Thrombosis Research 134 (2014) 1285–1291

Contents lists available at ScienceDirect

Thrombosis Research journal homepage: www.elsevier.com/locate/thromres

Regular Article

Platelet-free shear flow assay facilitates analysis of shear-dependent functions of VWF and ADAMTS13 Emma Kraus, Kristina Kraus, Tobias Obser, Florian Oyen, Ulrike Klemm, Reinhard Schneppenheim, Maria A. Brehm ⁎ Department of Pediatric Hematology and Oncology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany

a r t i c l e

i n f o

Article history: Received 10 January 2014 Received in revised form 1 August 2014 Accepted 19 August 2014 Available online 28 August 2014 Keywords: von Willebrand factor ADAMTS13 VWF strings shear flow assay

a b s t r a c t Introduction: The multimeric form of von Willebrand factor (VWF), is the largest soluble protein in mammals and exhibits a multidomain structure resulting in multiple functions. Upon agonist stimulation endothelial cells secrete VWF multimers from Weibel-Palade bodies into the blood stream where VWF plays an essential role in platelet-dependent primary hemostasis. Elongation of VWF strings on the cells’ surface leads to accessibility of VWF binding sites for proteins, such as platelet membrane glycoprotein Ib. The prothrombotic strings are sizeregulated by the metalloprotease ADAMTS13 by shear force-activated proteolytic cleavage. Material and Methods: VWF string formation was induced by histamine stimulation of HUVEC cells under unidirectional shear flow and VWF strings were detected employing the VWF binding peptide of platelet glycoprotein Ib coupled to latex beads. VWF strings were then used as substrate for kinetic studies of recombinant and plasma ADAMTS13. Results: To investigate specific aspects of the shear-dependent functions of VWF and ADAMTS13, we developed a shear flow assay that allows observation of VWF string formation and their degradation by ADAMTS13 without the need for isolated platelets. Our assay specifically detects VWF strings, can be coupled with fluorescent applications and allows semi-automated, quantitative assessment of recombinant and plasma ADAMTS13 activity. Conclusions: Our assay may serve as a valuable research tool to investigate the biochemical characteristics of VWF and ADAMTS13 under shear flow and could complement diagnostics of von Willebrand Disease and Thrombotic Thrombocytopenic Purpura as it allows detection of shear flow-dependent dysfunction of VWD-associated VWF mutants as well as TTP-associated ADAMTS13 mutants. © 2014 Elsevier Ltd. All rights reserved.

Introduction With a size of up to more than 10,000 kDa multimeric von Willebrand factor (VWF) is the largest soluble glycoprotein in mammals. It plays an essential role in platelet-dependent primary hemostasis and as a carrier protein that protects factor VIII from degradation. The multiple functions of VWF can be assigned to the VWF multidomain structure that contains binding sites, e.g. for platelet membrane glycoprotein Ib (GPIb), collagen, factor VIII, and Ca2+, sites for dimerization and multimerization and cleavage sites for furin and ADAMTS13 (a disintegrin and metalloproteinase with a thrombospondin type 1 Abbreviations: VWF, von Willebrand factor; WPB, Weibel-Palade Body; ADAMTS13, a disintegrin and metalloproteinase with a thrombospondin type 1 motif, member 13; GPIb, glycoprotein Ib; USS, Upshaw-Schulman Syndrome; TTP, Thrombotic Thrombocytopenic Purpura; RT, room temperature; PFA, paraformaldehyde; ROI, region of interest. ⁎ Corresponding author at: Department of Pediatric Hematology and Oncology, University Medical Center Hamburg-Eppendorf, Hs N21, Rm 111, Martinistrasse 52, 20246 Hamburg, Germany. Tel.: +49 40 741058523; fax: +49 40 741058931. E-mail address: [email protected] (M.A. Brehm).

http://dx.doi.org/10.1016/j.thromres.2014.08.013 0049-3848/© 2014 Elsevier Ltd. All rights reserved.

motif, member 13) [1,2]. The VWF monomers are subject to a multimerization cascade that starts with dimerization in the ER, glycosylation, multimerization, maturation by proteolytic cleavage of the pro-peptide in the trans-Golgi network, tubulation and storage in Weibel-Palade bodies (WPB’s) [3]. Upon agonist stimulation endothelial cells secrete high molecular weight VWF multimers from WPB’s. These multimers form VWF strings on the cells’ surface exposing the structural A1-domain that contains the binding site for the α-chain of glycoprotein Ib (GPIbα) on the plasma membrane of platelets [4]. This interaction has previously been used for visualization of VWF strings via attachment of platelets under shear flow [5,6]. The potentially prothrombotic VWF strings are size-regulated by ADAMTS13. Under physiological conditions, this metalloprotease cleaves VWF between Tyr1605 and Met1606 in its A2-domain only when this domain is unfolded by shear stress and the respective binding sites of both, VWF and ADAMTS13, adjoin each other [7]. This sophisticated substrate recognition involves distinct steps during which the TSP5-CUB domains of ADAMTS13 initially bind to VWF shear force-independent, then, upon shear-induced elongation

