The mechanism underlying thrombocythemia in patients with iron deficiency anemia remains unknown. Iron metabolism is highly conserved in humans and mice. We used Tmprss6-/- knockout (KO) mice as a genetic model of chronic iron deficiency anemia to study the effects of iron deficiency on lineage commitment. Tmprss6 is expressed in the liver, but not in hematopoietic cells, and is required for iron uptake from the diet. We observed that Tmprss6-/- mice have elevated platelet counts compared to littermate controls (average x1000/uL ± SD in wild type (WT): 477.9 ± 53.68, Tmprss6-/-: 915.9 ± 57; p<0.001; males and females 6-10 weeks old). Based on immunophenotype, the percentage of hematopoietic stem cells (HSC), megakaryocytic (Mk)-biased HSC (CD41+), megakaryocytic-erythroid progenitor (MEP), (Mk)- or erythroid (E)-committed progenitors in bone marrow (BM) is comparable between the two genotypes. To address whether thrombocythemia in iron deficiency anemia is related to altered commitment of MEP, we assayed the effects of low iron conditions on the function, transcriptome and signaling in MEP.

CFU assays of single MEPs showed a striking increase in CFU-Mk (unilineage committed Mk progenitors) within the MEP gate in Tmprss6-/- at the expense of the bipotent CFU-Mk/E compared to WT (p<0.001). Thus, MEP in KO mice are significantly biased toward the Mk lineage, suggesting that low iron content may reprogram cell commitment. Consistent with this reprogramming effect, MEPs from KO mice have a 50% decrease in the labile iron pool (p=0.03, measured by Calcein-AM), and a 30% decrease (p=0.04) in proliferation relative to WT as measured by CFSE.

To test the effect of altered iron content in vivo, we transplanted WT or KO BM into sublethally irradiated WT or Tmprss6-/- recipients. Tmprss6-/- recipients (of WT or KO BM) exhibited microcytic anemia and hypoferremia when compared to WT recipients (of WT and KO BM). Interestingly, the Mk-bias observed in non-transplanted Tmprss6-/- MEPs was recapitulated in Tmprss6-/-recipients (regardless of transplanted BM genotype) 6 months post-transplant (p < 0.009, n= 3 per group). We show using index single cell sorting followed by CFU assays that in WT mice, surface expression of transferrin receptor 1 (CD71) is a predictor of cell fate; cells from the MEP gate that grow as BFU-E have a higher CD71 mean fluorescence intensity (MFI) (179.9 ± 38.9) than cells that grow bipotent CFU-Mk/E colonies (65.9 ± 13.9) or Mk-only (CFU-Mk) colonies (24.9 ± 7.06; p<0.006). However, this correlation is lost in KO recipients, where the CD71 expression is elevated in all cells compared to WT (MFI: 279.9 ± 49.81, p<0.0001).

RNAseq of WT and KO MEPs showed that 137 genes are differentially expressed (fold change: 1.5; p<0.05) between the two genotypes; of these, 84 are up regulated (e.g. Tfrc and Mybl2) and 53 are downregulated in the KO compared to WT. Surprisingly, mRNA encoding transcription factors that are known to regulate E vs. Mk maturation such as Myb, Tal, Fli1, and Eklf were unchanged. Pathway analysis of the differentially expressed genes highlighted metabolic pathways (asparagine and cholesterol biosynthesis and b-alanine degradation, p<0.008).

We then asked if the iron regulation of the MEP commitment observed in iron deficient mice was also true for human MEP. In order to disrupt the iron sensing pathways in primary human MEP, we performed shRNA-based knockdown of Transferrin Receptor 2 (TFR2) using two different sequence targets and found a similar decrease in proliferation (p=0.03) and skewing toward Mk commitment (p<0.04) as was observed in Tmprss6-/- MEPs. Signal transduction analyses revealed that both human and murine MEPs in iron deficient conditions have lower levels of phospho-ERK (Thr202/Tyr204) compared to WT or scrambled controls.

Regulation of erythroid differentiation by iron is thought to prevent inappropriate iron utilization when body stores are limited. To function in a protective manner, this checkpoint must act at stages prior to erythropoietin-mediated expansion of erythroid progenitors. Here we propose a model where iron acts on the bipotent progenitor, one possible source of Mk and E cells, to bias lineage commitment. These data support that low iron in the environment affects MEP metabolism, leading to lower ERK phosphorylation, slower proliferation and increased Mk commitment.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution