Reducing agents induce thrombomodulin shedding in human endothelial cells

Thrombosis Research 126 (2010) e88–e93

Contents lists available at ScienceDirect

Thrombosis Research j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / t h r o m r e s

Regular Article

Reducing agents induce thrombomodulin shedding in human endothelial cells Mario Menschikowski ⁎, Albert Hagelgans, Graeme Eisenhofer, Oliver Tiebel, Gabriele Siegert Institute of Clinical Chemistry and Laboratory Medicine, Technical University of Dresden, Medical Faculty "Carl Gustav Carus", Dresden, Germany

a r t i c l e

i n f o

Article history: Received 25 February 2010 Received in revised form 26 April 2010 Accepted 6 May 2010 Keywords: Thrombomodulin Shedding Endothelial cells Thiols

a b s t r a c t The level of thrombomodulin (TM) on cell surfaces reflects its biosynthesis, intracellular turnover, proteolytic cleavage, and release in soluble form (sTM). In the present study we examined the mechanisms mediating and regulating sTM release. Inducers of endothelial protein C receptor (EPCR) shedding, such as proinflammatory cytokines, phorbol ester, and ionomycin did not affect sTM release from human umbilical endothelial cells (HUVECs). In contrast, several natural and synthetic reducing compounds (i.e., glutathione, dihydrolipoic acid, homocysteine, N-acetyl-L-cysteine, dithiothreitol, and non-thiol cell-impermeable reductant, tris-(2-carboxyethyl)phosphine), but not oxidized glutathione or α-lipoic acid effectively upregulated the release of sTM in endothelial cells. In addition, the direct activator of metalloproteases, 4-aminophenylmercuric acetate (APMA), was an effective inducer of TM shedding. Considerable inhibition of protein C activation was found with APMA, which is consistent with the effects of this agent on TM shedding. In addition to metalloproteases, serine proteases were shown by pharmacological inhibition studies to be involved in a similar degree in basal sTM release; however, serine proteases seem preferentially to be involved in thiol-induced TM proteolytic processing. From comparisons of non-thiol containing synthetic substrate with human recombinant TM it was demonstrated that disulfide bonds within TM are most likely modified by thiols making TM more susceptible to serine protease-mediated cleavage. In summary, the study shows that the extracellular redox state plays a crucial role in the regulation of TM shedding in HUVECs thereby offering new strategies to interfere with diminished activation of protein C during inflammatory diseases. © 2010 Elsevier Ltd. All rights reserved.

Introduction Thrombomodulin (TM), a cell-surface glycoprotein, is highly expressed in endothelial cells and numerous other cell types such as keratinocytes, osteoblasts, and mononuclear phagocytes [1,2]. Both cell-associated TM and the cleaved soluble form of TM (sTM) promote specific effects on coagulation and fibrinolysis through generation of activated protein C (aPC) and thrombin-mediated activation of fibrinolysis inhibitor [3–5]. Other activities of TM include antiAbbreviations: ADAM, a disintegrin and metalloprotease; α-LA, α-lipoic acid; aPC, activated protein C; APMA, 4-aminophenylmercuric acetate; BSA, bovine serum albumin; CM, cytokine mixture; DHLA, dihydrolipoic acid; DMSO, dimethyl sulfoxide; DTT, dithiothreitol; EPCR, endothelial protein C receptor; FCS, fetal calf serum; GSH, reduced glutathione; GSSG, oxidized glutathione; Hcys, homocysteine; HUVECs, human umbilical vein endothelial cells; IFN-γ, interferon-γ; IL-1β, interleukin-1β; MAPK, mitogen-activated protein kinase; MMP, matrix metalloprotease; NAC, N-acetylL-cysteine; PBS, phosphate-buffered saline; PC, protein C; PKC, protein kinase C; PMA, phorbol 12-myristate 13-acetate; rhTM, recombinant human thrombomodulin; sEPCR, soluble endothelial protein C receptor; TCEP, tris(2-carboxyethyl)phosphine hydrochloride; TM, thrombomodulin; sTM, soluble thrombomodulin; TNF-α, tumor necrosis factor-α. ⁎ Corresponding author. Fetscherstrasse 74, D-01307 Dresden, Germany. Tel.: + 49 351 458 2634; fax: + 49 351 458 4332. E-mail address: [email protected] (M. Menschikowski). 0049-3848/$ – see front matter © 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.thromres.2010.05.006

inflammatory and cytoprotective properties, influences on receptor mediated cell signaling, and anti-metastatic effects in malignant tumor progression [6,7]. Increased plasma levels of sTM have been described in patients with disseminated intravascular coagulation syndrome, pulmonary thromboembolism, acute respiratory distress syndrome, and chronic renal and hepatic failure [8–10]. Furthermore, elevated levels of sTM were observed in several autoimmune disorders, including systemic lupus erythematosis, Churg–Strauss syndrome, and Wegener's granulomatosis [11,12]. For this reason, the loss of TM from endothelial cell surfaces is thought to contribute to increased risk of thrombosis associated with inflammation. Induction of sTM release has also been observed in vitro after exposure of endothelial cells to hydrogen peroxide, homocysteine (Hcys), prostaglandin A2, lipopolysaccharide, and neutrophil derived proteases [13–18]. The cleavage of TM and release of its soluble fragments is thought to be mediated by neutrophil-derived proteases and accompanied by damage to cell membranes [13,19]. It was also reported that metalloproteases are implicated in TM ectodomain cleavage mediating the release of the lectin-like domain [20–22]. Apart from the above, little is known about the mechanisms underlying the induction of TM shedding or the types of proteases involved in TM cleavage. In view of the convincing evidence that both membrane-associated and soluble forms of TM play a crucial role in the endothelial protein C

