Cytotoxic Stress Transiently Increases the Contribution of Platelet-Biased Hematopoietic Stem Cells to Platelet Production

While hematopoietic stem cells (HSCs) and megakaryocyte (Mk)/platelet lineages exhibit substantial phenotypic and transcriptional overlap, direct Mk production from HSCs is still under investigation. In particular, the hypothesis that specific HSC clones might facilitate rapid megakaryopoiesis and their physiological roles are yet to be elucidated. Previous clonal studies using either single cell transplantations (Yamamoto, Cell 2013; Carrelha, Nature 2018) or barcoding assays (Pei, Cell Stem Cell 2020; Rodriguez-Fraticelli, Nature 2020) have identified subsets of HSCs with Mk-biased output, designated as Mk-biased HSCs or platelet-biased HSCs (P-HSCs), in contrast to multi-lineage HSCs (MUL-HSCs) with balanced output. Moreover, P-HSCs exhibit up-regulated platelet-lineage gene expression and chromatin accessibility, and they generate platelets at a faster rate than MUL-HSCs, pointing towards a direct platelet formation pathway (Rodriguez-Fraticelli, Nature 2020; Meng, 2023 Nat Cell Biol). Notably, under inflammatory stress, direct Mk progenitor (MkP) production and Mk formation from HSCs have been shown to accelerate (Haas, Cell Stem Cell, 2015; Morcos, Nat Commun 2022), suggesting that P-HSCs could drive rapid platelet production during hematopoietic emergencies.

To probe this premise, we re-analysed previously published RNA-seq data on HSC subtypes and discovered that the gene Vldlr is markedly expressed in MUL-HSCs relative to P-HSCs. We subsequently created a Vldlr-Cre transgenic mouse model for lineage tracing of platelet production from MUL-HSCs and P-HSCs. Flow cytometry analysis across the HSC-MPP2-preMegE-MkP-platelet developmental sequence revealed a notable increase in the percentages of Tomato- HSCs and platelets, in comparison to MPP2-preMegE-MkP, suggesting that Tomato- HSCs significantly bypass the conventional MPP2-preMegE-MkP pathway in platelet formation. Transplantation of labelled and non-labelled HSCs confirmed that Vldlr-Cre predominantly labelled MUL-HSCs, whilst Tomato- HSCs were rich in P-HSCs and largely generated Tomato- platelets, implying that platelets derived from the direct P-HSC pathway can be identified as Tomato- through Vldlr-Cre lineage tracing.

We exploited this model to examine the distinct roles of classical and direct pathways during emergency thrombopoiesis, induced by either antibody-mediated platelet depletion mimicking immune thrombocytopenia or a cytotoxic drug 5-fluorouracil (5-FU) that causes severe hematopoietic disruption. Both interventions precipitated a significant decrease in platelet counts, as previously demonstrated. However, the proportion of Tomato- platelets did not change during platelet recovery following antibody intervention, while the percentage of Tomato- platelets rose up to four-fold in early recovery stage following 5-FU treatment, suggesting an acceleration of the direct pathway under this stressor. The percentage of Tomato- platelets returned to baseline following platelet count normalization. Further analysis of bone marrow progenitors post-intervention revealed that the proportion of Tomato- MkP was maintained during recovery from antibody treatment but significantly expanded after 5-FU treatment. Nevertheless, the fraction of Tomato- HSCs remained unaltered under both interventions, indicating that neither reagent selectively activates P-HSCs or MUL-HSCs. We hypothesize that the 5-FU treatment, by significantly eliminating bone marrow progenitors, permits P-HSCs to rapidly replenish the MkP pool and platelets due to their faster platelet formation kinetics, whereas progenitors are not affected by the antibody, leaving both pathways to contribute to platelet recovery.

In summary, our findings suggest that P-HSCs provide a fast-track for platelet formation, enabling them to contribute to platelet recovery more rapidly following progenitor depletion induced by cytotoxic stress.