- Email: [email protected]
Shiga toxin 2a binds antithrombin and heparin, but does not directly activate platelets
International Journal of Medical Microbiology xxx (xxxx) xxx–xxx
Contents lists available at ScienceDirect
International Journal of Medical Microbiology journal homepage: www.elsevier.com/locate/ijmm
Shiga toxin 2a binds antithrombin and heparin, but does not directly activate platelets Ludwig Knabla,1, Michael Berktolda,1, Osama A. Hamadb, Karin Fromellb, Sneha Chatterjeea, Cornelia Spetha, Heribert Talaszc, Katharina Lindnere, Martin Hermanne, ⁎ ⁎ Kristina Nilsson-Ekdahlb, Bo Nilssonb, Werner Streifd, Judith Martinie, ,1, Reinhard Würznera, ,1, ⁎ Dorothea Orth-Höllera, ,1 a
Division of Hygiene and Medical Microbiology, Medical University of Innsbruck, Schöpfstrasse 41, 6020 Innsbruck, Austria Department of Immunology, Genetics and Pathology, Rudbeck Laboratory, Uppsala University, 75185 Uppsala, Sweden Division of Clinical Biochemistry, Biocentre, Medical University of Innsbruck, Innrain 80, 6020 Innsbruck, Austria d Department of Paediatrics I, Medical University of Innsbruck, Anichstrasse 35, 6020 Innsbruck, Austria e Univ. Clinic of Anesthesia and Intensive Care Medicine, Medical University of Innsbruck, Anichstrasse 35, 6020 Innsbruck, Austria b c
A R T I C LE I N FO
A B S T R A C T
Keywords: Stx2a Hemolytic uremic syndrome Plasmatic coagulation Antithrombin
Escherichia coli-induced hemolytic uremic syndrome (eHUS) is a life-threatening complication of infection with Shiga toxin (Stx), in particular Stx2a-producing Escherichia coli. Enhanced coagulation activation with formation of microthrombi seems to be a key event in development of eHUS. Platelet activation has been postulated as a possible, but controversially debated mechanism. The present study investigated the effect of Stx2a on plasmatic coagulation and platelets. Binding studies were initially performed with ELISA and co-immunoprecipitation and supported by quartz crystal microbalance with dissipation monitoring (QCM-D). Antithrombin (AT) activity was measured using the automated BCS XP® system. ROTEM® was used for functional coagulation testing. Platelet binding and activation was studied with FACS and light-transmission aggregometry. We found binding of Stx2a to AT, an important inhibitor of blood coagulation, but only a mild albeit significant reduction of AT activity against FXa in the presence of Stx2a. QCM-D analysis also showed binding of Stx2a to heparin and an impaired binding of AT to Stx2a-bound heparin. ROTEM® using Stx2a-treated plateletpoor plasma revealed a significant, but only moderate shortening of clotting time. Neither binding nor activation of platelets by Stx2a could be demonstrated. In summary, data of this study suggest that Stx2a binds to AT, but does not induce major effects on plasmatic coagulation. In addition, no interaction with platelets occurred. The well-known non-beneficial administration of heparin in eHUS patients could be explained by the interaction of Stx2a with heparin.
1. Introduction Hemolytic uremic syndrome (HUS) caused by infection with Shiga toxin (Stx), in particular Stx2a-producing enterohemorrhagic Escherichia coli (EHEC) is the leading cause of acute kidney injury (AKI) in pediatric patients (Karch et al., 2005). In the full clinical picture of
EHEC-induced HUS (recently termed eHUS) AKI is accompanied by microangiopathic hemolytic anemia and thrombocytopenia (Tarr, 1995). It is well accepted that hemostatic abnormalities play an important role in the development and progression of eHUS, especially in the development of AKI (Lee et al., 2013; Nevard et al., 1997; Proesmans,
Abbreviations: eHUS, Escherichia coli-induced hemolytic uremic syndrome; Stx, Shiga toxin; EHEC, Enterohemorrhagic Escherichia coli; AT, antithrombin; AKI, acute kidney injury; t-PA antigen, tissue plasminogen activator antigen; PAI-1, t-PA-plasminogen activator inhibitor type 1; TF, tissue factor; vWF, von Willebrand factor; NOACs, new oral anticoagulants; PRP, platelet-rich plasma; APC, allophycocyanin; PE, phycoerythrin; LTA, light-transmission aggregometry; ROTEM®, rotational thromboelastometry; MFI, mean fluorescence intensity; CT, clotting time; SAP, serum amyloid P; QCM-D, quartz crystal microbalance with dissipation monitoring; PAV, polyamine; CHC, charged heparin conjugate ⁎ Corresponding authors. E-mail addresses: [email protected] (J. Martini), [email protected] (R. Würzner), [email protected] (D. Orth-Höller). 1 Contributed equally to this work. https://doi.org/10.1016/j.ijmm.2018.07.008 Received 17 May 2018; Received in revised form 13 July 2018; Accepted 22 July 2018 1438-4221/ © 2018 Published by Elsevier GmbH.
Please cite this article as: Knabl, L., International Journal of Medical Microbiology (2018), https://doi.org/10.1016/j.ijmm.2018.07.008
International Journal of Medical Microbiology xxx (xxxx) xxx–xxx
L. Knabl et al.
(Linz, Austria); heparin from G.L. Pharma (Lannach, Austria); rabbit anti-FXa ab from US Biological Life Sciences (Salem, USA); sheep antihuman AT ab from Acris (Herford, Germany); mouse anti-Stx2 ab from Hycult Biotechnology (Uden, The Netherlands); goat anti-mouse APconjugated IgG, goat anti-rabbit AP-conjugated IgG, donkey anti-sheep IgG and rabbit anti-human AT abs, p-nitrophenyl phosphate from Sigma Aldrich. Platelet concentrates were obtained from the blood bank of the University Clinic Innsbruck. Platelets were checked to be not-activated by measuring CD62 P with FACS as detailed below. Blood samples for the preparation of plasma or platelet-rich plasma (PRP) were – after informed consent and with an existing ethical vote – collected from healthy volunteers. Blood was drawn in 10 mL tubes anticoagulated with buffered 3.2% citrate. PRP for FACS analysis and light-transmission aggregometry (LTA) were obtained by centrifugation (135 g for 15 min) and used within 4 h following preparation. For rotational thromboelastometry (ROTEM®) analysis and co-immunoprecipitation citrated blood was centrifuged at 2500 g for 10 min. The obtained plasma was then pooled, swiftly portioned into identical aliquots of 1.5 mL per microcentrifuge tube and stored at −78 °C until analyzed.
2001). Patients with eHUS have an occlusion of the renal microvasculature due to endothelial damage and subsequent formation of microthrombi leading to ischemic kidney damage (George and Nester, 2014). They usually present with a strong rise of laboratory parameters indicating global coagulation activation such as increased levels of DDimer, tissue plasminogen activator antigen (t-PA antigen), t-PA-plasminogen activator inhibitor type 1 (PAI-1) complex and prothrombin fragments 1 and 2 (Ay et al., 2009; Chandler et al., 2002). Plasma levels of tissue factor (TF), the most important initiator of coagulation in vivo (Camerer et al., 1996), are markedly elevated in children with eHUS, through Stx induced release of TF (Bhowmik, 2001; Kamitsuji et al., 2000; Stahl et al., 2009). Additionally, there is also evidence for Stxinduced alterations in the protein C pathway (Mayer et al., 2015), an important inhibitor of coagulation. On the other hand, there are data suggesting a major role of platelets and von Willebrand factor (vWF) in the development of eHUS; a widespread activation of platelets either through Stx itself or by the release of chemokines and other factors from endothelial cells has been repeatedly affirmed (Guessous et al., 2005; Karpman et al., 2001; Obrig and Karpman, 2012; Stahl et al., 2015; Sun et al., 2016). The exact pathomechanism however, and the question whether this hypercoagulopathic status is mainly initiated via activation of platelets or the plasmatic hemostasis or a combination of both, is currently under debate. Thrombus formation is the result of a highly complex interaction between different physiological systems and extensive cross talk between coagulation and inflammation exists (Levi and van der Poll, 2005). Furthermore, it has been shown that complement, an important part of the innate immune system, might be involved in the pathogenesis of eHUS (Ehrlenbach et al., 2013; Orth et al., 2009; Poolpol et al., 2014). Additionally, it is notable that only 15% of patients infected with EHEC develop eHUS; it can be assumed that specific host related factors, such as polymorphisms or deficiencies of factors involved in hemostasis may lead to a clinically relevant change of coagulation after infection with EHEC. Currently there is no consensus about how to treat coagulation abnormalities in affected patients or if they should be treated at all (Walsh and Johnson, 2018). In view of the clinically most relevant new oral anticoagulants (NOACs) with direct inhibition of either thrombin or factor Xa, new treatment strategies may emerge which could prevent or mitigate kidney failure in eHUS patients. This study was designed to investigate which steps of the coagulation process are mainly disturbed by the presence of Stx; it was therefore assessed whether Stx2a activates platelets and in a second set of experiments we evaluated the effect of Stx2a on various factors of the plasmatic coagulation.
2.2. Analysis of binding of Stx2a to either antithrombin (AT), thrombin or protein C and binding of Stx2a-bound AT to FXa by ELISA To establish whether Stx2a binds to AT, microtiter plates were coated with Stx2a (1 μg per well) in 100 μL coating buffer (12,430 mg NaHCO3, 5510 mg Na2CO3 in 1 l aqua dest.) overnight at 4 °C. After blocking with 1% (w/v) gelatine, each well was incubated with 1 μg of AT for 4 h at 37 °C. After additional washing steps, bound AT was detected with a primary rabbit anti-AT antibody (1:1000) followed by a goat anti-rabbit AP-conjugated IgG (1:1000), both diluted in PBS-T. The binding was detected with p-nitrophenyl phosphate as the substrate. AT alone was used as a positive control. As negative controls we coated BSA or serum amyloid P (SAP) (1 μg per well) in place of Stx2a followed by addition of AT (1 μg for 4 h) and detection with the above mentioned rabbit anti-AT antibody and goat anti-rabbit IgG. Interaction of Stx2a with AT was further analyzed at different pH values (pH 4, pH 4.5, pH 5, pH 5.5, pH 6, pH 10). After blocking, AT (1 μg per well) was added to the wells in 100 μL PBS-T adjusted to respective pH values. The remaining steps were performed as described above. To analyze the binding of Stx2a-bound AT to FXa, 1 μg of FXa was immobilized in 100 μL coating buffer. Simultaneously, 2.5 μg Stx2a and 2.5 μg AT were dissolved in 250 μL PBS-T and incubated in a micro centrifuge tube for 20 h at 37 °C. Afterwards, 100 μL of the Stx2a/ATsuspension was added to the immobilized FXa for 1 h at 37 °C. After washing, bound Stx2a/AT was detected with a primary rabbit antihuman AT antibody (1:1000) and a secondary anti-rabbit AP-conjugated IgG antibody (1:1000). Bound Stx2a/AT was detected by development with p-nitrophenyl phosphate as substrate. Controls were performed with BSA instead of Stx2a. To evaluate whether Stx2a binds to protein C or thrombin, the microtiter plates were coated with the respective proteins (1 μg per well, each) in 100 μL coating buffer. After blocking, Stx2a (1 μg) was added for 4 h at 37 °C. Bound Stx2a was detected with a primary mouse antiStx2 antibody (1:100) followed by an anti-mouse AP-conjugated IgG (1:1000), both diluted in PBS-T. For negative control BSA was added to protein C or thrombin in place of Stx2a, for positive control Stx2a (1 μg per well) alone was coated. Photometric readouts were done at 415 and 490 nm.