M. Menschikowski et al. / Thrombosis Research 126 (2010) e88–e93

receptor (EPCR) associated protein C pathway, the present study was carried out to assess the mechanisms regulating release of sTM in human endothelial cells and the role of redox status in this process.

e89

Materials and Methods

anti-mouse IgG FITC-conjugated antibody produced in goat, which were used at final dilution of 1:400 of the commercially supplied stock solution. Thereafter, cells were again washed twice, fixed with 4% paraformaldehyde in PBS, and analyzed on an EPICS XL flow cytometer (Beckman Coulter GmbH, Krefeld, Germany).

Reagents and antibodies

Cell-free proteolysis of recombinant human thrombomodulin

Human plasma-derived aPC (10.8-11.6 Units/mg) was from Hematological Technologies Inc. (Cell Systems, Biotechnologie Vertrieb GmbH, St. Katharinen, Germany) and human protein C was from Enzyme Research Laboratories Ltd (South Bend, IN, USA). Recombinant human TM (rhTM) and Immubind® Thrombomodulin ELISA kits were purchased from American Diagnostica GmbH (Pfungstadt, Germany). Monoclonal mouse antibodies to the modules EGF5 (clone PBS-01) and EGF2 (clone PBS-02) of epidermal growth factor (EGF)-like domains of TM, and goat polyclonal FITC-conjugated antibody to mouse IgG were purchased from Abcam (Cambridge, UK). PMA, GM6001, TAPI-0, and ionomycin were from Calbiochem (Schwalbach, Germany). Pefabloc SC, bovine serum albumin (BSA), reduced (GSH) and oxidized glutathione (GSSG), α-lipoic acid (α-LA), dihydrolipoic acid (DHLA), N-acetyl-L-cysteine (NAC), 4-aminophenylmercuric acetate (APMA), dithiothreitol (DTT), Tris(2-carboxyethyl)phosphine hydrochloride (TCEP), trypsin from bovine pancreas and elastase from hog pancreas were purchased from Sigma-Aldrich (Deisenhofen, Germany). Recombinant human IFN-γ, IL-1β, and TNF-α were from Roche Diagnostics GmbH (Mannheim, Germany). PMA, TAPI-0, ionomycin, APMA, and GM6001 were dissolved in dimethyl sulfoxide (DMSO). The final concentrations of DMSO were 0.3% or less, and controls using DMSO alone were run in all cases. Other agents were used as aqueous solutions.

To test the effect of reducing agents on proteolytic degradation of TM, the rhTM was dissolved in PBS at a final concentration of 50 ng/ml and incubated at 37o C in the presence of 10 µg/ml trypsin or 10 µg/ml elastase. In a separate panel, the incubation medium was supplemented with 0.5 mM DTT. After defined incubation time periods, the proteolysis was stopped by addition of a protease inhibitor cocktail (Sigma-Aldrich). The intensity of proteolysis was assessed by measuring the disappearance of immunochemical detectable rhTM in the incubation medium using Immubind® Thrombomodulin ELISA kit (American Diagnostica GmbH, Pfungstadt, Germany). The results were expressed as relative changes of rhTM levels as values taking the value at 0 min as 100%. The activity of elastase towards synthetic chromogenic substrate, N-succinyl-(Ala)3-p-nitroanilide (Sigma-Aldrich) was assayed using the methods recommended by the manufacturer. The reactive mixture containing 10 µg/ml of elastase and 1 mM chromogenic substrate was incubated at 37 °C in buffer solution containing 0.1 M HEPES, 0.5 M NaCl, and 10% DMSO. The reaction was followed by measuring the release of 4-nitroanilide every 5 min at 405 nm on Victor3 1420 Multilabel Counter reader (PerkinElmer LAS GmbH, Rodgau Jügesheim, Germany).

Cell culture

To test the influence of TM shedding on aPC generation, HUVECs were incubated in the presence of protein C and thrombin according to the method described by Grey et al. [23]. Briefly, cells were cultured in 24-well plates to confluence, treated for 2 h with APMA as modulator of TM shedding, and thereafter washed in buffer A containing 20 mM Tris-HCI (pH 7.5), 2 mM CaCl2, 150 mM NaCl, and 0.1% BSA. Washed cells were incubated for a further 1 h-period in the presence of human protein C (80 nM), thrombin (37.5 nM), and buffer A in a final volume of 200 µl/well at 37 °C and 5% CO2. Microplates were then centrifuged at 160 x g for 5 min, 150 µl of supernatants were transferred into 96-well plates and assayed for the generation of aPC using 0.8 mM chromogenic substrate S-2366. The extinction of reaction product was measured at 405 nm on Victor3 1420 Multilabel Counter reader (PerkinElmer LAS GmbH, Rodgau Jügesheim, Germany). To prevent nonspecific cleavage of substrate by thrombin, hirudin (10 antithrombin units per well) was added to each supernatant for 5 min at room temperature before testing for aPC activity. The amounts of generated aPC were calculated using aPC standards and normalized to cell protein content.

