BMI1 Regulates Proliferation of Human Late-Stage Erythroid Progenitors

Red blood cells (RBCs) are the most common form of cell-based therapy with ~11 million units transfused annually in the United States. Cultured (c)RBCs could supplement donor-derived units, particularly for alloimmunized patients requiring chronic transfusions. However, the limited proliferative capacity of erythroid precursors is a major obstacle to generate sufficient numbers of cRBCs for clinical purposes. We and others have determined that Bmi1, a member of the polycomb repressive complex 1 (PRC1), is both necessary and sufficient to drive the extensive self-renewal of immature erythroblasts (SREs), however, the mechanisms of BMI1 action remain poorly understood.

To identify potential downstream targets of BMI1 we performed CUT&RUN with BMI1 and RING1B on proliferating human SREs. Gene ontology analysis of the top 5% of BMI1 binding sites revealed significant enrichment of cell cycle pathway genes. Consistent with specific BMI1 binding at the INK/ARF locus in SREs, overexpression of BMI1 led to a significant reduction in the expression of the cell cycle inhibitors p14, p15, and p16. In addition, chemical inhibition and knockdown of BMI1 each significantly reduced the percentage of SREs in S-phase of the cell cycle.

Via comparative global transcriptomics, human SREs mapped nearest to CD34-derived proerythroblasts in UMAP space. Interestingly, BMI1 expression is normally upregulated as human erythroid progenitors differentiate, and then rapidly downregulated as immature erythroid precursors undergo terminal maturation, suggesting BMI1 may regulate normal erythropoiesis at the late progenitor/proerythroblast stage. In support of this hypothesis, treatment of human CD34+ cell erythroid cultures with the BMI1 inhibitors, PTC209 or PTC028, reduced late-stage (EP3 and EP4), but not immature (EP1 or EP2), erythroid progenitor numbers. Intriguingly, these inhibitors also reduced the percentage of cycling EP3 and EP4, but not EP1 and EP2, progenitors, suggesting that BMI1 regulates the cell cycle specifically in late-stage erythroid progenitors.

Our CUT&RUN experiments in human SREs also identified key cholesterol synthesis pathway genes bound by BMI1, including HMGCR, the rate-limiting enzyme in cholesterol synthesis. Since the self-renewal, but not terminal maturation, of avian SREs is dependent on cholesterol metabolism, we tested the function of cholesterol synthesis in human SREs. While short-term removal of exogenous cholesterol from the media did not alter SRE proliferation, concomitant treatment with Ro 48, which blocks the last step of cholesterol synthesis, led to a near complete loss of SREs. Interestingly, SRE proliferation was significantly, but not completely, rescued by the addition of exogenous lipids, highlighting the importance of cholesterol metabolism in human SRE proliferation. These findings, taken together, support the hypothesis that BMI1 regulates erythroid self-renewal through several mechanisms including the cell cycle and cholesterol biosynthesis.

No relevant conflicts of interest to declare.