NOTCHing down a win for megakaryocytes

By exploiting a congenic pair of induced pluripotent stem cell (iPSC) lines derived from a familial platelet disorder (FPD) patient, 1 harboring a monoallelic mutation in RUNX1 and the other a corrected allele, in this issue of Blood, Li et al have discovered that negative regulation of NOTCH4 by RUNX1 is required for normal human megakaryocyte (MK) development. Furthermore, the authors demonstrate that suppressing NOTCH signaling, by either genetic or chemical perturbation, significantly enhances MK yield.¹ 

In an elegant series of experiments, Li et al discover that a RUNX1-binding site in intron 29 of NOTCH4 negatively regulates NOTCH4 expression, the operation of which favors production of MKs in vitro. This study touches on multiple issues of interest to hematologists. At its heart, it provides critical insight into how loss of RUNX1 results in the lowered MK and platelet numbers associated with FPD,²  and then pushes past this issue and asks how this knowledge can be exploited to enhance the in vitro production of human MKs. From the perspective of translational research, interest in enhancing MK production is driven by the need to improve the cost efficiency of platelet production. This can be achieved by improved efficiency of MK production or enhancing platelet production per MK. Li et al have contributed a study that will likely aid refinement of the former consideration.

Previous work had demonstrated that correction of RUNX1 in FPD-iPSCs rescued the associated megakaryopoiesis defect.³  In their search for genes whose expression is directly regulated by RUNX1, Li et al compared differences in genome occupancy of RUNX1, and genome-wide transcriptional profiles between FPD-RUNX1^+/− cells and their RUNX1^+/+ corrected counterparts during hematopoietic induction. Multiple differentially regulated genes were found, among them was NOTCH4, whose expression was inversely correlated with RUNX1 rescue. The vast majority of what we know about MK development comes from studies using mouse models; in mice, NOTCH signaling is required for MK commitment from the hematopoietic stem and progenitor cell compartment, with expression of Notch4 being enriched in the MK/erythroid progenitor cell.⁴  This sits counter to the model put forward by Li et al, who demonstrate that deletion of the RUNX1-binding site in intron 29 of NOTCH4 results in increased NOTCH4 expression and subsequent impedance of MK differentiation. Using γ-secretase inhibitors (GSIs) to inhibit NOTCH signaling, the authors validate that suppression further enhances human megakaryopoiesis in vitro (see figure). The specificity for NOTCH4 suppression during megakaryopoiesis was confirmed using NOTCH1, 2, and 3 knockout iPSC lines. As explanation for this interspecies difference, the authors noted that the RUNX1-binding site is absent in the homologous region of mouse Notch4, highlighting the caution that we must exercise when relying solely on mouse models to understand human megakaryopoiesis.

One unresolved issue in this study is whether suppression of NOTCH signaling acts to expand MK-primed precursors or recruits new contributors from other developmental trajectories. In experiments where GSIs were used during iPSC differentiation, optimal MK output was achieved by suppressing NOTCH early in the differentiation protocol, raising the question of whether the trajectory of cells undergoing a nonhematopoietic differentiation had been coopted to megakaryopoiesis. The outcome of GSI inhibition experiments in cord blood (CB) experiments suggests that this diversion likely occurs posthematopoietic commitment, thus, MK outcome is being favored from a hematopoietic progenitor cell. Resolving this ambiguity will provide valuable insight into the fundamental principles that underpin human MK development.

In considering the future directions for this branch of experimental hematopoiesis, it seems timely that we now transition to asking how useful approaches such as these are for the production of platelets. Are platelets derived from an iPSC source equivalent to those produced from CB- or bone marrow–derived material? Are they capable of performing the increasing number of recognized functions required in an adult recipient?^5,6  The central consideration here is the question of how the proteome and functional capability of nascent platelets are affected by that of parental MKs. In mice, it is known that transcriptional differences exist between embryonic and fetal MKs⁷ ; whether these translate to functional differences in their respective platelets is unexplored. An evaluation of the capabilities of platelets produced by NOTCH-suppressed precursors, or by iPS-derived MK cell lines,⁸  is now necessary.