Human umbilical vein endothelial cells (HUVECs) were purchased from Promocell (Heidelberg, Germany) and cultivated according to the manufacturers’ instructions. Briefly, the cells were maintained in endothelial cell growth medium with supplement mix containing 2% fetal calf serum (FCS) and 0.4% endothelial cell growth supplement. For all experiments, exponentially growing subconfluent cells were used at passages 5 to 8. ELISA based quantitative determination of sTM and sEPCR The amounts of sTM and soluble endothelial protein C receptor (sEPCR) released by endothelial cells were determined using Asserachrom sTM and sEPCR ELISA kits (Diagnostica Stago, Asnieres, France) according to the manufacturer's instructions. For this purpose, cells were grown to confluence in 96-well microplates in complete medium with FCS and supplements. Cells were then further incubated for 1 h with fresh complete medium containing inducers or inhibitors of TM and EPCR shedding. At the end of incubation, the cell medium was removed, centrifuged at 800 g for 10 min to remove the cell debris, and used for analysis to determine the levels of sTM and sEPCR released by cells. Total cell protein was determined using Bicinchoninic Acid assay kit with BSA as internal standard (Sigma-Aldrich, Deisenhofen, Germany).

Protein C activation assay

Data analysis Data were analyzed by one-way analysis of variance coupled with Dunnet's post hoc test to compare each experimental group with a nominated control group using SPSS 14.0 software. Differences were considered significant at P b 0.05.

Flow cytometry Results Cells after incubation were trypsinized, washed in PBS and resuspended at 1·x 106 cells/ml in FACS buffer (PBS, pH 7.4, supplemented with 1% BSA and 0.1% sodium azide) and incubated with anti-TM monoclonal antibodies (clones PBS-01 and PBS-02) for 30 min at 4 °C. Cells were then washed twice with FACS-buffer and incubated under light protected conditions for 30 min at 4 °C with

Pharmacological regulation of sTM release in HUVEC Exposure of HUVEC for 1 h to pro-inflammatory cytokines, IL-1β, TNF-α, and IFN-γ, either individually (data not shown) or as mixture (CM), led to a considerable inhibition of basal sTM release (Fig. 1A).

e90

M. Menschikowski et al. / Thrombosis Research 126 (2010) e88–e93

Fig. 1. Effects of pharmacological inductors, thiols and non-thiol reducing agent, TCEP, on sTM release in HUVECs. (A) Cells were treated with a cytokine mix (CM) consisting of IL-1β, TNF-α, and IFN-γ at final concentrations of 6.25 ng/ml, respectively; 100 ng/ml PMA; 5 µM ionomycin (Ionom); 2 mM H2O2, 5 mM reduced (GSH); 5 mM oxidized glutathione (GSSG); 0.5 mM dihydrolipoic acid (DHLA), or 5 mM α-lipoic acid (α-LA) for 1 h. (B) Concentration-dependent effects of dithiothreitol (DTT) on the release of sTM and cellular concentrations of TM and effects of N-acetyl-L-cysteine (NAC) and tris-(2-carboxyethyl)-phosphine hydrochloride (TCEP) on sTM release from HUVECs after exposure for 1 h. Levels of released sTM into cell culture medium and concentrations of TM in cell lysates are expressed as percent changes relative to controls without additions. Results are presented as means ± SDs of analyses in quadruplicate and are representative of three independent experiments. * p b 0.05 versus control.

Simultaneously, TNF-α and IL-1β diminished TM levels in cell lysates of HUVECs (data not shown). Despite their potent actions on sEPCR release [24], PMA-mediated activation of protein kinase C (PKC) and ionomycin-mediated increases in cellular Ca2+ influx did not significantly influence levels of sTM. In contrast, hydrogen peroxide markedly increased the release of sTM in HUVEC (Fig. 1A). Previously we described that induction of EPCR shedding in endothelial cells was effectively blocked by a number of reduced thiol-group containing non-protein agents [24]. Therefore, we assessed whether TM shedding is also susceptible to these agents. Dihydrolipoic acid (DHLA) and reduced glutathione (GSH) strongly stimulated sTM release. In contrast, oxidised forms of α-lipoic acid (α-LA) and glutathione (GSSG) did not significantly influence release of sTM (Fig. 1A). Additionally, N-acetyl-cysteine (NAC) and dithiothreitol (DTT) distinctly stimulated the release of sTM into cell culture medium in a concentration-dependent manner (Fig. 1B). Concomitantly, decreased cellular concentrations of TM were observed after exposure of cells to DTT. Finally, TCEP as non-thiol cell-impermeable reductant induced the sTM release at increasing concentrations (Fig. 1B). A concentration-dependent biphasic effect on TM release was observed for Hcys (Fig. 2). At concentrations of 10-200 µM, Hcys decreased the release of sTM, but increase sEPCR release. However, at higher concentrations of 300-1000 µM, Hcys had distinct stimulatory effects on sTM release and suppressive effects on sEPCR release (Fig. 2).

