of ULVWF strings from endothelial cells in vitro. Exposure to the PP2B inhibitors or vehicle control did not damage HUVECs, as demonstrated by the failure of these agents to diminish cell metabolic/reductive capacity using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (data not shown). These results indicated that release of VWF multimers enriched in ULVWF forms by HUVECs is not the consequence of cell damage induced by the PP2B inhibitors. Protein phosphatase 2B (PP2B) inhibitors induced release of ultralarge von Willebrand factor (ULVWF)/ von Willebrand factor (VWF) multimers from human umbilical vein endothelial cells (HUVECs). (A) ULVWF/VWF multimeric patterns. HUVECs were treated with PP2B inhibitors, and ULVWF/VWF multimers from the cell supernatants were analyzed by unreduced 1% agarose/ sodium dodecylsulfate electrophoresis/anti-VWF antibody immunoblotting. VWF multimeric patterns from normal platelet-poor plasma or endothelial cell supernatant (EC sup) from histamine-treated cells are shown for comparison. ULVWF multimers are indicated by a vertical line. As compared with the normal platelet-poor plasma lane, ULVWF multimers were prominently enhanced in the lane containing the PP2B inhibitor-treated samples. Occasionally, the dimethylsulfoxide (DMSO)-treated lane also showed some VWF and ULVWF forms (n = 3). (B) Perfused platelets adhered instantaneously to ULVWF strings secreted by HUVECs in the presence of PP2B inhibitors under flow conditions. HUVECs were preincubated with the control reagents or PP2B inhibitors before the perfusion of normal fresh washed human platelets using a flow chamber (wall shear stress, 2.5 dyne cm−2). The number of ULVWF–platelet strings that formed was counted over the course of 2 min in 20 different microscope fields. Data are presented as mean ± standard error of the mean of two to six experiments. As compared with DMSO: ‡P = 0.02 [for 1 nmol L−1 cyclosporine A (CsA)]; *P = 0.0024 (for 10 nmol L−1 CsA); ∞P = 0.05 (for 1 nmol L−1 FK506); +P = 0.002 (for 10 nmol L−1 FK506); and #P = 0.0069 (for histamine). As compared with Tyrode's buffer-treated cells: †P = 0.05 [for 1 μmol L−1 autoinhibitory (AI) peptide]; ♣P < 0.01 for 2 μmol L−1 AI peptide). In order to determine whether CsA can enhance VWF secretion in vivo, we treated mice with CsA and analyzed plasma VWF antigen levels. As compared with the mice intraperitoneally injected with saline (NaCl) or DMSO vehicle control, CsA caused an approximately 1.6-fold increase in plasma VWF antigen within 15 min, suggesting that CsA enhances VWF secretion in vivo (Fig. 3). A functional effect of PP2B inhibition on HUVEC VWF release was further evaluated using siRNA to knock down the expression of endogenous PP2B-Aα and PP2B-Aβ in HUVECs. Knockdown of PP2B-Aα and PP2B-Aβ was confirmed by RT-PCR studies (Fig. 4A,C) and immunoblotting with isoform-specific antibodies (Fig. 4B,D). There was no evidence of any compensatory increase in PP2B-Aβ protein levels in PPP3CA knockdown in HUVECs and vice versa. As compared with control siRNA-treated cells, the phosphatase activities in PP2B-A α and PP2B-A β knockdown endothelial cells were decreased by approximately 48% (Fig. 4E). Targeting the PP2B-A α isoform leaves active PP2B-A β isoform that may account for the remaining approximately 50% phosphatase activity observed in PP2B-Aα knockdown cells and vice versa. More importantly, knockdown of PP2B-Aα and PP2B-Aβ resulted in a approximately 2.5 and 2-fold increased VWF release from HUVECs, respectively (Fig. 4F). Increased von Willebrand factor (VWF) levels in the plasma of mice treated with cyclosporine A (CsA). Data are mean ± standard error from three experiments. As compared with treatment with dimethylsulfoxide (DMSO) alone:*P = 0.042 for CsA treatment. Increased von Willebrand factor (VWF) levels in the supernatants of human umbilical vein endothelial cells (HUVECs) in protein phosphatase 2B (PP2B)-Aα and PP2B-Aβ knockdown cells. HUVECs