The role of ADAM proteolysis and mechanical damage in non-physiological shear stress induced platelet receptor shedding

Abstract

In order to explore the role of a disintegrin and metalloproteinase (ADAM) proteolysis and direct mechanical damage in non-physiological shear stress (NPSS)-caused platelet receptor shedding, the healthy donor blood treated with/without ADAM inhibitor was exposed to NPSS (150 Pa). The expression of the platelet surface receptors glycoprotein (GP) Ibα and GPVI in NPSS-damaged blood was quantified with flow cytometry. The impact of ADAM inhibition on adhesion of NPSS-damaged platelets on von Willibrand factor (VWF) and collagen was explored with fluorescence microscopy. The impact of ADAM inhibition on ristocetin- and collagen-caused aggregation of NPSS-damaged platelets was examined by aggregometry. The results showed that ADAM inhibition could lessen the NPSS-induced loss of platelet surface receptor GPIbα (12%) and GPVI (9%), moderately preserve adhesion of platelets on VWF (7.4%) and collagen (8.4%), and partially restore the aggregation of NPSS-sheared platelets induced by ristocetin (18.6 AU*min) and collagen (48.2 AU*min). These results indicated that ADAM proteolysis played a role in NPSS-induced receptor shedding. However, the ADAM inhibition couldn’t completely suppress the NPSS-caused loss of the platelet surface receptors (GPIbα and GPVI), only partially prevented the NPSS-induced reduction of platelet adhesion to VWF and collagen, and the agonist (ristocetin and collagen)-caused platelet aggregation. These results suggested that the direct mechanical damage is partially responsible for NPSS-induced receptor shedding in addition to the ADAM proteolysis. In conclusion, NPSS relevant to blood contacting medical devices can induce ADAM proteolysis and direct mechanical damage on the platelet receptor GPIbα and GPVI, leading to comprised hemostasis.

Introduction

Bleeding is the serious adverse event in patients implanted with ventricular assist devices (VADs), extracorporeal membrane oxygenation (ECMO), mechanical heart valves (MHVs) and other blood contacting medical devices (BCMDs)^1–3. Many factors are likely to contribute this adverse event in BCMD patients, such as excessive anticoagulation, the damage of high molecular weight multimers (HMWM) of von Willebrand factor (VWF), arteriovenous malformations ^4–7. Until now, bleeding remains an unsolved problem.

Platelets are key players in hemostasis. At the blood vessel injured area, the subendothelial matrix proteins, particularly VWF and collagen are exposed to blood stream⁸. Because of the bindings of platelet glycoprotein (GP) Ibα (GPIbα) to VWF and of platelet GPVI to collagen, platelets can quickly adhere and aggregate at the injured site leading to the initiation of normal hemostasis^9,10. The defect in these platelet adhesive receptors can lead to compromised adhesion of platelet with substrate proteins, impairing physiological hemostasis and increasing the risk for bleeding ^11,12.

In the context of assisted circulation with BCMDs, non-physiological shear stress (NPSS) usually appears in some areas within a device, for example, the bladed regions of the rotor in VADs ^13,14 and the regurgitant jet of MHVs¹⁵. The levels of NPSS can be higher than 100 Pascal (Pa) and even reach 1000 Pa¹³. This high level of NPSS (> 100 Pa) could induce damage to blood cells in patients implanted with BCMDs. Platelet receptor shedding had been found in patients supported with BCMDs. Lukito et al.¹⁶ found the elevated GPVI concentration in plasma and notably reduced surface expression levels of platelet receptor GPIbα and GPVI in ECMO and continuous flow (CF) VAD patients when compared to those in healthy subjects. They found that the mechanical circulation support was related with the shedding of receptor GPIbα and GPVI. Hu et al.¹⁷ compared the GPIbα shedding in the bleeding and non-bleeding groups of patients with CF-VADs and found that the level of GPIbα shedding was larger in the bleeding group after implantation. These results suggested that bleeding in CF-VAD patients was associated with platelet receptor shedding.

