In Vitro Large Scale Production of Human Mature Red Blood Cells from Hematopoietic Stem Cells by Coculturing with Human Fetal Liver Stromal Cells

In vitro models of human erythropoiesis are useful in studying the mechanisms of erythroid differentiation in normal and pathological conditions. Here we describe an erythroid liquid culture system starting from cord blood derived hematopoietic stem cells (HSCs). HSCs were cultured for more than 50 days in erythroid differentiation conditions and resulted in a more than 109-fold expansion within 50 days under optimal conditions. Homogeneous erythroid cells were characterized by cell morphology, flow cytometry, and hematopoietic colony assays. Furthermore, terminal erythroid maturation was improved by cosculturing with human fetal liver stromal cells. Cocultured erythroid cells underwent multiple maturation events, including decrease in size, increase in glycophorin A expression, and nuclear condensation. This process resulted in extrusion of the pycnotic nuclei in up to 80% of the cells. Importantly, they possessed the capacity to express the adult definitive β-globin chain upon further maturation. We also show that the oxygen equilibrium curves of the cord blood-differentiated red blood cells (RBCs) are comparable to normal RBCs. The large number and purity of erythroid cells and RBCs produced from cord blood make this method useful for fundamental research in erythroid development, and they also provide a basis for future production of available RBCs for transfusion.


Introduction
Erythropoiesis de�nes the process of differentiation and proliferation from hematopoietic stem cells (HSCs) to mature red blood cells (RBCs). In adult humans, erythroid differentiation produces about 2 × 10 11 red cells per day. Erythroid lineage development requires a delicate balance between the opposing effects of proliferation promoting factors and differentiation-inducing factors. e two most important cytokines are erythropoietin (EPO) and stem cell factor (SCF) [1]. EPO protects erythroid progenitor cells from apoptosis by activating antiapoptotic proteins, and it also stimulates hemoglobin (Hb) synthesis and is essential for terminal differentiation [2]. SCF acts synergistically with this lineage-speci�c factor by promoting proliferation of erythroid progenitor cells [1]. A variety of other factors, such as insulin-like growth factor 1 (IGF-1), insulin, and glucocorticoids, have been proposed to have a supportive effect on human RBCs development. In this way, IGF-1 enhances proliferation as well as terminal differentiation, by which it promotes nucleus condensation and enucleation [3]. Glucocorticoids, such as dexamethasone (Dex), seem to have a direct in�uence on the proliferation of erythroid progenitor cells [4]. In vivo, both renewal and maturation of human erythroid progenitors proceed in parallel in the bone marrow. It thus has been difficult to assess the contribution of particular signaling pathways and their deregulation during erythroid cell development.
Despite this knowledge about RBC development, most of the mechanisms underlying erythroid differentiation and maturation both at cellular and molecular levels still remain unknown. In particular, the mechanisms of enucleation, a unique phenomenon of human erythropoiesis, are still unknown. us, the availability of an in vitro erythropoiesis model, which exhibits all stages of RBC development, would be a powerful tool for investigating the factors and molecular mechanisms involved in proliferation and differentiation of human erythroid cells under normal and pathophysiologic conditions. Now, most information was obtained from established cell lines and primary animal cell models of chicken and mouse in de�ned media [2,5].
Most of the commonly available in vitro assays of erythropoiesis are based on cell lines, such as murine MEL or human K562, HEL, and UT-7 cell lines. However, cell lines usually do not recapitulate the entire process of erythropoiesis, as many regulatory pathways have been altered during the transformation process that led to their immortalization. erefore, for a long time, efforts have been made to establish in vitro unilineage differentiation of human primary erythroid cells.
Cord blood has received a great deal of attention as a source for HSCs as an alternative to bone marrow stem cells in transplantation medicine [6,7]. Furthermore, cord blood is highly enriched in committed hematopoietic progenitors, including those of the erythroid lineage [8]. We addressed this problem by in vitro large scale production of erythroid cells from umbilical cord blood derived HSCs.
Decades ago, most in vitro experiments were performed under semisolid culture conditions. is approach had the disadvantage that only immature differentiation stages were generated and the relatively small cell numbers derived in colonies severely limited subsequent investigations. In addition, the options for investigating the in�uence of speci�c growth hormones were restricted. To overcome these limitations, liquid cultures (LCs) have been developed over the past few years. Most LCs, however, showed only moderate proliferation or an absence of terminal differentiation and enucleation [9][10][11]. ese methods are difficult to test the effects of various factors at different maturation stages since it is difficult to add or subtract components to/from the culture. However, the use of liquid cultures is also limited either by the production of mixed erythroid and myeloid cells or by a weak or absent terminal enucleation [12].
Several reports suggest that hematopoietic niche cells may promote the terminal enucleation of erythroid cells in vitro [13,14]. Recently, Isern, et al. report that the fetal liver (FL) provides a previously unrecognized developmental niche for the maturation and enucleation of primitive erythroid cells (EryP). ese results demonstrate that the FL is a niche for maturation of primitive erythroid cells [15]. And the FL is also a ma�or site for the development of de�nitive erythroid cells (EryD), which matures within erythroblastic islands (EBIs) [16]. EBIs, �rst identi�ed in bone marrow and later in FL and spleen, are morphologically distinct 3D structures comprising a central macrophage surrounded by EryD at various stages of maturation [17]. Studies have revealed macrophage extensions that surround peripheral erythroblasts, providing intimate membrane contact between these cells. e central macrophages of the EBIs are thought to function as nurse cells during erythropoiesis [17][18][19]. erefore, we hypothesized that unknown growth proteins are produced as yet unidenti�ed populations of fetal liver stromal cells that stimulate the maturation of erythroid progenitor cells.
Here, we thoroughly describe optimized long-term serum-free culture conditions that allowed expansion into mass cultures of highly homogeneous human erythroid progenitors derived from umbilical cord blood. Cells could be routinely expanded for more than 50 days, undergoing up to 35 population doublings. e mature red blood cells could be produced by coculturing erythroid progenitors with human fetal liver stromal cells (hFLSCs). With this procedure, it is possible to reproduce different steps of human normal erythropoiesis. e large number and purity of erythroid cells produced from a small amount of peripheral blood make this method useful for studying either normal or pathological erythropoiesis. In addition, production of mature red blood cells from cord blood derived HSCs is of great potential and of importance in practice for it could eventually become an alternate source of cell for transfusion.

