• erythroid progenitors;
  • CD36;
  • ex vivo expansion ;
  • cell purification;
  • erythropoiesis


  1. Top of page
  2. Abstract
  6. Acknowledgements
  7. References

In humans, studies of the erythroid cell lineage are hampered by difficulties in obtaining sufficient numbers of erythroid progenitors. In fact, these progenitors in bone marrow or peripheral blood are scarce and no specific antibodies are available. We describe a new method which allows proliferation in liquid culture of large numbers of pure normal human erythroid progenitors. CD34+ cells were cultured for 7 d in serum-free conditions with the cytokine mixture interleukin (IL)-3/IL-6/stem cell factor (SCF). This resulted in cell expansion and the appearance of a high proportion of CD36+ cells which were purified on day 7. Methylcellulose clones from these cells were composed of 96.6% late BFU-E and 3.4% CFU-GM. These CD36+ cells could be recultured with the same cytokine mixture plus or minus erythropoietin (Epo) for a further 2–7 d. In both conditions further amplification of CD36+ cells was observed, but Epo induced a more dramatic cell expansion. Glycophorin-positive mature cells appeared only in the presence of Epo, and terminal red cell differentiation was observed after 7 d of secondary culture. Cells obtained from adult CD34+ progenitors mostly contained adult haemoglobin, whereas cord blood-derived cells contained equal proportions of adult and fetal haemoglobin. Activation of STAT5 and tyrosine phosphorylation of the Epo receptor and JAK2 were observed after Epo stimulation of these cells. This new method represents a straightforward alternative to the procedures previously described for the purification of normal erythroid progenitors and is useful in the study of erythropoietic regulation.

The permanent turnover of cells derived from the haemopoietic system is the crucial phenomenon which allows the renewal of blood cells. In adult humans the mean daily production of erythrocytes is estimated to be 2 × 1011 ( Erslev, 1983). Active erythropoiesis occurs after erythroid commitment of undifferentiated pluripotent stem cells which provide cells able to terminally differentiate to mature red blood cells. Erythropoietin (Epo) is the specific cytokine physiologically required for red cell production. However, the commitment of stem cells to both early and late erythroid progenitors (BFU-Es and CFU-Es) appears to be a stochastic process independent of the presence of Epo ( Ogawa, 1993; Papayannopoulou et al, 1993 ; Wu et al, 1995b ; Socolovsky et al, 1997 ), and red cell terminal differentiation itself can be induced in vitro independently of the presence of Epo ( Sui et al, 1996 ; Goldsmith et al, 1998 ). A variety of cytokines have been reported to stimulate primitive erythropoiesis, including stem cell factor (SCF) ( Dai et al, 1991 ; Papayannopoulou et al, 1993 ; Muta et al, 1995 ; Wu et al, 1995a , 1997; Broudy, 1997) and interleukin-6 (IL-6) ( Sui et al, 1996 ). Other synergistically acting cytokines such as IL-1 and IL-3 also seem to play a role in the generation and/or amplification of erythroid progenitors ( Brugger et al, 1993 ; Papayannopoulou et al, 1993 ).

The study of the molecular mechanisms leading to erythroid cell proliferation and differentiation in normal individuals is hampered by difficulties in obtaining sufficient numbers of purified erythroid progenitors. Indeed, biochemical and molecular studies require high numbers of cells whereas erythroid progenitors in bone marrow and in peripheral blood are scarce. Furthermore, no specific surface markers are currently available to enable easy immune selection of immature red cell progenitors.

In vitro cell culture methods using various cytokine mixtures have been described ( De Wolf et al, 1994 ; Mrug et al, 1997 ; Cippolleschi et al, 1997 ; Malik et al, 1998 ; Oda et al, 1998 ). These mixtures all included Epo, and resulted in the generation of erythroid cell populations of rather mature stages.

Other purification techniques requiring complex negative selections and serial depletions of non-erythroid cells ( Fibach et al, 1989 ; Sawada et al, 1989 ; Giampaolo et al, 1994 ; Muta et al, 1995 ; Miller et al, 1999 ; Oda et al, 1998 ; Dai et al, 1998 ) have been described for obtaining cell populations enriched in erythroid progenitors. Finally, fluorescent cell sorting using combinations of antibodies have defined lineage-specific subpopulations of cells among which CD34+/CD45RA/CD71high were reported to contain enriched erythroid progenitor cells ( Sauvageau et al, 1994 ; Lessard et al, 1998 ). However, these procedures are very expensive and time-consuming and cannot lead to purification of high numbers of erythroid progenitors.

We have developed a new method, based on a three-step purification and in vitro amplification technique, which produces high enrichment and proliferation in liquid culture of large numbers of normal human erythroid progenitors. Serum-free cultures of CD34+ cells obtained from cord blood or adult peripheral blood and grown with the cytokine combination: SCF, IL-6 and IL-3 resulted in the dramatic expansion of cultured cells and the appearance of high proportions of CD36+ cells. CD36 is a multifunctional glycoprotein receptor present on platelets, monocytes, mast cells, erythroid precursors, and some endothelial cells ( Van Schravendijk et al, 1992 ). It is involved in the adhesion of platelets, red blood cells infected by Plasmodium falciparum, and monocytes ( Kieffer et al, 1989 ; Oquendo et al, 1989 ; Huh et al, 1995 ). It is considered to be a receptor for thrombospondin, collagen, low-density lipoproteins, and phosphatidylserine ( Kieffer et al, 1989 ; Oquendo et al, 1989 ; Rigotti et al, 1995 ; Huh et al, 1996 ). In the present work our interest in the CD36 antigen was justified by the timing of its appearance during haemopoietic differentiation: indeed, CD36 is detected early on erythroid progenitors whereas it is a late marker of megakaryocytic and monocytic cells ( Edelman et al, 1986 ; Okumura et al, 1992 ; De Wolf et al, 1994 ; Nakahata & Okumura, 1994). CD36+ cells obtained after 7 d in our liquid culture conditions were isolated, and comprised a pure population of late BFU-Es and CFU-Es.

