|Home | About | Journals | Submit | Contact Us | Français|
Recent studies have indicated that antithrombin (AT) possesses both antiinflammatory and antiangiogenic properties.
The purpose of this study was to investigate the mechanism of the intracellular signaling activities of AT using wild-type and mutant serpins which have reduced anticoagulant activities due to mutations in either the reactive center loop (RCL) or the heparin-binding site.
Direct cellular effects of the AT derivatives were compared in the LPS-stimulated endothelial cells employing permeability and neutrophil adhesion assays in the absence and presence of pertussis toxin (PTX) and siRNAs for either syndecan-4 or sphingosine 1-phosphate receptor 1 (S1P1). Furthermore, the roles of prostacyclin and nuclear factor (NF)-κB in modulating these effects were investigated.
Both wild-type and the RCL mutant, AT/Proth-2, exhibited similar potent barrier protective activities and inhibited the adhesion of neutrophils to endothelial cells via inhibition of the NF-κB pathway. Indomethacin abrogated both activities. The heparin-binding site mutants, AT-K114E and AT-K125E, did not exhibit any protective activity in either one of these assays, but a potent pro-apoptotic activity was observed for the AT-K114E in endothelial cells. Both PTX and siRNA for syndecan-4 inhibited the protective effect of AT, but the siRNA for S1P1 was inconsequential.
The interaction of AT with syndecan-4 is required for its prostacyclin-dependent protective effect through a PTX-sensitive and non-S1P1-related Gi-protein coupled receptor. The RCL mutant, AT/Proth-2, with a markedly reduced anticoagulant but normal protective signaling properties, may potentially be developed as a safer antiinflammatory drug without increasing the risk of bleeding.
Antithrombin (AT) is a serine protease inhibitor of the serpin superfamily which regulates the proteolytic activities of the procoagulant proteases of both intrinsic and extrinsic pathways . Unlike the high reactivity of the cofactor-independent serpins α1-antitrypsin and α1-antichymotrypsin with their target proteases, AT is a slow inhibitor of its target coagulation proteases unless it binds to the heparin-like glycosaminoglycans (GAGs) that line the microvasculature . Similar to most other heparin-binding serpins, AT interacts with GAGs via the basic residues of the D-helix [3,4]. Mutagenesis and direct binding studies have indicated that among the basic residues of the heparin-binding site two Lysine residues at positions 114 and 125 play critical roles in the high-affinity interaction of AT with heparin . Heparin binding to the D-helix of AT is associated with a conformational change in the structure of the serpin that facilitates the active-site docking of RCL and optimal recognition of the serpin by the target coagulation proteases .
In addition to its anticoagulant activity through a direct inhibition of coagulation proteases , AT also possesses potent antiangiogenic [7,8] and antiinflammatory properties . Both of these properties of AT appear to be mediated through the interaction of the serpin with the vessel wall GAGs, thereby eliciting intracellular signaling responses in vascular endothelial cells [8,9]. Recent results have demonstrated that different GAG molecules may be involved in mediating the signaling specificity of AT. Thus, it has been discovered that cleaved and latent forms of AT with low affinities for heparin exhibit antiangiogenic activities in response to distinct growth factors [10,11], however, the native high affinity-heparin conformer of AT elicits antiinflammatory responses in cytokine stimulated endothelial cells [9,12,13]. The mechanism by which different conformers of AT elicit different signaling responses by a GAG-dependent manner is not understood. However, it has been demonstrated that the treatment of HUVECs with cleaved and latent conformers of AT results in the up-regulation of the expression of specific genes involved in cellular apoptosis and cell cycle arrest . On the other hand, native AT has been demonstrated to induce prostacyclin (PGI2) synthesis, thereby reducing neutrophil activation and cytokine production in endotoxin-stimulated endothelial cells [9,12,13]. The latter property, which is specific for the native conformer of AT and is not shared by either cleaved or the latent conformers of the serpin, may account for the protective antiinflammatory effect of AT in reducing mortality from sepsis in animal models [9,14].