In Vitro Culture and Characterization of Erythroid
Cells from Cord Blood HSCs. Aer informed consent had been obtained, cord blood from healthy volunteers was collected into sterile heparinized tubes. Mononuclear cells were obtained by centrifugation on Lymphoprep density gradient. e mononuclear cells were then enriched for HSCs by positive selection using anti-CD34-tagged magnetic beads using Mini-MACS columns (Miltenyi Biotech, Auburn, CA) according to the manufacturer's protocol. Purity was more than 95%; as determined by �ow cytometry. CD34 positive cells were cultured as described previously [20,21]. Brie�y, for initial expansion, 1 × 10 6 cells/mL were cultivated in serum-free medium (StemSpan; Stem Cell Technologies, Vancouver, BC, Canada) supplemented with Epo (2 U/mL; Sigma-Aldrich, St. Louis, MO), the synthetic glucocorticoid Dex (1 M; Sigma), IGF-1 (40 ng/mL; R&D Systems, Minneapolis, MN), SCF (100 ng/mL; R&D Systems), and ironsaturated human transferrin (400 mg/mL; Sigma-Aldrich).
Homogeneous cultures of erythroid progenitors established aer several days were kept in the same medium at 2 × 10 6 cells/mL by daily partial medium changes. As the cultured erythroblasts being over 5 × 10 7 (about 20 wells of 24-well plate), we puri�ed the erythroblasts by density centrifugation (Ficoll, 1.077 g/mL). And then only about 2 × 10 6 cells (about 20 wells of 24-well plate) were continued to be cultured. And the cells lost by density centrifugation were corrected when calculating the proliferation of erythroid cells. Proliferation kinetics of cells were daily monitored using an electronic cell counter (Vi-cell XR, Beckman Coulter, Germany); cumulative cell numbers were calculated as previously described [22]. Cumulative cell numbers were calculated and plotted against time.
Cell morphology was analyzed by light microscopy on cytocentrifuged smears stained with Wright-Giemsa. Hematopoietic colony assays were performed on Day 21. 5000 cells were plated in 35 mm plastic culture dishes containing methylcellulose-based medium (Methocult T GF H4435, StemCell Technologies). Cells were incubated for 14 days at 37 ∘ C and 5% CO 2 . Colonies were enumerated by microscopy. e surface marker expression of the cultured erythroid progenitor cells were analyzed by �ow cytometry. Cells was harvested on days 7, 14, and 21. All of the conjugated antibodies and the corresponding isotype controls were purchased from BD Pharmingen, Heidelberg, Germany. e antibodies against human cluster of differentiation (CD) molecules we used were CD3-allophycocyanin (APC), CD4-�uorescein isocyanate (FITC), CD8-Phycoerythrin (PE), CD13-FITC, CD14-APC, CD19-PE, CD33-FITC, CD34-FITC, CD38-PE, CD41-FITC, CD45-APC, CD71-FITC, CD117-PE, and glycophorin A (GPA)-PE. Erythroid cells were collected and washed twice in phosphate-buffered saline (PBS) with 0.1% BSA and stained in accordance with the manufacturer's suggested concentration of conjugated antibody for 30 min at 4 ∘ C. e stained cells were then washed 2× in PBS + 0.1% BSA and �xed with the wash buffer supplemented with 1% paraformaldehyde. e samples were then analyzed using a �ow cytometer (Becton Dickinson, Franklin Lakes, NJ). Cell populations were analyzed with the CellQuest program (Becton Dickinson).

