Establishment of a Spontaneous Stromal Microenvironment from Cord Blood Supports Human Dynamic Erythropoietic Temporal Maturation in Long-Term Serum- and Cytokine-Free 3D Cultures and Reveals a Distinct CD44^hi Population

Dynamic cultures which can represent physiologic erythropoiesis in vitro require a three-dimensional (3D) architecture with a supportive microenvironment and addition of erythropoietin (EPO). We have previously reported on a 3D bone marrow (BM) biomimicry using polyurethane scaffolds to expand cord blood mononuclear cells (CBMNCs) in a serum- and cytokine-free environment, without addition of dexamethasone, for 28 days (D). CBMNCs were seeded (4x10⁶ cells/scaffold), supplemented with 10ng/mL stem cell factor (SCF; D0-D28) and 100mU/mL EPO (D7-D28), with medium exchange every 3D and exposed to a hypoxia (5%)/normoxia (20%) schedule to mimic BM oxygen gradients. Hypoxia induced rapid erythroid commitment and established an early erythroid progenitor population in the absence of EPO. Normoxia and EPO was required at later maturational stages and enhanced the γ-globin to β-globin switch. We identified D7-D14 as crucial for endogenous cytokine production.

Herein, we extended cultures to D48 using two high-dose EPO-stimulation cycles (1U/mL; D20 and D44) to enhance erythropoiesis and further define the microenvironment. Proliferation was higher after EPO pulses (p<0.05) but did not result in enhanced erythropoiesis, suggesting the absence of erythroid precursors. An allogeneic CB unit was added to "recharge" the cultured scaffolds at D39 and a new cycle of erythropoietic differentiation was initiated. Cell proliferation was 4.5-fold higher at D68, compared with that at D28. From D53-D68, CD71⁺CD235a⁺ cells were constantly produced (25-54%), corresponding with the presence of erythroid precursors supporting CFU-E and BFU-E. Erythroblastic islands were identified and maturing and enucleated reticulocytes/RBCs were abundant (19±2%; 1±0.3x10⁶ cells) with expression of γ- and β-globin, band 3 and 4.1R RBC membrane proteins. To further evaluate the relative contributions of each CB unit to the "recharge" culture, seeded scaffolds were subjected to irradiation (or not) 48h prior to recharge. By D65, >30% of supernatant cells were CD71^+/modCD235a⁺ and supported BFU-E and CFU-E. Proliferation in long-term cultures was attributed to the second CB unit, regardless of irradiation, as shown by HLA-typing of D68 cells; the first CBMNCs only contributed to establishment of the microenvironment.

To further characterize erythroid differentiation dynamics, expression of CD44 vs CD235a was used to identify and sort three erythroid populations. Progenitors in CD44^-/modCD235a^- populations evolved to CD44^modCD235a^mod, and then to CD44^-/modCD235a⁺ cells, which constitute the most mature erythroid phenotype; CD44^modCD235a^mod erythroid precursors supported mainly BFU-E and CFU-E. A unique CD44^hiCD71^mod population increased during culture, displayed myeloid progenitor morphology and only supported CFU-GM. This population did not express CD34, CD33 or CD14 but expressed c-KIT, which suggests a hematopoietic population that provides essential culture support. Further characterization of the spontaneously created microenvironment by in situ quantitative analysis of scaffold mid-sections during the 68-day culture showed varying and high dynamic expression of Nestin, STRO-1, CD146, CD68 and CD169 in separate cell populations as well as expression of RUNX2 and Osx. Osteopontin was not detected.

In summary, a BM biomimicry composed of diverse stromal populations was spontaneously created in the 3D in vitro scaffold system. This microenvironment proved to effectively induce and sustain erythropoiesis to enucleation to at least D68 when supplied with a second source of CBMNCs, without addition of stroma-specific factors. A unique CD44^hi immature monocyte/macrophage population was identified, which contributes to the inductive microenvironment, and distinct stages of human erythroid maturation could be identified using CD44 and CD235a. This work presents a novel and dynamic ex vivo model that can (1) recapitulate physiologic human erythropoiesis in steady-state and stress conditions, (2) capture the fetal to adult hemoglobin transition, (3) explore the direct role of oxygen on erythropoiesis, (4) assess the microenvironment relevant to erythropoiesis in the absence of serum and exogenous factors, (5) sustain long-term erythropoiesis with terminal maturation and, (6) explore the different stromal niche environments spontaneously created for the support of erythropoiesis.