BACKGROUND. Hemolytic anemias (HA) such as autoimmune hemolytic anemia (AIHA), thalassemia, and sickle cell disease are associated with increased risk for thrombosis. Although exposure of procoagulant phospholipids on the RBC membrane in HA has been implicated, the pathogenesis remains to be elucidated. We recently showed that microparticles released from red cells (RMP) are procoagulant and significantly elevated in HA, suggesting a link of RMP to thrombotic complications. In this study, we report the release of RMP, heterogeneous in size, phenotypes and procoagulant activity by different methods.

METHODS. RMP were prepared from washed RBC (2×109/mL) by three methods, exposure to

  • 10 μM calcium ionophore A23187 (CaIo) for 60 min,

  • 20 μM lysophosphatidic acid (LPA) for 60 min, or

  • shear stress induced by a rotating Teflon pestle for 5 min at 160 rpm in a tissue grinder.

After removing residual whole cells, the supernatant was centrifuged at 15,000×g for 10 min (large RMP) and the supernatant was further centrifuged at 150,000×g for 30 min (small RMP). The resulting RMP were labeled for flow cytometry using PE-labeled anti-glycophorin (GlyP), FITC-anti-tissue factor (TF), FITC-annexin V (AnV), and/or FITC-lectin Ulex europeaus I (Ulex). Thrombin generation of RMP was measured by the calibrated automated thrombogram (CAT) system using fluorescent thrombin substrate on a fluorescence plate reader.


  1. The two-step centrifugation revealed two distinct populations, the large RMP expressing both GlyP and Ulex binding while the smaller RMP expressed only Ulex binding. Electron micrography showed diameters for the large RMP of 200 - 800 nM while the small RMP were 40 - 80 nm.

  2. Shear stress produced the greatest number of large GlyP + RMP (5.2 ± 1.3 × 106/μL), followed by CaIo (3.8 ± 0.7 × 106/μL), and LPA (2.5 ± 0.6 × 106/μL). However, CaIo produced the greatest number of small GlyP −/Ulex + RMP (4.6 ± 0.9 × 106/μL), followed by LPA (1.4 ±0.3 × 106/μL), and very few by shear tress (0.3 ±0.1 × 106/μL). The methods also gave different AnV binding: CaIo yielded >90% of RMP (large and small) that were positive, whereas RMP induced by LPA or shear stress gave only 45% and 18% positive, respectively. No TF + RMP were detected in any procedure.

  3. The RMP also differed in thrombin generation. Adjusting concentrations to equal numbers of Ulex + MP, CaIo RMP displayed the strongest activity (375 ± 62 nM) followed by LPA-induced (227 ± 58 nM) and shear-induced (136 ±33 nM).

Thrombin generation correlated well with degree of AnV binding.

CONCLUSIONS. We demonstrate that RBC release different species of RMP, heterogeneous in size, phenotypes and procoagulant activity. Both CaIo and LPA induced two distinct species of RMP, the larger expressing GlyP and Ulex binding while the smaller phenotype was negative for GlyP, possibly indicating that the small RMP are inside-out vesicles. In contrast, shear stress produced mainly large rightside-out vesicles with both GlyP expression and Ulex binding, and very low AnV binding. Thus, calcium influx appears necessary to release small inside-out RMP, and to induce the membrane flip-flop bringing negatively charged AnV binding sites to the plasma membrane. We postulated that highly procoagulant RMP that bind AnV may contribute to the thrombotic complications of HA and other RBC disorders associated with thrombosis.

Author notes

Disclosure: No relevant conflicts of interest to declare.