2. Materials and methods 2.1. Purification of Stx2a, proteins, antibodies, platelet concentrates platelet-rich plasma (PRP) and plasma Purification of Stx2a was done as described elsewhere (Zhang et al., 2008) and the purity level was evaluated by SDS-PAGE. Lipopolysaccharide contamination was determined to be 5.88 pg/μg using a method described elsewhere (Brigotti, 2012). Stx2a was labeled with FITC using an Oyster-488 antibody labeling kit (Luminartis, Münster, Germany). LPS from E. coli serotype O55B5 was obtained by Charles River Laboratories Inc. (Charleston, USA). The allophycocyanin (APC)-conjugated CD42b antibody (ab) and the phycoerythrin (PE)- conjugated CD62 P ab were purchased from BD Biosciences (San Jose, USA) and BioLegend (San Diego, USA), respectively; antithrombin (AT), factor Xa and protein C from Calbiochem (Darmstadt, Germany); thrombin and GPR peptide from Sigma Aldrich (Taufkirchen, Germany); ADP from Sigma Aldrich (Taufkirchen, Germany) and Chrono-Log (Haverton, USA); CaCl2 from Siemens (Munich, Germany); collagen from Nycomed
2.3. Analysis of binding of Stx2a to AT by co-immunoprecipitation To confirm the results of the ELISA test for Stx2a-AT-binding, coimmunoprecipitation using a kit from Thermo Fisher Scientific 2
International Journal of Medical Microbiology xxx (xxxx) xxx–xxx
L. Knabl et al.
(Waltham, USA) was performed. Anti-Stx2a antibody (100 μL) was coupled to the agarose resin as described in the manufacturer’s protocol. Simultaneously, 10 μg Stx2a and 10 μg AT or 20 μL plasma or plasma (20 μL) spiked with AT (10 μg) were incubated at 37 °C and continuous stirring for 4 h. After dilution in Dulbecco’s modified PBS at a ratio of 1:7.5, the suspensions were applied to the anti-Stx2a antibody-coupled gel and incubated overnight at 37 °C under gentle agitation. After washing, bound proteins were eluted according to the manufacturer’s instructions. Eluted protein complexes were analyzed by western blot using a 16% (w/v) gel (Bio-Rad, Vienna, Austria) under reducing conditions. Detection was done with a sheep anti-human ATantibody (1:3000) and a secondary anti-sheep IgG antibody (1:2000). As positive control AT was loaded on the gel, for the negative control BSA was loaded. Bands were visualized using the Clarity ECL blotting substrate (Bio-Rad) and the ImageQuant LAS4000 (GE Healthcare Life Sciences, Chalfont St Giles, UK).
2.5. Determination of AT activity in the presence of Stx2a To analyze AT activity directed against FIIa and FXa, whole blood was incubated with Stx2a (250 or 500 ng/ml) in a volume of 300 μL. For negative control HEPES buffer was added in place of Stx2a. The samples were incubated for 4 h at 37 °C. After incubation samples were centrifuged 15 min. at 1000 g at room temperature. AT activity was determined using the automated BCS XP® system (Siemens Healthcare Diagnostics GmbH, Eschborn, Germany). 2.6. Determination of plasmatic coagulation activation in the presence of Stx2a by rotational thromboelastometry (ROTEM®) Functional coagulation analysis was performed by ROTEM® (Tem International GmbH, Munich, Germany), a point-of-care monitoring system for severely bleeding patients (Kozek-Langenecker et al., 2013). Unlike conventional laboratory coagulation parameters (e.g. PT, aPTT), this technique analyses the entire clotting process from coagulation start until clot formation and determines clot firmness and clot stability over time. The exact measurement technique is described in detail elsewhere (Gorlinger et al., 2016). In this study, functional coagulation analysis was performed in pooled plasma. A continuous viscoelastic profile of clot formation was obtained (Ganter and Hofer, 2008). Due to our study hypothesis, main focus was on measurements of clotting time (CT). In order to detect the minutest details in changes of clot formation initiation, only calcium chloride was added to the plasma sample (star-tem® reagent), performing a so called NATEM® test. NATEM tests started with pipetting 20 μL star-tem reagent into the plastic cup, promptly followed by 300 μL pooled plasma. During the initialized ROTEM sequence of mixing by drawing and releasing the plasma probe with the pipette, 30 μL of the corresponding dose to 10 ng/mL or 100 ng/mL of HEPES-diluted Stx2a were additionally pipetted into the sample. Coagulation process starts immediately and is fully represented by the data graph obtained over 30 min. Control measurements were performed by adding 30 μL HEPES buffer to the plasma sample instead of Stx2a to ensure comparable hemodilution of the samples.
2.4. Evaluation of Stx2a binding to AT, heparin, factor Xa by quartz crystal microbalance with dissipation monitoring (QCM-D) QCM-D is a very useful technique for measuring biomolecule interactions at surfaces. It was therefore used to follow the interaction between Stx2a and AT, heparin and FXa. The principle of QCM-D has been described previously (Rodahl et al., 1997). Briefly, a QCM-D consists of a thin quartz disc clamped between two electrodes. The disc is set to oscillation by applying a voltage over the electrodes. The resonance frequency (f) of the oscillating quartz crystal depends on the mass assembled on its sensor surface. For very thin and rigid films, the decrease in frequency will be proportional to mass of the assembled layer and the mass can be calculated according to the Sauerbrey relation (Sauerbrey, 1959). The QCM-D measurements were performed using a QCM-D Omega Auto (Biolin Scientific AB, Göteborg, Sweden). All QCM-D experiments were carried out at 37 °C with a flow rate of 20 μL/min. PBS was used for priming and rinsing of the surfaces before, in between and after the addition of the reagents. 2.4.1. Evaluation of Stx2a binding to AT Two clean polystyrene coated QCM-D sensors were mounted in the QCM-D instrument. After priming with buffer, Stx2a (10 μg/mL) was allowed to adsorb for 30 min to one of the sensors, while AT (10 μg/mL) was adsorbed to the other. After wash and blocking (0.5% BSA) AT (10 μg/mL) was added to the first sensor and Stx2a (10 μg/mL) was added to the other. Binding was allowed to proceed for 20 min followed by rinsing with PBS.
2.7. Assessment of Stx2a binding to platelets by real time live confocal microscopy Platelets were placed in 8-well chambered cover glasses (Nalge Nunc International, Rochester, USA) and incubated with FITC-labeled Stx2a (20 ng/mL) for 4 h at 37 °C. The platelets were counterstained using WGA-AlexaFluor647 dye (Molecular Probes, Eugene, USA). Vero cells loaded with Stx2a (20 ng/mL) served as positive control. Images were taken at time points 0, 30 min, 1 h, 2 h and 4 h. Real Time Live Confocal Microscopy was performed with a spinning disk confocal system (Ultra VIEW VoX; Perkin Elmer, Waltham, USA) connected to a Zeiss AxioObserver Z1 microscope (Zeiss, Oberkochen, Germany). Images were acquired with the Volocity software (Perkin Elmer) using a 63x oil immersion objective with a numerical aperture of 1.4.
2.4.2. Evaluation of Stx2a binding to heparin Again two parallel experiments were performed using two polystyrene QCM-D sensors. Both sensors were first coated with positively charged amino groups i.e. polyamine (PAV) followed by the negatively charged heparin conjugate (CHC) according to manufacturer’s instructions (Corline Systems AB). The surfaces were rinsed and Stx2a (10 μg/ mL) was added to one of the surfaces. After incubation the surface was again rinsed and then AT (10 μg/mL) was added to the same surface. For comparison, AT, which is well known to bind to heparin, was added directly to the second heparin coated surface without the previous addition of Stx2a.
2.8. FACS analyses of Stx2a-mediated platelet activation and Stx2a/ platelet binding The antigen CD42b is present on both resting and activated platelets and was therefore used for platelet identification (Suzuki et al., 1987). To evaluate whether Stx2a enhanced platelet activity, PRP was incubated for 90 min at 37 °C with various concentrations of purified Stx2a (1 ng/mL, 10 ng/mL, 100 ng/mL). Platelet activation and subsequent secretion of α-granules was evaluated by FACS quantification of CD62 P (P-selectin) (Lu and Malinauskas, 2011). To check whether Stx2a interferes with ADP- or thrombin-mediated platelet stimulation, PRP was either mock-treated or pre-incubated with purified Stx2a for 90 min, followed by addition of either 0.1 U/mL
2.4.3. Evaluation of Stx2a binding to coagulation factor Xa The polystyrene sensor surfaces were first coated with anti-FXaantibody (10 μg/mL) and rinsed with buffer. This was followed by addition of FXa alone (1.7 × 10−7 M) or one of the following suspensions: FXa + AT (molar ratio 1:1), FXa + AT + Stx2a (1:1:1), FXa + AT + Stx2a (1:1:2) or FXa + Stx2a (1:1). 3
International Journal of Medical Microbiology xxx (xxxx) xxx–xxx
L. Knabl et al.
thrombin or 100 μM ADP. Again, platelet activity was quantified by FACS using CD62 P as marker. To inhibit fibrin polymerization 1.25 mM GPR peptide was added to PRP. Further experiments aimed to study whether Stx2a is bound to and/ or internalized by platelets. For that purpose PRP was incubated for 90 min with different concentrations (1 ng/mL, 10 ng/mL, 100 ng/mL) of FITC-labeled Stx2a, followed by FACS analysis. All experiments were also performed with whole blood instead of PRP. All FACS measurements were done by using FACScan apparatus (Becton Dickinson, Franklin Lakes, USA). Results were given as the mean fluorescence intensity (MFI).
directed against FIIa showed no difference when Stx2a was added compared to the negative control (data not shown). 3.3. Stx2a binding to heparin The Stx2a binding to a heparinized surface was measured with QCM-D. The recorded mass uptake of Stx2a to the heparinized surface was 140 ng/cm2 followed by 37 ng/cm2 of AT (Fig. 3a) whereas 120 ng/cm2 of AT was bound to the surface without the previous addition of Stx2a (Fig. 3b). These results clearly show that Stx2a binds to heparin and also occupies several of the AT binding sites at the heparin surface (Fig. 3c). In addition, this indicates that Stx2a has the same or competing binding sites for both AT and heparin, since AT clearly bound to Stx2a adsorbed directly to the surface in the absence of heparin (Fig. 1c).
2.9. Determination of Stx2a-mediated platelet aggregation by LTA LTA according to the method of Born was performed with the PAP4 (Mölab, Hilden, Germany) aggregometer to demonstrate the effects of Stx2a on platelet aggregation. The principle is described elsewhere (Paniccia et al., 2015). Stx2a (2 ng/mL, 1 μg/mL and 2 μg/mL) was added to PRP (platelet count adjusted to 250,000 platelets/μL) and aggregation tracing was monitored for 30 min. At this point, aggregation was triggered using collagen (final concentration 4 μg/mL) and ADP (final concentration 1.75 μM/L) to ensure maximum aggregation. In order to exclude any effects which may be caused by dilution, all experiments were repeated with PBS instead of Stx2a.
3.4. Stx2a does not bind to protein C, thrombin or FXa and has no influence on the FXa-binding to AT A microtiter plate was coated with protein C, thrombin, or FXa and incubated with either Stx2a or in case of FXa also with an AT/Stx2a suspension. The optical readout revealed that Stx2a does not bind to protein C, thrombin, or FXa and the binding of Stx2a to AT does not lead to alterations of AT binding to FXa. Thus, the binding capability of the Stx2a/AT-complex to FXa was not reduced compared to controls (data not shown). This was also confirmed with QCM-D where FXa was binding to AT. In contrast, neither Stx2a nor Stx2a/AT bound to FXa. From these results it could also be concluded that the presence of Shiga toxins did not inhibit the FXa binding to AT (data not shown).
2.10. Statistical analyses The results were analyzed by the use of GraphPad Prism (version 5) software. Student's t test was performed to compare the paired means of the two measurement groups. P values of < 0.05 were considered significant. P values of < 0.01 were considered highly significant.
3.5. Shortening of clotting time by Stx2a 3. Results
Functional coagulation tests performed with ROTEM® showed that addition of two different concentrations of Stx2a (10 ng/mL and 100 ng/mL) to pooled plasma samples resulted in a significant, but only moderate shortening of the clotting time (CT) (Fig. 4). Clotting time was 622 ± 94 s in pooled plasma samples diluted with HEPES buffer without Stx2a (n = 33; n = number of test repetitions with pooled plasma samples); CT was significantly shorter in plasma samples incubated with 10 ng/mL Stx2a (559 ± 91 s; n = 32) compared to the HEPES controls (p < 0.01 vs. HEPES controls); plasma samples incubated with 100 ng/mL Stx2a (n = 30) did not result in a more pronounced shortening (574 ± 77 s; p < 0.05). When adding LPS to plasma at the amount present in Stx2a (corresponding to the highest toxin concentration) there was no significant difference in CT compared to plasma only (data not shown). All other ROTEM® parameters (CFT, MCF and ML) were not significantly different and there was no difference between the groups.