1286

E. Kraus et al. / Thrombosis Research 134 (2014) 1285–1291

of VWF, additional exosite binding sites become exposed which subsequently can be bound by the complementary spacer domain of ADAMTS13 shear force-dependent [8]. Therefore, VWF activity, its susceptibility to proteolytic cleavage by ADAMTS13, and binding as well as catalytic activity of ADAMTS13 are shear force-regulated. Moreover, mutations in VWF and ADAMTS13 cause the potentially life threatening diseases von Willebrand Disease (VWD) [9] and Upshaw-Schulman Syndrome (USS), respectively. VWD is a bleeding disorder that arises from qualitative and quantitative deficiencies of VWF [1]. USS is the inherited form of Thrombotic Thrombocytopenic Purpura (TTP) [10], a severe microangiopathy characterized by platelet clumping, hemolytic anemia, and subsequent organ failure [11]. State-of-the-art diagnostic tests for both diseases use static approaches. Theoretically, in some cases the results of these static assays could be without pathological finding, e.g. there might be mutations that affect the protein’s function only under flow conditions. For VWF we recently showed that mutations can have unforeseeable effects under shear flow conditions [12]. The exact mechanisms of the shear-dependent functions of VWF and ADAMTS13 are not yet completely understood. Therefore, shear flow assays are essential to unravel these secrets. For example, are there mutations within VWF that specifically influence VWF string formation? Which mutations within ADAMTS13 influence only the shear-dependent binding to VWF? Are there forms of TTP with sheardependent symptoms that are not detectable with our standard static diagnostic tests? To address these questions it is useful to have an assay available that does not require isolated platelets. Platelets add to the complexity of the system and handling can be challenging. We developed a shear flow assay coupled with automated image recording that allows observation of VWF string formation and quantitative assessment of string degradation by ADAMTS13 without the need for isolated platelets. Instead, live visualization of these processes is facilitated employing GPIbα-coated latex particles. Material and Methods Materials Reagent I (Latex particles coated with mouse anti-GPIbα) and Reagent III (recombinant GPIbα peptide, carrying gain-of-function mutations). Both reagents are part of the INNOVANCE® VWF Ac assay kit (Siemens Healthcare Diagnostics Products GmbH, Marburg, Germany) which is commercially available and indicated to be used for the automatic determination for VWF activity in human plasma; not available for sale in the U.S. HUVEC cell culture under static and perfusion conditions HUVEC cells (Human Umbilical Vein Endothelial Cells, isolated in growth medium 2 from a single donor and cryopreserved) were purchased from Promocell (Heidelberg, Germany) and cultured in readyto-use endothelial cell growth medium 2 (#C-22011, Promocell). Aliquots of passage 3 were cryopreserved. Cells of each aliquot were only used for experiments until passage 10. Every 2-3 days cells grown in 75 cm2 cell culture flasks (Greiner, Frickenhausen, Germany) were washed twice with sterile HBRS buffer (140 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1 mM CaCl2, 10 mM HEPES, pH 7.4) and trypsinyzed by incubation with 1 ml 0.05% trypsin (Invitrogen, Darmstadt, Germany) at 37 °C and 5% CO2 for 1 min. 1/3 of the cells was transferred into a new cell culture flask and fresh medium was added. For seeding of HUVEC cells in μ-Slide I0.4 Luer, ibidiTreat®, tissue culture treated, sterile, channel slides (Ibidi, Martinsried, Germany) a concentration of 1.6x106 cells/ml was prepared and 1x105 cells/cm2 in 150 μl medium were filled into the channel of the slide, which then was incubated for 1-2 h at 37 °C and 5% CO2. Then the channel slides were connected to a perfusion set of an Ibidi pump system, that

