A Critical Role for ERO1α in Arterial Thrombosis and Ischemic Stroke

Abstract

Background: Platelet adhesion and aggregation play a crucial role in arterial thrombosis and ischemic stroke. Here, we identify platelet ERO1α (endoplasmic reticulum oxidoreductase 1α) as a novel regulator of Ca²⁺ signaling and a potential pharmacological target for treating thrombotic diseases.

Methods: Intravital microscopy, animal disease models, and a wide range of cell biological studies were utilized to demonstrate the pathophysiological role of ERO1α in arteriolar and arterial thrombosis and to prove the importance of platelet ERO1α in platelet activation and aggregation. Mass spectrometry, electron microscopy, and biochemical studies were used to investigate the molecular mechanism. We used novel blocking antibodies and small-molecule inhibitors to study whether ERO1α can be targeted to attenuate thrombotic conditions.

Results: Megakaryocyte-specific or global deletion of Ero1α in mice similarly reduced platelet thrombus formation in arteriolar and arterial thrombosis without affecting tail bleeding times and blood loss following vascular injury. We observed that platelet ERO1α localized exclusively in the dense tubular system and promoted Ca²⁺ mobilization, platelet activation, and aggregation. Platelet ERO1α directly interacted with STIM1 (stromal interaction molecule 1) and SERCA2 (sarco/endoplasmic reticulum Ca²⁺-ATPase 2) and regulated their functions. Such interactions were impaired in mutant STIM1-Cys49/56Ser and mutant SERCA2-Cys875/887Ser. We found that ERO1α modified an allosteric Cys49-Cys56 disulfide bond in STIM1 and a Cys875-Cys887 disulfide bond in SERCA2, contributing to Ca²⁺ store content and increasing cytosolic Ca²⁺ levels during platelet activation. Inhibition of Ero1α with small-molecule inhibitors but not blocking antibodies attenuated arteriolar and arterial thrombosis and reduced infarct volume following focal brain ischemia in mice.

Conclusions: Our results suggest that ERO1α acts as a thiol oxidase for Ca²⁺ signaling molecules, STIM1 and SERCA2, and enhances cytosolic Ca²⁺ levels, promoting platelet activation and aggregation. Our study provides evidence that ERO1α may be a potential target to reduce thrombotic events.

Keywords: calcium signaling; ischemic stroke; platelet activation; thrombosis.

Figure 1.. ERO1α has a crucial role in cremaster arteriolar and carotid arterial thrombosis without affecting tail bleeding times in mice.

(A-C) DyLight 649-conjugated anti-CD42c and Alexa 488-conjugated rabbit IgG or polyclonal anti-ERO1α antibodies (0.3 μg/g BW) were injected into WT mice for intravital microscopy. Platelet accumulation and extracellular Ero1α were detected at the site of laser-induced cremaster arteriolar injury. (A) Representative images. (B-C) The median integrated fluorescence intensities of anti-CD42c (F platelet) and anti-ERO1α antibodies (F Ero1α) (n = 41–43 arterioles in 6 WT mice/group). (D-F) Intravital microscopy with WT control and Ero1α CKO and KO mice. After laser-induced cremaster arteriolar injury, platelet accumulation and fibrin generation were detected by injection of DyLight 649-conjugated anti-CD42c and Alexa 488-conjugated anti-fibrin antibodies, respectively. (D) Representative images. (E-F) The median integrated fluorescence intensities of anti-CD42c (F platelet) and anti-fibrin (F fibrin) antibodies (n = 36–44 arterioles in 6 mice/group). Quantification of the antibody signal at a different time point after laser injury. (G-H) Ero1α KO mice were treated with recombinant wtERO1α or mERO1α (Cys94Ser), 4 μg/g BW. After laser injury, platelets and fibrin were detected as described above (n = 36–38 arterioles in 6 mice/group). (I) The time to occlusion (TTO) was measured by a Doppler blood flow meter in WT control, Ero1α CKO/KO, and β3 KO mice after the application of 6.5% FeCl₃ to a carotid artery (n = 8–11 mice/group). (J-K) After amputating the tail tip (5 mm), tail bleeding times and hemoglobin (Hb) contents were measured in WT control, Ero1α CKO/KO, and β3 KO mice (n = 8–9 mice/group). Bars represent the median value. P values determined by Mann-Whitney U-test (E-F and I-K) or ANOVA and Dunn’s test (G-H).