It has been known that the receptor cleavage (shedding) is the proteolytic release of ectodomain of the membrane receptor at the juxtamembrane region¹⁸. This process can produce a membrane-bound remnant fragment and a soluble bioactive fragment. These soluble fragments may be potential platelet-specific biomarkers that can bind to receptors on other cells, modify their behavior, and contribute to related biological processes¹⁹. Platelet receptor ectodomain cleavage mediated by metalloproteinase has been recently known as a mechanism for modifying platelet function²⁰. Shedding of surface receptor GPIbα and GPVI is believed to be dependent of a disintegrin and metalloproteinase (ADAM). GPVI is mainly cleaved by ADAM10, and ADAM17 plays a major role in the cleavage of GPIbα ^18,21,22.

Mechanisms of shear stress-induced platelet receptor shedding have not been well defined. Only a few investigations have been performed to examine the shear-mediated platelet receptor shedding either in the context of platelet-surface interaction or direct shear effect ^23,24. Cheng et al.²³ reported that shear-caused platelet-VWF binding leads to metalloproteinase-dependent GPIbα ectodomain cleavage. Al-Tamimi et al.²⁴ found that GPVI shedding was independent of VWF interacting to GPIbα, GPIIb/IIIa engagement, and platelet activation, suggesting a different signaling pathway to ADAM10. In these studies, shear stress-induced platelet receptor shedding was suggested to be caused by proteolysis, which was a biochemical process. Whether the high shear stress-induced receptor shedding can be caused by directly mechanical damage is still unclear. Additionally, in our previous study, we discovered that NPSS could cause platelet surface receptor (GPIbα and GPVI) shedding^25–27. It is also unknown that, when platelets are subjected to the NPSS condition related with BCMDs, whether the platelet receptor shedding is induced by ADAM proteolysis or direct mechanical damage or a combination of both. We hypothesize that the mechanical shear stress at a high level (> 100 Pa) as an extracellular stimulus may not only trigger a series of intracellular biochemical signaling pathways via surface receptor-mediated responses, but also cause direct mechanical shedding that is distinct from the intercellular signaling (Fig. 1). In the present study, the roles of the ADAM proteolysis and direct mechanical damage in the ectodomain cleavage of surface receptors (GPIbα, GPVI) caused by NPSS were investigated. Healthy donor blood was treated with or without ADAM inhibitor (GM6001) and subsequently exposed to NPSS using our blood-shearing device. Surface expression levels of receptor GPIbα and GPVI, platelet adhesion and agonist-induced platelet aggregation were characterized.

Figure 1:

The potential mechanism diagram for showing the role of ADAM proteolysis and mechanical damage in NPSS-caused platelet receptor shedding.

Materials and methods

Blood Collection

Blood used in this study was drawn from seven healthy adult donors (four men and three women with age from 24 to 31 years) without taken any antiplatelet or anticoagulant medicine for two weeks prior to blood donation was used in this study. The blood donation procedure was approved by the Institutional Review Board approval of the University of Maryland School of Medicine. After being informed of the purpose of this study, all donors signed a consent form before blood donation. Fresh whole blood (450 ml) was collected from donor’s antecubital vein collected into a sterilized bag containing 3.2% buffered sodium citrate solution (50ml).

Blood-shearing device

In this study, we used an axial flow-through Couette-type device to create the NPSS condition. The blood-shearing device consists of a stationary housing and a rotating spindle (Jarvik Heart, Inc., New York, NY). There is a narrow gap with a fixed width of 100 µm between the middle spindle surface and the housing wall. Our previous study has described this device detailly²⁸. Blood was pressure-driven by a syringe pump (PHD 2000, Harvard Apparatus, Holliston, MA) to move through the small gap in the axial direction. The blood was exposed to uniform NPSS in this narrow gap for a controlled duration (Fig. 2A).

Figure 2:

A) The blood-shearing system, which includes the syringe, syringe pump, blood-shearing device, connected tube, inlet and outlet port for collecting blood samples, and reservoir;

B) The physical picture of experimental setup.