contained 12 ml of endothelial growth medium MV2 (#C-22022, Promocell). The Ibidi pump system consists of an air pressure pump and a fluidic unit. Cells were treated according to the Ibidi application note 13: HUVECs under perfusion. In brief, the fluidic unit with the attached channel slide was connected to the pump and placed in an incubator (37 °C, 5% CO2). A shear stress of 5 dyne/cm2 was applied for up to 2 days. Alternatively, channel slides with HUVEC cells can be placed in a humidified chamber on a rocking shaker and incubated at 37 °C and 5% CO2 with medium changes twice a day. Shear flow assay and VWF string detection employing GPIbα − beads GPIbα − beads were produced by incubation of Reagent III (recombinant GPIbα peptide) and Reagent I (mouse anti-GPIbα coated Latex particles) (INNOVANCE® VWF Ac assay kit, Siemens Healthcare Diagnostics) in a ratio of 1:2.4, shaking for 1 hour at 4 °C. No washing step is required since an optimized peptide:latex particle ratio is provided within the kit. The ready-to-use GPIbα − beads can be stored at 4 °C for at least 6 months. Wildtype (wt) GPIbα − beads were produced by incubation of 0.5 ml beads and 1 ml wtGPIbα peptide containing cell culture supernatant for 1 h, rotating at 4 °C, followed by 3 wash steps (centrifugation for 10 min at 4000 xg, resuspension in 1 ml PBS). The reservoirs of a perfusion set (YELLOW-and-GREEN, length 50 cm, ID 1.6 mm, 10 ml reservoirs) were filled with 10 ml endothelial cell growth medium 2, prewarmed to 37 °C. HUVEC cells cultured in channel slides were washed with 1 ml pre-warmed (37 °C) HBRS buffer (140 mM NaCl, 5 mM KCl, 1 mM MgCl2 , 1 mM CaCl 2 , 10 mM HEPES, pH 7.4) and the channel slide was connected to the perfusion set air bubblefree and mounted in an incubation unit (Tokai-Hit, Osaka, Japan, 5% CO 2 , 37 °C) within a fluorescence microscope (BZ-9000, Keyence, Neu-Isenburg, Germany). The cells were adjusted to unidirectional shear flow by increasing shear stress from 2.5 to 3.75 to 5 dyne/cm2 in 4 min intervals. After 2 min at 5 dyne/cm2 shear stress 100 μl GPIbαbeads and 150 μM histamine were added to one reservoir and the VWF string formation, detected by binding of GPIbα − beads to VWF, was observed using a Plan Fluor ELWD DM x40 phase contrast objective with a numerical aperture of 0.6 or a Plan Fluor ELWD DM x20 phase contrast objective with a numerical aperture of 0.45. Fluorescent detection of VWF strings To observe string formation by fluorescence detection two approaches can be employed: 1) Fluorescent labeling of GPIbα beads prior to flow assay The GPIbα peptide was incubated for 1 h at 4 °C with anti-GPIbαantibody-latex-particles (1:2.4), then goat anti-mouse Alexa Fluor® 546 antibody (1:50, Invitrogen) was added and the beads were further incubated rotating for 1 h at 4 °C. Then, the beads were centrifuged for 10 min at 4000 xg, the supernatant was removed, and the beads were resuspended in 1 ml PBS. The washing procedure was repeated three times. Then the beads (resuspended in PBS) were used in the VWF string formation assay which was observed live using a standard fluorescence microscope (BZ-9000, Keyence) equipped with a TRITC filter and the same objectives as described above. 2) Direct immunofluorescent detection without beads For direct detection of VWF strings using anti-VWF antibodies HUVEC cells were perfused with 10 ml HBRS buffer containing 150 μm histamine, 1 μg anti-VWF antibody (polyclonal rabbit anti-human VWF (#A0082, DAKO, Hamburg, Germany) or monoclonal mouse anti-VWF (#MCA4682, clone RFF-VIII R/1, AbD Serotec, Puchheim, Germany)) and 3 μg of the respective secondary antibody (goat anti-rabbit Alexa Fluor® 488 or goat antimouse Alexa Fluor® 488 (Invitrogen)). String labeling was

E. Kraus et al. / Thrombosis Research 134 (2014) 1285–1291

observed live using a GFP filter and a Plan Fluor ELWD DM x40 phase contrast objective with a numerical aperture of 0.6.

1287

Hamburg, Germany, employing the TECHNOZYM® ADAMTS-13 Antigen ELISA by Technoclone (Vienna, Austria).

Immunofluorescence

Shear flow activity assay of recombinant and plasma ADAMTS13

For immunofluorescent staining of fixed VWF strings, a third reservoir was attached to the perfusion set at setup and filled with 4% paraformaldehyde (PFA) in PBS. After completion of the shear flow assay the reservoir with the medium was closed and the PFA reservoir was opened and cells were perfused with the PFA solution at 5 dyne/cm2 for 10 min. After this fixation step the flow was stopped, the slide was disconnected, and the entire medium was aspirated from the wells of the slide but not from the channel. Then the cells were washed by slowly applying 1 ml PBS into one empty well and aspirating from the opposite well (see also Ibidi Application Note 13). Subsequently, cells were incubated with 1% BSA in PBS for 20 min at 37 °C and rabbit anti-human VWF (1:1000, DAKO) was added to the channel. After incubation for 2 hours at 37 °C the cells were washed with 2 ml PBS as described above and secondary antibodies were applied. Anti-VWF was detected with goat anti-rabbit Alexa Fluor® 488 (1:5000, Invitrogen) and the mouse anti-GPIbα antibody on the surface of the Latex particles was directly detected using goat anti-mouse Alexa Fluor® 546 (1:5000, Invitrogen). Finally, after 1 hour incubation at RT in the dark, the cells were washed with 2 ml PBS. For additional visualization of intracellular VWF a permeabilization step (0.3% Triton X-100 in PBS, 3 min at 37 °C) is required after fixation.

VWF string formation and string detection on the surface of histamine-stimulated HUVEC cells was performed as described above. To perform enzyme kinetics of recombinant ADAMTS13 variants, the proteins (pre-incubated at 37 °C for 15 min) were directly added to one of the reservoirs after string formation. To measure plasma ADAMTS13 activity, a 1:10 dilution of plasma in endothelial growth medium 2, including a final concentration of 2 U/ml heparin, was injected from a third attached reservoir. VWF string cleavage was monitored by recording time-lapse multi-picture series as described below. Every experiment was performed at least 3 times.