were transfected with either control small interfering RNAs (siRNAs) or specific siRNAs targeting PPP3CA and PPP3CB, using DharmaFect. (A, C) reverse transcription polymerase chain reaction (RT-PCR) analysis. After 48 h, total cellular RNA was extracted, and RT-PCR was performed using primers specific for PP2B-Aα, PP2B-Aβ, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (loading control). (B, D) Imunoblotting. Lysates were immunoblotted with antibodies to PP2B-Aα, PP2B-Aβ, and actin (loading control). (E) PP2B-Aα and PP2B-Aβ knockdown in HUVECs inhibited PP2B activity. Data are the mean ± standard error of the mean (SEM) of three or four experiments. As compared with control siRNA-treated cells, knockdown of PP2B-Aα and PP2B-Aβ decreased PP2B activity by approximately 48%. (F) PP2B-Aα and PP2B-Aβ knockdown in HUVECs promoted VWF secretion. PP2B-Aα and PP2B-Aβ protein was depleted by an siRNA, and HUVEC supernatants were collected after 48 h, and centrifuged; VWF antigen levels were quantified with the Spectro VWF enzyme-linked immunosorbentassay kit (Ramco Laboratories). Results are mean ± SEM of four or five experiments performed in triplicates. As compared with control siRNA-treated cells: *P = 0.001 for PP2B-Aα knockdown, and †P = 0.05 for PP2B-Aβ knockdown. In order to investigate a potential mechanism by which CsA promotes endothelial cell secretion, we investigated the effect of PP2B inhibition by CsA on the serine phosphorylation of Munc18c. Munc18c phosphorylation is an essential post-translational modification of this vesicle trafficking protein that is associated with increased vesicle exocytosis in several other cell systems [14]. Treatment of HUVECs with 1 and 10 nmol L−1 of CsA resulted in approximately 1.5-fold or two-fold, respectively, increased serine phosphorylation of Munc18c, as compared with control DMSO vehicle-treated cells (Fig. 5A,B). When the catalytic subunit of PP2B (PP2B-Aα) was expressed as a GST fusion protein in E. coli, GST pull-down assays revealed that GST–PP2B-Aα, but not GST alone, interacted with Munc18c (Fig. 5C). Taken together, these data indicate that the catalytic subunit of PP2B can associate with Munc18c at least in vitro, and that increased serine phosphorylation of Munc18c is associated with both inhibition of PP2B activity by CsA and augmented secretion of ULVWF. Serine phosphorylation of the vesicle trafficking protein, Munc18c, was enhanced in cyclosporine A (CsA)-treated human umbilical vein endothelial cells (HUVECs). (A) CsA triggered serine phosphorylation of Munc18c. Lysates from HUVECs treated with either dimethylsulfoxide (DMSO) vehicle control or varying concentrations of CsA were immunoprecipitated (IP) with anti-phosphoserine antibody, separated by reducing sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE), and immunoblotted using anti-Munc18c antibody (upper panel). The position of the 50-kDa IgGγ heavy chain is indicated as IgG. Endothelial cell (EC) lysate is shown to indicate the position of Munc18c as a positive control reference for the immunoprecipitated phosphoserine samples (Lys; upper panel). Lysate used for immunoprecipitation experiments contains Munc18c (lower panel). The upper panel was obtained with a longer exposure than the lower panel, and is responsible for the difference in the intensity of Munc18c in the lysate of the upper panel as compared with the lower panel. (B) Densitometric quantification of serine-phosphorylated Munc18c. Phosphorylated Munc18c (pMunc18c) is the ratio of phosphorylated Munc18c to total Munc18c (in arbitrary units) formed in response to CsA (n = 3). As compared with control DMSO vehicle treatment: *P = 0.03 and ‡P = 0.01 for 1 and 10 nmol L−1 CsA, respectively. (C) Munc18c interacted with protein phosphatase 2B (PP2B)-Aα. PP2B-Aα– glutathione S-transferase (GST) coupled to glutathione beads was used to pull down Munc18c from the HUVEC lysate, as described in Materials and methods. Some Munc18c became