In this study, we used siRNA and known pharmacological modulators of signaling pathways to characterize the properties of wild-type and three variants of AT containing mutations at RCL, or the heparin-binding site, with the goals of better understanding the signaling mechanism of AT, its correlation to the protective signaling mechanism of APC and assessing the cellular activities of an RCL mutant exhibiting markedly reduced anticoagulant activity.
Expression, purification and characterization of the AT mutant in which the P3-P4’ residues of RCL were substituted with the corresponding residues of the second factor Xa (fXa) cleavage site on prethrombin-2 has been described [15,16]. The heparin-binding site mutants Lys-114 → Glu (AT-K114E) and Lys-125 → Glu (AT-K125E) were constructed in the same vector system and expressed in human HEK-293 cells and purified to homogeneity . Human plasma AT (AT-WT) was purchased from Haematologic Technologies Inc. (Essex Junction, VT). RCL-cleaved AT was prepared by digesting AT-WT with human neutrophil elastase followed by its separation on heparin-Sepharose column as an early eluting fraction [10,11]. The latent AT was prepared by incubating AT-WT in 0.25 M citrate buffer at 60 °C for 24 h followed by its separation by the heparin-Sepharose column chromatography [10,11].
Permeability was quantitated by the spectrophotometric measurement of the flux of Evans blue-bound albumin across immortalized human umbilical vein endothelial (EA.hy926) cell monolayers (courtesy of Dr. C. Edgell from University of North Carolina at Chapel Hill, NC) using a modified 2-compartment chamber model as described .
Neutrophil adherence to endothelial cells was evaluated by fluorescent labeling of neutrophils as described . The expression of vascular cell adhesion molecule-1 (VCAM-1), intercellular adhesion molecule-1 (ICAM-1) and E-selectin on EA.hy926 cells was evaluated by a whole-cell ELISA as described .
The effect of AT derivatives on stimulation of prostacyclin (PGI2) in EA.hy926 cells was analyzed by a commercial ELISA kit (Amersham Biosciences) by measuring the concentration of 6-keto-prostaglandin F1α (6-keto-PGF1α, a metabolite of PGI2) in the cell culture media according to the manufacturer’s protocol. Briefly, confluent cells were incubated in the absence and presence of AT (150 µg/mL) for 4 h at 37 °C in 5% CO2. The cell culture supernatants were collected, concentrated, and transferred to a 96-well assay plate coated with antiserum against 6-keto-PGF1α. After 30 min incubation on a shaker, the 6-keto-PGF1α-peroxidase conjugate (as a competitor of the unlabeled 6-keto-PGF1α) was added and incubated for another 30 min. The amount of labeled 6-keto-PGF1α (inversely proportional to concentration of the unlabeled ligand) was determined by addition of tetramethylbenzidine/hydrogen peroxide substrate and the resultant color is read at 450 nm. In the presence of indomethacin (50 µM), cells were first treated with the inhibitor for one hour before adding AT .
EA.hy926 cells were grown to confluence in 6-well culture dishes and subjected to starvation overnight in the serum free medium. The NF-κB pathway activation in the conditioned cell lysates (treated with AT before stimulation with 10 ng/mL LPS for 1 h) was monitored by Western-blotting employing two antibodies that are specific for either NF- κB p65 or its phosphorylated form.
The cellular effect of AT derivatives (0–150 µg/mL) on EA.hy926 cells was evaluated in an LPS (10 ng/mL for 4 h)-induced apoptosis assay [17,19]. Prior to LPS-stimulation, cells were incubated with AT for 24 h at 37 °C. The number of apoptotic cells was expressed as the percentage of TUNEL-positive cells of the total number of nuclei determined by Hoechst 33342 staining .
