Fig. 4: Functional characterization of CD48^–/lo and CD48^hi MkP subpopulations.

a, b Fgd5^{ZsGreen:CreERT2}/R26^LSL-tdRFP mice were TAM-induced, and indicated cell populations were analyzed at 10, 59, and 120 days (n = 4 mice/time point). a Representative example of enriched RFP labeling in CD48^–/lo MkPs 59 days after TAM induction. Quantification of label in MkP subpopulations (right data plot, individuals, and means are shown). b RFP labeling of cell populations displayed as fraction of ES HSCs. c R26^rtTA/Col1A1^H2B-GFP animals (n = 5) were DOX-induced and chased for 11 wks. Representative dot plot (left) and frequency of label retaining (H2B-GFP⁺) MkP subpopulations (right data plot, individual mice, and means are shown). d Single CD48^−/lo (n = 137) and CD48^hi (n = 122) MkPs as well as control cells (LS⁻E⁻K: lin⁻ CD201⁻ Sca-1⁻ CD117⁺ progenitors (n = 17); LSK: lin⁻ Sca-1⁺ CD117⁺ cells (n = 23)) were cultivated for 5 days and classified according to expansion and morphology. Bars represent mean values, and error bars SEMs. e 2500 CD48^hi or CD48^−/lo MkPs were purified from B6.RFP donor animals and transplanted into sublethally-irradiated B6.CD45.1/CD45.2 recipient mice (n = 7/condition). Contribution to platelets, erythrocytes, neutrophils was monitored for 3 weeks (mean and SD are shown). Transplanted LSK cells served as positive control (n = 1, see Supplementary Fig. 8h), transplanted B6.CD45.1/.2 carrier splenocytes as negative controls (n = 3); significance calculated with repeated measures 2-way ANOVA with Holm–Sidak error correction (ns, not significant; *, <0.05; **, <0.01). f Full model of thrombopoiesis. The direct (via CD48^−/lo MkP) and the conventional (via CMP-CD48^hi MkP) pathways equally contribute to native thrombopoiesis. g Mathematical inference on the fate mapping data based on the two-pathway model of thrombopoiesis fits the experimental data (n = 4 mice/time point; same mice as in a and b, mean and SEM are shown) and confirms the model prediction that MkPs consist of two distinct subpopulations (see Supplementary Fig. 8c, d). h CD48^−/lo MkPs differentiate more frequently than CD48^hi MkPs (left plot), while the cell number is lower in CD48^−/lo MkP (middle plot). Overall differentiating cell fluxes (number of cells becoming Mk per day) from both MkP subpopulations are balanced (right plot). i Contributions of the two megakaryopoiesis pathways are balanced in young adult mice, whereas CD48^−/lo MkPs supply more Mks in aged mice. In h and i error bars represent 95% confidence bounds.