3.1. Stx2a binds to AT (by ELISA, co-immunoprecipitation and QCM-D) To evaluate whether Stx2a binds to AT, ELISA tests were performed as screening assay. AT incubated with Stx2a showed a significantly higher OD value than both negative controls indicating a binding between AT and Stx2a (Fig. 1a). Moreover, it was shown that the extent of binding of Stx2a to AT was different when the pH in the test system was changed. At a pH range between 4 and 5.5 the binding of Stx2a to AT was significantly more pronounced compared to pH values between 6 and 10 (Fig. 1a). Co-immunoprecipitation confirmed the binding of Stx2a to AT which was previously observed by ELISA. In the binding control and in elution fractions 1 and 2, no band could be detected in the Western blot whereas in elution fraction 3 a clear band was visible at around 58 kDa indicating a binding of AT to the resin-bound Stx2a (Fig. 1b). The Stx2a binding to AT was finally corroborated by QCM-D. The sensogram showing the recorded mass uptake during the whole process is shown in Fig. 1c. There are interactions between Stx2a and AT, which appear more distinctly when Stx2a sits at the surface followed by the addition of AT, since 90 ng/cm2 of AT was bound to Stx2a adsorbed to the surface (50 ng/cm2) which gives a ratio of 2 mol AT captured/mole of Stx2a and only 20 ng/cm2 Stx2a was captured by AT at the surface (25 ng/cm2) corresponding to 0.7 mol of Stx2a bound/mole AT at the surface.
3.6. Stx2a does not bind to platelets Binding of FITC-labeled Stx2a to platelets was investigated by real time live confocal microscopy. Fluorescence was neither observed on the surface nor inside the platelets at any time point, suggesting that there is no binding of Stx2a to platelets. Additionally, no morphological changes in platelets were detected, implying an absent activation of platelets. In contrast, a strong fluorescence signal was found inside and on the membrane of Vero cells (serving as positive control) as a result of significant binding and uptake of Stx2a (data not shown). The observation that Stx2a does not bind to platelets was confirmed by FACS analyses. Incubation of PRP or whole blood with FITC-labeled Stx2a (1, 10 and 100 ng/mL) for 90 min did not lead to an increase in the fluorescence signal on the surface or inside the platelets compared to the negative control with medium only (data not shown).
3.2. AT activity directed against FXa is decreased in the presence of Stx2a Addition of two different concentrations of Stx2a (250 ng/mL and 500 ng/mL) to whole blood samples resulted in a significant but only moderate decrease of AT activity directed against FXa compared to the negative control with HEPES buffer instead of Stx2a (Fig. 2). AT activity 4
International Journal of Medical Microbiology xxx (xxxx) xxx–xxx
L. Knabl et al.
Fig. 1. (1a) Most prominent binding of Stx2a to AT measured by ELISA at a pH range between 4 and 5.5. Optical density (OD) was significantly higher at all pH values when AT (1 μg) was incubated with Stx2a (1 μg) compared to negative controls, BSA and serum amyloid P (SAP). For the positive control AT (1 μg) was coated. The results are the means +/- SD of five separate experiments; *, p < 0.05; **, p < 0.01; ****, p < 0.0001. (1b) Binding of Stx2a to AT by co-immunoprecipitation. Incubations of Stx2a (10 μg) and AT (10 μg) or plasma (20 μL, as natural source of AT) or plasma spiked with AT (10 μg, to increase the naturally occurring AT in plasma) were applied to an anti-Stx2a ab coupled agarose resin. All three elution fractions of the three experiments were analyzed by western blot using an anti-AT ab. Fractions 1 and 2 did not show any bands (data not shown). All three fractions 3, lanes 1 (Stx2a and AT) and 3 (Stx2a and plasma spiked with AT), and to a lesser extent lane 2 (Stx2a and plasma), showed bands of approx. 58 kDa (indicated by an arrow), representing AT, as shown in the positive control (lane 4, loaded with AT only). For the negative control (lane 5) BSA was loaded on the gel. M, marker. The results show a representative experiment of three. (1c) QCM-D measurement of the interaction between AT and Stx2a. The gradual buildup of several layers on a surface can be followed in real time using QCM-D. This was done by first monitoring the adsorption of AT (grey line) and Stx2a (black line) to a polystyrene coated sensor surface (1), followed by blocking (2), and finally the binding of Stx2a (grey line) or AT (black line) (3). The measurement was performed at 37 °C and the sensor surfaces were rinsed with PBS before, in between and after addition of each component.
or ADP resulted in a substantial increase of activated platelets, whereas the activation was more pronounced by the presence of thrombin than ADP. Simultaneous addition of Stx2a did not influence the platelet activation induced by thrombin or ADP alone. Similar results were obtained with whole blood instead of PRP (data not shown).
3.7. Stx2a does not activate platelets Flow cytometric analysis of Stx2a-treated PRP revealed that there is no increase of CD62 P on the surface of the platelets indicating that platelets are not activated by Stx2a, similar to the negative control with antibody-labeled PRP alone (Fig. 5). In contrast, addition of thrombin 5
International Journal of Medical Microbiology xxx (xxxx) xxx–xxx
L. Knabl et al.
Fig. 2. Significant, but only moderate decrease of AT activity directed against FXa by addition of Stx2a to whole blood samples compared to the HEPES buffer control. Two different concentrations of Stx2a (250 ng/mL or 500 ng/mL) were added to whole blood samples and AT activity directed against FXa (shown in % on the y axis) was measured by BCS XP® system. The results are the means +/- SD of ten separate experiments; *, p < 0.05.
3.8. Absence of platelet aggregation by Stx2a Addition of Stx2a to PRP did not lead to detectable aggregation (data not shown). However, addition of collagen or ADP to PRP pretreated with Stx2a resulted in a rapid and strong aggregation. Stx2a had no effect on platelet aggregation measured with light-transmission aggregometry (LTA). No differences were observed when PBS was used instead of Stx2a. 4. Discussion The first main finding of the present study is that Stx interacts with AT. AT is the most important inhibitor of the fibrin-building plasmatic coagulation process by irreversibly binding thrombin and factor Xa and thereby leading to anticoagulation (Patnaik and Moll, 2008). AT also blocks the active sites of factors IXa, XIa and XIIa by formation of stable equimolar complexes (Damus et al., 1973). The biological relevance of the interaction between Stx and AT is not clear at this point; in view of the fact that only 15% of patients infected with EHEC develop eHUS, it could be speculated that specific host factors, like a deficiency or polymorphism of AT is necessary for the interaction between Stx and AT to become relevant in vivo leading to the development of eHUS. Interestingly we also demonstrated that the extent of binding of Stx2a to antithrombin is pH-dependent, showing a significantly more pronounced binding of Stx2a to AT at a pH range between 4 and 5.5, compared to pH values between 6 and 10. Even though a pH of 4 seems to be quite outside the physiological range, it must be acknowledged, that infected tissue has a pH of 5–6 (Oberdisse et al., 2001; Schmuck et al., 2016). Possible explanations for the higher extent of binding of Stx2a to AT at acidic pH are (i) unfolding of Stx2a at acidic pH which can expose hidden binding sites or can render more accessible the existing binding sites and (ii) dissociation of the toxin allowing the release of the isolated B and A subunits which could be more suitable for binding to AT. Besides binding, AT activity directed against FXa was also demonstrated to be decreased in the presence of Stx2a, although only moderately. In the present study, another interesting observation was made: Stx2a was found to bind to heparin and the binding of heparin to AT was impaired in the presence of Stx2a. Heparin expresses its anticoagulatory action by forming a complex with antithrombin and thereby accelerating the activity of AT by a factor of 1000 (Bjork and Lindahl, 1982). The finding that Stx2a binds to heparin and impairs its binding to AT may partly explain the well-known non-beneficial administration of heparin in patients suffering from eHUS (Diekmann,
Fig. 3. (3a) Representative sensogram out of three experiments for the QCM-D analysis of Stx2a binding to heparin and impairment of binding of AT to heparin. Heparin was added to the QCM-D sensor surface via a polyamine core (PAV). Following addition of Stx2a resulted in a binding of Stx2a to heparin. The subsequent addition of AT to the same surface revealed an impaired binding of AT to Stx2a-bound heparin. (3b) Binding of heparin to AT shown by mass uptake in QCM-D. Heparin was added to the QCM-D sensor surface via a polyamine core (PAV). The addition of AT led to a significant mass uptake, which demonstrates the binding between AT and heparin. (3c) Scale-up of the final part of the sensograms presented in (a) and (b), showing the AT binding to heparin (1) subsequent to addition of Stx2a and (2) without previous addition of Stx2a. The lower mass uptake of AT bound to heparin pre-exposed to Stx2a indicates that AT and Stx2a competes for similar binding sites at the heparin surface.
1980; Van Damme-Lombaerts et al., 1988; Vitacco et al., 1973), however it has to be taken into consideration, that quantities of Stx2a and heparin are varying to a huge extent (pg quantities of toxin, mg quantities of heparin). Furthermore, it may be interesting to investigate the ability of Stx2a to form thrombin-antithrombin complexes in the absence or presence of heparin. Furthermore, functional coagulation analysis performed with ROTEM® showed that treatment of plasma with Stx2a leads to only a mild shortening of clotting time. Although we could demonstrate binding of Stx2a to antithrombin, ROTEM® analyses could not reveal a marked functional consequence of this binding. An interesting perspective might be to perform ROTEM® analysis on plasma samples of patients who developed eHUS. Another main finding of this study is that in contrast to various studies (Cooling et al., 1998; Ghosh et al., 6
International Journal of Medical Microbiology xxx (xxxx) xxx–xxx
L. Knabl et al.
Fig. 4. Significant, but only mild decrease of clotting time in ROTEM® analysis by addition of Stx2a compared to the HEPES buffer control. Various concentrations of Stx2a (10 ng/mL or 100 ng/mL) were added to plasma and time needed for clot formation (clotting time in seconds, y axis) was measured by ROTEM analysis. Analysis was performed with pooled citrated plasma samples. n = 30 and n = 32 indicates the number of test repetitions.
fibrinogen and platelets; regarding coagulation initiation, it cannot differentiate which factor is exactly missing. It is therefore a very insensitive methodology when it comes to detailed coagulation analysis. However, the large number of test repetitions (> 30 times) consistently led to the same results. Another limitation of the study is the range of Stx concentrations used for the experiments. In view of the fact that in any case the concentrations of the potential binding partners of Stx are much higher than the concentration of Stx, only a low percentage of AT would be bound by Stx2a in vivo. There are no reports on the concentration of Stx in plasma of patients suffering from eHUS, and thus, it is difficult to estimate the appropriate amount of Stx required for in vitro studies. In this light the marginal but significant results do not yield conclusive statements and no in vitro study may be able to disprove the idea that Stx2a affects eHUS development by affecting increasing plasmatic coagulation. Furthermore, one must acknowledge though, that this is an in vitro study, which might give a too simplistic view of the pathophysiological processes leading to the development of eHUS. Especially the interaction of Stx2a with endothelial cells and different flow conditions in microvessels, leading to a large heterogeneity of perfusion in the microvasculature in severely ill patients, seem to be events which need to be further investigated. In conclusion, the present study showed that Stx2a binds to AT, but functional analyses demonstrated only a mild change in AT activity and a moderate shortening of clotting time. Furthermore, we demonstrated that Stx2a does not interact with resting platelets in vitro. Importantly, Stx2a binds to heparin and impairs the binding of heparin to AT. This finding could explain why different trials with prophylactic or therapeutic administration of heparin had no influence on thrombotic microangiopathy in eHUS patients.
Fig. 5. Absence of P-selectin (CD62 P) expression as a measure of platelet activation in Stx2a treated-PRP. The vertical axis represents the mean fluorescence intensity (MFI). FACS analysis of platelet activation was performed by incubation of PRP with medium, ADP (weak positive control), thrombin (strong positive control), and various concentrations of Stx2a, Stx2a plus thrombin or Stx2a plus ADP. The results are the means +/- SD of five separate experiments.