Figure 1.. ERO1α has a crucial role in cremaster arteriolar and carotid arterial thrombosis without affecting tail bleeding times in mice.

(A-C) DyLight 649-conjugated anti-CD42c and Alexa 488-conjugated rabbit IgG or polyclonal anti-ERO1α antibodies (0.3 μg/g BW) were injected into WT mice for intravital microscopy. Platelet accumulation and extracellular Ero1α were detected at the site of laser-induced cremaster arteriolar injury. (A) Representative images. (B-C) The median integrated fluorescence intensities of anti-CD42c (F platelet) and anti-ERO1α antibodies (F Ero1α) (n = 41–43 arterioles in 6 WT mice/group). (D-F) Intravital microscopy with WT control and Ero1α CKO and KO mice. After laser-induced cremaster arteriolar injury, platelet accumulation and fibrin generation were detected by injection of DyLight 649-conjugated anti-CD42c and Alexa 488-conjugated anti-fibrin antibodies, respectively. (D) Representative images. (E-F) The median integrated fluorescence intensities of anti-CD42c (F platelet) and anti-fibrin (F fibrin) antibodies (n = 36–44 arterioles in 6 mice/group). Quantification of the antibody signal at a different time point after laser injury. (G-H) Ero1α KO mice were treated with recombinant wtERO1α or mERO1α (Cys94Ser), 4 μg/g BW. After laser injury, platelets and fibrin were detected as described above (n = 36–38 arterioles in 6 mice/group). (I) The time to occlusion (TTO) was measured by a Doppler blood flow meter in WT control, Ero1α CKO/KO, and β3 KO mice after the application of 6.5% FeCl₃ to a carotid artery (n = 8–11 mice/group). (J-K) After amputating the tail tip (5 mm), tail bleeding times and hemoglobin (Hb) contents were measured in WT control, Ero1α CKO/KO, and β3 KO mice (n = 8–9 mice/group). Bars represent the median value. P values determined by Mann-Whitney U-test (E-F and I-K) or ANOVA and Dunn’s test (G-H).

Figure 2.. Intracellular ERO1α is crucial for platelet activation, aggregation, and thrombus formation.

(A) Ex vivo platelet thrombus formation in a flow chamber assay. Adherent and aggregated platelets were stained with rhodamine-conjugated phalloidin and imaged by confocal microscopy. Representative images (bar = 10 μm). Surface coverage and thrombus volume were measured (mean ± SD, n = 4). (B-E) Aggregation and ATP secretion of WT and Ero1α-null platelets were induced by 0.025 U/ml thrombin, 0.05 μg/ml CRP, 0.5 μM A23187, or 2.5 μM ADP. (i) Platelet aggregation and (ii) ATP secretion. (F) Ero1α-null platelets were pretreated with 50 μg/ml wtERO1α or wtPDI and then with 0.025 U/ml thrombin. Quantification graphs of aggregation and ATP secretion are presented as the mean ± SD (n = 3–4). (G-H) P-selectin exposure and αIIbβ3 integrin activation of WT and Ero1α-null platelets were analyzed by flow cytometry. The data are presented as the geometric mean fluorescence intensity (MFI) value (mean ± SD, n = 3–4). (I) WT and Ero1α-null platelets were treated with or without 0.025 U/ml thrombin for 1 minute in an aggregometer. The lysates were immunoprecipitated with control IgG or anti-β3 antibodies and subjected to immunoblotting with antibodies against talin1 or β3 and densitometric analysis (mean ± SD, n = 3). (J) WT and Ero1α-null platelets were treated with or without 0.025 U/ml thrombin, followed by transmission electron microscopy. The original magnification was 34,800, and the extension was 313,000. Bar = 500 nm. (K-L) Intravital microscopy was performed to determine the effect of 15E9 (1–3 μg/g BW) and a blocking anti-PDI antibody (BD34, 3 μg/g BW) on platelet thrombus formation and fibrin generation at the site of laser-induced cremaster arteriolar injury (n = 27–40 arterioles in 6 mice/group). The n indicates biological replicates. P values determined by unpaired Student’s t-test (A-E and G-I), ANOVA and Dunnett’s test (F), or Mann-Whitney U-test (K-L).

Figure 2.. Intracellular ERO1α is crucial for platelet activation, aggregation, and thrombus formation.

