Both platelet- and endothelial cell-derived ERp5 support thrombus formation in a laser-induced mouse model of thrombosis

Abstract

Protein disulfide isomerase (PDI) and endoplasmic reticulum protein 57 (ERp57) are emerging as important regulators of thrombus formation. Another thiol isomerase, endoplasmic reticulum protein 5 (ERp5), is involved in platelet activation. We show here the involvement of ERp5 in thrombus formation using the mouse laser-injury model of thrombosis and a specific antibody raised against recombinant ERp5. Anti-ERp5 antibody inhibited ERp5-dependent platelet and endothelial cell disulfide reductase activity in vitro. ERp5 release at the thrombus site was detected after infusion of Alexa Fluor 488-labeled anti-ERp5 antibody at 0.05 μg/g body weight, a dose that does not inhibit thrombus formation. Anti-ERp5 at 3 μg/g body weight inhibited laser-induced thrombus formation in vivo by causing a 70% decrease in the deposition of platelets and a 62% decrease in fibrin accumulation compared to infusion of control antibody (P < .01). ERp5 binds to β3 integrin with an equilibrium dissociation constant (KD) of 21 µM, measured by surface plasmon resonance. The cysteine residues in the ERp5 active sites are not required for binding to β3 integrin. These results provide evidence for a novel role of ERp5 in thrombus formation, a function that may be mediated through its association with αIIbβ3.

Figure 1

Purification of wild-type ERp5, ERp5-AGHA, and polyclonal antibodies to ERp5. (A) Purified wild-type (WT) ERp5 (2 μg) and ERp5-AGHA (2 μg) were subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis and stained with coomassie blue dye. (B) Disulfide reductase activity of wild-type ERp5 (100 nM; □) and ERp5-AGHA (100 nM; ○). The relative increase in fluorescence produced by the reduction of di-E-GSSG is reported as a function of time. Di-E-GSSG probe plus DTT alone (×) serves as a negative control. (C) Immunoaffinity-purified anti-ERp5 antibody (0.5 μg/mL) detects wild-type ERp5 (50 ng) but does not detect PDI (500 ng) on western blot analysis. (D) ELISA of immunoaffinity-purified anti-ERp5 antibody (0.1 ng/mL) binding to recombinant His-tagged ERp5, His-tagged ERp72, or His-tagged PDI coated at 0.1 μg per well. Background indicates no bound thiol isomerases. N = 3 in triplicate; error bars represent 2 standard deviations. (E) ELISA of immunoaffinity-purified anti-ERp5 antibody (0.1 ng/mL) binding to recombinant His-tagged ERp5 or His-tagged ERp57. Background indicates no bound thiol isomerases. Proteins were coated at 0.1 μg per well. N = 3 in triplicate; error bars represent 2 standard deviations. These assays were developed with goat anti-rabbit IgG conjugated to HRP and to HRP chromogenic substrate tetramethylbenzidine, and OD was measured at 650 nm (D). In some experiments, the tetramethylbenzidine reaction was terminated with the addition of 50 μL of 0.16 M sulfuric acid, and OD was measured at 450 nm (E). MW, molecular weight; OD, optical density; RFU, relative fluorescence units.

Figure 2

Inhibition of ERp5 disulfide reductase and isomerase activity by anti-ERp5 antibody. (A) ERp5 disulfide reductase activity was measured by reduction of di-E-GSSG as a function of time. The inhibition of ERp5-catalyzed (50 nM) reduction of di-E-GSSG was determined in the presence of increasing amounts of anti-ERp5 antibody. His-PDI (50 nM) reductase activity (B) or His-ERp57 (100 nM) reductase activity (C) was measured in the di-E-GSSG assay in the presence of anti-ERp5 antibody. For panels A-C, anti-ERp5 antibody: none (□), 0.3 µM (▲), 0.6 µM (●), and 2.2 µM (○); preimmune IgG: 2.2 µM (▪); DTT, no enzyme: (△). N = 3 in triplicate. (D) The isomerase activity of ERp5 (1 μM) was measured in the RNase renaturation assay. The inhibition of isomerase activity was measured in the presence of increasing concentrations of anti-ERp5 antibody. (E) The isomerase activity of PDI (1 µM) was measured in the RNase renaturation assay in the absence and presence of anti-ERp5 antibody. (F) The isomerase activity of ERp57 (2.5 µM) was measured in the RNase renaturation assay in the absence and presence of anti-ERp5 antibody. For panels D-F, anti-ERp5 antibody: none (□), 0.3 µM (▲), 0.6 µM (●), and 2.2 µM (○); preimmune IgG: 2.2 µM (▪); denatured and reduced RNase: (△). N = 3 in triplicate. Abs, absorbance.

Figure 3

Secretion of ERp5 from platelets and endothelial cells. (A) Detection of ERp5 in lysate and supernatant of human and mouse (C57BL/6) platelets. Equal number of platelets were stimulated with thrombin (+IIa) at a dose of 0.5 U/mL or maintained in the resting state (−IIa). The supernatant (containing the platelet releasate) was separated from the platelets by centrifugation, and the platelets were lysed. Platelet lysates were probed for ERp5 and GAPDH. Platelet supernatants were probed for ERp5. (B) Expression of ERp5 on the surface of mouse platelets. Washed platelets were prepared from C57BL/6 mouse blood and activated with mouse thrombin (0.5 U/mL). Resting (−IIa) and activated (+IIa) platelets were incubated with monoclonal anti-human ERp5 antibody or isotype control (IgG) antibody, both labeled with Alexa Fluor 647. (C) Detection of ERp5 in the supernatant and lysate of cultured HUVECs before (−IIa) and after (+IIa) stimulation with 0.5 U/mL thrombin. HUVEC lysate was probed for ERp5 and GAPDH. HUVEC supernatant was probed for ERp5. (D) Release of ERp5 from thrombin-stimulated HUVECs over time as determined by densitometry compared to ERp5 control. N = 3; error bars represent standard deviation; **P < .005. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; pg, picograms.