Blood-shearing experiment

All the surfaces contacted with blood, such as blood-shearing device, tubing, and inner surface of syringe, were initially rinsed with 0.9% saline before experiment. A semi-micro viscometer with the shear rate range from 80.3 s⁻¹ to 321 s⁻¹ (Cannon Instrument Company, State College, PA) was used to measure the blood viscosity. The rotated speed of the spindle within the blood-shearing device was determined based on the desired level of NPSS and the measured blood viscosity. The measured blood viscosity values of our seven participants were range from 4 Pa·s to 4.5 Pa·s. The flow rate was chosen according to the desired exposure time. The high shear stress level of 150 Pa and exposure time of 1 sec relevant to BCMDs were selected in this study. To inhibit the ADAM activation, a broad-spectrum matrix metalloproteinase inhibitor, GM6001 (Millipore Sigma, Burlington, MA) was selected for the study based on previously reported studies^24,29. As reported by Al-Tamimi et al.²⁴, the metalloproteinase inhibitor GM6001 with the concentration of 100 µM can completely block the platelet receptor GPVI shedding under the pathologic shear rate of 10000 s⁻¹ (35 Pa) condition. In this study, GM6001 with the concentration of 200 µM was used to make sure that the activity of ADAM was blocked. Whole blood was first incubated with 200 µM GM6001 for ten minutes at room temperature (RT) before exposed to the high shear condition. Then, the blood preincubated with or without GM6001 was loaded into our blood shearing system and exposed to the NPSS of 150 Pa for 1 sec. For each shearing experiment, a baseline blood sample and a sheared blood sample were obtained from the inlet port and the outlet port of the shearing device, respectively (Fig. 2B). The blood shearing experiment was repeated 7 times.

Flow cytometry analysis of platelet surface receptor GPIbα and GPVI shedding

V450 Mouse Anti-Human CD41a antibody (CD41a-V450), BV510 Mouse Anti-Human CD42b antibody (CD42b-BV510) (both from BD Bioscience, San Jose, CA) and efluor 660-labeled anti-human GPVI (GPVI-efluor 660) (eBioscience, San Diego, CA) were used to identify platelets, and to quantify the receptor GPIbα and GPVI surface expression, respectively. The isotype controls for CD42b-BV510 and GPVI-efluor 660 were IgG1K-BV510 and IgG1K- efluor 660, respectively. To assess GPIbα and GPVI surface expression, 5 µl collected blood samples (baseline and sheared) were immediately incubated with 5 µl CD42b-BV510 or IgG1K-BV510, 5 µl GPVI-efluor 660 or IgG1K-eFluor 660, 5 µl CD41a-V450 and 25 µl HEPES buffer. After 30 min incubation at RT in the dark, the samples were fixed with 1 ml of 1% paraformaldehyde (PFA) in phosphate buffered saline (PBS) for 30 min at 4^oC in the dark. The flow cytometry data acquisition was performed on a flow cytometer (FACSVerse, BD Bioscience, San Jose, CA). FCS express software package (De Novo Software, Glendale, CA) was used to analyze flow cytometry data.

Analysis of platelet adhesion to VWF and collagen-immobilized surface

In this study, the VWF with concentration of 100 µg/ml (EMD Millipore, Billerica, MA) and collagen with the concentration of 1 mg/ml (Chrono-log, Havertown, PA) were coated in rectangular capillary tubes with the size of 0.2 mm × 2 mm ×25 mm (VitroCom, Mountain Lakes, NJ) overnight at 4 degree. The collected baseline and NPSS-damaged blood samples were first labeled with 10 μM mepacrine (Sigma, St. Louis, MO) and then perfused through the capillary tubes coated with proteins under the physiological shear rate of 500 s⁻¹ for five minutes in the dark. After the perfusion, the capillary tubes were rinsed with PBS and adherent platelets were fixed with 3.7% PFA. The platelet adhesion to VWF and collagen was quantified with the Olympus IX71 fluorescence microscope equipped with the Olympus DP80 digital camera. Adhesion images were obtained with the Olympus cellSens imaging software (Shinjuku-ku, Tokyo, Japan). For each capillary tube, we took 10 images with 2 mm interval along the center axis. The area coverage (in percent) of deposited platelets on VWF and collagen was computed from the 10 images using a custom-written program in MATLAB (MathWorks, Inc., Natick, MA, USA).