Recombinant human ADAMTS13 and wildtype GPIbα peptide The full-length ADAMTS13 cDNA was cloned into the expression vector pcDNA3.1 as described previously [13]. In vitro mutagenesis for generation of ADAMTS13 mutant p.Gly702Arg was performed using the QuikChange® kit (Stratagene, Waldbronn, Germany) according to the manufacturer’s instructions. For cloning of the wildtype (wt) GPIbα peptide (aa 1-298) into the expression vector pIRESneo2 first the coding sequence of a 3xFlag-6xHis-Tag was inserted by primer ligation between the restrictions sites BamHI und NotI. Then the coding sequence of the wtGPIb peptide was cloned into the vector using restrictions sites NheI and BamHI. Primer sequences are available upon request. All vectors were used to transform Top10 supercompetent cells (Invitrogen). After plasmid purification, 4 μg vector DNA were used to transiently transfect HEK293 EBNA cells (2x106) employing Lipofectamine 2000 (Invitrogen). The cells were grown for 72 hours (24 hours in Dulbecco modified Eagle medium (Invitrogen) with 10% [vol/vol] fetal bovine serum and 48 hours in serum-free OPTIPRO-SFM medium (Invitrogen)). ADAMTS13 protein secreted into the medium was concentrated employing Amicon Ultra 15 tubes (Millipore, Billerica, MA, USA) with a 100,000 Da cutoff. ADAMTS13 concentration was determined using the Imubind® ADAMTS13 ELISA kit (American diagnostics, Pfungstadt, Germany) according to the manufacturer’s instructions. The medium containing secreted wtGPIbα peptide was tested for wtGPIbα content by standard Western blotting techniques and detection using an antiFlag antibody (#F3165, clone M2, Sigma Aldrich). Preparation of ADAMTS13 containing plasma Blood was collected from healthy volunteers using Li-heparin blood vacuum collection tubes. The study was conducted in conformity to the Declaration of Helsinki [14] (version Fortaleza, Brazil, 2013) and to The International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) Guidelines, available at http://www.ich.org, accessed in October 2010. The whole blood was centrifuged for 12 min at 2300 x g at RT and the plasma was transferred into a fresh tube. ADAMTS13 antigen was determined by the MEDILYS Laborgesellschaft mbH, Hemostaseology, Asklepios Klinik Altona,

Imaging, data analysis and statistics For quantification of VWF string degradation by ADAMTS13 an automated time-lapse multi-picture series was captured starting 30 sec before addition of ADAMTS13. Using the Keyence software 4x3 images were taken starting at a set ROI every 30 sec for 10 min (1x3 images every 10 sec for 5 min when 700 ng/ml ADAMTS13 were investigated). Subsequently the 12 (or 3) images of each time point were automatically merged to one picture using the merge function of the BZ-II Analyzer software 1.42 (Keyence). A 100 μm scale bar was inserted into the merged picture that was used to calibrate the scale in ImageJ to mm. In some pictures brightness and contrast were adjusted. Employing the ImageJ software [15], the string length was measured for each string within the merged picture (either every visible string (n N 37) or a maximum of 100 strings) before and after indicated time points of addition of ADAMTS13. The overall string length of all measured strings per time point was calculated by adding up all string lengths. The overall string length at 0 min was defined as 0% cleavage and % cleavage was determined by calculating % residual string length after defined time points. Mean and SEM of three independent experiments are plotted using the GraphPad Software. In control experiments without ADAMTS13 we observed a mean decrease in the overall string length by 7% (which includes deattachment of strings as well as occasional breakage) over the observed time period of 3 min and 4% after 1 min. We therefore corrected respective data sets by these values. The initial activity of ADAMTS13 was determined by calculating % cleavage per minute and ng/ml protein using the initial slope of the % cleavage versus time curves between 0 and 15 sec. Unpaired t-test was performed using GraphPad Prism version 5.00 for Windows, GraphPad Software, San Diego, California, USA. Results Visualization of VWF strings employing GPIbα-coated latex particles The interaction between platelet GPIbα and VWF [4] has previously been used to visualize VWF strings by attachment of platelets to VWF under shear flow [5,6]. To visualize VWF strings without platelets, we developed a flow assay based on this interaction but only employing the minimal GPIbα peptide needed for VWF binding. Such a peptide harboring two gain-of-function mutations to enhance binding to VWF is part of the commercially available INNOVANCE® VWF Ac assay kit (Siemens Healthcare Diagnostics) for evaluation of VWF activity under static conditions. We used this peptide bound to mouse anti-GPIbαcoupled latex particles (also provided within the kit) and perfused these GPIbα-beads at 5 dyne/cm2 shear stress over histaminestimulated HUVEC cells. As shown in Fig. 1A VWF string formation was induced by histamine and the VWF strings were successfully visualized by attachment of the GPIbα-beads (experiments performed n N 100). For some applications the enhanced binding of the

1288

E. Kraus et al. / Thrombosis Research 134 (2014) 1285–1291

Verification of VWF and GPIbα-bead colocalization by immunofluorescence To prove that the observed structures were indeed formed by VWF and GPIbα-beads both components were visualized by immunofluorescence (Fig. 2A,B,D). Fixation under shear flow conditions prior immunofluorescence is crucial because unfixed strings will retract upon shear flow reduction and break during further downstream applications. As shown in Fig. 2, the pearls-on-a-string like structures do consist of the latex particles (Fig. 2B,C,D) and VWF (Fig. 2A,D). Colocalization of beads and VWF appears yellow in the overlay (Fig. 2D). These results confirm that the GPIbα-beads do bind to VWF under flow conditions and therefore can be used to detect and visualize VWF strings. Live fluorescent detection of VWF strings