(A) Ex vivo platelet thrombus formation in a flow chamber assay. Adherent and aggregated platelets were stained with rhodamine-conjugated phalloidin and imaged by confocal microscopy. Representative images (bar = 10 μm). Surface coverage and thrombus volume were measured (mean ± SD, n = 4). (B-E) Aggregation and ATP secretion of WT and Ero1α-null platelets were induced by 0.025 U/ml thrombin, 0.05 μg/ml CRP, 0.5 μM A23187, or 2.5 μM ADP. (i) Platelet aggregation and (ii) ATP secretion. (F) Ero1α-null platelets were pretreated with 50 μg/ml wtERO1α or wtPDI and then with 0.025 U/ml thrombin. Quantification graphs of aggregation and ATP secretion are presented as the mean ± SD (n = 3–4). (G-H) P-selectin exposure and αIIbβ3 integrin activation of WT and Ero1α-null platelets were analyzed by flow cytometry. The data are presented as the geometric mean fluorescence intensity (MFI) value (mean ± SD, n = 3–4). (I) WT and Ero1α-null platelets were treated with or without 0.025 U/ml thrombin for 1 minute in an aggregometer. The lysates were immunoprecipitated with control IgG or anti-β3 antibodies and subjected to immunoblotting with antibodies against talin1 or β3 and densitometric analysis (mean ± SD, n = 3). (J) WT and Ero1α-null platelets were treated with or without 0.025 U/ml thrombin, followed by transmission electron microscopy. The original magnification was 34,800, and the extension was 313,000. Bar = 500 nm. (K-L) Intravital microscopy was performed to determine the effect of 15E9 (1–3 μg/g BW) and a blocking anti-PDI antibody (BD34, 3 μg/g BW) on platelet thrombus formation and fibrin generation at the site of laser-induced cremaster arteriolar injury (n = 27–40 arterioles in 6 mice/group). The n indicates biological replicates. P values determined by unpaired Student’s t-test (A-E and G-I), ANOVA and Dunnett’s test (F), or Mann-Whitney U-test (K-L).

Figure 2.. Intracellular ERO1α is crucial for platelet activation, aggregation, and thrombus formation.

(A) Ex vivo platelet thrombus formation in a flow chamber assay. Adherent and aggregated platelets were stained with rhodamine-conjugated phalloidin and imaged by confocal microscopy. Representative images (bar = 10 μm). Surface coverage and thrombus volume were measured (mean ± SD, n = 4). (B-E) Aggregation and ATP secretion of WT and Ero1α-null platelets were induced by 0.025 U/ml thrombin, 0.05 μg/ml CRP, 0.5 μM A23187, or 2.5 μM ADP. (i) Platelet aggregation and (ii) ATP secretion. (F) Ero1α-null platelets were pretreated with 50 μg/ml wtERO1α or wtPDI and then with 0.025 U/ml thrombin. Quantification graphs of aggregation and ATP secretion are presented as the mean ± SD (n = 3–4). (G-H) P-selectin exposure and αIIbβ3 integrin activation of WT and Ero1α-null platelets were analyzed by flow cytometry. The data are presented as the geometric mean fluorescence intensity (MFI) value (mean ± SD, n = 3–4). (I) WT and Ero1α-null platelets were treated with or without 0.025 U/ml thrombin for 1 minute in an aggregometer. The lysates were immunoprecipitated with control IgG or anti-β3 antibodies and subjected to immunoblotting with antibodies against talin1 or β3 and densitometric analysis (mean ± SD, n = 3). (J) WT and Ero1α-null platelets were treated with or without 0.025 U/ml thrombin, followed by transmission electron microscopy. The original magnification was 34,800, and the extension was 313,000. Bar = 500 nm. (K-L) Intravital microscopy was performed to determine the effect of 15E9 (1–3 μg/g BW) and a blocking anti-PDI antibody (BD34, 3 μg/g BW) on platelet thrombus formation and fibrin generation at the site of laser-induced cremaster arteriolar injury (n = 27–40 arterioles in 6 mice/group). The n indicates biological replicates. P values determined by unpaired Student’s t-test (A-E and G-I), ANOVA and Dunnett’s test (F), or Mann-Whitney U-test (K-L).