We developed an LPS-mediated hyperpermeability assay in EA.hy926 cells and evaluated the barrier protective activities of wild-type AT (AT-WT) and an AT derivative (AT/Proth-2) which has dramatically reduced anticoagulant activity [15,16]. AT/Proth-2 has a normal affinity for heparin, but inhibits fXa with ~10-fold slower rate . However, because this mutant has three non-preferred residues for thrombin at P3, P1’ and P3’ sites of RCL (Fig. 1), it is essentially unreactive with thrombin . The dose-dependence of the LPS-mediated enhancement of permeability in EA.hy926 cells indicated that the maximum permeability is achieved at 10 nM LPS (Fig. 2A). The dose-dependence of the barrier protective effect of AT-WT in response to 10 nM LPS suggested that AT-WT exhibits a potent barrier protective effect at 150 µg/mL that corresponds to the physiological concentration of the serpin in plasma (Fig. 2B). AT/Proth-2 exhibited a barrier protective activity that was essentially identical to that of AT-WT (Fig. 2C). Neither one of the heparin-binding site mutants AT-K114E and AT-K125E (Fig. 1) exhibited any activity in this assay (Fig. 2C), suggesting that AT interaction with GAGs is required for the barrier protective of the serpin. Indomethacin abrogated the activity of both AT and AT/Proth-2, supporting the hypothesis that the protective effect of AT is mediated through induction of PGI2 (Fig. 2C). To provide further support for this hypothesis, the effect of AT derivatives on PGI2 stimulation in EA.hy926 cells was directly studied by measuring the concentration of 6-keto-prostaglandin F1α in cell culture media after cell treatment with AT. Both AT and AT/Proth-2 induced prostacyclin synthesis to similar extents, and indomethacin inhibited the protective effects of both serpins (Fig. 2D). Consistent with a GAG-dependent effect for AT, neither K114E nor K125E, exhibited any activity this assay (Fig. 2D). Since recombinant AT derivatives contain an N-terminal HPC4 epitope, we examined the activity of recombinant wild-type AT in the permeability assay and obtained essentially identical results (not shown). This is consistent with our previous results demonstrating that N-terminal HPC4 epitope does not interfere with either heparin-affinity or proteinase-reactivity of AT . Thus, all experiments were conducted with the plasma-derived AT (AT-WT). The uncatalyzed reactivity of both heparin-binding site mutants toward fXa and thrombin was normal.
APC exerts it barrier protective activity through EPCR-dependent activation of PAR-1 in lipid rafts of endothelial cells [17,19–21]. The protective activity of APC is inhibited by PTX and it also requires S1P1 as a co-receptor [17,19]. To determine whether a similar mechanism is involved in the barrier protective activity of AT, we compared the activities of both APC and AT in LPS-mediated hyperpermeability assay in the absence and presence of PTX, the cholesterol depleting molecule methyl-β-cyclodextrin (MβCD) and siRNA for S1P1. The results presented in Fig. 3A indicate that while the barrier protective activity of APC is inhibited by all treatments, the activity of AT is sensitive only to PTX. The barrier protective effect of AT is inhibited by full-length heparin, the AT-binding pentasaccharide (PS), and significantly diminished by the siRNA for syndecan-4 (Fig. 3B). Neither cleaved nor latent AT exhibited any barrier protective activity in response to LPS (Fig. 3C). The knockdown of syndecan-4 in endothelial cells by siRNA was very efficient with no detectable expression of the receptor (Fig. 3D).
AT-WT effectively inhibited the adhesion of freshly isolated neutrophils to LPS-stimulated endothelial cells in a concentration dependent manner with the inhibitory effect reaching its maximum level at a physiological concentration of the serpin (Fig. 4A). AT/Proth-2 also exhibited a potent inhibitory activity, but both heparin-binding site mutants were inactive (Fig. 4B). Indomethacin abrogated the activity of both AT-WT and AT/Proth-2, suggesting the involvement of PGI2 in the inhibitory effect (Fig. 4B). Both AT and AT/Proth-2 inhibited the adhesion of neutrophils by down-regulating LPS-mediated expression of cell adhesion molecules (Fig. 4C) by a PGI2-dependent manner as indomethacin inhibited this effect (Fig. 4D). Both AT-WT and AT/Proth-2, but not heparin-binding site mutants, effectively inhibited the activation of NF-κB in LPS-stimulated endothelial cells (Fig. 5). This property has been reported for AT-WT previously .