Fig. 1. VWF strings detected by GPIbα-beads. HUVEC cells, cultured in Ibidi channel slides under perfusion, were perfused at 5 dyne/cm2 shear stress with endothelial cell growth medium 2 containing 150 μM histamine and 100 μl of GPIbα-beads (A) or wildtype GPIbα-beads (B) for 10 minutes using an Ibidi pump system. VWF strings were visualized by binding of GPIbα-beads and imaged by phase contrast microscopy. (A) shows more than 15 strings; four are marked exemplary by black arrows. N N 100 (A) and n = 4 (B) independent experiments were performed. Scale bars represent 100 μm.

commercially supplied GPIbα-beads might be hindering, e.g. when VWF:GPIbα interaction should specifically be investigated. For such applications we also expressed the wildtype peptide and performed the assay with wtGPIbα-beads. As shown in Fig. 1B, VWF strings were detectable with the wtGPIbα peptide under the same assay conditions. No increase in shear rate was necessary for binding of the wtGPIbαbeads to VWF strings (n = 4). Therefore, our assay facilitates VWF string detection without platelets.

For some applications it might be useful to observe string formation live using fluorescence. In one approach we fluorescently labeled the GPIbα-beads, with a goat anti-mouse Alexa Fluor® 546 antibody prior to the assay. In a second approach we completely omitted the GPIbα-beads. Instead we added 1 μg of anti-VWF antibodies and 3 μg of respective fluorescently labeled secondary antibodies directly into the histamine-containing HBRS buffer. The attachment of the fluorescent beads (Fig. 2E) as well as binding of the antibodies (Fig. 2F, G) to VWF strings was successfully observed live using fluorescence microscopy and pictures were taken during the experiment. As expected the polyclonal anti-VWF antibody (Fig. 2F) gave a stronger signal than the monoclonal anti-VWF antibody (Fig. 2G). Concentration-dependent VWF proteolysis by wtADAMTS13 under flow conditions For investigation of the proteolytic activity of ADAMTS13; the metalloprotease that specifically cleaves VWF [10], in a reproducible and quantifiable manner, we developed an automated image recording protocol. VWF string formation was induced by histamine stimulation

Fig. 2. Fluorescent detection of VWF strings. A-D: HUVEC cells, cultured in Ibidi channel slides under perfusion, were perfused at 5 dyne/cm2 shear stress with 150 μM histamine and 100 μl of GPIbα-beads in endothelial cell growth medium for 10 minutes. VWF strings were fixed under perfusion with 4% paraformaldehyde and VWF was detected using rabbit antiVWF and goat anti-rabbit Alexa Fluor® 488 (A). Mouse anti-GPIbα on the surface of the GPIbα-beads was detected employing goat anti-mouse Alexa Fluor® 546 (B). Localization of GPIbα-beads (C, phase contrast) on VWF strings is shown in the overlay (D). (E): Detection of VWF strings using fluorescence microscopy and fluorescently labeled GPIbα-beads. Three independent experiments were performed. Scale bars represent 50 μm. F,G: HUVEC cells were perfused at 5 dyne/cm2 shear stress with 150 μM histamine, 1 μg anti-VWF antibodies (polyclonal rabbit anti-VWF (F), monoclonal mouse anti-VWF-A1 (G)) and 3 μg secondary antibodies (goat anti-rabbit Alexa Fluor® 488 (F), goat anti-mouse Alexa Fluor® 488 (G)). Pictures were taking during the experiment using a GFP filter. Three independent experiments were performed with each antibody. Scale bar represents 100 μm.

E. Kraus et al. / Thrombosis Research 134 (2014) 1285–1291

1289

Fig. 3. Proteolysis of VWF strings by recombinant wtADAMTS13 under flow. HUVEC cells were perfused at 5 dyne/cm2 shear stress with 150 μM histamine and 100 μl of GPIbα-beads in endothelial growth medium 2 for 5 min. An automated time-lapse multi-picture series was captured starting 30 sec before addition of 100 ng/ml recombinant ADAMTS13. Using the Keyence software 4x3 images were recorded starting at a set ROI every 30 sec for 10 min. Subsequently the 12 images of each time point were automatically merged to one picture using the merge function of the BZ-II Analyzer software 1.42 (Keyence). A 100 μm scale bar was inserted into the merged picture. The scale bar was used to calibrate the scale in ImageJ to mm. For time points 0 min (A) and 1 min after addition of ADAMTS13 (B) all strings were measured and labeled using ImageJ. Three independent experiments were performed. Scale bars represent 100 μm. The unlabeled images are provided as Supplemental Figs. S1 and S2. For proteolysis of VWF strings by 700 ng/ml recombinant wtADAMTS13 watch Supplemental Video 1.