bound to PP2B-Aα–GST beads, but not to beads coated with GST alone. The PP2B-bound Munc18c was then released by boiling in SDS buffer, separated by reducing SDS-PAGE, and detected by immunoblotting with anti-Munc18c. The GST and PP2B-Aα–GST proteins in the pull-down assays were detected by Coomassie Blue staining, and Munc18c in the lysates used for pull-down assays was detected by immunoblotting with anti-Munc18c. Protein kinases and phosphatases regulate signaling pathways initiated by the binding of agonists to endothelial cell receptors, and the phosphorylation status of vesicular trafficking proteins can regulate the secretion of granule contents [14]. Any specific role for serine/threonine phosphatases in the secretion of VWF from the WPBs of endothelial cells is, however, currently unknown. In this study, we demonstrate that inhibition of HUVEC PP2B activity by CsA activates the serine phosphorylation of a PP2B-associated vesicle trafficking protein, Munc18c. Concurrently, CsA inhibition of HUVEC PP2B promotes the secretion of hyperadhesive ULVWF multimeric strings. CsA has been reported previously to potentiate agonist (thrombin, histamine, and phorbol 12-myristate 13-acetate)-induced VWF release from HUVECs [21]. Other investigators have found increased quantities of VWF antigen in the supernatants of HUVECs exposed to CsA for 5–6 days [22]. It is not known whether this relatively long-term exposure of HUVECS to CsA caused cell damage and leakage of VWF, rather than CsA-triggered alterations in cell signaling and associated VWF secretion. We observed increased ULVWF/VWF secretion from HUVECs within minutes of cell exposure to CsA, FK506, or a PP2B AI peptide. These data suggest that inhibition of PP2B promotes ULVWF/VWF secretion from HUVECs, in the absence of any other endothelial cell-stimulating agonist. We also obtained compatible results from in vivo studies of mice treated with CsA. It is not yet known whether our findings are associated with the development of thrombotic microangiopathy in a subset of organ transplantation patients who are immunosuppressed with CsA or tacrolimus [23, 24]. In addition to inhibition of PP2B, CsA also inhibits: the mitochondrial permeability transition pore, a voltage sensitive and Ca2+-activated channel in the inner mitochondrial membrane; c-Jun N-terminal kinase; and P38 signaling in T cells [25, 26]. In order to determine whether inhibition of PP2B by CsA contributes to the promotion of ULVWF/VWF secretion, we utilized siRNAs to knock down the PP2B-Aα and PP2B-Aβ in HUVECs. Knockdown of PP2B-Aα or PP2B-Aβ resulted in a significant increase in ULVWF/VWF secretion by HUVECs as compared with control siRNA-treated cells (Fig. 4F). We conclude from these studies that CsA inhibition of PP2B activity helps to promote the secretion of ULVWF/VWF from endothelial cells. We investigated the molecular basis for the relationship between inhibition of PP2B and the release of ULVWF multimers from endothelial cells. The rapid secretion of ULVWF/VWF multimers from HUVECs upon CsA or AI peptide treatment precludes as the explanation increased VWF synthesis via transcriptional upregulation of the VWF gene. The absence of overt morphologic alterations or decreased metabolism in PP2B inhibitor-treated HUVECs indicates that increased release of ULVWF/VWF multimers also does not occur as a consequence of PP2B inhibitor-induced endothelial damage. On the other hand, a possible explanation for our observations is that PP2B inhibition by CsA or tacrolimus initiates (or participates in) ULVWF/VWF multimer secretion from endothelial cells via regulation of the phosphorylation status of a vesicle trafficking protein. We found that CsA increased the serine phosphorylation of Munc18c, a SNARE-associated protein involved in the phosphorylation-related release of granule contents from endothelial and other secretory cells [13, 27]. Consistent with the capacity of PP2B to regulate the serine phosphorylation