None of the AT derivatives exhibited cytoprotective activities in response to LPS in endothelial cells (Fig. 6A). Further studies revealed that AT induces cellular apoptosis independent of LPS (Fig. 6B). Interestingly, it was found that AT-K114E exhibits a potent pro-apoptotic activity that reaches a maximal response at ~15–20 µg/mL, contrasting the much higher concentration of 150 µg/mL for either AT-WT or AT-K125E (Fig. 6C). Recent studies have indicated that cleaved and latent forms of AT inhibit angiogenesis in response to specific growth factors [8,10,11]. Thus, this activity of AT may be related to the antiangiogenic activity of the serpin. To test this hypothesis, we compared the concentration-dependence of the pro-apoptotic activity of AT-WT with that of cleaved and latent forms of the serpin. Both cleaved and latent AT exhibited potent apoptotic activities at much lower concentrations of ~10 µg/mL (Fig. 6D). A previous study also reported that the Ala substitution mutant of Lys-114 (20 µg/mL), but not AT-WT, specifically elicited antiangiogenic activity in a basic fibroblast growth factor-induced angiogenesis assay [10,11].
Finally, we used the TUNEL assay to determine whether syndecan-4 is involved in the pro-apoptotic activity of either latent AT or AT-K114E. Unlike latent AT and AT-K114E, AT-WT (20 µg/mL) exhibited no pro-apoptotic activity (Fig. 7). The pro-apoptotic activity of both latent and AT-K114E also required D-helix interaction with GAGs since both heparins markedly inhibited the activity (Fig. 7). The syndecan-4 siRNA also markedly inhibited the pro-apoptotic activity of AT derivatives (Fig. 7), suggesting that interaction with syndecan-4 is also required for this effect of the serpin.
We have demonstrated in this study that both AT-WT and AT/Proth-2 elicit identical barrier protective and antiinflammatory activities in response to LPS in endothelial cells through the prostacyclin-dependent inhibition of the NF-κB pathway. Interactions with endothelial GAGs via D-helix were required for the protective signaling activities of AT since both heparin-binding site mutants were inactive in the permeability and cell adhesion assays. The observation that siRNA for syndecan-4 partially inhibited the activity of AT suggests that AT interaction with syndecan-4 is essential for the protective activity of the serpin. Syndecan-4 has been demonstrated to play key roles in PKCα-dependent regulation of a variety of cellular responses . Neither the cleaved nor the latent forms of AT exhibited any activity in the LPS-induced hyperpermeability assay, suggesting that the activity of AT requires the native conformation of the serpin. These protective properties of AT are essentially identical to those observed with APC in similar cellular models [18,20,21]. However, APC exerts its barrier protective activity through EPCR-dependent cleavage of PAR-1 localized to caveolae-rich lipid-rafts in endothelial cells [17,19]. Moreover, the protective effect of APC is inhibited by PTX, suggesting that the signaling effect of APC requires activation of the Gi/o-subfamily of G-proteins . The activation of PAR-1 by the EPCR-bound APC has been reported to activate phosphatidylinositol 3-kinase which leads to phosphorylation of the Gi-protein coupled receptor S1P1 and subsequent activation of the protective Rac1 signaling pathway . The observations that neither MβCD nor siRNA for S1P1 affected the barrier protective activity of AT suggest that, unlike APC, the protective activity of AT does not require lipid-raft localization nor is it mediated through the cross-activation of S1P1. However, the PTX sensitivity of the protective activity of AT suggests that, similar to APC, the activation of a Gi/o-protein is required for the protective signaling mechanism of the serpin.