of HUVEC cells at a shear stress of 5 dyne/cm2 as described above. 5 min after stimulation an automated time-lapse multi-picture series was recorded. The images of each time point were automatically merged to one big picture using the BZ-II Analyzer software. Fig. 3A and B show the merged pictures of time points 0 min (Fig. 3A) and 1 min after addition of 100 ng/ml recombinant wtADAMTS13 (Fig. 3B). VWF strings were measured and labeled using the ImageJ software (the original unlabeled pictures are shown in Suppl. Figs. S1, S2). Exemplary, the respective measured data of Fig. 3A and B and the overall string lengths at time points 0 min and 1 min are given in Suppl. Table 1. We performed the string cleavage assay employing different concentrations of recombinant wtADAMTS13 and found a clear correlation between ADAMTS13 concentration and cleavage rate. Fig. 4 shows the reduction of the overall strings length (in % cleavage) after 1 min by 5, 20, 100, 350, and 700 ng/ml recombinant wtADAMTS13 (means and SEM of three independent experiments for each concentration are shown). 5 ng/ml was the lowest concentration that gave reliable data using

our experimental setup, presumably, due to methodological problems with addition and proper mixing of such small volumes within the reservoirs. After further optimization of these procedures it might be possible to investigate even lower concentrations. The degradation of VWF strings by ADAMTS13 at a physiological concentration of 700 ng/ml can be watched in Suppl. Video 1. These experiments show that the concentration-dependent degradation of VWF by recombinant wtADAMTS13 can accurately be determined using our assay offering the possibility to perform enzyme kinetic studies of ADAMTS13 under shear flow conditions. Comparison of enzyme kinetics of wildtype and mutant ADAMTS13 under flow conditions We further employed our assay to determine enzyme kinetics of mutant ADAMTS13 by measuring the residual activity of an ADAMTS13 mutant carrying mutation p.Gly702Arg, which we identified in a patient

1290

E. Kraus et al. / Thrombosis Research 134 (2014) 1285–1291

observed a strongly reduced secretion in the patient’s plasma as well as in the cell culture medium of HEK EBNA cells used to produce the recombinant mutant [data not shown]. Measurement of plasma ADAMTS13 activity under flow conditions

Fig. 4. VWF string cleavage in dependence of ADAMTS13 concentration. HUVEC cells were perfused at 5 dyne/cm2 shear stress with 150 μM histamine and 100 μl of GPIbαbeads in endothelial growth medium 2 for 5 min. Capturing of an automated time-lapse multi-picture series of 12 images every 30 sec for 10 min was started before addition of 5, 20, 100, 350 ng/ml recombinant wtADAMTS13. When 700 ng/ml wtADAMTS13 were investigated 3 images every 10 sec for 5 min were captured. The images of each time point were merged, a 100 μm scale bar was used to calibrate the scale in ImageJ to mm and the length of all strings was measured. The overall string length at 0 min was defined as 0 % cleavage and % cleavage after 1 min was determined by reduction of the overall string length in % after 1 min. % cleavage of VWF strings was plotted versus concentration of recombinant wtADAMTS13. Mean and SEM of three independent experiments for each concentration corrected by values measured in control experiments (4%, n = 4) without ADAMTS13 are shown.

with hereditary TTP. We measured the cleavage rate of VWF strings by 100 ng/ml of recombinant mutant p.Gly702Arg and wtADAMTS13 over a 3 min time period and plotted the cleavage rate versus time (Fig. 5, mean and SEM of three independent experiments for both proteins are shown). Using the initial slope of these curves between data points 0 and 15 sec we determined the initial activity of p.Gly702Arg and wtADAMTS13 to be 0.24% and 1% cleavage per min and ng/ml protein, respectively. Therefore, mutant p.Gly702Arg exhibits a residual initial activity of 24 % compared to wtADAMTS13. TTP symptoms are only observed in patients when less than 5% ADAMTS13 activity is present in their plasma. We have found that this strong reduction in activity is caused by a combination of reduced proteolytic activity and disturbed secretion in most patients with hereditary TTP [own unpublished data]. This also seems to be the case for p.Gly702Arg for which we

Fig. 5. Enzyme kinetics of ADAMTS13 mutant p.Gly702Arg. HUVEC cells were perfused at 5 dyne/cm2 shear stress with 150 μM histamine and 100 μl of GPIbα-beads in endothelial growth medium 2 for 5 min. Capturing of an automated time-lapse multi-picture series of 12 images every 30 sec for 10 min was started before addition of 100 ng/ml recombinant wtADAMTS13 or mutant p.Gly702Arg. The 12 images of each time point were merged, a 100 μm scale bar was used to calibrate the scale in ImageJ to mm and the length of all strings was measured for time points 0, 1, 2, and 3 min. The reduction in the overall string length is plotted as % cleavage versus time. Mean and SEM of three experiments for each protein corrected by values measured in control experiments (n = 4) without ADAMTS13 are shown. The initial slope of the two curves between time points 0 and 15 sec were used to calculate the initial activities of wtADAMTS13 and p.Gly702Arg to be 1% and 0.24% cleavage per min and ng/ml protein, respectively. Unpaired t-test showed that the reduction of the initial activity of p.Gly702Arg compared to wtADAMTS13 is statistically significant, p b 0.005.

