In vitro assessment and phase I randomized clinical trial of anfibatide a snake venom derived anti-thrombotic agent targeting human platelet GPIbα

Abstract

The interaction of platelet GPIbα with von Willebrand factor (VWF) is essential to initiate platelet adhesion and thrombosis, particularly under high shear stress conditions. However, no drug targeting GPIbα has been developed for clinical practice. Here we characterized anfibatide, a GPIbα antagonist purified from snake (Deinagkistrodon acutus) venom, and evaluated its interaction with GPIbα by surface plasmon resonance and in silico modeling. We demonstrated that anfibatide interferds with both VWF and thrombin binding, inhibited ristocetin/botrocetin- and low-dose thrombin-induced human platelet aggregation, and decreased thrombus volume and stability in blood flowing over collagen. In a single-center, randomized, and open-label phase I clinical trial, anfibatide was administered intravenously to 94 healthy volunteers either as a single dose bolus, or a bolus followed by a constant rate infusion of anfibatide for 24 h. Anfibatide inhibited VWF-mediated platelet aggregation without significantly altering bleeding time or coagulation. The inhibitory effects disappeared within 8 h after drug withdrawal. No thrombocytopenia or anti-anfibatide antibodies were detected, and no serious adverse events or allergic reactions were observed during the studies. Therefore, anfibatide was well-tolerated among healthy subjects. Interestingly, anfibatide exhibited pharmacologic effects in vivo at concentrations thousand-fold lower than in vitro, a phenomenon which deserves further investigation.Trial registration: Clinicaltrials.gov NCT01588132.

Conflict of interest statement

The authors declare the following competing interests: this research and the work in Lei et al. and Hou et al. was partially funded by Lee’s Pharmaceutical Holdings limited. B.X.L., X.D., Z.Y., F.Q., and M.L.L. are supported by Lee’s Pharmaceutical Holdings Limited and/or Zhaoke Pharmaceutical Co. Limited. The other authors declare they have no actual or potential competing interests.

Figures

Figure 1

Purification and structure of anfibatide with GPIbα. (A) MALDI-TOF mass spectrometry showed a mass to charge ratio (m/z) of 29,799.7. Three-dimensional models of (B) structure of anfibatide (purple) integrated with GPIb (orange) and (C) VWF-A1 domain (green) and GPIbα complex from PDB entry 1SQ0 (https://www.rcsb.org/structure/1SQ0). (D) α-Thrombin (blue) and GPIbα complex from PDB entry 1OOK (https://www.rcsb.org/structure/1OOK). Red arrows point to the sulfotyrosine region of GPIbα, the α-thrombin binding site. Protein-complex figures generated using Schrodinger PyMol 2 software (https://pymol.org/2/).

Figure 2

Anfibatide binding to GPIbα prevented VWF and thrombin binding as well as platelet aggregation induced by ristocetin/VWF and low-dose thrombin, but had no effect on blood clotting. (A) SPR analysis of anfibatide binding to disulfide-linked dimeric GPIbαN-Long (a chimera of human GPIbα residues − 2 to 288 with 133 residues of SV40 large T antigen) either with wild-type (WT) GPIbα sequence or with Tyr to Phe substitution of residues 276, 278 and 279 (3Y/F). GPIbαN-Long was captured onto the SPR chip by an immobilized anti-SV40-T mAb, followed by anfibatide at increasing concentrations. Data are presented as the ratio of anfibatide/GPIbαN-Long mass bound to the SPR chip during the equilibrium phase (prior to dissociation) and are fit to a one-site ligand binding model. Mass ratio ± SD, N = 3, Black: GPIbαN-Long WT; Orange: GPIbαN-Long 3Y/F. (B) Inhibition of α-thrombin (BP-αFIIa), VWF A1 domain or mAb LJ-Ib10 binding to GPIbα on human washed platelets by increasing anfibatide concentrations. Data are presented as % binding, relative to the binding of the ligand at a concentration equal to the KD of the ligand in the absence of anfibatide. The data was fit to the Cheng–Prusoff transformation. Ligand binding % ± SD, N = 3, Black: BP-αFIIa; Orange: LJ-Ib10; Green: rVWF. (C) Platelet aggregation was induced by ristocetin (1.2 mg/mL) (P < 0.001), ADP (20 μM) (P > 0.05), TRAP (500 μM) (P > 0.05), or collagen (10 µg/mL) (P > 0.05) in anfibatide-treated (6 µg/mL) PRP or by thrombin (0.1–1 U/mL) in gel-filtered platelets. Curves are from representative light transmission aggregometry plots. (D) Clot formation was measured by thromboelastography in anfibatide-treated whole blood. Anfibatide (6 µg/mL) did not significantly alter the time to initial clot formation (R time, min ± SEM) or maximum clot strength (MA ± SEM). NS, no significant difference. Red: anfibatide-treated plasma. Black: untreated plasma (N = 6).

Figure 3

Anfibatide inhibited platelet adhesion, aggregation and thrombus formation and dissolved preformed thrombi under flow conditions. Platelets in heparin-anticoagulated whole blood from healthy volunteers were fluorescently labeled with DiOC6 before perfusion over collagen at the wall shear rate of 1500 s−1 (A) or 300 s−1 (B) with or without anfibatide (6 µg/mL). (C) Control blood was first perfused at 1500 s−1 for 4 min to form thrombi; then, perfusion was continued with control or anfibatide-treated (6 µg/mL) blood. Representative images of fluorescent platelets (Left) are shown along with plots of the platelet mean (± SEM) fluorescence intensity as a function of time (Right; N = 12). P < 0.01 between control and treatment groups in all three figures. Two-tailed Student’s t-test was used to test for significant differences between 2 groups.

Figure 4

Study design. All randomly-assigned participants were included in the primary outcome analysis. Dose units are μg per 60 kg body weight. N: number of volunteers per group.

Figure 5

Plasma concentration–time curves of anfibatide in healthy volunteers after single intravenous bolus injection at dose levels of 1, 1.5, 2, 3, 4 and 5 µg/60 kg, respectively (Time = 0–8 h, A; Time = 0–48 h, B). Mean ± SEM.

Figure 6

Anfibatide inhibited ristocetin-induced platelet aggregation. Human platelet aggregation induced by ristocetin was studied by platelet aggregometry in single (A, N = 8–9/each) and multiple dose groups (B, N = 6–12/each). The mean inhibitory rate of anfibatide on platelet aggregation over time has been shown. CRI, constant rate infusion; Mean ± SEM.

Figure 7

Anfibatide had no significant effect on coagulation measured by prothrombin time (A), thrombin time (B), activated thromboplastin time (C), and international normalized ratio (D). anfibatide did not significantly change circulating d-dimers (E), representing fibrinolysis. (CRI indicates constant rate infusion). Mean ± SD.

Figure 8

Anfibatide did not significantly prolong bleeding time. Bleeding time was monitored in both single (A, N = 8–9/each) and multiple dose groups (B, N = 6–12/each), and each bleeding time measured was determined to the nearest 30 s. Majority of the subjects in the single dose groups (except 5 subjects) had a lower bound close to 4 min, while those in the multiple dose groups had a lower bound ranging from 2 to 4 min. None of the subjects in the single and multiple dose groups had a bleeding time of more than 9 min. Bleeding time of all subjects was within the normal range of 2 to 9 min. CRI: constant rate infusion.