Embryonic MK activation: a delicate balance

In this issue of Blood, Hoover et al identified a new role for podoplanin (PDPN) in regulating embryonic megakaryocyte (eMk) activation and found a link between increased eMk activation and midgestation brain hemorrhages in mice lacking PDPN.¹ 

In both mice and humans, embryonic megakaryocytes (eMks) generated in the yolk sac are found in the circulation at early stages of development. C-type lectin-like receptor 2 (CLEC-2) is expressed at high levels on Mks and platelets and signals through a cytosolic YxxL sequence known as a hemi-immunotyrosine–based activation motif (hemITAM). HemITAM receptors activate a Syk-dependent signaling pathway upstream of PLCγ2. The only known endogenous ligand for CLEC-2 is the O-glycoprotein PDPN, which is expressed on a wide variety of cell types outside of the vasculature, including the neuroepithelium during midgestation. Prior studies in mice with constitutive deletions of CLEC-2 (Clec-2^−/−) or PDPN (Pdpn^−/−) have described spontaneous brain hemorrhages starting at embryonic day 11.5 (E11.5) associated with abnormal blood vessels, demonstrating a critical role for this pathway in brain vascular development.^2,3  Logically, these findings were thought to be secondary to decreased platelet activation in the setting of absent PDPN/CLEC-2 signaling.

Contrary to this hypothesis, Hoover and collaborators found that the brain hemorrhages in Pdpn^−/− and Clec-2^−/− mice were preceded by the development of abnormal aneurysmlike angiogenic sprouts within the diencephalon, which contained an increased number of eMks with features of increased activation and degranulation. Even more surprisingly, eMks in the angiogenic sprouts from both Pdpn^−/− and Clec-2^−/− mice exhibited phosphorylated Syk and PLCγ2, the downstream effectors of hemITAM signaling, suggesting that this pathway was activated by an alternative mechanism. In addition to PDPN/CLEC-2, collagen also signals through the hemITAM pathway in Mks and platelets. During midgestation, the researchers found high levels of collagen-1 expression in the diencephalon, which provided a potential alternative for the eMk activation pattern observed in the brains of Pdpn^−/− and Clec-2^−/− embryos.

In a series of elegant experiments, Hoover et al demonstrated that collagen-1 strongly stimulated angiopoietin-1 secretion in fetal Mks and that PDPN reduced this response. Consistently, the activated form of TIE-2 (the endothelial angiopoietin-1 receptor) was significantly increased in the aneurysm-like angiogenic sprouts of mice lacking PDPN or CLEC-2. Because TIE-2 hyperphosphorylation has been shown to cause venous malformations,⁴  these combined findings strongly suggest that the unrestrained collagen-1–induced activation of eMks in the diencephalon of E10.5 Pdpn^−/− and Clec-2^−/− embryos leads to excessive angiopoietin-1 secretion, increased TIE-2 signaling in angiogenic sprouts, and abnormal growth of aneurysms to the point of rupture and hemorrhage. Further supporting the role of increased Mk activation in the pathogenesis of these hemorrhages, treatment of Pdpn^−/− pregnant dams with both aspirin and ticagrelor from E8.5 to E12.5 normalized the angiogenic sprouting in the diencephalon and almost completely reversed the bleeding phenotype.

These findings, which challenge traditional notions about platelet function and bleeding, have intriguing parallels with clinical observations in neonates born extremely prematurely. Infants born between 22 and 27 weeks of gestation have a high incidence of intraventricular hemorrhages (IVHs) in the first week of life. IVHs almost always originate in the highly vascularized germinal matrix, which is vulnerable to hemodynamic changes during this developmental period.⁵  If the hemorrhage in the germinal matrix is substantial, the immature ependyma ruptures and the cerebral ventricles fill with blood, a presentation that is associated with poor neurodevelopmental outcomes. Similar to the observations from Hoover et al, early administration of indomethacin, a nonsteroidal agent that inhibits platelet function, significantly reduces the incidence of severe IVH in these infants.⁶ 

Neonatal platelets are significantly hyporeactive to multiple agonists, compared with adult platelets. Because of this hyporeactivity and the high incidence of IVH, in clinical practice thrombocytopenic preterm neonates are transfused at higher platelet thresholds than older children or adults. However, the platelet hyporeactivity in neonates is effectively counteracted by factors in neonatal blood that stimulate clotting, including high levels of von Willebrand factor, high hematocrits, and high mean corpuscular volumes, with the net result of adequate primary hemostasis in preterm and term neonates.⁷  Thus, the neonatal platelet hyporeactivity should not be viewed as a deficiency, but rather as an integral part of a developmentally unique, but well balanced, neonatal hemostatic system.

In 2019, in the largest multicenter randomized trial to date, liberal vs restrictive prophylactic platelet transfusion thresholds were compared in thrombocytopenic preterm neonates. Surprisingly, neonates receiving transfusions at higher platelet counts had a significantly higher rate of mortality and/or major bleeding within 28 days of randomization (the primary outcome), compared with those receiving transfusions at lower platelet counts.⁸  The mechanisms underlying these findings are unknown, but it has been hypothesized that they are related to acute hemodynamic changes associated with the rapid increase in intravascular volume and/or to the “developmental mismatch” that occurs when adult, comparatively hyperreactive, platelets are transfused into a preterm neonate. In support of the latter, a prior study showed that in vitro mixing of adult platelets with thrombocytopenic neonatal blood led to a prothrombotic phenotype, although this finding provided no explanation for the increased incidence of bleeding.⁹ 

The findings of Hoover et al raise the intriguing possibility that comparatively hyperreactive adult platelets transfused into preterm infants during a vulnerable period could alter the delicate angiogenic balance of the developing brain and increase the risk of hemorrhage by providing inappropriate angiogenic signals. Although the patterns of PDPN and collagen expression in the developing human brain are not known, it is known that human adult platelets exhibit increased activation compared with neonatal platelets in response to collagen-related peptide and rhodocytin, which activate the collagen receptor and CLEC-2, respectively.¹⁰  Additional studies are needed to better understand the interactions of adult vs neonatal platelets with the unique environment of the developing brain and the potential effects of transfused adult platelets on vascular development and integrity during vulnerable periods. In this case, too much of a good thing may be a bad thing.