Introduction

Immune thrombocytopenia (ITP) is a common clinical bleeding disorder characterized by an immune-mediated clearance of autologous platelets, predominantly through autoantibodies targeting platelet surface receptors GPIIbIIIa and/or GPIb-IX and clearance by phagocytic cells in the reticuloendothelial system via Fcγ-receptors [1–4]. Low platelet counts place ITP patients at risk for severe bleeding including fatal intracranial hemorrhage. Most therapies for ITP including first-line corticosteroids and immunoglobulin G (IVIG), and last resort splenectomy, mainly target the Fc-dependent clearance pathway via blocking/attenuating Fc-Fcγ-R interaction or removal of putative site of platelet clearance [4]. However, the pathogenesis and mechanisms of therapies remain poorly understood and around 15–20% of ITP patients are inexplicably refractory to first-line therapies, and around 10% are refractory to splenectomy [5, 6]. In recent years, murine models and large cohort human studies report antibody specificity (i.e., anti-GPIIbIIIa versus anti-GPIb-IX) may play a significant role in dictating response to therapy in ITP [7–9]; whereby presence of anti-GPIb-IX antibodies results in decreased response to corticosteroids and IVIG [7–9]. Most recently, we reported that anti-GPIbα and some anti-GPIIbIIIa antibodies in humans induced platelet desialylation leading to Fc-independent platelet clearance in the liver via hepatic asialoglycoprotein Ashwell-Morell receptors [10], suggesting antibody-mediated desialylation may be one of the underlying mechanisms behind resistance to standard ITP therapies [8, 9, 11].

In the present study, we sought to address whether increased platelet desialylation was correlated with decreased response to treatment in ITP patients (Additional file 1: Supplementary Material). The platelets of randomly and consecutively enrolled 61 patients diagnosed with primary ITP were tested for desialylation prior to the indicated treatments (Table 1). Fluorescein-conjugated lectins Ricinus communis agglutinin I (RCA-1) and Erythrina cristagalli lectin (ECL) were used to detect desialylated galactose and β-GlcNAc residues via flow cytometry. We found the platelets of ITP patients had significantly higher desialylation as measured by both RCA-1 and ECL binding compared to those of healthy blood donors (p < 0.05) (Fig. 1). The 61 ITP patients subsequently underwent standard first-line therapy independent of platelet desialylation and MAIPA assays. After 1 month of treatment, there were 26 complete responders (CR), 21 responders (R), and 14 non-responders (NR) (Table 1). Retrospective data analysis using Kruskal-Wallis rank sum test revealed NR patients had significantly higher platelet desialylation, as compared to the CR and R groups (p < 0.01). Correlation analysis indicated that efficacy and the desialylation level are related (RCA-1 r = 0.395, p < 0.01; ECL r = 0.391, p < 0.01). The higher desialylation, the poorer the efficacy of therapy observed.

Table 1 Platelet desialylation of different groups [M (P 25, P 75)]
Full size table
Fig. 1
figure 1

A Platelet desialylation in representative patients and healthy controls. The RCA-1 and ECL binding to platelets from healthy blood donors (a, b), responder (c, d), and non-responder patients (e, f) were examined by flow cytometry. The representative dot plots from each group are shown. B Platelet desialylation in different patient groups and healthy controls. The RCA-1 and ECL levels (Mean ± SEM) in complete responders (CR), responders (R) and non-responders (NR) (a, b); anti-GPIbα antibody positive (+) group, single anti-GPIIb/IIIa antibody (+) group, and double negative group (c, d); and in the thrombocytopenias/controls (e, f) (ITP, CTD, MDS, AA, AML, and healthy controls) were examined by flow cytometry. Each point represents the level of platelet desialylation of an individual patient or healthy blood donor (**p < 0.01; ***p < 0.001)

Full size image

To test whether the presence of anti-GPIbα antibodies is associated with the platelet desialylation, we detected antibody using MAIPA in the available 33 patient samples collected prior to treatment. We observed a two- to sixfold increased platelet desialylation in patients with anti-GPIbα antibodies (n = 9) compared to that in patients with anti-GPIIbIIIa (n = 14) or without detectable antibodies (n = 10) (Table 1 and Fig. 1B). However, statistical significance was not reached, which is likely due to small sample size. Future larger studies should be useful in determining direct correlation between anti-GPIbα antibody positivity with platelet desialylation.

Interestingly, we also observed significant platelet desialylation in patients with non-ITP thrombocytopenias including connective tissue diseases (CTD), myelodysplastic syndrome (MDS), and aplastic anemia (AA) (p < 0.001) but not acute myeloid leukemia (AML) compared to healthy controls (Table 1 and Fig. 1B). Notably, although RCA-1 and ECL measures different types of deglycosylation, we did not observe significant difference between these two assays, suggesting either of them can be used for the potential diagnosis and prognosis.

In summary, our data demonstrates for the first time that the higher level of platelet desialylation is correlated with non-response to the first-line ITP therapies (likely also splenectomies; Additional file 1: Supplementary Material). These findings not only suggest that platelet desialylation is a useful biomarker in predicting response to treatment in clinical ITP but positions sialidase inhibitors, such as Tamiflu [12], as a potential novel therapeutic in the treatment of ITP as well as other thrombocytopenias.