Thrombocytopenia affects 20-30% of infants admitted to Neonatal Intensive Care Units. In 20-25% of affected neonates, the platelet count drops to <50x109/L, and 9% of those infants experience clinically significant bleeding. Currently, platelet transfusions are the only therapeutic option for thrombocytopenic neonates. Two novel thrombopoietin (TPO) mimetics, romiplostim (ROM) and eltrombopag (ELT), are in clinical use for the treatment of adults with ITP. Based on the severity and duration of thrombocytopenia, 10% of thrombocytopenic neonates could benefit from this therapy. We have previously described significant cellular and molecular differences between neonatal and adult megakaryocytes (MKs) in their response to TPO (Pastos et al., Blood 2006; Liu et al., Blood 2011). Thus, developmental stage-specific studies are needed prior to considering the use of these agents in neonates. This study was designed to compare the responses of human neonatal vs. adult MK progenitors to TPO, ROM, and ELT in vitro.


Hematopoietic progenitor (CD34+) cells isolated from cord blood (CB) and adult mobilized peripheral blood (PB) were cultured in a serum-free culture system in the presence of escalating concentrations of rTPO (1-100 ng/ml), ROM (1-200 ng/ml), or ELT (1-30 µM). In each dose-response experiment, cells were also cultured with 50 ng/mL of rTPO as an internal control. After 14 days, cells were counted and their differentiation (CD41 expression), maturation (CD42b expression) and ploidy were evaluated by three-color flow cytometry. To adjust for inter-individual differences, the responses to escalating concentrations of ROM or ELT were expressed as percentages of the number of MKs cultured with 50 ng/mL of TPO (“percent of TPO50”).


Consistent with our prior studies, CB MK progenitors proliferated at a higher rate than adult PB progenitors in response to TPO, and generated approximately 10 times more MKs (CD41+ cells) per CD34+ cell than adult PB (CB n=3, PB n=4). To compare the sensitivity of CB and PB MK progenitors to ROM and ELT, percent of TPO50 vs. ROM or ELT concentration curves were generated. CB progenitors (n=3) generated the highest number of MKs at 200 ng/mL of ROM, which was nearly equivalent to TPO50. PB progenitors (n=2), in contrast, generated the maximal MK number at 30 ng/mL of ROM, and this represented approximately 80% of TPO50. At all concentrations of TPO and ROM, CB MKs exhibited lower ploidy levels than PB MKs (4-10% vs. 16-34% MKs ≥8n, respectively), but higher percentages of CD42+ MKs (85-95% vs. 63-80% in CB and PB, respectively). Next, we generated percent of TPO50 vs. ELT concentration curves. CB progenitors (n=5) reached the highest MK number at 6 µM ELT, and this represented approximately 40% of TPO50. For PB MKs (n=4), the maximal number of MKs was generated at 10 µM ELT, and this represented only 10% of TPO50. Nevertheless, these MKs had ploidy and CD42 expression levels similar to those cultured in TPO50. ELT concentrations higher than 10 µM did not support the survival of MK progenitors. To evaluate whether this toxicity was TPO receptor-dependent or not, we then cultured CD34+ cells with escalating ELT concentrations together with a physiological concentration of TPO (CB n=3, PB n=2). These studies revealed a strong synergism between TPO and low concentrations of ELT, particularly in CB cultures. However, PB cells proliferated less and CB cells died in the presence of higher ELT concentrations (30 µM) despite the addition of TPO, suggesting that the dose-dependent ELT toxicity is TPO receptor-independent.


Overall, we confirmed our previous findings that CB progenitors are hyperproliferative compared to adult progenitors, and generate small, low-ploidy, but mature MKs. Dose-response curves were similar for TPO and ROM. ELT generated less MKs than TPO or ROM, although these MKs were fully mature. CB MKs were more sensitive than PB MKs to the thrombopoietic as well as to the toxic effects of ELT. The dose-dependent toxicity of ELT in vitro is TPO-receptor independent. The mechanisms underlying this effect are unknown, but might be related to the structural similarities between ELT and metal chelators, potentially inducing intracellular iron depletion as has been shown in leukemic cells (Roth et al., Blood 2012).


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.