Desialylation of platelets induced by Von Willebrand Factor is a novel mechanism of platelet clearance in dengue

Abstract

Thrombocytopenia and platelet dysfunction are commonly observed in patients with dengue virus (DENV) infection and may contribute to complications such as bleeding and plasma leakage. The etiology of dengue-associated thrombocytopenia is multifactorial and includes increased platelet clearance. The binding of the coagulation protein von Willebrand factor (VWF) to the platelet membrane and removal of sialic acid (desialylation) are two well-known mechanisms of platelet clearance, but whether these conditions also contribute to thrombocytopenia in dengue infection is unknown. In two observational cohort studies in Bandung and Jepara, Indonesia, we show that adult patients with dengue not only had higher plasma concentrations of plasma VWF antigen and active VWF, but that circulating platelets had also bound more VWF to their membrane. The amount of platelet-VWF binding correlated well with platelet count. Furthermore, sialic acid levels in dengue patients were significantly reduced as assessed by the binding of Sambucus nigra lectin (SNA) and Maackia amurensis lectin II (MAL-II) to platelets. Sialic acid on the platelet membrane is neuraminidase-labile, but dengue virus has no known neuraminidase activity. Indeed, no detectable activity of neuraminidase was present in plasma of dengue patients and no desialylation was found of plasma transferrin. Platelet sialylation was also not altered by in vitro exposure of platelets to DENV nonstructural protein 1 or cultured DENV. In contrast, induction of binding of VWF to glycoprotein 1b on platelets using the VWF-activating protein ristocetin resulted in the removal of platelet sialic acid by translocation of platelet neuraminidase to the platelet surface. The neuraminidase inhibitor oseltamivir reduced VWF-induced platelet desialylation. Our data demonstrate that excessive binding of VWF to platelets in dengue results in neuraminidase-mediated platelet desialylation and platelet clearance. Oseltamivir might be a novel treatment option for severe thrombocytopenia in dengue infection.

Fig 1. Platelet number and platelet activation with impaired reactivity in patients with dengue infection.

(A, B) Platelet number at enrolment in cohorts from Bandung and Jepara. Binding of fibrinogen (C, D) to platelets and platelet P-selectin expression (E, F) in unstimulated samples and after ex vivo stimulation with two concentrations of ADP. Platelet P-selectin expression and binding of fibrinogen were measured using flow cytometry and are expressed as median fluorescence intensity (MFI) in arbitrary units. Data depicted as geometric mean with 95% confidence interval. Differences between groups were analyzed using the Mann-Whitney U test, *P < 0.05, ** P<0.01, ***P<0.001.

Fig 2. Increased VWF-platelet binding and plasma VWF:Ag and active VWF levels in dengue patients.

(A) Platelet-VWF binding measured by flow cytometry in unstimulated samples (line is geometric mean with 95% confidence interval) and (B) after ex vivo stimulation with increasing concentrations of ristocetin in participants with acute dengue and healthy volunteers. (C) Time course of plasma VWF:Ag levels and (D) active VWF levels in the same participants. (E) Pearson correlation between platelet-VWF binding and platelet number in enrolment samples and (F) fibrinogen binding after ex vivo stimulation with high dose (156μM) TRAP in enrolment samples. Data shown are individual values or aggregated data as geometric mean with 95% confidence interval. Differences between groups were analyzed using the Mann-Whitney U test, *P < 0.05, ** P<0.01, ***P<0.001.

Fig 3. Platelet desialylation in acute dengue and plasma sialidase activity.

Binding of the lectins (A) SNA and (B) MAL-II to platelet sialic acid residues measured by flow cytometry, in patients with acute dengue (n = 40), non-dengue febrile illness (n = 15) and healthy controls (n = 20) in Jepara. (C) Plasma sialidase activity measured as moles of substrate (2-(4-methylumbelliferyl)α-D-N-acetylneuraminic acid) hydrolyzed per hour. (D) Binding of SNA and MAL-II lectins to platelets following incubation of platelet rich plasma from healthy Dutch volunteers with platelet poor plasma from patients with dengue (acute and convalescence sample), non-dengue febrile illness and healthy controls (n = 6 per group). Data expressed as geometric mean with 95% CI. Samples depicted in panels A-B were analyzed on a BD FACS Calibur in Jepara, and samples in panel D on a Beckman coulter FC500 flow cytometry in The Netherlands, which explains the differences in MFI values. Differences between groups were analyzed using the Mann-Whitney U test, *P < 0.05, ** P<0.01, ***P<0.001. NS depicts statistically non-significant result across groups using Kruskal-Wallis test.

Fig 4. Dengue non-structural protein 1 (NS1) nor dengue virus (DENV) induce platelet desialylation.

Binding of (A) SNA or (B) MAL-II lectins to washed platelets or (C) the expression of P-selectin after incubation with two concentrations of DENV2 NS1 protein for 4 hrs at 37°C (n = 7 platelet donors). (D) shows expression of platelet P-selectin and (E-G) the binding of SNA, MAL-II, or RCA lectins to washed platelets following incubation with dengue virus type 2 (DENV 2) for 3 hrs at 37°C (n = 8 platelet donors). Purified neuraminidase from C. perfringens (100 mU) and the platelet agonist adenosine diphosphate (ADP) at 125 μM were used as positive controls. Mock infection with supernatant of uninfected C6/36 cells harvested at the same time as DENV2 stocks was used as negative control. Data are shown as geometric mean with 95% confidence interval. Samples depicted in panels A-C were analyzed on a Beckman coulter Cytoflex, and samples in panel D-G on a FC500 flow cytometer, which explains the differences in MFI values. Differences between groups were analyzed using the Mann-Whitney U test, *P < 0.05, ** P<0.01, ***P<0.001.

Fig 5. Binding of VWF to platelets induces platelet desialylation which can be inhibited by oseltamivir.

Binding of (A) von Willebrand factor (VWF), (B) SNA and (C) MAL-II lectins to platelets and (D) platelet membrane expression of Neuraminidase-1 (Neu-1) after exposing platelet rich plasma (PRP) of healthy volunteers to ristocetin for 1 hr at 37°C in presence and absence of GPIb receptor blocking antibodies. (E) Pearson correlation between the expression of Neu-1 and the lysosomal marker CD63 on the platelet membrane after incubation PRP with ristocetin (1.6mg/ml for 1 hour). (F) Binding of RCA lectin to galactose or N-acetylgalactosamine residues on platelets after exposing PRP to ristocetin (G) with and without the addition of increasing concentration of oseltamivir. Purified neuraminidase from C. perfringens (100mU) was used as a positive control. Data were analyzed using Student’s T-tests and presented as geometric mean with 95% confidence interval. *** P < 0.001, ** P < 0.005, * P < 0.05; ^### P < 0.001 when samples were compared with ristocetin-treated samples that were unexposed to oseltamivir. Data were from on 5–7 platelet donors in ≥ 2 independent experiments.