Figure 4

Inhibition of function of platelet and endothelial ERp5 with anti-ERp5 antibody. Inhibition of disulfide reductase activity on the activated cell surface of human platelets and HUVECs measured in the di-E-GSSG reduction assay. (A) Percent reductase activity of platelets in the presence of anti-ERp5 antibody (0.3 µM, gray; 1.2 μM, white) compared to reductase activity of platelets in the presence of control IgG (1.2 μM, black). (B) Percent reductase activity of HUVECs in the presence of anti-ERp5 antibody (0.3 µM, gray; 2.2 μM, white) compared to reductase activity of HUVEC in the presence of control IgG (2.2 μM, black). N = 3 in triplicate; *P < .05. (C) Inhibition of mouse platelet aggregation with anti-ERp5 antibody. Platelets were incubated with control IgG or anti-ERp5 antibody (0.2 μM) and subsequently stimulated with thrombin (0.2 U/mL).

Figure 5

ERp5 is expressed in vivo in the developing thrombus. Anti-ERp5 antibody labeled with Alexa Fluor 488 (0.05 µg/g body weight) or preimmune IgG labeled with Alexa Fluor 488 (0.05 µg/g body weight) and anti CD-42b antibody labeled with DyLight 649 (0.1 µg/g body weight) were infused into a mouse 5 to 10 minutes prior to arteriolar injury. (A) Median total integrated fluorescence for anti-ERp5 (black; 24 thrombi from 3 mice) compared to median of the total integrated fluorescence for control IgG (gray; 27 thrombi from 3 mice); **P < .005. (B) Mice were injected intravenously with eptifibatide (10 µg/g body weight) 10 minutes prior to vessel injury and every 20 minutes after initial injection, and the experiment in panel A was repeated. Median total integrated fluorescence vs time after vessel injury for anti-ERp5 antibody (black) and IgG control (gray) is shown. Platelet accumulation (C) and fibrin generation (D) were measured before (1) and after the injection of eptifibatide alone (2) or eptifibatide followed by anti-ERp5 antibody (3 µg/g body weight) (3). *P < .05;***P < .001. F, median fluorescence intensity.

Figure 6

Anti-ERp5 antibody inhibits thrombus formation in vivo. (A) Anti-CD42b antibody labeled with DyLight 649 (0.1 µg/g body weight) and anti-fibrin-specific antibody labeled with Alexa Fluor 488 (0.5 µg/g body weight) were infused into a mouse 5 to 10 minutes prior to arteriolar injury. Preimmune IgG at 3 µg/g body weight (left) or anti-ERp5 antibody at 1 µg/g body weight (middle) and 3 µg/g body weight (right) were infused intravenously 20 minutes prior to injury. Representative binarized images of the fluorescent signal from platelets (red) and fibrin (green) over 180 seconds after laser-induced vessel wall injury show inhibition of thrombus formation with increasing concentrations of anti-ERp5 antibody. Median integrated platelet (B) and fibrin (C) fluorescence over time is shown before (1) vs after infusion of control IgG, 3 µg/g body weight (2); after anti-ERp5 antibody, 1 µg/g body weight (3); and after anti-ERp5 antibody, 3 µg/g body weight (4). **P < .01.

Figure 7

Binding of ERp5 and ERp5-AGHA to αIIbβ3 and directly to the β3 subunit. (A) Binding of αIIbβ3 to immobilized ERp5 or ERp5-AGHA was measured in the presence or absence of MnCl₂ (Mn⁺⁺; 2 mM), in the presence of EDTA (5 mM), or in the absence of added αIIbβ3 (background), as indicated. Bound αIIbβ3 was detected with an anti-CD41 conformation-independent antibody. ERp5 (black) and ERp5-AGHA (gray) binding to αIIbβ3 were compared by nonparametric Student t tests: no Mn⁺⁺ vs Mn⁺⁺ and no Mn⁺⁺ vs background,***P < .0001. There is no significant difference between no Mn⁺⁺ vs EDTA. Results are the average of 3 experiments. (B) Binding of ERp5 (100 nM) or ERp5-AGHA (100 nM) to αIIbβ3 (black) or glycoprotein (GP)Ibα (gray), both coated at 20 nM on 96-well plates in the absence of Mn⁺⁺ or in the absence of bound ERp5 or ERp5-AGHA (background; white). Bound ERp5 or ERp5-AGHA was detected with anti-ERp5 antibody. ERp5 binding to αIIbβ3 vs GPIba, ERp5-AGHA binding to αIIbβ3 vs GPIbα, ERp5 binding to αIIbβ3 vs background, and ERp5-AGHA binding to αIIbβ3 vs background, **P < .005. Binding of ERp5 or ERp5-AGHA to GPIbα vs background was not statistically significant. Results are the average of 3 experiments. Surface plasmon resonance was used to determine the binding constant for ERp5 with recombinant β3 integrin tagged with calmodulin in the presence (C) or absence of Mn⁺⁺ (D). ERp5 was coated on the Biacore chip. Lines represent the fitted curves of ERp5 at concentrations of 0, 0.33, 0.66, 1.31, 2.63, 5.25, 10.25, 21, and 42 µM. The K_D values for ERp5 interaction with the β3 subunit in the presence and absence of Mn⁺⁺ were 21.7 μM and 19.1 μM, respectively. ERp5 did not bind to calmodulin alone. n.s., not significant; RU, relative units.