Platelet aggregation test

The Multiplate® analyzer (Verum Diagnostica GmbH, Munich, Germany) was used to assess the platelet aggregation capacity of baseline and NPSS-damaged blood samples induced by ristocetin and collagen. Ristocetin can promote platelet aggregation by inducing VWF binding to platelet receptor GPIbα. Collagen can cause the platelet aggregation via interaction with receptor GPVI. Briefly, 300 µl whole blood was first mixed with 300 µl of 0.9% NaCl solution or saline solution with 3 mM calcium chloride (NaCl/CaCl₂) in a test cells. These samples were stirred and incubated for 3 minutes at 37^oC. After incubation, the ristocetin solution (50 µl, final concentration 0.77 mg/ml) was added to the mixed blood sample and collagen solution (25 µl, final concentration 3.2 μg/ml) was added to the mixed blood sample with NaCl/CaCl_2. The agonist-induced platelet adhesion and aggregation on the electrode wires can cause the impedance change between the electrode wires in the test cell. The change in electrical impedance was transformed to arbitrary aggregation units (AU) and shown as 2 separated aggregation (impedance) curves against time. The aggregation was continuously recorded for 6 minutes. Platelet aggregation measurement was quantified by the area under the aggregation curve (AUC) with the arbitrary unite (AU*min). The difference from the averaged curve was assessed based on the AUC values of the 2 separated aggregation curves. The analysis was acceptable when the difference was within 20%. In our study, the different between two curves for all our experimental conditions was within 10%. The specific steps of this aggregation test were reported by Tóth O’s study³⁰.

Statistical Analysis

Results are expressed as mean ± SE (standard error) (n=7). Statistical differences between two sample groups are determined by using the one-way ANOVA with Tukey multiple comparison post-test (SPSS statistics software package version 18.0, Chicago, IL). A *P value < 0.05 is considered to be statistically significant.

Results

To investigate the role of ADAM proteolysis and direct mechanical damage in NPSS-induced platelet receptor shedding, the whole blood treated with and without GM6001 was exposed to 150 Pa NPSS for 1 sec. The surface expression levels of receptor GPIbα and GPVI in the blood samples after exposure were quantified with flow cytometry. Figure 3 shows the representative histograms of the channel fluorescence intensity (FI) representing the expression level of platelet surface receptor GPIbα (Fig. 3A) or GPVI (Fig. 3C) in the baseline, NPSS-sheared, and GM6001 pretreated + NPSS-sheared blood samples and the normalized mean fluorescence intensity (MFI) of platelet GPIbα (Fig. 3B) and GPVI surface expression (Fig. 3D). The numbers of platelet receptors (GPVI and GPIbα) on the surface are directly related with their channel FI. As shown in the histograms, the FI peaks for receptor GPIbα and GPVI in the NPSS-damaged blood sample shifted to the left (low FI) compared to those in the baseline blood sample, implying NPSS-caused shedding of receptor GPIbα and GPVI. When the blood sample was pretreated with GM6001 and then exposed to NPSS, the FI peaks of receptor GPIbα and GPVI shifted back to the right (high FI) compared to those in the NPSS-damaged blood sample. This indicated that ADAM inhibition lessened the NPSS-induced cleavage of receptor GPIbα and GPVI. However, the FI peaks of receptor GPIbα and GPVI in the GM6001 treated and NPSS-damaged blood sample did not shift back to the locations of the baseline blood sample, indicating that the ADAM inhibition could not completely suppress the NPSS-caused cleavage of receptor GPIbα and GPVI. The quantitative comparison of the relative MFI levels of the platelet receptor GPIbα (Fig. 3B) and GPVI (Fig. 3D) showed that the ADAM inhibition lessened the NPSS-caused cleavage of the receptor GPIbα and GPVI. When compared to the MFI levels for platelet receptor GPIbα and GPVI of baseline blood, the MFI levels of the NPSS-sheared blood significantly (*P < 0.01) decreased to 78% for GPIbα and 85% for GPVI, respectively. When the blood was pretreated with GM6001 and then exposed to NPSS, the MFI levels reverted to 90% for GPIbα and 94% for GPVI, respectively. Because of the ADAM inhibition, there were certain degree reversion for the MFI levels of receptor GPIbα and GPVI in GM6001 pretreated and NPSS-sheared blood. However, the MFI levels of GPIbα and GPVI in GM6001 pretreated and NPSS-sheared blood were still significantly lower (*P < 0.05) compared to baseline blood. These results implied that the ADAM inhibition could revert the MFI levels of GPIbα and GPVI to certain degree, but could not completely eliminate the NPSS-caused platelet receptor cleavage.