Isolation, Culture, and Characterization of hFLSCs.
hFLSCs were obtained from 24-week human fetal liver tissues as described previously [23]. Human tissue collection for research purpose was approved by Research Ethics Committee of Beijing Institute of Transfusion Medicine. Pregnant women gave written consent for clinical procedure and research use of the embryonic tissues in accordance with the Declaration of Helsinki. hFLSCs were cultured in a medium comprising 45% Dulbecco's modi�ed Eagle's medium (DMEM, Sigma-Aldrich), 45% DMEM/F12 (Sigma-Aldrich), and supplemented with 10% FCS (fetal calf serum; Gibco, Grand Island, N�). Brie�y, hFLSCs were prepared using a ceiling culture method. Human fetal liver tissues were minced into 1 mm 3 pieces and transferred onto human gelatin-coated 25-cm 2 �asks. 4 mL of hFLSCs culture medium was added onto the ceiling surface, and aer 5 h the �asks were turned back over. e tissue pieces were incubated undisturbed for 7 days at 37 ∘ C, allowing the hFLSCs to migrate and adhere to the gelatin-coated surface. Once the primary hFLSCs reached con�uency, the hFLSCs were digested with 0.25% Trypsin (Sigma-Aldrich) for 5 min at 37 ∘ C. Aer adding hFLSCs medium, cultures were pipetted into single-cell suspensions, centrifuged, resuspended, and replated onto new �asks.
e surface marker expression of hFLSCs was analyzed by �ow cytometry. Stained cells were analysed on a FACS Calibur using Cell Quest soware (Becton Dickinson). All of the conjugated antibodies and the corresponding isotype controls were purchased from BD Pharmingen, Heidelberg, Germany. e antibodies we used were CD11b-PE, CD29-PE, CD34-FITC, CD44-FITC, CD45-FITC, CD90-FITC, CD105-PE, and CD144-PE.