2004; Viisoreanu et al., 2000) demonstrating binding and activation of platelets by Stx, we were not able to confirm this effect using life confocal microscopy or FACS analysis or LTA. However, a few points need to be taken into account; first, different types of Stx (Stx1 vs. Stx2a) (Cooling et al., 1998) and second, different preparations of platelets (activated vs. resting platelets) (Ghosh et al., 2004) have been used in the different studies. Our data are in line with Viisoreanu and colleagues (Viisoreanu et al., 2000) who showed an absent platelet aggregation using Stx1 and Stx2a under static and flow conditions. Karpman and colleagues (Karpman et al., 2001) demonstrated aggregation of platelets by Stx1; however, data on platelet preparation are missing. Eight years later the same research group (Stahl et al., 2009) observed activation of platelets by Stx2a in monocyte-platelet- or neutrophil-platelet-aggregates, but not in resting platelets (using whole blood as source of platelets). The observations by Stahl and colleagues (Stahl et al., 2015) of an increase of platelet and leukocyte- derived microvesicles in patients suffering from eHUS led to the hypothesis that platelet activation is not mediated by a direct effect of Stx on platelets but rather due to a Stx-induced release of chemokines and other factors. These findings may explain the non-beneficial effect of a therapeutic platelet inhibitor such as dipyridamol in eHUS patients (Van Damme-Lombaerts et al., 1988). The present study has a few limitations. It must be noted that the reduction in clotting time seen in ROTEM® was statistically significant, but only moderate. ROTEM® is a functional coagulation testing method, designed for bedside use in severely bleeding patients. It can mainly differentiate between a lack of plasmatic coagulation factors or
Acknowledgements The authors are grateful to the Austrian Science Fund (FWF; doctoral programme of excellence HOROS, W1253 to S.C. and R.W.) for financial support. The manuscript is dedicated to our colleague and mentor, Prof. Helge Karch, on the occasion of his 65th birthday. References Ay, C., Vormittag, R., Dunkler, D., Simanek, R., Chiriac, A.L., Drach, J., Quehenberger, P., Wagner, O., Zielinski, C., Pabinger, I., 2009. D-dimer and prothrombin fragment 1 + 2 predict venous thromboembolism in patients with cancer: results from the Vienna cancer and thrombosis study. J. Clin. Oncol. 27, 4124–4129. Bhowmik, D., 2001. Elevated tissue factor levels in children with hemolytic uremic syndrome. Clin. Nephrol. 55, 262. Bjork, I., Lindahl, U., 1982. Mechanism of the anticoagulant action of heparin. Mol. Cell. Biochem. 48, 161–182. Brigotti, M., 2012. The interactions of human neutrophils with shiga toxins and related plant toxins: danger or safety? Toxins (Basel) 4, 157–190. Camerer, E., Kolsto, A.B., Prydz, H., 1996. Cell biology of tissue factor, the principal initiator of blood coagulation. Thromb. Res. 81, 1–41. Chandler, W.L., Jelacic, S., Boster, D.R., Ciol, M.A., Williams, G.D., Watkins, S.L.,
7
International Journal of Medical Microbiology xxx (xxxx) xxx–xxx
L. Knabl et al.
Oberdisse, E., Hackenthal, E., Kukhinsky, K., 2001. Pharmakologie und Toxikologie. Springer. Obrig, T.G., Karpman, D., 2012. Shiga toxin pathogenesis: kidney complications and renal failure. Curr. Top. Microbiol. Immunol. 357, 105–136. Orth, D., Khan, A.B., Naim, A., Grif, K., Brockmeyer, J., Karch, H., Joannidis, M., Clark, S.J., Day, A.J., Fidanzi, S., Stoiber, H., Dierich, M.P., Zimmerhackl, L.B., Würzner, R., 2009. Shiga toxin activates complement and binds factor H: evidence for an active role of complement in hemolytic uremic syndrome. J. Immunol. 182, 6394–6400. Paniccia, R., Priora, R., Liotta, A.A., Abbate, R., 2015. Platelet function tests: a comparative review. Vasc. Health Risk Manag. 11, 133–148. Patnaik, M.M., Moll, S., 2008. Inherited antithrombin deficiency: a review. Haemophilia 14, 1229–1239. Poolpol, K., Orth-Höller, D., Speth, C., Zipfel, P.F., Skerka, C., de Cordoba, S.R., Brockmeyer, J., Bielaszewska, M., Würzner, R., 2014. Interaction of Shiga toxin 2 with complement regulators of the factor H protein family. Mol. Immunol. 58, 77–84. Proesmans, W., 2001. The role of coagulation and fibrinolysis in the pathogenesis of diarrhea-associated hemolytic uremic syndrome. Semin. Thromb. Hemost. 27, 201–205. Rodahl, M., Hook, F., Fredriksson, C., Keller, C.A., Krozer, A., Brzezinski, P., Voinova, M., Kasemo, B., 1997. Simultaneous frequency and dissipation factor QCM measurements of biomolecular adsorption and cell adhesion. Faraday Discuss. 229–246. Sauerbrey, G., 1959. Verwendung von Schwingquarzen zur Wägung dünner Schichten und zur Mikrowägung. Zeitschrift fuer Physik 155, 206–222. Schmuck, C., Engels, B., Schirmeister, T., Reinhold, F., 2016. Chemie für Medizner. Pearson, München. Stahl, A.L., Sartz, L., Nelsson, A., Bekassy, Z.D., Karpman, D., 2009. Shiga toxin and lipopolysaccharide induce platelet-leukocyte aggregates and tissue factor release, a thrombotic mechanism in hemolytic uremic syndrome. PLoS One 4, e6990. Stahl, A.L., Arvidsson, I., Johansson, K.E., Chromek, M., Rebetz, J., Loos, S., Kristoffersson, A.C., Bekassy, Z.D., Morgelin, M., Karpman, D., 2015. A novel mechanism of bacterial toxin transfer within host blood cell-derived microvesicles. PLoS Pathog. 11, e1004619. Sun, D.S., Ho, P.H., Chang, H.H., 2016. Soluble P-selectin rescues viper venom-induced mortality through anti-inflammatory properties and PSGL-1 pathway-mediated correction of hemostasis. Sci. Rep. 6, 35868. Suzuki, H., Yamamoto, N., Tanoue, K., Yamazaki, H., 1987. Glycoprotein Ib distribution on the surface of platelets in resting and activation states: an electron microscope study. Histochem. J. 19, 125–136. Tarr, P.I., 1995. Escherichia coli O157:H7: clinical, diagnostic, and epidemiological aspects of human infection. Clin. Infect. Dis. 20, 1–8. Van Damme-Lombaerts, R., Proesmans, W., Van, D.B., Eeckels, R., Binda ki, M.P., Mercieca, V., Vlietinck, R., Vermylen, J., 1988. Heparin plus dipyridamole in childhood hemolytic-uremic syndrome: a prospective, randomized study. J. Pediatr. 113, 913–918. Viisoreanu, D., Polanowska-Grabowska, R., Suttitanamongkol, S., Obrig, T.G., Gear, A.R., 2000. Human platelet aggregation is not altered by Shiga toxins 1 or 2. Thromb. Res. 98, 403–410. Vitacco, M., Sanchez, A.J., Gianantonio, C.A., 1973. Heparin therapy in the hemolyticuremic syndrome. J. Pediatr. 83, 271–275. Walsh, P.R., Johnson, S., 2018. Treatment and management of children with haemolytic uraemic syndrome. Arch. Dis. Child. 103, 285–291. Zhang, W., Bielaszewska, M., Pulz, M., Becker, K., Friedrich, A.W., Karch, H., Kuczius, T., 2008. New immuno-PCR assay for detection of low concentrations of shiga toxin 2 and its variants. J. Clin. Microbiol. 46, 1292–1297.
Igarashi, T., Tarr, P.I., 2002. Prothrombotic coagulation abnormalities preceding the hemolytic-uremic syndrome. N. Engl. J. Med. 346, 23–32. Cooling, L.L., Walker, K.E., Gille, T., Koerner, T.A., 1998. Shiga toxin binds human platelets via globotriaosylceramide (Pk antigen) and a novel platelet glycosphingolipid. Infect. Immun. 66, 4355–4366. Damus, P.S., Hicks, M., Rosenberg, R.D., 1973. Anticoagulant action of heparin. Nature 246, 355–357. Diekmann, L., 1980. [Treatment of the hemolytic-uremic syndrome with streptokinase and heparin (author's transl)]. Klin. Padiatr. 192, 430–435. Ehrlenbach, S., Rosales, A., Posch, W., Wilflingseder, D., Hermann, M., Brockmeyer, J., Karch, H., Satchell, S.C., Würzner, R., Orth-Höller, D., 2013. Shiga toxin 2 reduces complement inhibitor CD59 expression on human renal tubular epithelial and glomerular endothelial cells. Infect. Immun. 81, 2678–2685. Ganter, M.T., Hofer, C.K., 2008. Coagulation monitoring: current techniques and clinical use of viscoelastic point-of-care coagulation devices. Anesth. Analg. 106, 1366–1375. George, J.N., Nester, C.M., 2014. Syndromes of thrombotic microangiopathy. N. Engl. J. Med. 371, 1847–1848. Ghosh, S.A., Polanowska-Grabowska, R.K., Fujii, J., Obrig, T., Gear, A.R., 2004. Shiga toxin binds to activated platelets. J. Thromb. Haemost. 2, 499–506. Gorlinger, K., Bhardwaj, V., Kapoor, P.M., 2016. Simulation in coagulation testing using rotational thromboelastometry: a fast emerging, reliable point of care technique. Ann. Card. Anaesth. 19, 516–520. Guessous, F., Marcinkiewicz, M., Polanowska-Grabowska, R., Kongkhum, S., Heatherly, D., Obrig, T., Gear, A.R., 2005. Shiga toxin 2 and lipopolysaccharide induce human microvascular endothelial cells to release chemokines and factors that stimulate platelet function. Infect. Immun. 73, 8306–8316. Kamitsuji, H., Nonami, K., Murakami, T., Ishikawa, N., Nakayama, A., Umeki, Y., 2000. Elevated tissue factor circulating levels in children with hemolytic uremic syndrome caused by verotoxin-producing E. coli. Clin. Nephrol. 53, 319–324. Karch, H., Tarr, P.I., Bielaszewska, M., 2005. Enterohaemorrhagic Escherichia coli in human medicine. Int. J. Med. Microbiol. 295, 405–418. Karpman, D., Papadopoulou, D., Nilsson, K., Sjogren, A.C., Mikaelsson, C., Lethagen, S., 2001. Platelet activation by Shiga toxin and circulatory factors as a pathogenetic mechanism in the hemolytic uremic syndrome. Blood 97, 3100–3108. Kozek-Langenecker, S.A., Afshari, A., Albaladejo, P., Santullano, C.A., De, R.E., Filipescu, D.C., Fries, D., Gorlinger, K., Haas, T., Imberger, G., Jacob, M., Lance, M., Llau, J., Mallett, S., Meier, J., Rahe-Meyer, N., Samama, C.M., Smith, A., Solomon, C., Van der Linden, P., Wikkelso, A.J., Wouters, P., Wyffels, P., 2013. Management of severe perioperative bleeding: guidelines from the European Society of Anaesthesiology. Eur. J. Anaesthesiol. 30, 270–382. Lee, B.C., Mayer, C.L., Leibowitz, C.S., Stearns-Kurosawa, D.J., Kurosawa, S., 2013. Quiescent complement in nonhuman primates during E coli Shiga toxin-induced hemolytic uremic syndrome and thrombotic microangiopathy. Blood 122, 803–806. Levi, M., van der Poll, T., 2005. Two-way interactions between inflammation and coagulation. Trends Cardiovasc. Med. 15, 254–259. Lu, Q., Malinauskas, R.A., 2011. Comparison of two platelet activation markers using flow cytometry after in vitro shear stress exposure of whole human blood. Artif. Organs 35, 137–144. Mayer, C.L., Parello, C.S., Lee, B.C., Itagaki, K., Kurosawa, S., Stearns-Kurosawa, D.J., 2015. Pro-coagulant endothelial dysfunction results from EHEC Shiga toxins and host damage-associated molecular patterns. Front. Immunol. 6, 155. Nevard, C.H., Jurd, K.M., Lane, D.A., Philippou, H., Haycock, G.B., Hunt, B.J., 1997. Activation of coagulation and fibrinolysis in childhood diarrhoea-associated haemolytic uraemic syndrome. Thromb. Haemost. 78, 1450–1455.