To test the suitability of our assay for diagnostics, we then measured plasma ADAMTS13 activity. Platelet poor plasma was prepared from whole blood of three healthy individuals and ADAMTS13 antigen was determined. Plasma ADAMTS13 concentrations were 1160 ng/ml, 1800 ng/ml, and 690 ng/ml in samples 1, 2, and 3, respectively. The plasma was supplemented with heparin to avoid clotting, diluted 1:10 in medium, and perfused over VWF strings. We determined the initial proteolytic activity within each plasma sample from the initial slope between data points 0 and 15 sec of the % cleavage versus time curves (Fig. 6). The initial activities of plasma ADAMTS13 were 0.41%, 0.15% and 0.81% cleavage per min and ng/ml protein in samples 1, 2 and 3, respectively. As a consequence the assay is suitable to detect differences in ADAMTS13 activity in plasma and could be used to diagnose ADAMTS13 dysfunction directly from TTP patient plasma. An activity b5% U/ml (or converted to the units used in our assay b0.05% cleavage per min and ng/ml protein), which is normal for TTP patients [16], would be distinguishable from normal activity values. Discussion We developed a shear flow assay to investigate VWF as well as ADAMTS13 functions and deficiencies under shear flow conditions. Our assay allows detection of VWF strings with commercially available GPIbα-beads without requirement of preparation and fluorescent labeling of platelets, which significantly simplifies handling. The advantages of this approach compared to string detection with either isolated platelets or platelets in whole blood are: 1) no blood as a source of platelets is required, 2) no time consuming platelet isolation and staining is necessary, 3) the assay can be performed in medium or buffer, therefore, no special equipment for reflection interference contrast microscopy is required; which would be necessary if the experiment was performed in whole blood. We confirmed the specific detection of VWF strings by VWF immunofluorescence thereby also showing that the method is compatible with downstream applications as long as the VWF strings are fixed under flow conditions.

Fig. 6. Proteolysis of VWF strings by plasma ADAMTS13 under flow conditions. HUVEC cells were perfused at 5 dyne/cm2 shear stress with 150 μM histamine and 100 μl of GPIbα-beads in endothelial growth medium 2 for 5 min. Then VWF strings were perfused with a 1:10 dilution of plasma in medium including 2 U/ml heparin. Capturing of an automated time-lapse multi-picture series of 12 images every 30 sec for 10 min was started 30 sec before perfusion with the plasma. The 12 images of each time point were merged, a 100 μm scale bar was used to calibrate the scale in ImageJ to mm and the length of all strings was measured for time points 0, 1, 2, and 3 min. The reduction of the overall string length in % was plotted versus time. Mean and SEM of three independent experiments for each plasma sample corrected by values measured in control experiments (n = 4) without ADAMTS13 are shown.

E. Kraus et al. / Thrombosis Research 134 (2014) 1285–1291

For in-flow fluorescent applications we optimized labeling of the GPIbα-beads with fluorescently labeled antibodies. To even further simplify handling we omitted the beads completely and performed direct detection of VWF strings employing anti-VWF antibodies and respective fluorescently labeled secondary antibodies only. When we used a polyclonal antibody a very strong fluorescent signal was observed due to multiple epitopes. The assay also worked with a monoclonal antibody with a single epitope in the A1-domain of VWF. In the latter case the signal was weaker, presumably because only one antibody molecule can bind to one monomer of VWF. In the future, the assay could be employed for a wide range of applications in VWD and TTP research and diagnostics. There is no standardized diagnostic test available to determine VWF function under shear flow conditions. Therefore, unknown types of VWD with shear-dependent dysfunction of VWF might exist that are currently undetectable. Recently, we have shown that shear flow assays are essential to fully comprehend the effects of VWF mutations, since shear-dependent dysfunctions do not correlate with the domain that is affected by a mutation, and are therefore adding to the VWD phenotype in unpredictable ways [12]. When used with transfected cells that overexpress VWD mutants, our assay would allow the investigation of shear-dependent effects of these mutants, e.g. on GPIbα binding or the ability to form strings. In VWF research usually whole blood or platelet preparations are used in shear flow assays to investigate VWF function. While this is necessary for certain applications, here GPIbα as the only target allows selectively studying GPIbα binding as the initial step of platelet interaction with VWF. For such purpose either the wildtype GPIbα peptide or the commercially available reagents, which are of consistently high quality, could be employed. State-of-the-art diagnostic tests to determine plasma ADAMTS13 activity use unphysiological static approaches based on proteolysis of either mutated VWF peptides or full-length VWF under denaturing conditions [17]. Plasma ADAMTS13 activity and antigen levels poorly correlate when these static assays are employed. For example, it has been shown that the same concentration of ADAMTS13 in the plasma of single healthy donors can exhibit activities within a range of 0.51.75 U/ml and that high activities can be measured in plasma with low ADAMTS13 concentration and vice versa [18]. That recombinant ADAMTS13 exhibits a clear concentration dependency in our assay and plasma ADAMTS13 in static assays does not, indicates that environmental and biological factors influence plasma ADAMTS13 activity in vivo. These facts underline the necessity of shear-flow assays under close to physiological conditions to further investigate these factors. We showed here, that, within a time frame of 10-15 min, our assay facilitates performance of kinetic studies of recombinant and plasma ADAMTS13 variants. The apparent reciprocal correlation between antigen and activity that we found for plasma ADAMTS13 of healthy donors might be a coincidence but it could further be investigated in the future in a bigger cohort of healthy donors. Further, shear-dependent dysfunctions of ADAMTS13 could not be diagnosed under static conditions. The activity of ADAMTS13 in TTP patient plasma is usually N5% U/ml [16]. Converted to the units used in our assay this is equal to less than 0.05% cleavage per min and ng/ml protein. As a consequence, our assay is suitable to diagnose ADAMTS13 dysfunction directly from patient plasma because such values would be distinguishable from the normal activity values we measured. Hence, our assay now provides a method for fast, semi-automated and quantitative assessment of shear-dependent ADAMTS13 function and dysfunction. It could be indicated for patients who show inconclusive results in the static tests. For example, there might be

