Adult megakaryopoiesis: when taking a short-cut results in a different final destination

It is now well-established that there are 2 distinct pathways for megakaryopoiesis. One is defined by the traditional stepwise differentiation pathway that progresses from hematopoietic stem and progenitor cells (HSPCs) to common myeloid progenitors (CMPs), megakaryocyte-erythroid progenitors (MEPs), megakaryocyte progenitors (MkPs), and then to megakaryocytes (MKs), which give rise to platelets.1,2 The other pathway has more recently been reported as the direct differentiation pathway from HSPCs to MkPs3,4 (Fig. 1A). Recent studies by Li et al5 have revealed that the 2 pathways respond differently to inflammation and myeloablation stressors, producing inflammatory or niche-supporting MKs, respectively, in addition to platelet-producing MKs.5Figure 1.: Differential MkP/megakaryocyte maturation pathways produce functionally distinct megakaryocytes. (A) The Cd48 dre;R26 rox-tdTomato reporter mouse model differentiates between MkPs and MKs that are formed from direct differentiation (unlabeled; white) and those produced through the classical stepwise differentiation process (tdTomato+; orange). The direct differentiation pathway preferentially produced niche-supporting MKs while the stepwise process produced immune-associated MKs, both processes equally produced platelet-producing MKs. The direct differentiation pathway was increased upon myeloablation (B; a single dose of chemotherapy agent 5-fluorouracil) while inflammation increased the stepwise differentiation pathway (C; a single dose of lipopolysaccharide). HSPC = hematopoietic stem and progenitor cell, MK = megakaryocyte, MkP = megakaryocyte progenitor.Murine lineage-tracing models have been prominent tools for understanding hematopoiesis.2,5,6 Li et al5 generated Cd48dre;R26rox-tdTomato mice by introducing a constitutively active Dre recombinase behind the endogenous Cd48 gene stop codon and crossed them with Rosa26rox-STOP-rox-tdTomato(R26rox-tdTomato) mice. Flow cytometry analyses of 8-week-old mice confirmed that, in this model, hematopoietic stem cells (HSCs; defined as lineage negative, c-Kit+Sca-1+ [LKS+] CD34−CD135−) and multipotent progenitors (MPPs; LKS+CD34+CD135−) were tdTomato−. The majority of downstream progenitors were tdTomato+, with the exception of MkPs and MKs, which comprised approximately 50% tdTomato− cells, indicating that they emerged from the direct differentiation pathway5 (Fig. 1A). To generate a model to fate map Cd48-Dre-targeted cells, Li et al5 generated the tamoxifen-inducible Ubc-creER;Cd48dre;R26ZT1 mice, which labeled Cd48-targeted lineages with tdTomato and lineages generated from cells that never expressed Cd48 with ZsGreen at the time of tamoxifen induction.5 Li et al5 injected tamoxifen intraperitoneally into 2-month-old Ubc-creER;Cd48dre;R26ZT1 mice for 3 consecutive days and analyzed them 1 day later. Flow cytometry revealed that 80% to 90% of HSCs and MPPs expressed ZsGreen, and 80% to 90% of CMPs, GMPs, MEPs, common lymphoid progenitors (CLPs), myeloid, erythroid, B and T cells were labeled with tdTomato. There were equal proportions of MKs that expressed tdTomato or ZsGreen.5 Longer term analyses of the Ubc-creER;Cd48dre;R26ZT1 mice provided valuable insight into the turnover cycles of each of the Cd48-targeted hematopoietic cell populations, because the tdTomato+ cells were replaced by ZsGreen cells. This revealed that it took approximately 30 days for Cd48-targeted tdTomato+ MKs to disappear in the mice, with the majority being depleted between 5 and 14 days post-tamoxifen. Similar observations were obtained using Kit-creER in place of Ubc-creER.5 Li et al5 then performed scRNA-seq on CD41+ MKs isolated from the bone marrow of Cd48dre;R26rox-tdTomato mice. Differential gene expression analysis defined four MK clusters that were termed platelet-producing MKs, cycling MKs, niche-supporting MKs and immune MKs, consistent with their previous report.7 The Cd48-targeted tdTomato+ cells (generated by stepwise