Functional Analysis of Hemoglobin.
Cocultured erythroid cells collected at 8 days were used to characterize the function of hemoglobin. Oxygen equilibrium curves were determined using a Hemox-Analyzer, Model B. e gas phase gradients were obtained using nitrogen and room air, and the curves were run in both directions.

In Vitro Large Scale Production of Human Erythroid
Progenitors. e purity of the isolated CD34 + HSCs was always greater than 95% (Figure 1(a)). CD34 + cells derived from cord blood were induced to erythroid progenitors in serum-free medium (StemSpan) by using a combination of Epo, SCF, Dex, and IGF-1. From day 7 on, morphologically homogeneous erythroid progenitors could be expanded into mass cultures in a rapid time (Figures 1(b)-1(c)), which proliferated exponentially for over 50 days, demonstrating a clear capacity for long-term self-renewal (Figure 1(d)). At regular intervals, more mature or apoptotic cells were removed by density gradient centrifugation. Aer 50 days, the majority of progenitors was undergoing gradual proliferation arrest before eventually undergoing apoptosis. During the entire expansion, an apparently overall increase in cell number of 10 8 -fold to more than 10 9 -fold was obtained ( ). Under optical conditions, the erythroid cells resulted in a more than 10 9 -fold expansion within 50 days (Figure 1(d)).

Characterization of Cord Blood-Derived Erythroid
Progenitors. Morphologically, the erythroid progenitors obtained using the above protocol were nucleated and substantially larger than de�nitive erythocytes with an average diameter of approximately 10 m. Giemsa-Wright staining showed the morphology of the cultured erythroid progenitor cells (Figures 2(a)-2(c)). In order to characterize the morphological changes of the cultured cells through the stages of erythropoiesis, cell morphology was assessed by microscopic examination, at day 14, greater than 95% of the cells in culture were erythroid cells, and less than 5% of cells morphologically appeared to be promyelocytes and monocytes.

Flow Cytometric Analysis of Differentiation into Erythroid
Progenitor Cells. Differentiation over time of CD34 + cells was documented by FACS analysis of cell surface markers ( Figure 3 and Table 1). At day 21, 95% were CD71 positive, 90% were CD117 positive, and 15% of the cells expressed GPA, whereas the majority of the cells did not express myelomonocytic or megakaryocytic or lymphatic antigens (1% of cells expressed CD14; 2.4% of cells expressed CD41; 3% of cells expressed CD4; 2.5% of cells expressed CD8) and progenitor antigens (0.7% cells were positive for CD34; 0.5% cells expressed CD38) ( Table 1). CD34 + cells rapidly declined, to be quite absent on day 14 ( Figure 3). is is expected since CD34 is a marker of early progenitor cells and, in the erythroid lineage, is not expressed aer the BFU-E stages. In contrast, GPA, which is a marker of more mature erythroid cells, �rst appears at the basophilic erythroblast stage and its expression increases throughout the rest of erythroid differentiation [16]. As expected, GPA expression increased almost linearly by the culture time.

Isolation and Characterization of hFLSCs.
Because fetal liver is an important hematopoietic organ for erythroid development, we hypothesized that hFLSCs might produce protein(s) that support in vitro terminal maturation of erythroid progenitors. We isolated and cultured hFLSCs from 24-week human fetal liver. Aer one week culture, the stromal cells migrated out the border of the tissue clumps (Figures 4(a) and 4(b)), and it took two weeks for the cells to become con�uent. Aer the primary and three subsequent passages, hFLSCs were adherent to be fusiform shape and highly uniform in morphology (Figure 4(a)). As hFLSCs are propagated in culture for more than 30 passages, there are no obviously cell morphology changes. Flow cytometry analysis was used to characterize cell surface markers of hFLSCs, showing that the hFLSCs were positive for some stromal progenitor markers, such as CD29, CD44, CD71, CD90, and CD105, but negative for CD34 and CD45 (Figure 4(c)). hFLSCs maintained normal karyotypes aer 20 passages (Figures 4(d) and 4(e)).

Enucleation and Maturation of Cord Blood-Derived Erythroid Cells In Vitro. A critical scienti�c and clinical issue
is over whether cultured erythroid cells can be matured in vitro to generate enucleated erythrocytes. We tried to induce erythroid progenitors to terminal maturation by coculturing with hFLSCs. Erythroid progenitor cells cultured in this condition with hFLSCs to induce terminal maturation. And the cells expanded one or more times when cocultured with hFLSCs (Figures 5(a) and 5(b)). Approximately 60-80% of erythroid cells were enucleated when these cells were transferred to hFLSCs layers from nonstromal cultures and cocultured from days 8 (Figures 5(c) and 5(d)). Erythroid progenitor cells kept in liquid conditions (without transfer to hFLSCs) could enucleate 10%-30% suggest that enucleation could not be achieved completely in feeder-free condition.
To further investigate the events associated with enucleation, we examined cell surface marker expression related to the process of erythrocyte maturation. We observed a progressive decrease of CD117, an early erythroblast marker, on day 8 and decreased their expression over time. And they showed low to negligible level of GPA protein, a mature erythrocyte marker, at the beginning, but increased their expression dramatically with their maturation (Figure 5(e)).

Gene Expression during Erythropoiesis.
Erythroid progenitors cultured at day 10 and 15 and enucleated cells at day 4 and 8 were analyzed for erythroid gene expression ( , , and globins, GATA1, PU-1, and SCL/TAL-1) (Figures 6(a)  and 6(b)). e expression level of transcription factor GATA1 mRNA was upregulated during the early stages of erythroid differentiation and then downregulated when cocultured with hFLSCs. At an early stage of differentiation, andglobin mRNA was hardly expressed; during maturation the expression level of -globin mRNA increased. At cocultured day 8, the enucleated cells only expressed the -globin.

Oxygen Equilibrium
Analysis. e oxygen equilibrium curves of the cord blood-derived erythroid cells (cultured with hFLSCs for 8 days) were either very similar to or somewhat rightward shied (Figure 6(c)), relative to those of human cord blood and normal adult RBCs.

