Endocytosis of the thrombopoietin receptor Mpl regulates megakaryocyte and erythroid maturation in mice

Abstract

Dnm2^fl/fl Pf4-Cre (Dnm2^Plt-/- ) mice lacking the endocytic GTPase dynamin 2 (DNM2) in platelets and megakaryocytes (MKs) develop hallmarks of myelofibrosis. At the cellular level, the tyrosine kinase JAK2 is constitutively active but decreased in expression in Dnm2^Plt-/- platelets. Additionally, Dnm2^Plt-/- platelets cannot endocytose the thrombopoietin (TPO) receptor Mpl, leading to elevated circulating TPO levels. Here, we assessed whether the hyperproliferative phenotype of Dnm2^Plt-/- mice was due to JAK2 constitutive activation or to elevated circulating TPO levels. In unstimulated Dnm2^Plt-/- platelets, STAT3 and, to a lower extent, STAT5 were phosphorylated, but their phosphorylation was slowed and diminished upon TPO stimulation. We further crossed Dnm2^Plt-/- mice in the Mpl^-/- background to generate Mpl^-/-Dnm2^Plt-/- mice lacking Mpl ubiquitously and DNM2 in platelets and MKs. Mpl^-/- Dnm2^Plt-/- platelets had severely reduced JAK2 and STAT3 but normal STAT5 expression. Mpl^-/- Dnm2^Plt-/- mice had severely reduced bone marrow MK and hematopoietic stem and progenitor cell numbers. Additionally, Mpl^-/- Dnm2^Plt-/- mice had severe erythroblast (EB) maturation defects, decreased expression of hemoglobin and heme homeostasis genes and increased expression of ribosome biogenesis and protein translation genes in spleen EBs, and developed anemia with grossly elevated plasma erythropoietin (EPO) levels, leading to early fatality by postnatal day 25. Mpl^-/- Dnm2^Plt+/+ mice had impaired EB development at three weeks of age, which normalized with adulthood. Together, the data shows that DNM2-dependent Mpl-mediated endocytosis in platelets and MKs is required for steady-state hematopoiesis and provides novel insights into a developmentally controlled role for Mpl in normal erythropoiesis, regulating hemoglobin and heme production.

Keywords: dnm2; erythropoiesis; hematopoiesis; megakaryopoeiesis; mpl.

Figure 1

JAK2-STAT signaling defects in Dnm2^Plt–/– platelets. (A) Dnm2^Plt+/+ and Dnm2^Plt–/– platelets were activated with 50 ng/ml TPO for 10 min at 37°C, lysed, subjected to SDS-PAGE, and probed for total and phosphorylated STAT3 (pSTAT3; Tyr705), total and phosphorylated STAT5 (pSTAT5; Tyr694 in STAT5A, Tyr699 in STAT5B), and β-actin as loading control, as indicated. Densitometry analysis of STAT3 (B) and STAT5 (C) phosphorylation. Results represent mean ± SD of 4 independent experiments and are compared statistically by two-way ANOVA (**,P <.01; ***, P <.001). (D) Platelet lysates of Dnm2^Plt+/+ , Dnm2^Plt–/– , Mpl^–/– Dnm2^Plt+/+ , and Mpl^–/– Dnm2^Plt–/– mice at P24 corresponding to 2 µg of protein were subjected to SDS-PAGE and probed for JAK2, STAT3, STAT5, and GAPDH as loading control, as indicated. Results are representative of 3 independent experiments. Densitometry analysis of JAK2 (E), STAT3 (F), and STAT5 (G) expression. Results represent mean ± SD of 3 independent experiments and are compared statistically by one-way ANOVA (ns, not significant, *P <.05; **P <.01; ***P <.001).

Figure 2

Mpl^–/– Dnm2^Plt–/– mice exhibit early lethality, failure to thrive, splenomegaly, and severe anemia. (A) Survival of Mpl^+/– Dnm2^Plt+/+ , Mpl^+/– Dnm2^Plt–/– , Mpl^–/– Dnm2^Plt+/+ , and Mpl^–/– Dnm2^Plt–/– littermates. Results are estimated using the Kaplan-Meier method and are compared statistically using the Log-rank test (n = 20 mice in each group: ***, Log-rank P <.001). Body weights (B) and spleen/body ratio (C) at P14, P19, and P24. Platelet (D), RBC (E), reticulocyte (F), and WBC (G) counts from birth (P0) to postnatal day 24 (P24). Results represent mean ± SD of 5-19 independent experiments and are compared statistically by two-way ANOVA (**P <.01; ***P <.001).