differentiation) contained most of the immune MKs (Cd53 and Lsp1 high) but few niche-supporting MKs; and Cd48− tdTomato− cells (derived by direct differentiation) contained most of the niche-supporting MKs (Igf1 and Mylk4 high) but few immune MKs.5 The platelet-producing and cycling MKs were distributed equally between both groups (Fig. 1A). In comparison to Cd48− tdTomato− MKs, the Cd48-targeted tdTomato+ MKs contained more 2N and fewer 32N MKs and had 10-fold greater phagocytic activity, a feature of immune MKs.7,8 The authors also found that 13% of Cd48-targeted tdTomato+ MKs expressed major histocompatibility complex (MHCII) compared to 1% of Cd48− tdTomato− MKs. Furthermore, confocal imaging of bone marrow sections showed Cd48− tdTomato− MKs localized closer to Ctnnal1GFP HSCs.5 The 2 MK differentiation pathways responded differently to physiological stressors. There was an increase in the frequency of Cd48− tdTomato− MKs produced by the direct differentiation pathway in response to myeloablation by 5-fluorouracil (5-FU) (Fig. 1B). In contrast, Cd48-targeted tdTomato+ MKs produced by stepwise differentiation were increased after lipopolysaccharide (LPS)-induced inflammation (Fig. 1C). Notably, scRNA-seq studies revealed that the niche-supporting MK cluster obtained from 5-FU–treated mice (8 days post-5-FU) increased from 12% to 27% compared to the control. Furthermore, the immune MK cluster obtained from the LPS-treated mice (12 hours post-treatment) increased from 34% to 50% compared to the control. Collectively, the studies by Li et al5 have provided additional evidence that megakaryopoiesis is complex, involving distinct differentiation pathways that produce functionally different MKs with diverse physiological roles. Using murine lineage-tracing models, the authors showcased the kinetics and functional differences between the stepwise and direct pathways. It is worth noting that the Itga2b (CD41) expression level in the immune cluster was lower than in the other three subpopulations. If this is also observed for cell surface expression, CD41 expression levels can potentially be used to separate the immune MKs from the other MK subpopulations. Surprisingly, there were very few Cd48+ tdTomato+ cells observed in the HSC (LKS+CD34−CD135−) and MPP (LKS+CD34+CD135−) populations. It has previously been shown that the MPP population contains significant proportions of both MPP2 (LKS+CD135−CD150+CD48+) and MPP3 (LKS+CD135−CD150−CD48+), cells that express high levels of CD48.9 This discrepancy is likely due to a difference in gating strategies used to identify the cells in the current study and it would be beneficial to revise the flow cytometry gating strategies to be consistent with those validated by previous publications. This may also explain some discrepancies to the studies of platelet production in young FlkSwitch reporter mice (Cd135Cre;R26mTmG) generated by the Forsberg group.2 In the FlkSwitch reporter mice, the majority of MkPs and platelets in young mice expressed GFP, indicating that they were generated via Cd135-expressing HSPCs (the majority which co-express CD489). Intriguingly, a recent study using the FlkSwitch reporter mice identified a distinct subpopulation of Cd135− tdTomato+ MkPs in aged (24-month-old) mice and showed that the Cd135− tdTomato+ platelets were hyper-reactive.10 It would be worth determining if the Cd135− tdTomato+ aged MkPs preferentially produce the MK subpopulations identified in the Cd48dre mice.5 Furthermore, it would be of interest to determine if the platelets produced by Cd48− tdTomato− MKs are hyper-reactive in both young and aged mice. Additionally, the 2N CD41+ MKs isolated in the current study would include CD41+ HSCs6 and other cells that express CD41, including MkPs, and it is important to exclude those cells in future studies. Irrespective, the findings reported by Li et al5 improve our understanding of megakaryopoiesis and provide novel mouse models that will be very useful for further studies of hematopoiesis, including megakaryopoiesis.