8
Contents lists available at ScienceDirect
International Journal of Medical Microbiology journal homepage: www.elsevier.com/locate/ijmm
Shiga toxin 2a binds antithrombin and heparin, but does not directly activate platelets Ludwig Knabla,1, Michael Berktolda,1, Osama A. Hamadb, Karin Fromellb, Sneha Chatterjeea, Cornelia Spetha, Heribert Talaszc, Katharina Lindnere, Martin Hermanne, ⁎ ⁎ Kristina Nilsson-Ekdahlb, Bo Nilssonb, Werner Streifd, Judith Martinie, ,1, Reinhard Würznera, ,1, ⁎ Dorothea Orth-Höllera, ,1 a
Division of Hygiene and Medical Microbiology, Medical University of Innsbruck, Schöpfstrasse 41, 6020 Innsbruck, Austria Department of Immunology, Genetics and Pathology, Rudbeck Laboratory, Uppsala University, 75185 Uppsala, Sweden Division of Clinical Biochemistry, Biocentre, Medical University of Innsbruck, Innrain 80, 6020 Innsbruck, Austria d Department of Paediatrics I, Medical University of Innsbruck, Anichstrasse 35, 6020 Innsbruck, Austria e Univ. Clinic of Anesthesia and Intensive Care Medicine, Medical University of Innsbruck, Anichstrasse 35, 6020 Innsbruck, Austria b c
A R T I C LE I N FO
A B S T R A C T
Keywords: Stx2a Hemolytic uremic syndrome Plasmatic coagulation Antithrombin
Escherichia coli-induced hemolytic uremic syndrome (eHUS) is a life-threatening complication of infection with Shiga toxin (Stx), in particular Stx2a-producing Escherichia coli. Enhanced coagulation activation with formation of microthrombi seems to be a key event in development of eHUS. Platelet activation has been postulated as a possible, but controversially debated mechanism. The present study investigated the effect of Stx2a on plasmatic coagulation and platelets. Binding studies were initially performed with ELISA and co-immunoprecipitation and supported by quartz crystal microbalance with dissipation monitoring (QCM-D). Antithrombin (AT) activity was measured using the automated BCS XP® system. ROTEM® was used for functional coagulation testing. Platelet binding and activation was studied with FACS and light-transmission aggregometry. We found binding of Stx2a to AT, an important inhibitor of blood coagulation, but only a mild albeit significant reduction of AT activity against FXa in the presence of Stx2a. QCM-D analysis also showed binding of Stx2a to heparin and an impaired binding of AT to Stx2a-bound heparin. ROTEM® using Stx2a-treated plateletpoor plasma revealed a significant, but only moderate shortening of clotting time. Neither binding nor activation of platelets by Stx2a could be demonstrated. In summary, data of this study suggest that Stx2a binds to AT, but does not induce major effects on plasmatic coagulation. In addition, no interaction with platelets occurred. The well-known non-beneficial administration of heparin in eHUS patients could be explained by the interaction of Stx2a with heparin.
1. Introduction Hemolytic uremic syndrome (HUS) caused by infection with Shiga toxin (Stx), in particular Stx2a-producing enterohemorrhagic Escherichia coli (EHEC) is the leading cause of acute kidney injury (AKI) in pediatric patients (Karch et al., 2005). In the full clinical picture of
EHEC-induced HUS (recently termed eHUS) AKI is accompanied by microangiopathic hemolytic anemia and thrombocytopenia (Tarr, 1995). It is well accepted that hemostatic abnormalities play an important role in the development and progression of eHUS, especially in the development of AKI (Lee et al., 2013; Nevard et al., 1997; Proesmans,
Abbreviations: eHUS, Escherichia coli-induced hemolytic uremic syndrome; Stx, Shiga toxin; EHEC, Enterohemorrhagic Escherichia coli; AT, antithrombin; AKI, acute kidney injury; t-PA antigen, tissue plasminogen activator antigen; PAI-1, t-PA-plasminogen activator inhibitor type 1; TF, tissue factor; vWF, von Willebrand factor; NOACs, new oral anticoagulants; PRP, platelet-rich plasma; APC, allophycocyanin; PE, phycoerythrin; LTA, light-transmission aggregometry; ROTEM®, rotational thromboelastometry; MFI, mean fluorescence intensity; CT, clotting time; SAP, serum amyloid P; QCM-D, quartz crystal microbalance with dissipation monitoring; PAV, polyamine; CHC, charged heparin conjugate ⁎ Corresponding authors. E-mail addresses: [email protected] (J. Martini), [email protected] (R. Würzner), [email protected] (D. Orth-Höller). 1 Contributed equally to this work. https://doi.org/10.1016/j.ijmm.2018.07.008 Received 17 May 2018; Received in revised form 13 July 2018; Accepted 22 July 2018 1438-4221/ © 2018 Published by Elsevier GmbH.
Please cite this article as: Knabl, L., International Journal of Medical Microbiology (2018), https://doi.org/10.1016/j.ijmm.2018.07.008
International Journal of Medical Microbiology xxx (xxxx) xxx–xxx
L. Knabl et al.
(Linz, Austria); heparin from G.L. Pharma (Lannach, Austria); rabbit anti-FXa ab from US Biological Life Sciences (Salem, USA); sheep antihuman AT ab from Acris (Herford, Germany); mouse anti-Stx2 ab from Hycult Biotechnology (Uden, The Netherlands); goat anti-mouse APconjugated IgG, goat anti-rabbit AP-conjugated IgG, donkey anti-sheep IgG and rabbit anti-human AT abs, p-nitrophenyl phosphate from Sigma Aldrich. Platelet concentrates were obtained from the blood bank of the University Clinic Innsbruck. Platelets were checked to be not-activated by measuring CD62 P with FACS as detailed below. Blood samples for the preparation of plasma or platelet-rich plasma (PRP) were – after informed consent and with an existing ethical vote – collected from healthy volunteers. Blood was drawn in 10 mL tubes anticoagulated with buffered 3.2% citrate. PRP for FACS analysis and light-transmission aggregometry (LTA) were obtained by centrifugation (135 g for 15 min) and used within 4 h following preparation. For rotational thromboelastometry (ROTEM®) analysis and co-immunoprecipitation citrated blood was centrifuged at 2500 g for 10 min. The obtained plasma was then pooled, swiftly portioned into identical aliquots of 1.5 mL per microcentrifuge tube and stored at −78 °C until analyzed.
2001). Patients with eHUS have an occlusion of the renal microvasculature due to endothelial damage and subsequent formation of microthrombi leading to ischemic kidney damage (George and Nester, 2014). They usually present with a strong rise of laboratory parameters indicating global coagulation activation such as increased levels of DDimer, tissue plasminogen activator antigen (t-PA antigen), t-PA-plasminogen activator inhibitor type 1 (PAI-1) complex and prothrombin fragments 1 and 2 (Ay et al., 2009; Chandler et al., 2002). Plasma levels of tissue factor (TF), the most important initiator of coagulation in vivo (Camerer et al., 1996), are markedly elevated in children with eHUS, through Stx induced release of TF (Bhowmik, 2001; Kamitsuji et al., 2000; Stahl et al., 2009). Additionally, there is also evidence for Stxinduced alterations in the protein C pathway (Mayer et al., 2015), an important inhibitor of coagulation. On the other hand, there are data suggesting a major role of platelets and von Willebrand factor (vWF) in the development of eHUS; a widespread activation of platelets either through Stx itself or by the release of chemokines and other factors from endothelial cells has been repeatedly affirmed (Guessous et al., 2005; Karpman et al., 2001; Obrig and Karpman, 2012; Stahl et al., 2015; Sun et al., 2016). The exact pathomechanism however, and the question whether this hypercoagulopathic status is mainly initiated via activation of platelets or the plasmatic hemostasis or a combination of both, is currently under debate. Thrombus formation is the result of a highly complex interaction between different physiological systems and extensive cross talk between coagulation and inflammation exists (Levi and van der Poll, 2005). Furthermore, it has been shown that complement, an important part of the innate immune system, might be involved in the pathogenesis of eHUS (Ehrlenbach et al., 2013; Orth et al., 2009; Poolpol et al., 2014). Additionally, it is notable that only 15% of patients infected with EHEC develop eHUS; it can be assumed that specific host related factors, such as polymorphisms or deficiencies of factors involved in hemostasis may lead to a clinically relevant change of coagulation after infection with EHEC. Currently there is no consensus about how to treat coagulation abnormalities in affected patients or if they should be treated at all (Walsh and Johnson, 2018). In view of the clinically most relevant new oral anticoagulants (NOACs) with direct inhibition of either thrombin or factor Xa, new treatment strategies may emerge which could prevent or mitigate kidney failure in eHUS patients. This study was designed to investigate which steps of the coagulation process are mainly disturbed by the presence of Stx; it was therefore assessed whether Stx2a activates platelets and in a second set of experiments we evaluated the effect of Stx2a on various factors of the plasmatic coagulation.
2.2. Analysis of binding of Stx2a to either antithrombin (AT), thrombin or protein C and binding of Stx2a-bound AT to FXa by ELISA To establish whether Stx2a binds to AT, microtiter plates were coated with Stx2a (1 μg per well) in 100 μL coating buffer (12,430 mg NaHCO3, 5510 mg Na2CO3 in 1 l aqua dest.) overnight at 4 °C. After blocking with 1% (w/v) gelatine, each well was incubated with 1 μg of AT for 4 h at 37 °C. After additional washing steps, bound AT was detected with a primary rabbit anti-AT antibody (1:1000) followed by a goat anti-rabbit AP-conjugated IgG (1:1000), both diluted in PBS-T. The binding was detected with p-nitrophenyl phosphate as the substrate. AT alone was used as a positive control. As negative controls we coated BSA or serum amyloid P (SAP) (1 μg per well) in place of Stx2a followed by addition of AT (1 μg for 4 h) and detection with the above mentioned rabbit anti-AT antibody and goat anti-rabbit IgG. Interaction of Stx2a with AT was further analyzed at different pH values (pH 4, pH 4.5, pH 5, pH 5.5, pH 6, pH 10). After blocking, AT (1 μg per well) was added to the wells in 100 μL PBS-T adjusted to respective pH values. The remaining steps were performed as described above. To analyze the binding of Stx2a-bound AT to FXa, 1 μg of FXa was immobilized in 100 μL coating buffer. Simultaneously, 2.5 μg Stx2a and 2.5 μg AT were dissolved in 250 μL PBS-T and incubated in a micro centrifuge tube for 20 h at 37 °C. Afterwards, 100 μL of the Stx2a/ATsuspension was added to the immobilized FXa for 1 h at 37 °C. After washing, bound Stx2a/AT was detected with a primary rabbit antihuman AT antibody (1:1000) and a secondary anti-rabbit AP-conjugated IgG antibody (1:1000). Bound Stx2a/AT was detected by development with p-nitrophenyl phosphate as substrate. Controls were performed with BSA instead of Stx2a. To evaluate whether Stx2a binds to protein C or thrombin, the microtiter plates were coated with the respective proteins (1 μg per well, each) in 100 μL coating buffer. After blocking, Stx2a (1 μg) was added for 4 h at 37 °C. Bound Stx2a was detected with a primary mouse antiStx2 antibody (1:100) followed by an anti-mouse AP-conjugated IgG (1:1000), both diluted in PBS-T. For negative control BSA was added to protein C or thrombin in place of Stx2a, for positive control Stx2a (1 μg per well) alone was coated. Photometric readouts were done at 415 and 490 nm.