Fig. 2. Effects of homocysteine on sTM and sEPCR releases in HUVECs. Cells were treated for 1 h with different concentrations of homocysteine. Levels of sTM and sEPCR released into cell medium are expressed as percent changes relative to controls without Hcys. Results are presented as means ± SDs of analyses in triplicate and are representative of three independent experiments.

shedding is exerted by metalloprotease inhibitors, TAPI-0 and GM6001, whereas pefabloc SC was almost ineffective (Fig. 3B). Effects of metalloprotease activation on thrombomodulin shedding In accordance with the effects of metalloprotease inhibitors on basal TM shedding, the direct activation of metalloproteases through 4-aminophenylmercuric acetate (APMA) dose-dependently increased the sTM release after exposure of HUVECs (Fig. 4A). This APMAmediated effect on TM shedding was confirmed by flow cytometry analysis using monoclonal antibodies, clones PBS-02 and PBS-01, against EGF2 and EGF5 repeats of TM. Here, a significant loss of both antigens was observed after a 1 h incubation of HUVECs with 90 µM APMA (Fig. 4B) indicating the possible implication of metalloproteases in TM shedding. Because protein C activation in endothelial cells is dependent on TM and EPCR exposition on cell surfaces [3], we analyzed the effect of APMA on the generation of aPC. According to the results APMA strongly inhibited aPC generation and metalloprotease inhibitor, TAPI-0, nearly completely blocked the APMA-mediated inhibition of aPC generation (Fig. 4C).

Effects of metalloprotease and serine protease inhibitors on sTM release in comparison to those on sEPCR release TAPI-0 and GM6001, broad spectrum metalloprotease inhibitors (against a disintegrin and metalloproteases, ADAMs, and matrix metalloproteases, MMPs), considerably reduced the basal sTM release in HUVECs (Fig. 3A). A similar inhibition was observed using the broad range serine protease inhibitor, pefabloc SC. In contrast, TM shedding induced by treatment of HUVECs with 0.5 mM DTT was strongly suppressed by pefabloc SC, whereas TAPI-0 and GM6001 had significant but minimal effects (Fig. 3B). When compared with EPCR shedding induced by 100 ng/ml PMA, a distinct difference in efficiency of protease inhibitors appeared. In this case, powerful inhibition of

Fig. 3. Effects of protease inhibitors on sTM and sEPCR releases in HUVECs. Cells were incubated with 30 µM TAPI-0 (metalloprotease inhibitor), 20 µM GM6001 (MMP inhibitor), or 1 mM pefabloc SC (serine protease inhibitor) alone (A) and together with 0.5 mM DTT in case of sTM release or 100 ng/ml PMA in case of sEPCR release, respectively (B). After exposure for 1 h, the levels of sTM (A and B) and sEPCR (B) were determined. Bars show levels of sTM and sEPCR as percent changes relative to controls without additions. Results are presented as means ± SDs of analyses in triplicate and are representative of three independent experiments. * p b 0.05 versus controls without protease inhibitors.

M. Menschikowski et al. / Thrombosis Research 126 (2010) e88–e93

e91

Discussion

Fig. 4. Effects of APMA on sTM release (A, B) and protein C activation (C) in HUVECs. (A) Cells were treated for 1 h with increasing concentrations of APMA. Levels of sTM released into medium are expressed as percent changes relative to control without APMA. (B) Flow cytometry analysis of TM exposition on cell surfaces of HUVECs using monoclonal antibodies against EGF-5 and EGF-2 domains before and after treatment with 90 µM APMA. Bars show the mean fluorescence intensity (MFI). (C) After exposure of cells to 30 µM APMA alone and together with metalloprotease inhibitor, TAPI-0 (30 µM) for 2 h, cells were washed and incubated with human protein C (80 nM) and thrombin (37.5 nM) for 1 h. Cell culture supernatants were assayed for aPC generation using chromogenic substrate S-2366. Bars show levels of aPC relative to control (without additions), which was assigned a value of 100%. Results are presented as means ± SDs of four samples of each group in one experiment representative of three performed analyses with similar results. * p b 0.05 versus controls without APMA (B) and with APMA but without TAPI-0 (C).

Effects of thiol reduction on the proteolysis of recombinant human thrombomodulin (rhTM) in a cell-free system In order to further examine the mechanism underling the stimulatory effects of thiols on TM shedding, the proteolytic cleavage of rhTM by serine proteases was studied. Trypsin and elastase at concentrations of 10 µg/ml rapidly decreased the level of rhTM in time-dependent manner (Fig. 5A and B). In the presence of 0.5 mM DTT, the proteolytic cleavage of rhTM distinctly accelerated. To asses whether the effect of thiols is based on modification of substrates or enzyme activities, Nsuccinyl-(Ala)3-p-NA as non-thiol containing synthetic substrate was applied. Here, DTT exerted no stimulating effects on elastase-mediated proteolysis (Fig. 5C) suggesting that the stimulatory effect of DTT on elastase action dependents on distinct sites of rhTM molecule.