1291

patients with shear-dependent TTP symptoms who would show normal ADAMTS13 activity in currently employed tests. Our assay now provides new possibilities to specify and simplify shear flow assays for certain applications and therefore fosters and complements VWD and TTP research and diagnostics. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.thromres.2014.08.013. Conflict of interest None Acknowledgements This study was financially supported by the DFG within the Research Group FOR1543 (SHENC): “Shear flow regulation of hemostasis - bridging the gap between nanomechanics and clinical presentation” (TO, UK, MAB, RS). We thank Paula Ciporina and Gesa König for technical assistance and Torsten Wundenberg for helpful discussions about enzyme kinetics. References [1] Schneppenheim R, Budde U. von Willebrand factor: the complex molecular genetics of a multidomain and multifunctional protein. J Thromb Haemost 2011;9(Suppl. 1): 209–15. [2] Zhou M, Dong X, Baldauf C, Chen H, Zhou Y, Springer TA, et al. A novel calciumbinding site of von Willebrand factor A2 domain regulates its cleavage by ADAMTS13. Blood 2011;117:4623–31. [3] Wagner DD. Cell biology of von Willebrand factor. Annu Rev Cell Biol 1990;6: 217–46. [4] Huizinga EG, Tsuji S, Romijn RA, Schiphorst ME, de Groot PG, Sixma JJ, et al. Structures of glycoprotein Ibalpha and its complex with von Willebrand factor A1 domain. Science 2002;297:1176–9. [5] De Ceunynck K, Rocha S, Feys HB, De Meyer SF, Uji-i H, Deckmyn H, et al. Local elongation of endothelial cell-anchored von Willebrand factor strings precedes ADAMTS13 protein-mediated proteolysis. J Biol Chem 2011;286:36361–7. [6] Goerge T, Kleineruschkamp F, Barg A, Schnaeker EM, Huck V, Schneider MF, et al. Microfluidic reveals generation of platelet-strings on tumor-activated endothelium. Thromb Haemost 2007;98:283–6. [7] Gao W, Anderson PJ, Sadler JE. Extensive contacts between ADAMTS13 exosites and von Willebrand factor domain A2 contribute to substrate specificity. Blood 2008; 112:1713–9. [8] Crawley JT, de Groot R, Xiang Y, Luken BM, Lane DA. Unraveling the scissile bond: how ADAMTS13 recognizes and cleaves von Willebrand factor. Blood 2011;118: 3212–21. [9] Schneppenheim R. The pathophysiology of von Willebrand disease: therapeutic implications. Thromb Res 2011;128(Suppl. 1):S3–7. [10] Levy GG, Nichols WC, Lian EC, Foroud T, McClintick JN, McGee BM, et al. Mutations in a member of the ADAMTS gene family cause thrombotic thrombocytopenic purpura. Nature 2001;413:488–94. [11] Coppo P, Veyradier A. Thrombotic microangiopathies: towards a pathophysiologybased classification. Cardiovasc Hematol Disord Drug Targets 2009;9:36–50. [12] Brehm MA, Huck V, Aponte-Santamaria C, Obser T, Grassle S, Oyen F, et al. von Willebrand disease type 2A phenotypes IIC, IID and IIE: A day in the life of shearstressed mutant von Willebrand factor. Thromb Haemost 2014:112. [13] Hassenpflug WA, Budde U, Obser T, Angerhaus D, Drewke E, Schneppenheim S, et al. Impact of mutations in the von Willebrand factor A2 domain on ADAMTS13dependent proteolysis. Blood 2006;107:2339–45. [14] Rickham PP. Human Experimentation. Code of Ethics of the World Medical Association. Declaration of Helsinki. Br Med J 1964;2:177. [15] Schneider CA, Rasband WS, Eliceiri KW. NIH Image to ImageJ: 25 years of image analysis. Nat Methods 2012;9:671–5. [16] Bianchi V, Robles R, Alberio L, Furlan M, Lammle B. Von Willebrand factorcleaving protease (ADAMTS13) in thrombocytopenic disorders: a severely deficient activity is specific for thrombotic thrombocytopenic purpura. Blood 2002; 100:710–3. [17] Peyvandi F, Palla R, Lotta LA, Mackie I, Scully MA, Machin SJ. ADAMTS-13 assays in thrombotic thrombocytopenic purpura. J Thromb Haemost 2010;8:631–40. [18] Rieger M, Ferrari S, Kremer Hovinga JA, Konetschny C, Herzog A, Koller L, et al. Relation between ADAMTS13 activity and ADAMTS13 antigen levels in healthy donors and patients with thrombotic microangiopathies (TMA). Thromb Haemost 2006; 95:212–20.