Figure 3:

The comparison of platelet surface receptor GPIbα (A, B) and GPVI (C, D) shedding for baseline, NPSS-sheared (150 Pa/1 sec), and GM6001 pretreated (200 µM) + NPSS sheared (150 Pa/1 sec) blood samples. Representative flow cytometry histograms of surface GPIbα (A) and GPVI (C) levels of all blood samples. The quantification of surface GPIbα (B) and GPVI (D) levels. The normalized ratio of mean fluorescence intensity (MFI) was calculated by using sheared samples to compare with baseline sample (n=7, *P<0.05).

To confirm the role of ADAM inhibition and direct mechanical damage in the NPSS-caused loss of the receptor GPIbα and GPVI and its impact on platelet adhesion function, the baseline, NPSS-sheared and GM6001 pretreated + NPSS sheared blood samples were perfused through VWF- and collagen-coated capillary tubes at a physiological shear rate (500 s⁻¹). Figure 4A and 4B show typical fluorescence images of adhered platelets on VWF and the percentage of area coverage after perfusing the blood samples through the VWF coated glass tubes, respectively. Figure 5A and 5B show typical fluorescence images of adhered platelets on collagen and the percentage of area coverage after perfusing the blood samples through the capillary tubes coated with protein, respectively. For the baseline blood samples, the averaged area coverage of adherent platelets on VWF and collagen was 23.9% and 30.4%, respectively. When blood samples were exposed to NPSS condition, the averaged area coverage of NPSS-damaged platelet adhesion to VWF and collagen was 9.7% and 11%, respectively. The numbers of platelets adhered on VWF and collagen in the NPSS-damaged blood decreased compared with those of the unsheared blood samples (baseline). When the blood was pretreated with GM6001 and then exposed to NPSS, the averaged area coverage of GM6001-treated and NPSS-sheared platelets to VWF and collagen was 17.1% and 19.4%, respectively. The numbers of platelets adhered to VWF and collagen in both GM6001-treated and NPSS-sheared condition rebounded compared to those of the non-treated NPSS sheared blood, but were still significantly less (*P < 0.05) than those of the baseline blood. This observation implied that the ADAM inhibition could preserve the adhesion capacity of platelets on VWF and collagen to certain degree, but could not completely eliminate the impact of NPSS on platelet adhesion to VWF and collagen.

Figure 4:

The comparison of platelet adhesion to VWF for baseline, NPSS-sheared (150 Pa/1 sec), and GM6001 pretreated (200 µM) + NPSS sheared (150 Pa/1 sec) blood samples. (A) Typical fluorescent images of platelet attachment to VWF (magnification, X400); (B) The percentage of area covered by fluorescent platelet (n=7, *P<0.05).

Figure 5:

The comparison of platelet adhesion to collagen for baseline, NPSS-sheared (150 Pa/1 sec), and GM6001 pretreated (200 µM) + NPSS sheared (150 Pa/1 sec) blood samples. (A) Typical fluorescent images of platelet attachment to collagen (magnification, X400); (B) The percentage of area covered by fluorescent platelet (n=7, *P<0.05).

