Platelet-VWF complexes are preferred substrates of ADAMTS13 under fluid shear stress

Abstract

Endothelial cells secrete prothrombotic ultralarge von Willebrand factor (VWF) multimers, and the metalloprotease ADAMTS13 cleaves them into smaller, less dangerous multimers. This reaction is stimulated by tensile force applied to the VWF substrate, which may occur on cell surfaces or in the circulating blood. The cleavage of soluble VWF by ADAMTS13 was accelerated dramatically by a combination of platelets and fluid shear stress applied in a cone-plate viscometer. Platelet-dependent cleavage of VWF was blocked by an anti-GPIbalpha monoclonal antibody or by a recombinant soluble fragment of GPIbalpha that prevents platelet-VWF binding. Multimeric gel analysis showed that shear and platelet-dependent cleavage consumed large VWF multimers. Therefore, ADAMTS13 preferentially acts on platelet-VWF complexes under fluid shear stress. This reaction is likely to account for a majority of VWF proteolysis after secretion and to determine the steady-state size distribution of circulating VWF multimers in vivo.

Figure 1

Fluid shear stress and platelet-VWF interactions increase VWF cleavage by ADAMTS13. Reactions containing VWF (30 μg/mL) were treated with (+) or without (−) shear stress and other components as indicated. Samples were analyzed by SDS-PAGE and Western blotting with polyclonal anti-VWF to detect the 350-kDa homodimeric cleavage product containing C-terminal fragments of VWF. The corresponding 280-kDa N-terminal homodimeric product is recognized less efficiently by anti-VWF and was visible in some reactions. (A) Cleavage of VWF by ADAMTS13 (5 U/mL) was increased slightly by fluid shear stress (50 dyne/cm² for 10 minutes, lane 7) or formalin-fixed platelets (10⁶/μL, lane 4). Maximal cleavage required both shear stress and platelets (lane 8) and was blocked by 5 mM EDTA (lane 9). (B) Monoclonal antibody 6D1 against GPIbα (30 μg/mL) blocked the cleavage of VWF exposed to fluid shear stress (50 dyne/cm² for 10 minutes) in reactions containing platelets (lane 8) but not in reactions without platelets (lane 4). Control mouse IgG1 had no effect (lanes 3 and 7). (C) Recombinant GPIbα-Ig/2V (30 μg/mL) increased the cleavage of VWF by ADAMTS13 (2.5 U/mL) in reactions exposed to fluid shear stress (16 dyne/cm² for 5 minutes) without platelets (lane 4) and partially inhibited the cleavage of VWF in reactions with platelets (lane 8). (D) The cleavage of endogenous VWF in fresh platelet-rich plasma subjected to fluid shear stress (20 dyne/cm² for 5 minutes) was markedly greater than the cleavage of VWF in platelet-poor plasma. Cleavage was prevented by 50 mM EDTA.

Figure 2

VWF cleavage depends on shear stress, time, platelets, and ADAMTS13. (A) Reactions containing VWF (30 μg/mL) and ADAMTS13 (2.5 U/mL), without platelets (□) or with platelets (1 × 10⁶/μL) (■), were subjected to different levels of fluid shear stress for 3 minutes and analyzed for VWF cleavage by SDS-PAGE and Western blotting with polyclonal anti-VWF. Error bars indicate SD (n = 5, ■) or the range of duplicates (□). (B) Reactions containing VWF (30 μg/mL) and ADAMTS13 (2.5 U/mL), without (●, ○) or with formalin-fixed platelets (10⁶/μL) (■, □) were performed without (□, ○) or with (■, ●) 50 dyne/cm² shear stress for different times and analyzed for VWF cleavage. ■, 50 dyne/cm² shear stress and 10⁶ platelets/μL (n = 5); ●, 50 dyne/cm² and no platelets (n = 2); □, no shear stress and 10⁶ platelets/μL (n = 2); ○, no shear stress and no platelets (n = 1). Error bars indicate SD (■) or the range of duplicates (●, ○). (C) Reactions containing the indicated concentration of ADAMTS13 (U/mL) with or without platelets (10⁶/μL) were exposed to fluid shear stress (16 dyne/cm² for 5 minutes) and analyzed for VWF cleavage by SDS-PAGE and Western blotting as described in the legend to Figure 1.

Figure 3

Sigmoidal dependence of VWF cleavage on platelet count. Reactions containing VWF (30 μg/mL), without (−) or with (+) ADAMTS13 (2.5 U/mL), and various concentrations of platelets were exposed to fluid shear stress (50 dyne/cm², 3 minutes) and analyzed by SDS-PAGE and Western blotting. (A) Examples of platelet dose response results with formalin-fixed platelets (top) and fresh washed platelets (bottom). (B) Data with fixed platelets were combined by normalizing the density of the 350-kDa product band to the density of a cleavage product standard (Std). Error bars indicate SE (n = 9).

Figure 4

Platelets increase the cleavage of large VWF multimers under fluid shear stress. Reactions containing VWF (30 μg/mL) were performed as indicated with (+) or without (−) ADAMTS13 (2.5 U/mL), platelets (2 × 10⁶/μL), and fluid shear stress (16 dyne/cm², 10 minutes). The distribution of VWF multimers was assessed by 1.5% SDS-agarose gel electrophoresis.

Figure 5

Shear stress induces large VWF multimers to bind platelets. Reactions were performed as indicated with (+) or without (−) VWF (30 μg/mL), formalin-fixed platelets (10⁶/μL), and fluid shear stress (16 dyne/cm², 10 minutes). Samples were diluted with an equal volume of reaction buffer and centrifuged for 5 minutes at 6400g. The platelet pellets (or any other pelleted material for reactions without platelets) were washed 3 times by resuspension in reaction buffer and centrifugation. The washed platelets or other contents were dispersed in Laemmli sample buffer, heated at 80°C for 5 minutes, and centrifuged, and supernatant corresponding to 20 μL of the initial reaction volume was analyzed by Western blotting after 1.5% SDS-agarose gel electrophoresis. The 20 ng of input VWF analyzed for reference corresponds to approximately 3% of the VWF initially present in the samples of reactions analyzed in lanes 1 to 4.

Figure 6

Binding of platelets to VWF modeled as a function of platelet count and multimer length. The fractional occupancy of platelet binding sites on VWF multimers was calculated from the K_d for platelet GPIbα-A1 domain binding (∼4 μmol/L),^– plasma VWF concentration (∼10 μg/mL), VWF subunit mass (∼250 kDa), and number of GPIbα per platelet (∼28 000). The probability that a multimer has bound a particular number x of platelets is given by the binomial formula for P(x,n), where n is the number of VWF A1 domains (or subunits) per multimer and y is the fractional occupancy of VWF domain A1 sites by GPIbα ([GPIbα-A1]/[A1]_total): P(0,n) = (1 − y)ⁿ and P(1,n) = ny(1 − y)ⁿ⁻¹. Therefore, the probability of having at least 1 platelet bound per multimer is 1 − (1 − y)ⁿ, and the probability of finding at least 2 platelets bound per multimer is 1 − {(1 − y)ⁿ + ny(1 − y)ⁿ⁻¹}. At a fixed multimer length (n = 20 subunits), P is greater than or equal to 2 platelets/multimer increases exponentially with increasing platelet count (A), whereas P is greater than or equal to 1 platelet/multimer increases approximately linearly with increasing platelet count (B). At a fixed platelet count (250 000/μL), P is greater than or equal to 2 platelets/multimer increases exponentially with increasing multimer length (C), whereas P is greater than or equal to 1 platelet/multimer increases approximately linearly with increasing multimer length (D).