Figure 3

The MK hyperplasia of Dnm2^Plt–/– mice requires Mpl expression. Platelet count at P24 (A) and P56 (E). Results represent mean ± SD of 7-13 independent experiments and are compared statistically by one-way ANOVA (ns, not significant, ***, P <.001). Seven-µm frozen bone marrow sections of Dnm2^Plt+/+ , Dnm2^Plt–/– , Mpl^–/– Dnm2^Plt+/+ , and Mpl^–/– Dnm2^Plt–/– mice at P24 (B) and P56 (F) were probed for resident MKs (GPIbα, green) and bone marrow vasculature (laminin, red). Sections shown are representative of 3-4 mice in each genotype. Scale bars represent 150 µm. Bone marrow MK numbers at P24 (C) and P56 (G). Results represent mean ± SD of 3-4 independent experiments and are compared statistically by one-way ANOVA (ns, not significant, **P <.01; ***P <.001). (D) Plasma TPO levels at P24. Results represent mean ± SD of 5-16 independent experiments and are compared statistically by one-way ANOVA (*P <.05; **P <.01; ***P <.001).

Figure 4

Mpl^–/– Dnm2^Plt–/– mice exhibit impaired erythroid development at P24. RBC count at P24 (A) and P56 (F). Reticulocyte count at P24 (B) and P56 (G). Results represent mean ± SD of 4-18 independent experiments and are compared statistically by one-way ANOVA (ns, not significant, ***P <.001). Flow cytometry profiles of spleen EBs at P24 (C) and P56 (H) using the erythroid markers CD71 and TER-119. Data shown are representative of 3-13 mice in each genotype. Percentage of mature spleen CD71^low/TER-119^high EBs at P24 (D) and P56 (I). Results represent mean ± SD of 5-10 independent experiments and are compared statistically by one-way ANOVA (ns, not significant, *P <.05; **P <.01; ***P <.001). Plasma EPO levels at P24 (E). Results represent mean ± SD of 5-15 independent experiments and are compared statistically by one-way ANOVA (ns, not significant, ***P <.001). (J) Genomic Dnm2 expression in isolated spleen CD71^high EBs was evaluated by qPCR and normalized to Dnm2^Plt+/+ cells. Results represent mean ± SD of 4 independent experiments and are compared statistically by Student’s t-test. (ns, not significant).

Figure 5

The increased HSPC expansion of Dnm2^Plt–/– mice requires Mpl expression. Frequency of bone marrow Lin^–/Sca1⁺/Kit⁺ (LSK) (A), long-term (LT)-HSC (B), short-term (ST)-HSC (C), Pre-Meg-E (D), MK progenitor (MKP) (E), Pre-CFU-E (F), and CFU-ProE (G) in Dnm2^Plt+/+ , Dnm2^Plt–/– , Mpl^–/– Dnm2^Plt+/+ , and Mpl^–/– Dnm2^Plt–/– mice at P24. Results represent mean ± SD of 4-12 independent experiments and are compared statistically by one-way ANOVA (**P <.01; ***P <.001). ns, not significant.

Figure 6

Transcriptional profiling of spleen EBs in Dnm2^Plt+/+ , Dnm2^Plt–/– , Mpl^–/– Dnm2^Plt+/+ , and Mpl^–/– Dnm2^Plt–/– mice. Volcano plots showing genes differentially expressed in EBs of Mpl^–/– Dnm2^Plt+/+ + Mpl^–/– Dnm2^Plt–/– versus Dnm2^Plt+/+ + Dnm2^Plt–/– mice (Mpl effect), displayed as adjusted P-value (A) and baseMean (B). Volcano plots showing genes differentially expressed in EBs of Dnm2^Plt–/– + Mpl^–/– Dnm2^Plt–/– versus Dnm2^Plt+/+ + Mpl^–/– Dnm2^Plt+/+ mice (DNM2 effect), displayed as adjusted P-value (C) and baseMean (D). The log2FoldChange indicates the mean expression level change for each gene. Dashed lines indicate fold changes of -1.5 and 1.5 (x-axis) and adjusted P-value of.05 (y-axis). Each dot denotes one gene. Heat maps showing relevant genes identified using the Mpl (E) and DNM2 (F) effects in all four groups.