Discussion
e expansion of hematopoietic progenitors from umbilical cord blood ex vivo is of great interest, both for basic research and the exploitation of clinical potential. is work described a method which enabled the puri�cation and in vitro expansion of large numbers of erythroid progenitors from cord blood-derived HSCs. is system could be used to reproducibly generate large numbers of erythroid cells under serum-free conditions. In the presence of Epo, SCF, IGF-1, and Dex added aer 2 week of culture, almost all the cells were erythroid progenitor cells positive to erythroid speci�c cell surface markers. Aer long time culture, the culture conditions resulted in a great cell expansion and accumulation of erythroid cells. So, Epo, SCF, IGF-1, and Dex represented the most potent factors required for inducing proliferation of erythroid progenitor cell [1,20,21,25]. e complete erythroid differentiation and maturation were documented by the morphological analysis and by changes in surface markers expression. More than 90% of the cells expressed GPA aer 8 days of being cocultured with hFLSCs. CD71 is highly expressed on both proliferating cells of various cell types and early erythroid cells with high iron uptake for Hb synthesis [26,27]. Aer day 14 in the erythroid progenitor cells expansion conditions, CD71 expression is up to 97%. And on the last culture days with hFLSCs almost all cells expressed GPA. Furthermore, most GPA + cells became CD71 − , which is known as a sign of cell cycle arrest and terminal maturation to erythrocytes [28].
Previously used procedures failed to yield long-term proliferating but immature erythroid progenitors, due to massive spontaneous differentiation [12,29]. Using serum-free media together with Epo, SCF, IGF-1, and Dex we could expand human erythroid progenitors from cord blood derived HSCs for more than 50 days. is yielded mass populations with an in vitro lifespan of approximately 30-35 generations, almost absent from spontaneous differentiation largely facilitating future biochemical and molecular characterization. Importantly, optimization of differentiation conditions allowed induction of synchronous terminal maturation, resulting in enucleated cells virtually indistinguishable from erythrocytes of peripheral blood.
In our hands, the serum-free medium used here was critical for generating long-term cultures. Other serum-free media proved to be less efficient. Compared with other current procedures, it should also be noted that our methods did not require negative selection (depletion) against lymphoid or granulocytic cells [30]. Routinely, a (calculated) 10 8 -10 9fold erythroid cells expansion was obtained, in some cases even a (calculated) 10 10 -fold increase was obtained (about 35 generations). us, the procedure described here enables the production of large numbers of homogeneous primary erythroid cells for detailed molecular characterization of mechanisms.
In vivo, stromal cells form a complex microenvironment for HSCs that controls their multiple fates, including apoptosis, migration, and the cell divisions that lead to formation of lineage-committed progenitors [31,32]. Different types of stromal and other cell types can have either positive or negative effects on fates of HSCs. In previous studies, terminal differentiation and enucleation of erythrocytes in vitro were reported to be dependent on macrophages presence in cultures [33,34]. Giarratana et al. [13,35] recently described a protocol to achieve in vitro terminal erythroid maturation (90%-100% of enucleated erythrocytes) using the murine stromal cell lines MS5 or human stromal cell lines to reproduce ex vivo the microenvironment existing in vivo. In our model, terminal erythroid maturation was satisfactorily achieved in coculturing with hFLSCs, making the culture system more safe.
However, the underlining mechanism of the hFLSCs during erythroid differentiation was not fully explained in the present study. ere are some other studies showed that the mouse and human FLSCs can promote the in vitro generation of erythroid cells from different sources of stem cells [36][37][38]. All these studies suggested that the fetal liver stromal cells might mimic the hematopoietic supportive environment of the fetal liver in vitro through many different pathways [39].
Erythropoiesis at molecular level is driven by a combination of transacting factors that act in concert to direct the genes expression for erythroid-speci�c proteins [40,41]. We established that culture conditions of cord blood-derived HSCs allow to reproduce normal erythroid differentiation con�rmed also at the molecular level. Genes required for primitive erythropoiesis were expressed at high levels in early days of cultures. Globin gene expression in the maturing erythrocytes in vitro recapitulated expression patterns which occur in adult hemoglobin expression well [42,43]. A more recent study has used the in vitro culture model to study the molecular mechanisms of erythropoiesis. e results show that KLF1 is a critical activator of the BCL11A gene, encoding a suppressor of HbF expression [44]. And, little has been known so far about enucleation mechanism of erythroblasts. Our methods also provided a new model to investigate molecular and cellular mechanisms involved in the enucleation of human RBCs.
We demonstrated here that erythroid cells derived from cord blood HSCs possessed oxygen equilibrium curves comparable to normal cord blood and adult RBCs. is �nding demonstrated that the cord blood-derived RBCs have oxygen carrying properties that are comparable to those of normal adult erythrocytes. Another critical issue for clinical utilization of cord blood-derived RBCs is about whether they can be enucleated in vitro. hFLSCs coculture improved efficiency of enucleation, but even without these cells enucleation could not be achieved for a completely feeder-free system. We showed that 60%-80% of the RBCs underwent multiple differentiation events, including a progressive decrease in size and increase in GPA expression (a mature RBC marker), which is similar to normal RBCs. However, further study will be necessary to investigate cord blood-derived erythroid cells in vivo function.
Taken together, culture method described here could facilitate extensive biochemical analyses of erythropoiesis, which were technically demanding so far due to lack of material. Our system also provided a useful basis for future production of available RBCs for transfusion. blood CD36 + erythroid progenitors: temporal changes in gene and protein expression, " Experimental Hematology, vol. 31, no. 9, pp. 760-769, 2003.