Erythropoiesis is sensitive to inflammation and its associated mediators. We previously identified the damage-associated molecular pattern, high mobility group box 1 protein (HMGB1), as a causative factor involved in the development of anemia using a murine model of chronic inflammation, and recently, showed that HMGB1 negatively affects human erythropoiesis, in part, through altered erythropoietin (EPO) signaling and hematopoietic stem and progenitor cell (HSPC) lineage commitment. Despite these findings, a mechanistic understanding of erythroid dysfunction in the presence of HMGB1 is lacking.

We first documented a dose-dependent logarithmic difference in erythroid cell growth at Days 7 and 11 of culture of CD34+ HSPCs during erythroid differentiation as well as a relatively smaller loss of cells at Day 14 between HMGB1-treated and vehicle-treated cultures. These results suggest that HMGB1 exerts its effects at distinct stages of erythroid differentiation: at both the progenitor and the precursor stages. Indeed, flow analysis of erythroblast differentiation showed that HMGB1 caused an accumulation of precursors at the basophilic erythroblast stage. The loss of poly- and orthochromatic erythroblasts resulted from increased apoptosis at these stages of erythroid differentiation as assessed by annexin-V staining. Induction of apoptosis was confirmed by a dose-dependent increase in apoptosis and 3-fold increase in cleaved caspase-3 in HUDEP-2 cells, an erythroid cell line. Furthermore, a mutant form of HMGB1 (3S-HMGB1) in which critical cysteine residues were mutated to abolish the activity of the protein failed to alter erythroid proliferation and survival.

HMGB1 appears to deprive differentiating erythroid cells of growth and survival signals through attenuated phosphorylation of downstream EPO effectors including pJAK2 and pSTAT5, independent of SHP1/2 activity. As a consequence of perturbed EPO signaling, erythroid cells contained reduced levels of nuclear HSP70 and GATA1, which ultimately led to apoptosis. Importantly, EPO co-administrated with either tumor necrosis factor (TNF), interleukin-1β (IL-1β), or IL-6 led to phosphorylation levels of JAK2-STAT5 comparable to controls, suggesting that direct inhibition of EPO signaling is specific to HMGB1.

To document the effect of HMGB1 on EPO-dependent and independent progenitors, erythroid progenitor growth and differentiation were monitored at 24hr intervals in culture. CFU-E but not BFU-E numbers were decreased following treatment with HMGB1. Moreover, HMGB1-treated cells expressed high levels of GATA2 indicating deregulation of GATA transcription factor switching. Myelo-erythroid culture systems demonstrated that HMGB1 strongly inhibited erythroid commitment of CD34+ HSPCs while it favored the generation of increased numbers of CD11b+ and CD13+ expressing myeloid cells. The induced switch in lineage commitment is consistent with increased GATA2 expression and decreased GATA1 expression. Further, we observed that erythroid and hematopoietic progenitor cultures became more acidic in the presence of HMGB1 suggesting that HMGB1 may also influence progenitor metabolic processes.

Our data demonstrate that HMGB1 negatively affects erythropoiesis at EPO-dependent and -independent stages of erythroid differentiation through decreased CFU-E numbers and apoptosis of erythroblasts, and also favor increased myeloid differentiation of HSPCs. Contrary to classical inflammatory mediators, HMGB1 relies on a novel mechanism, involving downstream EPO effectors, to mediate erythroid dysfunction. Ongoing pharmacologic and genetic screens of EPO signaling axes, putative HMGB1 receptor(s), and progenitor metabolic profiling are currently underway to better understand how HMGB1 causes erythropoietin refractory anemia in the setting of inflammation.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.