Figure 3.. ERO1α regulates the function of Ca²⁺ signaling molecules, STIM1 and SERCA2, and enhances cytosolic Ca²⁺ levels during platelet activation.

(A) Mass spectrometric analysis was performed to identify proteins interacting with ERO1α in resting and thrombin-activated human platelets. Volcano plots show proteins whose binding to ERO1α was significantly increased or decreased during platelet activation. (B-C) Immunoprecipitation of ERO1α with lysates of resting and thrombin (Thr)-activated human platelets, followed by immunoblotting and densitometry. (D-F) Ca²⁺ release and influx of WT and Ero1α-null platelets (CKO) were assessed in response to 0.1 U/ml thrombin, 2 μM A23187, or 20 μM thapsigargin (TG), followed by the addition of 1 mM CaCl₂. The Ca²⁺ signal was quantified by the AUC. (G) Immunogold electron microscopy using resting and SFLLRN-activated human platelets. Ultrathin platelet sections were incubated with antibodies against PDI or ERO1α, and bound antibodies were labeled with two different immunogold colloids (15 nm for PDI and 10 nm for ERO1α). (H) Bio-layer interferometry was performed using a biotinylated ERO1α biosensor. After incubation with a different concentration of wtSTIM1 and mSTIM1, the specific interaction between ERO1α and STIM1 was measured by subtracting the non-specific binding. The dissociation constant, K_D, was calculated from the K_on and K_off values. The data represent the mean ± SD (n = 3). (I-K) Human platelets (hplt) and WT and Ero1α-null platelets were treated with or without 0.025 U/ml thrombin for 0–5 minutes. After labeling the lysates with MPB, proteins were pulled down with streptavidin magnetic beads. The bound fraction was used for immunoblotting with anti-STIM1 antibodies, followed by densitometry. (L) A schematic of mass spectrometric analysis. (M) The representative tandem unmodified and NEM-modified Cys49-containing STIM1 peptides, ATGTSSGANSEESTAAEFCR (988.42 m/z) and ATGTSSGANSEESTAAEFC[+125]R (1,050.95 m/z), respectively. (N) The peak areas for ATGTSSGANSEESTAAEFCR and ATGTSSGANSEESTAAEFC[+125]R were normalized to that of the STIM1 control peptide, EDLNYHDPTVK, which was observed at consistent levels in all samples. (O) WTERO1α was incubated with wtSERCA2 or mSERCA2 (Mut). After immunoprecipitation with polyclonal anti-ERO1α antibodies, the bound fraction was blotted with anti-SERCA2 antibodies, followed by densitometry. (P) The representative tandem NEM-modified Cys875-containing SERCA2 peptide, VSFYQLSHFLQC[+125]K (575.62 m/z). Please note that the unmodified peptide was not detected. (Q) The peak area for the NEM-modified peptide was normalized to that of the SERCA2 control peptide, DIVPGDIVEIAVGDK, which was observed at consistent levels in all samples. The data represent the mean ± SD (n = 3–4 except N and Q). The n indicates biological replicates. P values determined by unpaired Student’s t-test.

Figure 3.. ERO1α regulates the function of Ca²⁺ signaling molecules, STIM1 and SERCA2, and enhances cytosolic Ca²⁺ levels during platelet activation.

(A) Mass spectrometric analysis was performed to identify proteins interacting with ERO1α in resting and thrombin-activated human platelets. Volcano plots show proteins whose binding to ERO1α was significantly increased or decreased during platelet activation. (B-C) Immunoprecipitation of ERO1α with lysates of resting and thrombin (Thr)-activated human platelets, followed by immunoblotting and densitometry. (D-F) Ca²⁺ release and influx of WT and Ero1α-null platelets (CKO) were assessed in response to 0.1 U/ml thrombin, 2 μM A23187, or 20 μM thapsigargin (TG), followed by the addition of 1 mM CaCl₂. The Ca²⁺ signal was quantified by the AUC. (G) Immunogold electron microscopy using resting and SFLLRN-activated human platelets. Ultrathin platelet sections were incubated with antibodies against PDI or ERO1α, and bound antibodies were labeled with two different immunogold colloids (15 nm for PDI and 10 nm for ERO1α). (H) Bio-layer interferometry was performed using a biotinylated ERO1α biosensor. After incubation with a different concentration of wtSTIM1 and mSTIM1, the specific interaction between ERO1α and STIM1 was measured by subtracting the non-specific binding. The dissociation constant, K_D, was calculated from the K_on and K_off values. The data represent the mean ± SD (n = 3). (I-K) Human platelets (hplt) and WT and Ero1α-null platelets were treated with or without 0.025 U/ml thrombin for 0–5 minutes. After labeling the lysates with MPB, proteins were pulled down with streptavidin magnetic beads. The bound fraction was used for immunoblotting with anti-STIM1 antibodies, followed by densitometry. (L) A schematic of mass spectrometric analysis. (M) The representative tandem unmodified and NEM-modified Cys49-containing STIM1 peptides, ATGTSSGANSEESTAAEFCR (988.42 m/z) and ATGTSSGANSEESTAAEFC[+125]R (1,050.95 m/z), respectively. (N) The peak areas for ATGTSSGANSEESTAAEFCR and ATGTSSGANSEESTAAEFC[+125]R were normalized to that of the STIM1 control peptide, EDLNYHDPTVK, which was observed at consistent levels in all samples. (O) WTERO1α was incubated with wtSERCA2 or mSERCA2 (Mut). After immunoprecipitation with polyclonal anti-ERO1α antibodies, the bound fraction was blotted with anti-SERCA2 antibodies, followed by densitometry. (P) The representative tandem NEM-modified Cys875-containing SERCA2 peptide, VSFYQLSHFLQC[+125]K (575.62 m/z). Please note that the unmodified peptide was not detected. (Q) The peak area for the NEM-modified peptide was normalized to that of the SERCA2 control peptide, DIVPGDIVEIAVGDK, which was observed at consistent levels in all samples. The data represent the mean ± SD (n = 3–4 except N and Q). The n indicates biological replicates. P values determined by unpaired Student’s t-test.

Figure 3.. ERO1α regulates the function of Ca²⁺ signaling molecules, STIM1 and SERCA2, and enhances cytosolic Ca²⁺ levels during platelet activation.

(A) Mass spectrometric analysis was performed to identify proteins interacting with ERO1α in resting and thrombin-activated human platelets. Volcano plots show proteins whose binding to ERO1α was significantly increased or decreased during platelet activation. (B-C) Immunoprecipitation of ERO1α with lysates of resting and thrombin (Thr)-activated human platelets, followed by immunoblotting and densitometry. (D-F) Ca²⁺ release and influx of WT and Ero1α-null platelets (CKO) were assessed in response to 0.1 U/ml thrombin, 2 μM A23187, or 20 μM thapsigargin (TG), followed by the addition of 1 mM CaCl₂. The Ca²⁺ signal was quantified by the AUC. (G) Immunogold electron microscopy using resting and SFLLRN-activated human platelets. Ultrathin platelet sections were incubated with antibodies against PDI or ERO1α, and bound antibodies were labeled with two different immunogold colloids (15 nm for PDI and 10 nm for ERO1α). (H) Bio-layer interferometry was performed using a biotinylated ERO1α biosensor. After incubation with a different concentration of wtSTIM1 and mSTIM1, the specific interaction between ERO1α and STIM1 was measured by subtracting the non-specific binding. The dissociation constant, K_D, was calculated from the K_on and K_off values. The data represent the mean ± SD (n = 3). (I-K) Human platelets (hplt) and WT and Ero1α-null platelets were treated with or without 0.025 U/ml thrombin for 0–5 minutes. After labeling the lysates with MPB, proteins were pulled down with streptavidin magnetic beads. The bound fraction was used for immunoblotting with anti-STIM1 antibodies, followed by densitometry. (L) A schematic of mass spectrometric analysis. (M) The representative tandem unmodified and NEM-modified Cys49-containing STIM1 peptides, ATGTSSGANSEESTAAEFCR (988.42 m/z) and ATGTSSGANSEESTAAEFC[+125]R (1,050.95 m/z), respectively. (N) The peak areas for ATGTSSGANSEESTAAEFCR and ATGTSSGANSEESTAAEFC[+125]R were normalized to that of the STIM1 control peptide, EDLNYHDPTVK, which was observed at consistent levels in all samples. (O) WTERO1α was incubated with wtSERCA2 or mSERCA2 (Mut). After immunoprecipitation with polyclonal anti-ERO1α antibodies, the bound fraction was blotted with anti-SERCA2 antibodies, followed by densitometry. (P) The representative tandem NEM-modified Cys875-containing SERCA2 peptide, VSFYQLSHFLQC[+125]K (575.62 m/z). Please note that the unmodified peptide was not detected. (Q) The peak area for the NEM-modified peptide was normalized to that of the SERCA2 control peptide, DIVPGDIVEIAVGDK, which was observed at consistent levels in all samples. The data represent the mean ± SD (n = 3–4 except N and Q). The n indicates biological replicates. P values determined by unpaired Student’s t-test.