Fig. 5. Effects of DTT on proteolytic cleavage of rhTM by trypsin and elastase. In a cellfree system, 50 ng/ml of rhTM were incubated with 10 µg/ml trypsin (A) and 10 µg/ml elastase (B) alone and together with 0.5 mM DTT at 37 °C for defined time periods. After incubation, proteolysis was stopped by addition of protease inhibitor cocktail and the intensity of proteolysis was assessed by disappearance of rhTM estimated using Immubind® thrombomodulin ELISA kit. Results are expressed as relative changes in rhTM levels taking the value at 0 min in controls without proteases as 100%. (C) Instead of rhTM, 1 mM chromogenic substrate, N-succinyl-(Ala)3-p-nitroanilide, were incubated with 10 µg/ml of elastase alone and together with 0.5 mM DTT at 37 °C for defined time periods. Release of 4-nitroanilide was measured at 405 nm and results are expressed as changes of extinction.

Previously, we demonstrated that the shedding of EPCR in HUVEC is induced by pro-inflammatory cytokines; by pharmacological agents such as PMA, ionomycin, anisomycin, and thiol oxidants and alkylators; and that this process is inhibited by reduced non-protein thiol compounds [24]. In the present study we show that the shedding of TM is regulated in different manner compared with regulation of EPCR shedding. First, pro-inflammatory cytokines, alone or as a mixture, inhibit the shedding of TM, whereas the shedding of EPCR is induced by IL-1β and TNF-α [24]. In addition to reduced levels of released sTM, TNF-α and IL-1β also diminish TM levels in cell lysates of HUVECs. This observation is consistent with previous studies showing a time- and dose-dependent reduction of TM gene expression after exposure of endothelial cells to TNF-α and IL-1β [25–29]. Second, the release of TM is insensitive to PMA and ionomycin, agents that were mostly effective in respect to the induction of EPCR release [24,30]. Third, GSH, DHLA, and NAC effectively stimulated release of sTM into the cell culture medium of HUVEC whereas the same agents effectively blocked release of sEPCR. The different actions of nonprotein thiols on TM and EPCR shedding suggest that different proteases are involved in the shedding of both receptors and/or that the reduced thiols seem to interact with the receptors in a diverse manner. Whether the different regulation of TM and EPCR shedding may have pathophysiological relevance is still unknown but may represent a protective mechanism preventing a complete loss of both EPCR and TM connected with an inadequate aPC generation under proinflammatory conditions. Circulating subspecies of TM released from cell surfaces are normally detectable in plasma and urine indicating that TM processing occurs under physiological conditions and involves a variety of cleavage sites [31–33]. Our findings that sTM release in HUVECs under basal conditions was inhibited by TAPI-0 and GM6001, both broad-spectrum metalloprotease inhibitors, underscore the role of these proteases in TM ectodomain cleavage in HUVEC. The observation that a marked increase in sTM release by the mercurial compound, APMA, which is frequently used to activate matrix metalloproteases in vitro [34], provides additional supportive evidence for a contribution of metalloproteases to TM cleavage. In parallel, treatment of HUVECs with APMA led to a reduced exposition of EGF2 and EGF5 modules of TM on the cell surface of HUVECs. This shows that metalloproteases can attack TM not only in regions of the N-terminal lectin-like domain, as recently reported [20], but also in the vicinity of membranes where the epitope for anti-EGF5 antibody is localized. The EGF-like domain consisting of six EGF-like repeats confers the TM molecule activities both in cell proliferation and coagulation/ fibrinolysis [35,36], whereas the lectin-like domain of TM is effective in protection against sepsis and exerts distinct antiinflammatory properties [37,38]. The mechanisms responsible for differential proteolytic release of these TM domains and realization of related activities, however, remain to be elucidated. For thiol-induced sTM release, however, the role of metalloproteases seems to be less distinct suggesting that, in addition to metalloproteases, other proteases are involved in the process of TM shedding. Our finding that the broad range serine protease inhibitor, pefabloc SC, produced strong inhibitory effects on sTM releases indicate that serine proteases play a major role in thiol-induced TM proteolytic processing. This conclusion is supported by our additional data that treatment with exogenous serine proteases, such as trypsin and elastase, results in a rapid proteolytic cleavage of rhTM and agrees with previous reports [39,40]. In order to further examine the mechanism underlying the stimulatory effects of thiols on TM shedding, we studied the effect of DTT on elastase-mediated proteolytic cleavage of the non-thiol containing synthetic substrate, N-succinyl-(Ala)3-p-NA. DTT failed to increase the proteolytic cleavage of synthetic substrate by elastase suggesting that reducing compounds modify intra-molecular disulfide