2. Materials and methods 2.1. Purification of Stx2a, proteins, antibodies, platelet concentrates platelet-rich plasma (PRP) and plasma Purification of Stx2a was done as described elsewhere (Zhang et al., 2008) and the purity level was evaluated by SDS-PAGE. Lipopolysaccharide contamination was determined to be 5.88 pg/μg using a method described elsewhere (Brigotti, 2012). Stx2a was labeled with FITC using an Oyster-488 antibody labeling kit (Luminartis, Münster, Germany). LPS from E. coli serotype O55B5 was obtained by Charles River Laboratories Inc. (Charleston, USA). The allophycocyanin (APC)-conjugated CD42b antibody (ab) and the phycoerythrin (PE)- conjugated CD62 P ab were purchased from BD Biosciences (San Jose, USA) and BioLegend (San Diego, USA), respectively; antithrombin (AT), factor Xa and protein C from Calbiochem (Darmstadt, Germany); thrombin and GPR peptide from Sigma Aldrich (Taufkirchen, Germany); ADP from Sigma Aldrich (Taufkirchen, Germany) and Chrono-Log (Haverton, USA); CaCl2 from Siemens (Munich, Germany); collagen from Nycomed
2.3. Analysis of binding of Stx2a to AT by co-immunoprecipitation To confirm the results of the ELISA test for Stx2a-AT-binding, coimmunoprecipitation using a kit from Thermo Fisher Scientific 2
International Journal of Medical Microbiology xxx (xxxx) xxx–xxx
L. Knabl et al.
(Waltham, USA) was performed. Anti-Stx2a antibody (100 μL) was coupled to the agarose resin as described in the manufacturer’s protocol. Simultaneously, 10 μg Stx2a and 10 μg AT or 20 μL plasma or plasma (20 μL) spiked with AT (10 μg) were incubated at 37 °C and continuous stirring for 4 h. After dilution in Dulbecco’s modified PBS at a ratio of 1:7.5, the suspensions were applied to the anti-Stx2a antibody-coupled gel and incubated overnight at 37 °C under gentle agitation. After washing, bound proteins were eluted according to the manufacturer’s instructions. Eluted protein complexes were analyzed by western blot using a 16% (w/v) gel (Bio-Rad, Vienna, Austria) under reducing conditions. Detection was done with a sheep anti-human ATantibody (1:3000) and a secondary anti-sheep IgG antibody (1:2000). As positive control AT was loaded on the gel, for the negative control BSA was loaded. Bands were visualized using the Clarity ECL blotting substrate (Bio-Rad) and the ImageQuant LAS4000 (GE Healthcare Life Sciences, Chalfont St Giles, UK).
2.5. Determination of AT activity in the presence of Stx2a To analyze AT activity directed against FIIa and FXa, whole blood was incubated with Stx2a (250 or 500 ng/ml) in a volume of 300 μL. For negative control HEPES buffer was added in place of Stx2a. The samples were incubated for 4 h at 37 °C. After incubation samples were centrifuged 15 min. at 1000 g at room temperature. AT activity was determined using the automated BCS XP® system (Siemens Healthcare Diagnostics GmbH, Eschborn, Germany). 2.6. Determination of plasmatic coagulation activation in the presence of Stx2a by rotational thromboelastometry (ROTEM®) Functional coagulation analysis was performed by ROTEM® (Tem International GmbH, Munich, Germany), a point-of-care monitoring system for severely bleeding patients (Kozek-Langenecker et al., 2013). Unlike conventional laboratory coagulation parameters (e.g. PT, aPTT), this technique analyses the entire clotting process from coagulation start until clot formation and determines clot firmness and clot stability over time. The exact measurement technique is described in detail elsewhere (Gorlinger et al., 2016). In this study, functional coagulation analysis was performed in pooled plasma. A continuous viscoelastic profile of clot formation was obtained (Ganter and Hofer, 2008). Due to our study hypothesis, main focus was on measurements of clotting time (CT). In order to detect the minutest details in changes of clot formation initiation, only calcium chloride was added to the plasma sample (star-tem® reagent), performing a so called NATEM® test. NATEM tests started with pipetting 20 μL star-tem reagent into the plastic cup, promptly followed by 300 μL pooled plasma. During the initialized ROTEM sequence of mixing by drawing and releasing the plasma probe with the pipette, 30 μL of the corresponding dose to 10 ng/mL or 100 ng/mL of HEPES-diluted Stx2a were additionally pipetted into the sample. Coagulation process starts immediately and is fully represented by the data graph obtained over 30 min. Control measurements were performed by adding 30 μL HEPES buffer to the plasma sample instead of Stx2a to ensure comparable hemodilution of the samples.
2.4. Evaluation of Stx2a binding to AT, heparin, factor Xa by quartz crystal microbalance with dissipation monitoring (QCM-D) QCM-D is a very useful technique for measuring biomolecule interactions at surfaces. It was therefore used to follow the interaction between Stx2a and AT, heparin and FXa. The principle of QCM-D has been described previously (Rodahl et al., 1997). Briefly, a QCM-D consists of a thin quartz disc clamped between two electrodes. The disc is set to oscillation by applying a voltage over the electrodes. The resonance frequency (f) of the oscillating quartz crystal depends on the mass assembled on its sensor surface. For very thin and rigid films, the decrease in frequency will be proportional to mass of the assembled layer and the mass can be calculated according to the Sauerbrey relation (Sauerbrey, 1959). The QCM-D measurements were performed using a QCM-D Omega Auto (Biolin Scientific AB, Göteborg, Sweden). All QCM-D experiments were carried out at 37 °C with a flow rate of 20 μL/min. PBS was used for priming and rinsing of the surfaces before, in between and after the addition of the reagents. 2.4.1. Evaluation of Stx2a binding to AT Two clean polystyrene coated QCM-D sensors were mounted in the QCM-D instrument. After priming with buffer, Stx2a (10 μg/mL) was allowed to adsorb for 30 min to one of the sensors, while AT (10 μg/mL) was adsorbed to the other. After wash and blocking (0.5% BSA) AT (10 μg/mL) was added to the first sensor and Stx2a (10 μg/mL) was added to the other. Binding was allowed to proceed for 20 min followed by rinsing with PBS.
2.7. Assessment of Stx2a binding to platelets by real time live confocal microscopy Platelets were placed in 8-well chambered cover glasses (Nalge Nunc International, Rochester, USA) and incubated with FITC-labeled Stx2a (20 ng/mL) for 4 h at 37 °C. The platelets were counterstained using WGA-AlexaFluor647 dye (Molecular Probes, Eugene, USA). Vero cells loaded with Stx2a (20 ng/mL) served as positive control. Images were taken at time points 0, 30 min, 1 h, 2 h and 4 h. Real Time Live Confocal Microscopy was performed with a spinning disk confocal system (Ultra VIEW VoX; Perkin Elmer, Waltham, USA) connected to a Zeiss AxioObserver Z1 microscope (Zeiss, Oberkochen, Germany). Images were acquired with the Volocity software (Perkin Elmer) using a 63x oil immersion objective with a numerical aperture of 1.4.
2.4.2. Evaluation of Stx2a binding to heparin Again two parallel experiments were performed using two polystyrene QCM-D sensors. Both sensors were first coated with positively charged amino groups i.e. polyamine (PAV) followed by the negatively charged heparin conjugate (CHC) according to manufacturer’s instructions (Corline Systems AB). The surfaces were rinsed and Stx2a (10 μg/ mL) was added to one of the surfaces. After incubation the surface was again rinsed and then AT (10 μg/mL) was added to the same surface. For comparison, AT, which is well known to bind to heparin, was added directly to the second heparin coated surface without the previous addition of Stx2a.
2.8. FACS analyses of Stx2a-mediated platelet activation and Stx2a/ platelet binding The antigen CD42b is present on both resting and activated platelets and was therefore used for platelet identification (Suzuki et al., 1987). To evaluate whether Stx2a enhanced platelet activity, PRP was incubated for 90 min at 37 °C with various concentrations of purified Stx2a (1 ng/mL, 10 ng/mL, 100 ng/mL). Platelet activation and subsequent secretion of α-granules was evaluated by FACS quantification of CD62 P (P-selectin) (Lu and Malinauskas, 2011). To check whether Stx2a interferes with ADP- or thrombin-mediated platelet stimulation, PRP was either mock-treated or pre-incubated with purified Stx2a for 90 min, followed by addition of either 0.1 U/mL
2.4.3. Evaluation of Stx2a binding to coagulation factor Xa The polystyrene sensor surfaces were first coated with anti-FXaantibody (10 μg/mL) and rinsed with buffer. This was followed by addition of FXa alone (1.7 × 10−7 M) or one of the following suspensions: FXa + AT (molar ratio 1:1), FXa + AT + Stx2a (1:1:1), FXa + AT + Stx2a (1:1:2) or FXa + Stx2a (1:1). 3
International Journal of Medical Microbiology xxx (xxxx) xxx–xxx
L. Knabl et al.
thrombin or 100 μM ADP. Again, platelet activity was quantified by FACS using CD62 P as marker. To inhibit fibrin polymerization 1.25 mM GPR peptide was added to PRP. Further experiments aimed to study whether Stx2a is bound to and/ or internalized by platelets. For that purpose PRP was incubated for 90 min with different concentrations (1 ng/mL, 10 ng/mL, 100 ng/mL) of FITC-labeled Stx2a, followed by FACS analysis. All experiments were also performed with whole blood instead of PRP. All FACS measurements were done by using FACScan apparatus (Becton Dickinson, Franklin Lakes, USA). Results were given as the mean fluorescence intensity (MFI).
directed against FIIa showed no difference when Stx2a was added compared to the negative control (data not shown). 3.3. Stx2a binding to heparin The Stx2a binding to a heparinized surface was measured with QCM-D. The recorded mass uptake of Stx2a to the heparinized surface was 140 ng/cm2 followed by 37 ng/cm2 of AT (Fig. 3a) whereas 120 ng/cm2 of AT was bound to the surface without the previous addition of Stx2a (Fig. 3b). These results clearly show that Stx2a binds to heparin and also occupies several of the AT binding sites at the heparin surface (Fig. 3c). In addition, this indicates that Stx2a has the same or competing binding sites for both AT and heparin, since AT clearly bound to Stx2a adsorbed directly to the surface in the absence of heparin (Fig. 1c).
2.9. Determination of Stx2a-mediated platelet aggregation by LTA LTA according to the method of Born was performed with the PAP4 (Mölab, Hilden, Germany) aggregometer to demonstrate the effects of Stx2a on platelet aggregation. The principle is described elsewhere (Paniccia et al., 2015). Stx2a (2 ng/mL, 1 μg/mL and 2 μg/mL) was added to PRP (platelet count adjusted to 250,000 platelets/μL) and aggregation tracing was monitored for 30 min. At this point, aggregation was triggered using collagen (final concentration 4 μg/mL) and ADP (final concentration 1.75 μM/L) to ensure maximum aggregation. In order to exclude any effects which may be caused by dilution, all experiments were repeated with PBS instead of Stx2a.
3.4. Stx2a does not bind to protein C, thrombin or FXa and has no influence on the FXa-binding to AT A microtiter plate was coated with protein C, thrombin, or FXa and incubated with either Stx2a or in case of FXa also with an AT/Stx2a suspension. The optical readout revealed that Stx2a does not bind to protein C, thrombin, or FXa and the binding of Stx2a to AT does not lead to alterations of AT binding to FXa. Thus, the binding capability of the Stx2a/AT-complex to FXa was not reduced compared to controls (data not shown). This was also confirmed with QCM-D where FXa was binding to AT. In contrast, neither Stx2a nor Stx2a/AT bound to FXa. From these results it could also be concluded that the presence of Shiga toxins did not inhibit the FXa binding to AT (data not shown).
2.10. Statistical analyses The results were analyzed by the use of GraphPad Prism (version 5) software. Student's t test was performed to compare the paired means of the two measurement groups. P values of < 0.05 were considered significant. P values of < 0.01 were considered highly significant.
3.5. Shortening of clotting time by Stx2a 3. Results
Functional coagulation tests performed with ROTEM® showed that addition of two different concentrations of Stx2a (10 ng/mL and 100 ng/mL) to pooled plasma samples resulted in a significant, but only moderate shortening of the clotting time (CT) (Fig. 4). Clotting time was 622 ± 94 s in pooled plasma samples diluted with HEPES buffer without Stx2a (n = 33; n = number of test repetitions with pooled plasma samples); CT was significantly shorter in plasma samples incubated with 10 ng/mL Stx2a (559 ± 91 s; n = 32) compared to the HEPES controls (p < 0.01 vs. HEPES controls); plasma samples incubated with 100 ng/mL Stx2a (n = 30) did not result in a more pronounced shortening (574 ± 77 s; p < 0.05). When adding LPS to plasma at the amount present in Stx2a (corresponding to the highest toxin concentration) there was no significant difference in CT compared to plasma only (data not shown). All other ROTEM® parameters (CFT, MCF and ML) were not significantly different and there was no difference between the groups.