Figure 4.. Novel small-molecule inhibitors recapitulate the defect in Ero1α-null platelets.

(A) The structure of B12 and its inhibitory effect on ERO1α activity. (B-E) The inhibitory effect of B12 on P-selectin exposure and αIIbβ3 integrin activation in (B-C) mouse and (D-E) human platelets. (F) The structure of B12–5 and its inhibitory effect on ERO1α activity. (G) A docking model of B12–5-ERO1α (3AHQ) and FAD-ERO1α complex. There are pi-sulfur interactions of the phenothiazine head group (B12–5) or a tricyclic ring (FAD) with Met 389 and Cys397. Also, there are H bond interactions between the tail/alkyl group and His255, Asn259, and Arg287. (H) Bio-layer interferometry was performed using a biosensor conjugated with biotinylated ERO1α. After incubation with a different concentration of B12–5, the specific interaction between B12–5 and ERO1α was measured by subtracting the non-specific binding. The K_D value was calculated from the K_on and K_off values. The represent trace (n = 3). (I-L) The inhibitory effect of B12–5 on P-selectin exposure and αIIbβ3 integrin activation in (I-J) mouse and (K-L) human platelets. (M-N) The inhibitory effect of B12–5 (10 μM) on WT and Ero1α-null platelet aggregation induced by thrombin or CRP. (O-R) The effect of B12–5 on Ca²⁺ mobilization in (O-P) mouse and (Q-R) human platelets after stimulation with thrombin or A23187. The data represent the mean ± SD (n = 3–4). The n indicates biological replicates. P values determined by ANOVA and Dunnett’s test (B-E, I-L, and O-R (for B12–5)) or Student’s t-test (M-N and O-R (for BAPTA)).

Figure 4.. Novel small-molecule inhibitors recapitulate the defect in Ero1α-null platelets.

(A) The structure of B12 and its inhibitory effect on ERO1α activity. (B-E) The inhibitory effect of B12 on P-selectin exposure and αIIbβ3 integrin activation in (B-C) mouse and (D-E) human platelets. (F) The structure of B12–5 and its inhibitory effect on ERO1α activity. (G) A docking model of B12–5-ERO1α (3AHQ) and FAD-ERO1α complex. There are pi-sulfur interactions of the phenothiazine head group (B12–5) or a tricyclic ring (FAD) with Met 389 and Cys397. Also, there are H bond interactions between the tail/alkyl group and His255, Asn259, and Arg287. (H) Bio-layer interferometry was performed using a biosensor conjugated with biotinylated ERO1α. After incubation with a different concentration of B12–5, the specific interaction between B12–5 and ERO1α was measured by subtracting the non-specific binding. The K_D value was calculated from the K_on and K_off values. The represent trace (n = 3). (I-L) The inhibitory effect of B12–5 on P-selectin exposure and αIIbβ3 integrin activation in (I-J) mouse and (K-L) human platelets. (M-N) The inhibitory effect of B12–5 (10 μM) on WT and Ero1α-null platelet aggregation induced by thrombin or CRP. (O-R) The effect of B12–5 on Ca²⁺ mobilization in (O-P) mouse and (Q-R) human platelets after stimulation with thrombin or A23187. The data represent the mean ± SD (n = 3–4). The n indicates biological replicates. P values determined by ANOVA and Dunnett’s test (B-E, I-L, and O-R (for B12–5)) or Student’s t-test (M-N and O-R (for BAPTA)).

Figure 5.. B12–5 has an antithrombotic effect without prolonging tail bleeding times in mice.