e92

M. Menschikowski et al. / Thrombosis Research 126 (2010) e88–e93

bridges within TM ectodomains. This may be connected with conformational changes allowing a more efficient access of the relevant proteases to cleavage sites of TM. Such a mechanism of redox regulation has been proposed previously for shedding of the complement receptor upon NAC and GSH treatment [41]. Protein unfolding induced by oxidation or reduction of proteins can increase the exposure of peptide bonds to reactions with proteases by 10- to 100-fold [42,43]. Oxidative processes play important roles in inflammatory diseases [44,45]. For TM, the methionine-388 residue linking the fourth and fifth EGF-like domains is an important target for redox-regulation of proteolytic cleavage. Both domains are essential for the anticoagulant function of TM [4,6]. Oxidation of methionine-388 strongly alters both the structure and functional activity of TM and represents a mechanism by which the anticoagulant potential of endothelium is decreased, especially after the damage of endothelial cells [45]. Such oxidative modification of the methionine-388 residue can be forced through reactive oxygen species generated by activated neutrophils during inflammatory diseases such sepsis and septic shock as well as diabetes, thyrotoxicosis, and arthritis [45,46]. In diabetes, oxidative stress is caused by the generation of hydrogen peroxide and superoxide due to increased oxidation of glucose, NADPH-oxidase activity and metabolic fluxes through the mitochondrial respiratory chain (for review [47]). Therefore, a general link between oxidative stress, oxidation of the methionine-388 residue in TM and increased tendency to coagulation during inflammation has been suggested [45,46]. Another mechanism for redox-regulation of TM processing may involve a thiol-mediated reduction of disulfide bonds in the EGFlike domains of TM. In this way, important structural properties of these domains are irreversibly disrupted and may result in new cleavage sites for proteases [16]. Together, our data indicate that a series of different proteases and redox-mechanisms are involved in TM shedding; in contrast, only a minor selection of proteases, such as ADAM17, seem to play a pivotal role in EPCR shedding [48]. This finding explains the presence of multiple types of TM fragments in cell culture incubates, serum and urine samples compared to the single sEPCR fragment of 43 kDa in human serum and plasma [8,32,33,49]. To explain the observed biphasic properties of Hcys, it should be emphasized that this compound, in contrast to other reduced thiols such DTT, is able to provoke oxidative stress and activate both MAPK signaling [50–52] and matrix metalloproteases [53]. All these effects may contribute to the intensified release of sEPCR induced by physiological and supra-physiological concentrations of Hcys, where the level of sTM decreased or remained rather unaffected at least in vitro. Whether the inductive effect of Hcys on sEPCR release has a biological relevance and whether herein an alternative mechanism may exist explaining the thrombotic properties and risk potency of increasing serum concentrations of Hcys for coronary heart diseases (for review [54]), requires further investigation. In conclusion, this study shows that the shedding of TM in HUVEC is differentially regulated in comparison to that of EPCR. These data elucidating the mechanisms influencing TM release from endothelial cells have implications for development of novel therapeutic strategies to counter decreased aPC generation during inflammation or associated with malignancies. Since the shedding of TM and/ or EPCR appears critical to decreased aPC generation, one opportunity to interfere with the process involves the use of specific protease inhibitors, such TAPI-0, that can target the processes of TM and EPCR shedding and thereby reverse reductions in protein C activation.

Conflict of interest statement All authors disclose any financial and personal relationships with other people or organisations that can inappropriately influence (bias) their work in relation to this article.

Acknowledgments The authors are grateful to Mrs Margot Vogel for her expert technical assistance. This work is supported by research grant from GTH (Gesellschaft für Thrombose- und Hämostaseforschung e.V.).