of Munc18c, we also observed that PP2B-Aα interacts with Munc18c in vitro. We suggest that increased serine phosphorylation of Munc18c in the presence of CsA-induced (or tacrolimus-induced) PP2B inhibition may lead to decreased association of Munc18c with syntaxin-4. This may enable syntaxin-4 to associate with vesicle-associated membrane protein 3 (VAMP-3) on WPBs to form a SNARE complex that facilitates the secretion of WPB contents, including ULVWF/VWF multimers (Fig. 6). Proposed mechanism for the regulation of von Willebrand factor (VWF) secretion by cyclosporine A (CsA). The mechanism described is an adaption of a model used for other secretory system [35]. In endothelial cells, granule exocytosis is likely to occur through a core complex formed between the target (t) soluble N-ethylmaleimide-sensitive factor attachment protein receptors (SNAREs), called syntaxin-4 (Syx-4) and soluble N-ethylmaleimide-sensitive factor attachment protein 23 (SNAP-23), and a vesicle (v) SNARE called vesicle-associated membrane protein 3 (VAMP-3). Serine/threonine phosphatases may interact with SNARE-associated proteins such as Munc18c and possibly other vesicle trafficking proteins, directly or indirectly, to maintain a dephosphorylated state. In the absence of CsA, Munc18c interacts with Syx-4, and prevents VAMP-3 binding. CsA treatment results in the formation of the CsA–cyclophilin complex, which inhibits protein phosphatase 2B (PP2B) and triggers serine phosphorylation of Munc18c. This induces a conformational change in Munc18c such that Munc18c has decreased affinity for Syx-4. Munc18c dissociates from this complex, thereby enabling VAMP-3 on the Weibel–Palade body to interact with the Syn-4 and SNAP-23 complex to form a core complex that promotes the exocytosis of VWF from the endothelial cells. It is unlikely that endothelial cell vesicle trafficking proteins are the only targets of PP2B. Cytoskeletal contractile forces facilitate the secretion of granule contents by centralizing, or clustering, organelles. The phosphorylation status of cytoskeletal proteins may be involved in the regulation of granule content secretion. This latter possibility is compatible with observations that generic inhibitors of protein phosphatase type 1, type 2A and type 4 (such as by okadaic acid) enhance WPB clustering, which precedes the secretion of WPB contents by cAMP-stimulating secretagogs [11]. Studies using CsA and tacrolimus indicate that PP2B may also participate in the regulation of granular exocytosis in other cells, including neurons [28-31] pancreatic islet cells [32], mast cells [33], and T lymphocytes [34]. In summary, our studies demonstrate that inhibition of endothelial cell PP2B by CsA or tacrolimus promotes the secretion of ULVWF/VWF multimers. These observations are compatible with the clinically observed association between CsA treatment and increased plasma VWF levels in humans, as well as the association between prolonged exposure to CsA and thrombotic microangiopathy in a subset of exposed patients. This result implies that uninhibited PP2B is normally involved in suppressing ULVWF/VWF secretion. Chemical or pharmacologic manipulation of PP2B activity designed to control ULVWF/VWF secretion by endothelial cells may have potential clinical utility in the treatment of thrombotic disorders. The authors acknowledge V. Patel for his excellent technical help in experiments involving the phosphorylation of Munc18c and measuring phosphatase activity in HUVECs. This work was supported in part by grants HL081613 (K. V. Vijayan) and HL071895 (J. F. Dong) from the National Institute of Health (NIH) and the Mary Gibson Foundation (J. L. Moake). K. V. Vijayan and F. C. Gushiken are supported by National Scientist Development grants 0435017N and 0630180N, respectively, from the American Heart Association. S. Pradhan is supported in part by NIH grant T-32HL072754. The other authors state that they have no conflict of interest.