(A-C) Ex vivo effects of B12–5. WT (C57BL/6) mice were treated with intravenous injection of vehicle (20% PEG + 3% DMSO) or B12–5 (5 μg/g BW). One hour later, blood was drawn, followed by isolating platelets. (A) P-selectin exposure and (B) αIIbβ3 integrin activation were assessed in flow cytometry. (C) Platelet aggregation was measured after treatment with 0.025 U/ml thrombin. The data represent the mean ± SD (n = 4). (D) The plasma levels of B12–5 were analyzed by LC-MS/MS after intravenous injection of the compound (5 μg/g BW) into WT mice and quantified by comparison with a standard curve of B12–5 (mean ± SD, n = 3). (E-F) After laser-induced cremaster arteriolar injury, intravital microscopy was conducted with WT mice pretreated with intravenous injection of saline, eptifibatide (5 μg/g BW), vehicle, or B12–5 (5 μg/g BW). The median integrated fluorescence intensities of anti-CD42c (F platelet) and anti-fibrin (F fibrin) antibodies (n = 37–42 arterioles in 6 mice/group). Quantification of the antibody signal at a different time point after laser injury. (G) Comparison of B12–5 (5 μg/g BW) with eptifibatide (5 μg/g BW) and their respective controls in inhibiting FeCl₃-induced carotid artery thrombosis in WT mice. (H-I) WT mice were treated with intravenous injection of vehicle control, eptifibatide (5 μg/g BW), or B12–5 (5 μg/g BW). Thirty minutes later, tail bleeding times and Hb contents were measured after amputation of the tail tip. Bars represent the median value (E-I). (J-M) WT control and Ero1α CKO and KO mice were subjected to tMCAO. Twenty-three hours later, neurological deficits were assessed by the Bederson score and grip test, and infarct volume was measured as described in Methods. (N-Q) WT mice were subjected to tMCAO and then treated with saline, vehicle, or B12–5 (5 μg/g BW). Twenty-three hours later, the neurological deficit and infarct volume were measured as described above. (K and O) The data represent the mean ± SD (n = 7–8). (L-M and P-Q) The bar indicates the median with interquartile range. P values determined by unpaired Student’s t-test (A-C, K, and O) or Mann-Whitney U-test (E-I, L-M, and P-Q).

Figure 5.. B12–5 has an antithrombotic effect without prolonging tail bleeding times in mice.

(A-C) Ex vivo effects of B12–5. WT (C57BL/6) mice were treated with intravenous injection of vehicle (20% PEG + 3% DMSO) or B12–5 (5 μg/g BW). One hour later, blood was drawn, followed by isolating platelets. (A) P-selectin exposure and (B) αIIbβ3 integrin activation were assessed in flow cytometry. (C) Platelet aggregation was measured after treatment with 0.025 U/ml thrombin. The data represent the mean ± SD (n = 4). (D) The plasma levels of B12–5 were analyzed by LC-MS/MS after intravenous injection of the compound (5 μg/g BW) into WT mice and quantified by comparison with a standard curve of B12–5 (mean ± SD, n = 3). (E-F) After laser-induced cremaster arteriolar injury, intravital microscopy was conducted with WT mice pretreated with intravenous injection of saline, eptifibatide (5 μg/g BW), vehicle, or B12–5 (5 μg/g BW). The median integrated fluorescence intensities of anti-CD42c (F platelet) and anti-fibrin (F fibrin) antibodies (n = 37–42 arterioles in 6 mice/group). Quantification of the antibody signal at a different time point after laser injury. (G) Comparison of B12–5 (5 μg/g BW) with eptifibatide (5 μg/g BW) and their respective controls in inhibiting FeCl₃-induced carotid artery thrombosis in WT mice. (H-I) WT mice were treated with intravenous injection of vehicle control, eptifibatide (5 μg/g BW), or B12–5 (5 μg/g BW). Thirty minutes later, tail bleeding times and Hb contents were measured after amputation of the tail tip. Bars represent the median value (E-I). (J-M) WT control and Ero1α CKO and KO mice were subjected to tMCAO. Twenty-three hours later, neurological deficits were assessed by the Bederson score and grip test, and infarct volume was measured as described in Methods. (N-Q) WT mice were subjected to tMCAO and then treated with saline, vehicle, or B12–5 (5 μg/g BW). Twenty-three hours later, the neurological deficit and infarct volume were measured as described above. (K and O) The data represent the mean ± SD (n = 7–8). (L-M and P-Q) The bar indicates the median with interquartile range. P values determined by unpaired Student’s t-test (A-C, K, and O) or Mann-Whitney U-test (E-I, L-M, and P-Q).