References [1] Fink LM, Eidt JF, Johnson K, Cook JM, Cook CD, Morser J, et al. Thrombomodulin activity and localization. Int J Dev Biol 1993;37:221–6. [2] Boffa MC, Karmochkine M. Thrombomodulin: an overview and potential implications in vascular disorders. Lupus 1998;7:120–5. [3] Esmon CT. The protein C pathway. Chest 2003;24:26S–32S. [4] Esmon CT, Esmon NL, Harris KW. Complex formation between thrombin and thrombomodulin inhibits both thrombin-catalyzed fibrin formation and factor V activation. J Biol Chem 1982;257:7944–7. [5] Mosnier LO, Bouma BN. Regulation of fibrinolysis by thrombin activatable fibrinolysis inhibitor, an unstable carboxypeptidase B that unites the pathways of coagulation and fibrinolysis. Arterioscler Thromb Vasc Biol 2006;26:2445–53. [6] Van de Wouwer M, Conway EM. Novel functions of thrombomodulin in inflammation. Crit Care Med 2004;32:S254–61. [7] Hanly AM, Winter DC. The role of thrombomodulin in malignancy. Semin Thromb Hemost 2007;33:673–9. [8] Takano S, Kimura S, Ohdama S, Aoki N. Plasma thrombomodulin in health and diseases. Blood 1990;76:2024–9. [9] Matsuyama T, Tokuda M, Izumi Y. Significance of thrombomodulin release from gingival epithelial cells in periodontitis patients. J Periodontal Res 2008;43: 379–85. [10] Ware LB, Fang X, Matthay MA. Protein C and thrombomodulin in human acute lung injury. Am J Physiol Lung Cell Mol Physiol 2003;285:L514–21. [11] Boehme MW, Schmitt WH, Youinou P, Stremmel WR, Gross WL. Clinical relevance of elevated serum thrombomodulin and soluble E-selectin in patients with Wegener's granulomatosis and other systemic vasculitides. Am J Med 1996;101: 387–94. [12] Boehme MW, Raeth U, Galle PR, Stremmel W, Scherbaum WA. Serum thrombomodulin — a reliable marker of disease activity in systemic lupus erythematosus (SLE): advantage over established serological parameters to indicate disease activity. Clin Exp Immunol 2000;119:189–95. [13] Boehme MW, Deng Y, Raeth U, Bierhaus A, Ziegler R, Stremmel W, et al. Release of thrombomodulin from endothelial cells by concerted action of TNF-α and neutrophils: in vivo and in vitro studies. Immunology 1996;87:134–40. [14] Archipoff G, Beretz A, Freyssinet JM, Klein-Soyer C, Brisson C, Cazenave JP. Heterogeneous regulation of constitutive thrombomodulin or inducible tissuefactor activities on the surface of human saphenous-vein endothelial cells in culture following stimulation by interleukin-1, tumour necrosis factor, thrombin or phorbol ester. Biochem J 1991;273:679–84. [15] Dittman WA, Kumada T, Sadler JE, Majerus PW. The structure and function of mouse thrombomodulin. Phorbol myristate acetate stimulates degradation and synthesis of thrombomodulin without affecting mRNA levels in hemangioma cells. J Biol Chem 1988;263:15815–22. [16] Lentz SR, Sadler JE. Inhibition of thrombomodulin surface expression and protein C activation by the thrombogenic agent homocysteine. J Clin Invest 1991;88: 1906–14. [17] Sawada K, Yamamoto H, Yago H, Suehiro S. A simple assay to detect endothelial cell injury; measurement of released thrombomodulin from cells. Exp Mol Pathol 1992;57:116–23. [18] Abe H, Okajima K, Okabe H, Takatsuki K, Binder BR. Granulocyte proteases and hydrogen peroxide synergistically inactivate thrombomodulin of endothelial cells in vitro. J Lab Clin Med 1994;123:874–81. [19] Boehme MW, Galle P, Stremmel W. Kinetics of thrombomodulin release and endothelial cell injury by neutrophil-derived proteases and oxygen radicals. Immunology 2002;107:340–9. [20] Wang L, Bastarache JA, Wickersham N, Fang X, Matthay MA, Ware LB. Novel role of the human alveolar epithelium in regulating intra-alveolar coagulation. Am J Respir Cell Mol Biol 2007;36:497–503. [21] Van den Berg CW, Gonçalves-de-Andrade RM, Magnoli FC, Tambourgi DV. Loxosceles spider venom induces the release of thrombomodulin and endothelial protein C receptor: implications for the pathogenesis of intravascular coagulation as observed in loxoscelism. J Thromb Haemost 2007;5:989–95. [22] Wu HL, Lin CI, Huang YL, Chen PS, Kuo CH, Chen MS, et al. Lysophosphatidic acid stimulates thrombomodulin lectin-like domain shedding in human endothelial cells. Biochem Biophys Res Commun 2008;367:162–8. [23] Grey ST, Hancock WW. A physiologic anti-inflammatory pathway based on thrombomodulin expression and generation of activated protein C by human mononuclear phagocytes. J Immunol 1996;156:2256–63. [24] Menschikowski M, Hagelgans A, Eisenhofer G, Siegert G. Regulation of endothelial protein C receptor shedding by cytokines is mediated through differential activation of MAP kinase signaling pathways. Exp Cell Res 2009;315:2673–82. [25] Conway EM, Rosenberg RD. Tumor necrosis factor suppresses transcription of the thrombomodulin gene in endothelial cells. Mol Cell Biol 1988;8:5588–92. [26] Lentz SR, Tsiang M, Sadler JE. Regulation of thrombomodulin by tumor necrosis factor-α: comparison of transcriptional and posttranscriptional mechanisms. Blood 1991;77:542–50.

