The Final Cellular Output in Human Erythroid Massive Amplification Culture (HEMA) Is Determined by Dynamic Interactions Between Immature and Mature Cell Populations.

Abstract 3156

Poster Board III-93

Hematopoietic progenitor cells cultured with stem cell factor (SCF), interleukin-3 (IL-3) and erythropoietin (EPO) generate erythroid cells that mature in 2 weeks. Addition of dexamethasone and estradiol retards maturation providing the basis for the development of HEMA, a system that generates ex vivo human erythroblasts (EBs) in numbers sufficient for transfusion. In HEMA, EBs at different stages of maturation coexist for > 21 days. FACS analyses for thrombospondin receptor (CD36) and Glycophorin A (CD235a) divides EBs into 4 maturation classes: CD36^highCD235a^neg (CFU-E) and CD36^highCD235a^low (pro-EBs) (immature EBs, iEBs); CD36^highCD235a^high (basophilic-polychromatic EBs) (mature EBs) and CD36^lowCD235a^high (orthochromatic EBs) cells. At day 10, 5-19% of the cells are non-erythroid CD36^negCD235a^neg cells while orthochromatic EBs are barely detectable at any time. The aim of this study was to identify the relative contribution of different populations to the final EBs output of HEMA. The cells present at day 10 in HEMA established with adult mononuclear cells were separated by sorting, labeled with CSFE and cultured for 4 additional days for fate determination (proliferation, maturation and/or death). To identify which population had the ability to generate more EBs, unfractionated cells and cells isolated by sorting (non-erythroid, iEBs and mEBs) were cultured either in colony assay or HEMA. Ninety-three percent of the colony forming ability (all CFU-GM) of unfractionated cells was recovered in the CD36^negCD235a^neg fraction which did not grow in HEMA. However, iEBs did not contain colony forming cells but generated twice as many EBs as unfractionated cells [3.9±1.0 vs 2.38±0.32 fold increase (FI), respectively). In cultures maintained for 16 days, iEBs sequentially sorted every 2 days generated 5-fold more EBs than the corresponding unfractionated EBs (FI=25 vs 5, respectively). mEBs did not generate colonies or proliferate but died in HEMA. Calculations of hemi CSFE staining decrements indicated that the division index of iEBs and mEBs in the first 24 hrs of HEMA was 1.32±0.08 and 0.4±0.08 (p<0.01), respectively. The higher number of cells generated by iEBs with respect to unfractionated cells suggested that mEBs may inhibit iEBs proliferation. To characterize this effect, iEBs and mEBs were co-cultured for 48 hrs at ratios of 100/0, 5/1, 5/3 and 1/2. One of the two populations was labelled by CSFE staining. To take into account a possible toxic effect of CSFE-labeling, experiments were repeated with either CSFE-labelled iEBs or CSFE-labelled mEBs measuring proliferation/maturation/apoptosis of labeled and unlabelled cells, respectively. iEBs alone increased in numbers by 2-fold over 4 days. By contrast, iEB-derived cells did not increase in number in co-cultures at 5/1 and 5/3 mEB ratios and were undetectable in those at 1/2 mEB ratio. Although cell number did not increase in the co-cultures, the proliferation index (1.3), levels of maturation (50% newly generated mEBs) and apoptosis (barely detectable Annexin V^pos cells) of iEBs in co-cultures remained similar to those observed when these cells were cultured alone. The failure to detect cells after 4 days in iEB/mEB co-cultures at 1/2 ratio suggests that mEBs induce non-apoptotic iEB death.

In conclusion, three cells populations were present at day 10 in HEMA of adult mononuclear cells: a CD36^negCD235a^neg population that contains myeloid progenitor cells and does not generate EBs; iEBs, that generate additional iEBs as well as mEBs; and mEBs that have low proliferation activity, limited life-span and induce non-apototic iEB death. Therefore, the final cellular output of HEMA is determined by a delicate balance between proliferation, maturation and cell death of iEBs and mEBs.