Platelet aggregation caused by ristocetin and collagen in the baseline, NPSS-damaged and GM6001 pretreated + NPSS sheared blood samples were examined to further confirm the role of ADAM proteolysis and mechanical damage in NPSS-caused loss of receptor GPIbα and GPVI. Figure 6 shows representative impedance (aggregation) time histories of the baseline, NPSS-sheared, and GM6001 pretreated + NPSS-sheared blood samples when simulated by ristocetin (Fig. 6A) and collagen (Fig. 6C) and the quantitation of the platelet aggregation caused by agonist (Fig. 6B and Fig. 6D). The averaged AUC values of platelet aggregation caused by ristocetin and collagen in baseline blood were 62.8 AU*min and 111.86 AU*min, respectively. When the blood was subjected to the NPSS of 150 Pa for 1 sec, the averaged AUC values of the ristocetin- and collagen-caused platelet aggregation were 14.4 AU*min and 45 AU*min, respectively, which was apparently reduced compared to that of unsheared blood (baseline). When the blood was pretreated with GM6001 and then exposed to the same level of NPSS, the averaged AUC values of the ristocetin- and collagen-caused platelet aggregation were 33 AU*min and 93.2 AU*min, respectively. For the GM6001-pretreated and NPSS-sheared blood, the platelet aggregation capacities induced by ristocetin and collagen were stronger than those of the NPSS-sheared blood, but weaker than those of the baseline blood. Again, the result implied that the ADAM inhibition suppressed, but not completely abolish the NPSS-induced reduction in platelet aggregation when stimulated by ristocetin and collagen.

Figure 6:

The comparison of platelet aggregation caused by ristocetin (A, B) and collagen (C, D) in baseline, NPSS-sheared (150 Pa/1 sec), and GM6001 pretreated (200 µM) + NPSS sheared (150 Pa/1 sec) blood samples. (A) and (C) show the typical impedance (aggregarion) curves of ristocetin- and collagen- caused platelet aggregation, respectively. The Y-axis is the aggregation, expressed in arbitrary aggregation units (AU) and the X-axis is the time, expressed in seconds. (B) and (D) show the quantitative assessment of ristocetin- and collagen- caused platelet aggregation, respectively (n=7, *P<0.05). The ristocetin- and collagen-caused platelet aggregation was assessed by the area under the aggregation curve (AUC) with arbitrary units (AU*min).

Discussion

When patients received medical treatment with BCMDs, their blood passes through devices repeatedly and blood components are inevitably subjected to high mechanical shear stress, i.e., NPSS. This high mechanical shear condition can induce damage on blood components. The shear-caused damage to erythrocytes and VWF had been found in many studies ^31–34. It had been shown that NPSS can cause erythrocytes deformation and membrane rupture, and NPSS can fragment HMWM-VWF. Recently, platelet receptor shedding had been found in patients implanted with BCMDs. A few studies had demonstrated that NPSS can induce platelet receptor shedding ^25,27,35. However, mechanisms of the NPSS-induced platelet receptor shedding remain unclear. Some experimental data had suggested that ADAM proteolysis might be the main reason to cause platelet receptor shedding (GPVI and GPIbα) ^22,36. It is still unknown whether ADAM proteolysis and direct mechanical damage occur in the NPSS-induced platelet receptor shedding. Our results showed that the ADAM inhibition could lessen the NPSS-induced loss of GPIbα and GPVI, and partially restored the adhesion capacities of platelets on VWF and collagen and the agonist-induced platelet aggregation capacities (ristocetin and collagen). These findings supported the notion that ADAM proteolysis played an important role in NPSS-induced platelet receptor shedding. However, the ADAM inhibition could not completely suppress the NPSS-induced platelet GPVI and GPIbα shedding. These findings suggested that the direct mechanical damage could be partially responsible for NPSS-induced platelet receptor shedding in addition to the ADAM proteolysis.