3.1. Stx2a binds to AT (by ELISA, co-immunoprecipitation and QCM-D) To evaluate whether Stx2a binds to AT, ELISA tests were performed as screening assay. AT incubated with Stx2a showed a significantly higher OD value than both negative controls indicating a binding between AT and Stx2a (Fig. 1a). Moreover, it was shown that the extent of binding of Stx2a to AT was different when the pH in the test system was changed. At a pH range between 4 and 5.5 the binding of Stx2a to AT was significantly more pronounced compared to pH values between 6 and 10 (Fig. 1a). Co-immunoprecipitation confirmed the binding of Stx2a to AT which was previously observed by ELISA. In the binding control and in elution fractions 1 and 2, no band could be detected in the Western blot whereas in elution fraction 3 a clear band was visible at around 58 kDa indicating a binding of AT to the resin-bound Stx2a (Fig. 1b). The Stx2a binding to AT was finally corroborated by QCM-D. The sensogram showing the recorded mass uptake during the whole process is shown in Fig. 1c. There are interactions between Stx2a and AT, which appear more distinctly when Stx2a sits at the surface followed by the addition of AT, since 90 ng/cm2 of AT was bound to Stx2a adsorbed to the surface (50 ng/cm2) which gives a ratio of 2 mol AT captured/mole of Stx2a and only 20 ng/cm2 Stx2a was captured by AT at the surface (25 ng/cm2) corresponding to 0.7 mol of Stx2a bound/mole AT at the surface.
3.6. Stx2a does not bind to platelets Binding of FITC-labeled Stx2a to platelets was investigated by real time live confocal microscopy. Fluorescence was neither observed on the surface nor inside the platelets at any time point, suggesting that there is no binding of Stx2a to platelets. Additionally, no morphological changes in platelets were detected, implying an absent activation of platelets. In contrast, a strong fluorescence signal was found inside and on the membrane of Vero cells (serving as positive control) as a result of significant binding and uptake of Stx2a (data not shown). The observation that Stx2a does not bind to platelets was confirmed by FACS analyses. Incubation of PRP or whole blood with FITC-labeled Stx2a (1, 10 and 100 ng/mL) for 90 min did not lead to an increase in the fluorescence signal on the surface or inside the platelets compared to the negative control with medium only (data not shown).
3.2. AT activity directed against FXa is decreased in the presence of Stx2a Addition of two different concentrations of Stx2a (250 ng/mL and 500 ng/mL) to whole blood samples resulted in a significant but only moderate decrease of AT activity directed against FXa compared to the negative control with HEPES buffer instead of Stx2a (Fig. 2). AT activity 4
International Journal of Medical Microbiology xxx (xxxx) xxx–xxx
L. Knabl et al.
Fig. 1. (1a) Most prominent binding of Stx2a to AT measured by ELISA at a pH range between 4 and 5.5. Optical density (OD) was significantly higher at all pH values when AT (1 μg) was incubated with Stx2a (1 μg) compared to negative controls, BSA and serum amyloid P (SAP). For the positive control AT (1 μg) was coated. The results are the means +/- SD of five separate experiments; *, p < 0.05; **, p < 0.01; ****, p < 0.0001. (1b) Binding of Stx2a to AT by co-immunoprecipitation. Incubations of Stx2a (10 μg) and AT (10 μg) or plasma (20 μL, as natural source of AT) or plasma spiked with AT (10 μg, to increase the naturally occurring AT in plasma) were applied to an anti-Stx2a ab coupled agarose resin. All three elution fractions of the three experiments were analyzed by western blot using an anti-AT ab. Fractions 1 and 2 did not show any bands (data not shown). All three fractions 3, lanes 1 (Stx2a and AT) and 3 (Stx2a and plasma spiked with AT), and to a lesser extent lane 2 (Stx2a and plasma), showed bands of approx. 58 kDa (indicated by an arrow), representing AT, as shown in the positive control (lane 4, loaded with AT only). For the negative control (lane 5) BSA was loaded on the gel. M, marker. The results show a representative experiment of three. (1c) QCM-D measurement of the interaction between AT and Stx2a. The gradual buildup of several layers on a surface can be followed in real time using QCM-D. This was done by first monitoring the adsorption of AT (grey line) and Stx2a (black line) to a polystyrene coated sensor surface (1), followed by blocking (2), and finally the binding of Stx2a (grey line) or AT (black line) (3). The measurement was performed at 37 °C and the sensor surfaces were rinsed with PBS before, in between and after addition of each component.
or ADP resulted in a substantial increase of activated platelets, whereas the activation was more pronounced by the presence of thrombin than ADP. Simultaneous addition of Stx2a did not influence the platelet activation induced by thrombin or ADP alone. Similar results were obtained with whole blood instead of PRP (data not shown).
3.7. Stx2a does not activate platelets Flow cytometric analysis of Stx2a-treated PRP revealed that there is no increase of CD62 P on the surface of the platelets indicating that platelets are not activated by Stx2a, similar to the negative control with antibody-labeled PRP alone (Fig. 5). In contrast, addition of thrombin 5
International Journal of Medical Microbiology xxx (xxxx) xxx–xxx
L. Knabl et al.
Fig. 2. Significant, but only moderate decrease of AT activity directed against FXa by addition of Stx2a to whole blood samples compared to the HEPES buffer control. Two different concentrations of Stx2a (250 ng/mL or 500 ng/mL) were added to whole blood samples and AT activity directed against FXa (shown in % on the y axis) was measured by BCS XP® system. The results are the means +/- SD of ten separate experiments; *, p < 0.05.
3.8. Absence of platelet aggregation by Stx2a Addition of Stx2a to PRP did not lead to detectable aggregation (data not shown). However, addition of collagen or ADP to PRP pretreated with Stx2a resulted in a rapid and strong aggregation. Stx2a had no effect on platelet aggregation measured with light-transmission aggregometry (LTA). No differences were observed when PBS was used instead of Stx2a. 4. Discussion The first main finding of the present study is that Stx interacts with AT. AT is the most important inhibitor of the fibrin-building plasmatic coagulation process by irreversibly binding thrombin and factor Xa and thereby leading to anticoagulation (Patnaik and Moll, 2008). AT also blocks the active sites of factors IXa, XIa and XIIa by formation of stable equimolar complexes (Damus et al., 1973). The biological relevance of the interaction between Stx and AT is not clear at this point; in view of the fact that only 15% of patients infected with EHEC develop eHUS, it could be speculated that specific host factors, like a deficiency or polymorphism of AT is necessary for the interaction between Stx and AT to become relevant in vivo leading to the development of eHUS. Interestingly we also demonstrated that the extent of binding of Stx2a to antithrombin is pH-dependent, showing a significantly more pronounced binding of Stx2a to AT at a pH range between 4 and 5.5, compared to pH values between 6 and 10. Even though a pH of 4 seems to be quite outside the physiological range, it must be acknowledged, that infected tissue has a pH of 5–6 (Oberdisse et al., 2001; Schmuck et al., 2016). Possible explanations for the higher extent of binding of Stx2a to AT at acidic pH are (i) unfolding of Stx2a at acidic pH which can expose hidden binding sites or can render more accessible the existing binding sites and (ii) dissociation of the toxin allowing the release of the isolated B and A subunits which could be more suitable for binding to AT. Besides binding, AT activity directed against FXa was also demonstrated to be decreased in the presence of Stx2a, although only moderately. In the present study, another interesting observation was made: Stx2a was found to bind to heparin and the binding of heparin to AT was impaired in the presence of Stx2a. Heparin expresses its anticoagulatory action by forming a complex with antithrombin and thereby accelerating the activity of AT by a factor of 1000 (Bjork and Lindahl, 1982). The finding that Stx2a binds to heparin and impairs its binding to AT may partly explain the well-known non-beneficial administration of heparin in patients suffering from eHUS (Diekmann,
Fig. 3. (3a) Representative sensogram out of three experiments for the QCM-D analysis of Stx2a binding to heparin and impairment of binding of AT to heparin. Heparin was added to the QCM-D sensor surface via a polyamine core (PAV). Following addition of Stx2a resulted in a binding of Stx2a to heparin. The subsequent addition of AT to the same surface revealed an impaired binding of AT to Stx2a-bound heparin. (3b) Binding of heparin to AT shown by mass uptake in QCM-D. Heparin was added to the QCM-D sensor surface via a polyamine core (PAV). The addition of AT led to a significant mass uptake, which demonstrates the binding between AT and heparin. (3c) Scale-up of the final part of the sensograms presented in (a) and (b), showing the AT binding to heparin (1) subsequent to addition of Stx2a and (2) without previous addition of Stx2a. The lower mass uptake of AT bound to heparin pre-exposed to Stx2a indicates that AT and Stx2a competes for similar binding sites at the heparin surface.
1980; Van Damme-Lombaerts et al., 1988; Vitacco et al., 1973), however it has to be taken into consideration, that quantities of Stx2a and heparin are varying to a huge extent (pg quantities of toxin, mg quantities of heparin). Furthermore, it may be interesting to investigate the ability of Stx2a to form thrombin-antithrombin complexes in the absence or presence of heparin. Furthermore, functional coagulation analysis performed with ROTEM® showed that treatment of plasma with Stx2a leads to only a mild shortening of clotting time. Although we could demonstrate binding of Stx2a to antithrombin, ROTEM® analyses could not reveal a marked functional consequence of this binding. An interesting perspective might be to perform ROTEM® analysis on plasma samples of patients who developed eHUS. Another main finding of this study is that in contrast to various studies (Cooling et al., 1998; Ghosh et al., 6
International Journal of Medical Microbiology xxx (xxxx) xxx–xxx
L. Knabl et al.
Fig. 4. Significant, but only mild decrease of clotting time in ROTEM® analysis by addition of Stx2a compared to the HEPES buffer control. Various concentrations of Stx2a (10 ng/mL or 100 ng/mL) were added to plasma and time needed for clot formation (clotting time in seconds, y axis) was measured by ROTEM analysis. Analysis was performed with pooled citrated plasma samples. n = 30 and n = 32 indicates the number of test repetitions.
fibrinogen and platelets; regarding coagulation initiation, it cannot differentiate which factor is exactly missing. It is therefore a very insensitive methodology when it comes to detailed coagulation analysis. However, the large number of test repetitions (> 30 times) consistently led to the same results. Another limitation of the study is the range of Stx concentrations used for the experiments. In view of the fact that in any case the concentrations of the potential binding partners of Stx are much higher than the concentration of Stx, only a low percentage of AT would be bound by Stx2a in vivo. There are no reports on the concentration of Stx in plasma of patients suffering from eHUS, and thus, it is difficult to estimate the appropriate amount of Stx required for in vitro studies. In this light the marginal but significant results do not yield conclusive statements and no in vitro study may be able to disprove the idea that Stx2a affects eHUS development by affecting increasing plasmatic coagulation. Furthermore, one must acknowledge though, that this is an in vitro study, which might give a too simplistic view of the pathophysiological processes leading to the development of eHUS. Especially the interaction of Stx2a with endothelial cells and different flow conditions in microvessels, leading to a large heterogeneity of perfusion in the microvasculature in severely ill patients, seem to be events which need to be further investigated. In conclusion, the present study showed that Stx2a binds to AT, but functional analyses demonstrated only a mild change in AT activity and a moderate shortening of clotting time. Furthermore, we demonstrated that Stx2a does not interact with resting platelets in vitro. Importantly, Stx2a binds to heparin and impairs the binding of heparin to AT. This finding could explain why different trials with prophylactic or therapeutic administration of heparin had no influence on thrombotic microangiopathy in eHUS patients.
Fig. 5. Absence of P-selectin (CD62 P) expression as a measure of platelet activation in Stx2a treated-PRP. The vertical axis represents the mean fluorescence intensity (MFI). FACS analysis of platelet activation was performed by incubation of PRP with medium, ADP (weak positive control), thrombin (strong positive control), and various concentrations of Stx2a, Stx2a plus thrombin or Stx2a plus ADP. The results are the means +/- SD of five separate experiments.
