The observation that CD34pos and mononuclear cells (MNC) from cord (CB) and adult (AB) blood generate great numbers of erythroid cells (EBs) in human erythroid massive amplification (HEMA) cultures (cultures stimulated with human SCF, IL-3, erythropoietin, dexamethasone and estradiol) led to the hypothesis that EBs generated ex-vivo from currently discarded stem cell sources (low volume CB and MNC from leukoreduced AB) may serve as transfusion products. Studies on ex-vivo EB expansion performed primarily with cells from de-identified Caucasian donors have demonstrated donor variability on the number of EBs generated. Alloimmunized patients of African descendent often require antigen-matched transfusion products which can only be found among ethnically-matched donors. The aim of this study was to determine how the variability in ex-vivo EB expansion under HEMA conditions is affected by ontogeny, ethnicity, gender and loss of companion cells during preparation procedures.
EBs expansions in 3 replicate cultures from the same AB (3 donors) differed only by 10%. By contrast, average day 14 FI in HEMAser cultures from 27 AB (26 Caucasians and 1 African-american) and 7 CB (all Caucasians) varied widely. By day 15, FI ranged from 1.7–30 for AB (FI=30 for the African-american AB) and 7.5–337 for CB. The average FI observed with CB was greater than that observed with AB (51±63 vs 12±10, p<0.05) in spite of similar colony forming cell (CFC) content [123±82 (n=5) vs 73±28 (n=18) CFC/105 MNC).
AB MNC from females (n=3) contained lower numbers of CFC than those from males (n=3)(79±35 vs 164±36 total CFC/105 MNC, p<0.05). However, day 15 FI observed in the corresponding HEMAser (11±3 vs 11±7) and HEMAdef (media formulated with human components, 80±20 vs 82±10) cultures were similar. The FI observed with AB under HEMAdef conditions are not statistically different from those obtained from CB.
Both AB and CB MNC generate ∼10-times more EBs than CD34pos cells. This effect may be ascribed either to loss of CD34neg erythroid precursors (van den Akker et al. Haematologica, 95:1594, 2010) and/or to the fact that current purification procedures recover only 30–50% of the total number of CD34pos cells present in MNC. To clarify the effect of loss of companion cells (CD34pos and/or CD34neg) on EB expansion, the frequency and EB expansion potential of cells sorted according to the CD34/CD36 profile from AB and CB were compared. Total MNC were cultured in parallel as control. In AB, 3 populations (A, CD34posCD36neg, 0.1±0.05%; B, CD34posCD36pos, barely detectable-0.1% and C, CD34negCD36low, 2±0.8%) with EBs expansion potential were identified. Populations A, B and C generated EBs with day 15 FI of ∼9,500, 113 and 60, respectively which matured (>10% CD235ahigh) by day 8, 7 and <7. The frequency of B and C cells transiently increased (up to 0.3–0.9%) during the first 6 days of HEMAser culture of population A. These results suggest that population A, B and C are linked in a mother-daughter relationship and that transition from A to C involves major losses in expansion potential. Due to this loss of expansion potential, the contribution of CD34neg cells to the total number of EBs generated from AB MNC is negligible. Calculations of the total EB numbers generated from an entire AB unit and from individual purified fractions indicated that AB MNC generated more EBs than the purified fractions in combination (total EBs numbers=2.7×109 vs 7×108). Therefore, for AB, loss of CD34pos cells, rather than of CD34neg precursors, is the main reason for the lower numbers of EBs generated by CD34pos cells than MNC. Population A (1.3±0.9%), B (0.4±0.3%) and C (6.8±4.8%) were also recognized in CB but they generated similar numbers of EBs (day 16 FI ∼600 in all cases). Since in CB, transition from population A-B and C was not associated with restriction in proliferation potential, the greater numbers of EBs generated by MNC than by CD34pos cells are due both to loss of CD34neg (CD36low) precursors and of CD34pos cells during purification.
These studies indicate that, in spite of individual variability in EBs expansion, cell losses during sample processing and media used for expansion have greater impact on the number of EBs expanded ex-vivo than demographic properties (ontogeny, ethnicity and/or gender) of the donor.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.