APC has been approved by the FDA as a therapeutic drug for treating patients with severe sepsis. However, unlike APC, the recent KyberSept human clinical trial with AT did not show a beneficial effect on the mortality rate in patients with severe sepsis, though this study also used a low-dose of heparin concomitant with AT in the trial [9,25]. Since the interaction of AT with GAGs on the endothelium is necessary for the antiinflammatory effect of AT, it is believed that heparin may have antagonized this effect in the septic setting [9,25]. There is some support for this hypothesis as a post hoc analysis of a subgroup of patients receiving a high-dose AT without heparin showed benefit in the 90-day mortality rate [9,25]. Thus further studies are warranted to evaluate the role of AT-therapy in severe sepsis. Nevertheless, due to the requirement for a high-dose of AT in treating sepsis [9,25], bleeding remains a serious drawback of AT-therapy. AT/Proth-2, which is capable of slowly inhibiting fXa but exhibits no reactivity with thrombin [15,16], may have superior therapeutic utility in severe sepsis without impairing hemostasis, provided it lacks immunogenecity. Since increased incidence of bleeding has been also observed with the APC-therapy in severe sepsis, APC variants have been constructed which have reduced anticoagulant, but normal signaling activities . One of these variants has been shown to be effective in down-regulating the proinflammatory responses in an animal model of sepsis . Further studies are required to determine whether AT/Proth-2 has normal antiinflammatory activities in similar models.
The observation that AT elicited a pro-apoptotic activity in endothelial cells was surprising. This activity of AT is not shared by APC which is known for its potent cytoprotective activity [20,21]. Nevertheless, noting the metastable nature of AT, which can spontaneously convert to a latent conformation upon long term storage and/or incubation at higher temperatures, there is a possibility that the pro-apoptotic activity of AT-WT is due to the presence of some latent or cleaved forms of the serpin in the preparation, though we observed the same effect with the freshly purified serpin from the heparin-column. This observation raises an important question regarding the potential therapeutic use of AT in severe sepsis since an effective therapy appears to require a high concentration of the serpin [9,25]. In light of the unstable nature of the molecule and the possible requirement for its pasteurization for use in clinical settings, it may not 14 be possible to prepare and maintain AT at higher concentrations devoid from the latent conformer. If so, additional mutations on the structure of AT, similar to those in plasminogen activator inhibitor 1 , may be required to prevent AT from spontaneously converting to a latent conformation. However, further studies are required to determine whether the pro-apoptotic activity of AT in cultured endothelial cells has any physiological relevance. The observation that AT-K114E has a potent pro-apoptotic but no antiinflammatory activity may suggest that different AT receptors are involved in these different responses. Thus, it is possible that the pro-apoptotic activity of AT is a cell cycle-dependent physiological response specific for only dividing endothelial cells during wound healing and angiogenesis and that AT plays a key regulatory role in this process. This property of AT may have a beneficial effect in treating inflammatory problems in certain cancer patients.
The observations that AT-K114E as well as the cleaved and latent conformers of AT exhibit potent pro-apoptotic activity but lack barrier protective activity suggest that distinct GAGs are involved in these different cellular responses. Nevertheless, syndecan-4 appears to contribute to both pro-apoptotic and barrier protective activities of AT in cultured endothelial cells. These results point toward possible involvement of co-receptors for syndecan-4 in endowing signaling specificity for various conformers of the serpin in these two disparate pathways. In support of this hypothesis, it was recently demonstrated that the antiangiogenic activity of cleaved and latent conformers of AT is mediated through the variant serpins blocking the interaction with and/or down-regulation of the expression of a co-receptor required for the fibroblast growth factor-dependent proliferation of endothelial cells and that this property is not shared by the native serpin [10,11]. A similar mechanism could also be responsible for the pro-apoptotic activity of AT variants observed in this study, but this hypothesis needs to be confirmed. Further studies will be required to determine whether similar co-receptor(s) exists on the membrane surface that is involved in mediating the barrier protective activity of AT in endothelial cells.
We would like to thank Audrey Rezaie for proofreading of the manuscript. The research discussed herein was supported by grants awarded by the National Heart, Lung, and Blood Institute of the National Institute of Health HL 62565 and HL 68571 to ARR.