2004; Viisoreanu et al., 2000) demonstrating binding and activation of platelets by Stx, we were not able to confirm this effect using life confocal microscopy or FACS analysis or LTA. However, a few points need to be taken into account; first, different types of Stx (Stx1 vs. Stx2a) (Cooling et al., 1998) and second, different preparations of platelets (activated vs. resting platelets) (Ghosh et al., 2004) have been used in the different studies. Our data are in line with Viisoreanu and colleagues (Viisoreanu et al., 2000) who showed an absent platelet aggregation using Stx1 and Stx2a under static and flow conditions. Karpman and colleagues (Karpman et al., 2001) demonstrated aggregation of platelets by Stx1; however, data on platelet preparation are missing. Eight years later the same research group (Stahl et al., 2009) observed activation of platelets by Stx2a in monocyte-platelet- or neutrophil-platelet-aggregates, but not in resting platelets (using whole blood as source of platelets). The observations by Stahl and colleagues (Stahl et al., 2015) of an increase of platelet and leukocyte- derived microvesicles in patients suffering from eHUS led to the hypothesis that platelet activation is not mediated by a direct effect of Stx on platelets but rather due to a Stx-induced release of chemokines and other factors. These findings may explain the non-beneficial effect of a therapeutic platelet inhibitor such as dipyridamol in eHUS patients (Van Damme-Lombaerts et al., 1988). The present study has a few limitations. It must be noted that the reduction in clotting time seen in ROTEM® was statistically significant, but only moderate. ROTEM® is a functional coagulation testing method, designed for bedside use in severely bleeding patients. It can mainly differentiate between a lack of plasmatic coagulation factors or
Acknowledgements The authors are grateful to the Austrian Science Fund (FWF; doctoral programme of excellence HOROS, W1253 to S.C. and R.W.) for financial support. The manuscript is dedicated to our colleague and mentor, Prof. Helge Karch, on the occasion of his 65th birthday. References Ay, C., Vormittag, R., Dunkler, D., Simanek, R., Chiriac, A.L., Drach, J., Quehenberger, P., Wagner, O., Zielinski, C., Pabinger, I., 2009. D-dimer and prothrombin fragment 1 + 2 predict venous thromboembolism in patients with cancer: results from the Vienna cancer and thrombosis study. J. Clin. Oncol. 27, 4124–4129. Bhowmik, D., 2001. Elevated tissue factor levels in children with hemolytic uremic syndrome. Clin. Nephrol. 55, 262. Bjork, I., Lindahl, U., 1982. Mechanism of the anticoagulant action of heparin. Mol. Cell. Biochem. 48, 161–182. Brigotti, M., 2012. The interactions of human neutrophils with shiga toxins and related plant toxins: danger or safety? Toxins (Basel) 4, 157–190. Camerer, E., Kolsto, A.B., Prydz, H., 1996. Cell biology of tissue factor, the principal initiator of blood coagulation. Thromb. Res. 81, 1–41. Chandler, W.L., Jelacic, S., Boster, D.R., Ciol, M.A., Williams, G.D., Watkins, S.L.,
7
International Journal of Medical Microbiology xxx (xxxx) xxx–xxx
L. Knabl et al.
Oberdisse, E., Hackenthal, E., Kukhinsky, K., 2001. Pharmakologie und Toxikologie. Springer. Obrig, T.G., Karpman, D., 2012. Shiga toxin pathogenesis: kidney complications and renal failure. Curr. Top. Microbiol. Immunol. 357, 105–136. Orth, D., Khan, A.B., Naim, A., Grif, K., Brockmeyer, J., Karch, H., Joannidis, M., Clark, S.J., Day, A.J., Fidanzi, S., Stoiber, H., Dierich, M.P., Zimmerhackl, L.B., Würzner, R., 2009. Shiga toxin activates complement and binds factor H: evidence for an active role of complement in hemolytic uremic syndrome. J. Immunol. 182, 6394–6400. Paniccia, R., Priora, R., Liotta, A.A., Abbate, R., 2015. Platelet function tests: a comparative review. Vasc. Health Risk Manag. 11, 133–148. Patnaik, M.M., Moll, S., 2008. Inherited antithrombin deficiency: a review. Haemophilia 14, 1229–1239. Poolpol, K., Orth-Höller, D., Speth, C., Zipfel, P.F., Skerka, C., de Cordoba, S.R., Brockmeyer, J., Bielaszewska, M., Würzner, R., 2014. Interaction of Shiga toxin 2 with complement regulators of the factor H protein family. Mol. Immunol. 58, 77–84. Proesmans, W., 2001. The role of coagulation and fibrinolysis in the pathogenesis of diarrhea-associated hemolytic uremic syndrome. Semin. Thromb. Hemost. 27, 201–205. Rodahl, M., Hook, F., Fredriksson, C., Keller, C.A., Krozer, A., Brzezinski, P., Voinova, M., Kasemo, B., 1997. Simultaneous frequency and dissipation factor QCM measurements of biomolecular adsorption and cell adhesion. Faraday Discuss. 229–246. Sauerbrey, G., 1959. Verwendung von Schwingquarzen zur Wägung dünner Schichten und zur Mikrowägung. Zeitschrift fuer Physik 155, 206–222. Schmuck, C., Engels, B., Schirmeister, T., Reinhold, F., 2016. Chemie für Medizner. Pearson, München. Stahl, A.L., Sartz, L., Nelsson, A., Bekassy, Z.D., Karpman, D., 2009. Shiga toxin and lipopolysaccharide induce platelet-leukocyte aggregates and tissue factor release, a thrombotic mechanism in hemolytic uremic syndrome. PLoS One 4, e6990. Stahl, A.L., Arvidsson, I., Johansson, K.E., Chromek, M., Rebetz, J., Loos, S., Kristoffersson, A.C., Bekassy, Z.D., Morgelin, M., Karpman, D., 2015. A novel mechanism of bacterial toxin transfer within host blood cell-derived microvesicles. PLoS Pathog. 11, e1004619. Sun, D.S., Ho, P.H., Chang, H.H., 2016. Soluble P-selectin rescues viper venom-induced mortality through anti-inflammatory properties and PSGL-1 pathway-mediated correction of hemostasis. Sci. Rep. 6, 35868. Suzuki, H., Yamamoto, N., Tanoue, K., Yamazaki, H., 1987. Glycoprotein Ib distribution on the surface of platelets in resting and activation states: an electron microscope study. Histochem. J. 19, 125–136. Tarr, P.I., 1995. Escherichia coli O157:H7: clinical, diagnostic, and epidemiological aspects of human infection. Clin. Infect. Dis. 20, 1–8. Van Damme-Lombaerts, R., Proesmans, W., Van, D.B., Eeckels, R., Binda ki, M.P., Mercieca, V., Vlietinck, R., Vermylen, J., 1988. Heparin plus dipyridamole in childhood hemolytic-uremic syndrome: a prospective, randomized study. J. Pediatr. 113, 913–918. Viisoreanu, D., Polanowska-Grabowska, R., Suttitanamongkol, S., Obrig, T.G., Gear, A.R., 2000. Human platelet aggregation is not altered by Shiga toxins 1 or 2. Thromb. Res. 98, 403–410. Vitacco, M., Sanchez, A.J., Gianantonio, C.A., 1973. Heparin therapy in the hemolyticuremic syndrome. J. Pediatr. 83, 271–275. Walsh, P.R., Johnson, S., 2018. Treatment and management of children with haemolytic uraemic syndrome. Arch. Dis. Child. 103, 285–291. Zhang, W., Bielaszewska, M., Pulz, M., Becker, K., Friedrich, A.W., Karch, H., Kuczius, T., 2008. New immuno-PCR assay for detection of low concentrations of shiga toxin 2 and its variants. J. Clin. Microbiol. 46, 1292–1297.
Igarashi, T., Tarr, P.I., 2002. Prothrombotic coagulation abnormalities preceding the hemolytic-uremic syndrome. N. Engl. J. Med. 346, 23–32. Cooling, L.L., Walker, K.E., Gille, T., Koerner, T.A., 1998. Shiga toxin binds human platelets via globotriaosylceramide (Pk antigen) and a novel platelet glycosphingolipid. Infect. Immun. 66, 4355–4366. Damus, P.S., Hicks, M., Rosenberg, R.D., 1973. Anticoagulant action of heparin. Nature 246, 355–357. Diekmann, L., 1980. [Treatment of the hemolytic-uremic syndrome with streptokinase and heparin (author's transl)]. Klin. Padiatr. 192, 430–435. Ehrlenbach, S., Rosales, A., Posch, W., Wilflingseder, D., Hermann, M., Brockmeyer, J., Karch, H., Satchell, S.C., Würzner, R., Orth-Höller, D., 2013. Shiga toxin 2 reduces complement inhibitor CD59 expression on human renal tubular epithelial and glomerular endothelial cells. Infect. Immun. 81, 2678–2685. Ganter, M.T., Hofer, C.K., 2008. Coagulation monitoring: current techniques and clinical use of viscoelastic point-of-care coagulation devices. Anesth. Analg. 106, 1366–1375. George, J.N., Nester, C.M., 2014. Syndromes of thrombotic microangiopathy. N. Engl. J. Med. 371, 1847–1848. Ghosh, S.A., Polanowska-Grabowska, R.K., Fujii, J., Obrig, T., Gear, A.R., 2004. Shiga toxin binds to activated platelets. J. Thromb. Haemost. 2, 499–506. Gorlinger, K., Bhardwaj, V., Kapoor, P.M., 2016. Simulation in coagulation testing using rotational thromboelastometry: a fast emerging, reliable point of care technique. Ann. Card. Anaesth. 19, 516–520. Guessous, F., Marcinkiewicz, M., Polanowska-Grabowska, R., Kongkhum, S., Heatherly, D., Obrig, T., Gear, A.R., 2005. Shiga toxin 2 and lipopolysaccharide induce human microvascular endothelial cells to release chemokines and factors that stimulate platelet function. Infect. Immun. 73, 8306–8316. Kamitsuji, H., Nonami, K., Murakami, T., Ishikawa, N., Nakayama, A., Umeki, Y., 2000. Elevated tissue factor circulating levels in children with hemolytic uremic syndrome caused by verotoxin-producing E. coli. Clin. Nephrol. 53, 319–324. Karch, H., Tarr, P.I., Bielaszewska, M., 2005. Enterohaemorrhagic Escherichia coli in human medicine. Int. J. Med. Microbiol. 295, 405–418. Karpman, D., Papadopoulou, D., Nilsson, K., Sjogren, A.C., Mikaelsson, C., Lethagen, S., 2001. Platelet activation by Shiga toxin and circulatory factors as a pathogenetic mechanism in the hemolytic uremic syndrome. Blood 97, 3100–3108. Kozek-Langenecker, S.A., Afshari, A., Albaladejo, P., Santullano, C.A., De, R.E., Filipescu, D.C., Fries, D., Gorlinger, K., Haas, T., Imberger, G., Jacob, M., Lance, M., Llau, J., Mallett, S., Meier, J., Rahe-Meyer, N., Samama, C.M., Smith, A., Solomon, C., Van der Linden, P., Wikkelso, A.J., Wouters, P., Wyffels, P., 2013. Management of severe perioperative bleeding: guidelines from the European Society of Anaesthesiology. Eur. J. Anaesthesiol. 30, 270–382. Lee, B.C., Mayer, C.L., Leibowitz, C.S., Stearns-Kurosawa, D.J., Kurosawa, S., 2013. Quiescent complement in nonhuman primates during E coli Shiga toxin-induced hemolytic uremic syndrome and thrombotic microangiopathy. Blood 122, 803–806. Levi, M., van der Poll, T., 2005. Two-way interactions between inflammation and coagulation. Trends Cardiovasc. Med. 15, 254–259. Lu, Q., Malinauskas, R.A., 2011. Comparison of two platelet activation markers using flow cytometry after in vitro shear stress exposure of whole human blood. Artif. Organs 35, 137–144. Mayer, C.L., Parello, C.S., Lee, B.C., Itagaki, K., Kurosawa, S., Stearns-Kurosawa, D.J., 2015. Pro-coagulant endothelial dysfunction results from EHEC Shiga toxins and host damage-associated molecular patterns. Front. Immunol. 6, 155. Nevard, C.H., Jurd, K.M., Lane, D.A., Philippou, H., Haycock, G.B., Hunt, B.J., 1997. Activation of coagulation and fibrinolysis in childhood diarrhoea-associated haemolytic uraemic syndrome. Thromb. Haemost. 78, 1450–1455.
8