Bmi-1 Regulates Ex Vivo Self-Renewal of Human Erythroblasts

About twelve million units of red blood cells (RBCs) are transfused yearly in the United States. Ex vivo cultured RBCs could serve as a supplemental source to treat alloimmunized patients requiring chronic transfusions. The limited proliferative capacity of erythroid precursors is a major obstacle to generating sufficient numbers of RBCs to constitute even one unit of blood. We previously determined that erythroblasts derived from the murine embryo have a unique ability to self-renew extensively ex vivo, and that Bmi-1, a member of the polycomb repressive complex 1 (PRC1), is both necessary and sufficient to drive the extensive self-renewal of murine bone marrow-derived erythroblasts. Here, we tested the hypothesis that Bmi-1 regulates the ex vivo self-renewal of human erythroid cells. Bmi-1 overexpression increased the proliferative capacity of adult human peripheral blood mononuclear cell-derived erythroblasts, which normally have restricted ex vivo self-renewal, more than 10 billion-fold. Bmi-1-induced self-renewing erythroblasts (iSREs) retained an immature erythroid morphology and cell surface phenotype throughout culture. Chemical inhibition of Bmi-1 led to collapse of the culture, with a massive reduction in cycling cells and an increase in apoptosis. Taken together, these data indicate that Bmi-1 in iSREs is both necessary and sufficient to increase the self-renewal capacity of human erythroid precursors. Importantly, Bmi-1 overexpression does not interfere with the ability of the iSREs to mature into reticulocytes in vitro. Serological analysis of multiple blood group antigens demonstrated that iSRE-derived reticulocytes express the same antigens as the donor's RBCs. Bmi-1 can act as part of canonical or non-canonical PRC1 and different complex members have been associated with hematopoietic self-renewal versus differentiation. Expression studies of self-renewing human erythroblasts revealed that the non-canonical PRC1 member, RYBP, was expressed at higher levels than canonical PRC1 members. In addition, maturation of erythroid precursors is associated with down-regulation of Bmi-1 and non-canonical PRC1 members, as well as the upregulation of several canonical members. Knockdown of RYBP reduced the proliferative capacity of the iSREs, suggesting that Bmi-1 regulates erythroid self-renewal through non-canonical PRC1 interactions. Consistent with this hypothesis, preliminary studies inhibiting canonical PRC1 members led to an additional several trillion-fold increase in the ex vivo proliferative capacity of human iSREs compared to vehicle-treated control cultures. These data suggest that Bmi-1 regulates erythroid self-renewal through non-canonical PRC1. Increasing the ex vivo self-renewal capacity of human erythroblasts ultimately paves the way for the generation of sufficient numbers of cultured RBCs for blood typing and transfusion therapy, as well as the establishment of in vitro models of RBC-intrinsic disorders.