M. Menschikowski et al. / Thrombosis Research 126 (2010) e88–e93 [27] Ishii H, Horie S, Kizaki K, Kazama M. Retinoic acid counteracts both the downregulation of thrombomodulin and the induction of tissue factor in cultured human endothelial cells exposed to tumor necrosis factor. Blood 1992;80:2556–62. [28] Nan B, Lin P, Lumsden AB, Yao Q, Chen C. Effects of TNF-α and curcumin on the expression of thrombomodulin and endothelial protein C receptor in human endothelial cells. Thromb Res 2005;115:417–26. [29] Yeh CH, Hung LY, Hsu C, Le SY, Lee PT, Liao WL, et al. RNA-binding protein HuR interacts with thrombomodulin 5'untranslated region and represses internal ribosome entry site-mediated translation under IL-1β treatment. Mol Biol Cell 2008;19:3812–22. [30] Xu J, Qu D, Esmon NL, Esmon CT. Metalloproteolytic release of endothelial cell protein C receptor. J Biol Chem 2000;275:6038–44. [31] Jackson DE, Tetaz TJ, Salem HH, Mitchell CA. Purification and characterization of two forms of soluble thrombomodulin from human urine. Eur J Biochem 1994;221:1079–87. [32] Uehara S, Gotoh K, Handa H. Separation and characterization of the molecular species of thrombomodulin in the plasma of diabetic patients. Thromb Res 2001;104:325–32. [33] Nakano M, Furutani M, Hiraishi S, Ishii H. Characterization of soluble thrombomodulin fragments in human urine. Thromb Haemost 1998;79:331–7. [34] Merlos-Suárez A, Ruiz-Paz S, Baselga J, Arribas J. Metalloprotease-dependent protransforming growth factor-α ectodomain shedding in the absence of tumor necrosis factor-α-converting enzyme. J Biol Chem 2001;276:48510–7. [35] Koutsi A, Papapanagiotou A, Papavassiliou AG. Thrombomodulin: from haemostasis to inflammation and tumourigenesis. Int J Biochem Cell Biol 2008;40: 1669–73. [36] Hamada H, Ishii H, Sakyo K, Horie S, Nishiki K, Kazama M. The epidermal growth factor–like domain of recombinant human thrombomodulin exhibits mitogenic activity for Swiss 3T3 cells. Blood 1995;86:225–33. [37] Conway EM, Van de Wouwer M, Pollefeyt S, Jurk K, Van Aken H, De Vriese A, et al. The lectin-like domain of thrombomodulin confers protection from neutrophilmediated tissue damage by suppressing adhesion molecule expression via nuclear factor κB and mitogen-activated protein kinase pathways. J Exp Med 2002;196: 565–77. [38] Shi CS, Shi GY, Hsiao SM, Kao YC, Kuo KL, Ma CY, et al. Lectin-like domain of thrombomodulin binds to its specific ligand Lewis Y antigen and neutralizes lipopolysaccharide-induced inflammatory response. Blood 2008;112:3661–70. [39] Kurosawa S, Galvin JB, Esmon NL, Esmon CT. Proteolytic formation and properties of functional domains of thrombomodulin. J Biol Chem 1987;262:2206–12. [40] Abe H, Okajima K, Okabe H, Takatsuki K, Binder BR. Granulocyte proteases and hydrogen peroxide synergistically inactivate thrombomodulin of endothelial cells in vitro. J Lab Clin Med 1994;123:874–81.

e93

[41] Aichem A, Masilamani M, Illges H. Redox regulation of CD21 shedding involves signaling via PKC and indicates the formation of a juxtamembrane stalk. J Cell Sci 2006;119:2892–902. [42] Lockwood TD. Redox control of protein degradation. Antioxid Redox Signal 2000;2:851–78. [43] Martínez-Sánchez G, Giuliani A, Pérez-Davison G, León-Fernández OS. Oxidized proteins and their contribution to redox homeostasis. Redox Rep 2005;10:175–85. [44] Halliwell B. Oxidants and human disease: some new concepts. FASEB J 1987;1: 358–64. [45] Glaser CB, Morser J, Clarke JH, Blasko E, McLean K, Kuhn I, et al. Oxidation of a specific methionine in thrombomodulin by activated neutrophil products blocks cofactor activity. A potential rapid mechanism for modulation of coagulation. J Clin Invest 1992;90:2565–73. [46] Stites WE, Froude JW. Does the oxidation of methionine in thrombomodulin contribute to the hypercoaguable state of smokers and diabetics? Med Hypotheses 2007;68:811–21. [47] Newsholme P, Haber EP, Hirabara SM, Rebelato EL, Procopio J, Morgan D, et al. Diabetes associated cell stress and dysfunction: role of mitochondrial and nonmitochondrial ROS production and activity. J Physiol 2007;583:9–24. [48] Qu D, Wang Y, Esmon NL, Esmon CT. Regulated endothelial protein C receptor shedding is mediated by tumor necrosis factor-α converting enzyme/ADAM17. J Thromb Haemost 2007;5:395–402. [49] Kurosawa S, Stearns-Kurosawa DJ, Hidari N, Esmon CT. Identification of functional endothelial protein C receptor in human plasma. J Clin Invest 1997;100:411–8. [50] Tyagi N, Gillespie W, Vacek JC, Sen U, Tyagi SC, Lominadze D. Activation of GABA-A receptor ameliorates homocysteine-induced MMP-9 activation by ERK pathway. J Cell Physiol 2009;220:257–66. [51] Akasaka K, Akasaka N, Di Luozzo G, Sasajima T, Sumpio BE. Homocysteine promotes p38-dependent chemotaxis in bovine aortic smooth muscle cells. J Vasc Surg 2005;41:517–22. [52] Dong F, Zhang X, Li SY, Zhang Z, Ren Q, Culver B, et al. Possible involvement of NADPH oxidase and JNK in homocysteine-induced oxidative stress and apoptosis in human umbilical vein endothelial cells. Cardiovasc Toxicol 2005;5:9–20. [53] Bescond A, Augier T, Chareyre C, Garcon D, Hornebeck W, Charpiot P. Influence of homocysteine on matrix metalloproteinase-2: activation and activity. Biochem Biophys Res Commun 1999;263:498–503. [54] Antoniades C, Antonopoulos AS, Tousoulis D, Marinou K, Stefanadis C. Homocysteine and coronary atherosclerosis: from folate fortification to the recent clinical trials. Eur Heart J 2009;30:6–15.