There may be two distinct mechanisms involved in the NPSS-induced platelet receptor shedding. NPSS may mechanically remove the ectodomains of platelet receptor GPIbα and GPVI and also trigger an intracellular signaling leading to ADAM activation. According to previous research findings, a possible signaling pathway of shear-induced ADAM activation may be involved with NPSS-mediated interaction of GPIbα with VWF which triggers the intracellular signaling cascades, such as calcium elevation ³⁷. The increase in intracellular calcium concentration caused the activation of protein kinase C³⁸, resulting in activation of ADAM by phosphorylating its 735 threonine ³⁹. The above signaling pathway of shear-induced ADAM activation will be tested in our future studies.

Although the shear-caused platelet receptor shedding had been had been carried out by a few previous studies ^23,24, these studies normally used viscometers or parallel chambers to create shear condition with relatively low shear stress levels and long exposure time. Such as, the shear stress levels from 0.875 to 7 Pa and exposure time from 10 to 60 seconds were used in the study of Cheng et al.²³, and the shear stress levels from 1 to 35 Pa and exposure time from 5 to 10 minutes were utilized in the study of Al-Tamimi et al.²⁴ These shear stress levels and exposure time were different from the NPSS condition generated by BCMDs^13,14. Different from those studies, our study used special design blood shearing devices to create the NPSS (150 Pa) with short exposure time (1 sec) which mimics the shear condition in BCMD patients. Of note, for the first time, this study found that the direct mechanical shear damage can cause platelet receptor shedding when platelets were exposed to NPSS with short exposure time. This direct mechanical shear damage which is different from the metalloproteinase proteolysis (biochemical reaction) may also be included in shear-induced damage of other blood cells in patients implanted with BCMDs. The finding of this study may help us find an effect method to prevent shear-induced receptor shedding and reduce the hemorrhage risk in BCMD patients.

This study has some limitations. (1) The shear vulnerability of platelet receptor in heart failure patients may be different from that of healthy donors. (2) The real clinical mechanical shear condition in patients implanted with BCMDs is complicated. In this study, this clinical shear condition was simplified to one representative NPSS condition (150 Pa / 1 sec) created by our axial flow-through Couette-type device (blood-shearing device). (3) Only one concentration of metalloproteinase inhibitor GM6001 (200 µM) was used in this study. Using different concentrations of GM6001 would improve our experimental design.

Conclusion

Our findings demonstrated the important role of ADAM proteolysis and direct mechanical damage in NPSS-induced platelet receptor shedding. NPSS can induce the ADAM proteolysis and cause direct mechanical damage on key platelet hemostatic receptors (GPIbα and GPVI). This study for the first time showed that the direct mechanical damage played a role in NPSS-induced platelet receptor shedding.

Abbreviations

VAD

ventricular assist device

ECMO

extracorporeal membrane oxygenation

MHV

mechanical heart valve

BCMD

blood contacting medical device

HMWM

high molecular weight multimer

VWF

von Willebrand factor

GPIbα

glycoprotein Ibα

GPVI

glycoprotein VI

NPSS

non-physiological shear stress

Pa

Pascal

CF-VAD

continuous flow ventricular assist device

ADAM

a disintegrin and metalloproteinase

GPIIb/IIIa

glycoprotein IIb/IIIa

RT

room temperature

CD41a-V450

V450 Mouse Anti-Human CD41a antibody

CD42b-BV510

BV510 Mouse Anti-Human CD42b antibody

GPVI-efluor 660

efluor 660-labeled anti-human GPVI antibody

IgG1K-BV510

BV510 Mouse IgG1K Isotype Control

IgG1K-efluor 660

efluor 660 Mouse IgG1K Isotype Control

PFA

paraformaldehyde

PBS

phosphate buffered saline

AU

arbitrary aggregation units

AUC

area under the aggregation (impedance) curve

FI

fluorescence intensity

MFI

mean fluorescence intensity