We demonstrated by biochemical and immunological analysis that these erythroid progenitors responded to Epo stimulation and behaved normally in semi-solid culture conditions. Furthermore, terminal red cell differentiation could be obtained, and the analysis of cell haemoglobin content revealed that erythroid cells derived from adult CD34+ cells mostly contained adult haemoglobin, whereas cord blood-derived cells exhibited equal percentages of both adult and fetal haemoglobin.


  1. Top of page
  2. Abstract
  6. Acknowledgements
  7. References

Cell retrieval

Umbilical cord blood (CB) units from normal full-term deliveries were obtained, after informed consent of the mothers, from the Obstetrics Unit of Hôpital Saint-Vincent de Paul, Paris, and collected in placental blood collection bags (Maco Pharma, Tourcoing, France) by the Centre d'Hémobiologie Périnatal. The mean CB total volume collected was 85 ml. Samples of peripheral blood mononuclear cells (PBMC) from cytapheresis collection of patients suffering from lymphoma or multiple myeloma were also obtained, after informed consent of the patients. After ablative chemotherapy, the stem cells were mobilized by G-CSF treatment (5μg/kg/d) from day-7 post-chemotherapy for 10–12 d then collected by cytapheresis with a COBE-Spectra machine (COBE, Denver, Col., U.S.A.).

Cell separation and culture conditions

CB units were diluted with 50 ml phosphate-buffered saline (PBS, Gibco-BRL, Life Technologies, Cergy-Pontoise, France) containing 0.8% bovine serum albumin (BSA, Stem Cell Technologies, TEBU, Le Perray-en-Yvelines, France). Diluted CB units as well as PBMC samples were submitted to Ficoll density gradient separation (Lymphocytes separation medium, Eurobio, Les Ulis, France). Low-density cells were recovered and enriched for CD34+ cells by two cycles of positive selection using anti-CD34 antibody (QBEND/10) and magnetic cell sorting on Midi-MACS then Mini-MACS columns (CD34 isolation kit, Miltenyi Biotech, TEBU). CD34+ enriched cells were numbered and viability was determined by trypan blue exclusion. Cells were then suspended in 6 ml serum-free IMDM (Gibco-BRL) in 25 cm2 flasks (ATGC, Noisy-le-Grand, France), at a final concentration of 0.5–1 × 105/ml, in the presence of 15% BIT 9500 (mixture of 5% BSA + 50 μg/ml bovine pancreatic Insulin + 1 mg/ml human Transferrin) (Stem Cell Technologies, TEBU), 100 U/ml penicillin-streptomycin, and 2 m ML-glutamine. A cytokine mixture containing 10 ng/ml recombinant human (rh) IL-3, 10 ng/ml rhIL-6 and 25 ng/ml rhSCF was immediately added. IL-3 and IL-6 were gifts from Dr Shimosaka (Kirin Brewery Co., Tokyo, Japan) and SCF from Amgen (Thousand Oaks, Calif., U.S.A.). Cells were incubated in 5% CO2 in air, at 37°C for 7 d. In order to ensure good cell proliferation, an equal volume of fresh medium containing the three cytokines was added on days 3 and 5 of culture.

On day 7, cells were pelleted by centrifugation, numbered and resuspended in PBS with 0.8% BSA. Monoclonal CD36 IgG1 antibody (clone FA6-152) (Immunotech, Marseille, France) was added to the cell pellet at a final concentration of 1 μg/106 cells and incubated for 20 min at 4°C. Cells were washed in PBS-BSA, then incubated 15 min at 4°C with rat anti-mouse IgG1 antibody coupled to magnetic microbeads (Miltenyi Biotech). Immunomagnetic separation was performed again on either Midi- or Mini-MACS columns, depending on the numbers of cells treated. Samples of cells eluted from the columns were kept for further immunophenotypic analysis and cloning assays. These cells could be cultured again for 1–8 d. The culture medium was the same as described above (IMDM containing 15% BIT 9500 + IL-3 + IL-6 + SCF)±rhEpo (2 U/ml) (gift from Dr M. Brandt, Boehringer-Mannheim, Germany).

Cytological analysis of cells

In order to determine morphological evolution of cells, cytospins were performed throughout the cultures (Shandon Inc., Pittsburg, Pa., U.S.A.) and coloured by May-Grünwald-Giemsa. The presence of haemoglobin in the cells was determined by benzidine staining.

Semisolid culture assays

Samples of (i) CB cells or PBMCs after Ficoll separation (25 000/ml), (ii) cells eluted from CD34 separation columns (500/ml), (iii) cells after 7 d culture (500/ml), (iv) cells eluted from CD36 separation columns (500/ml), and (v) cells cultured again after CD36 separation (1000/ml) were plated in semisolid medium containing methylcellulose, 20% fetal calf serum (FCS) and a cocktail of cytokines comprising Epo, IL-3, IL-6, SCF, G-CSF, GM-CSF and IL-11 which allows growth of all types of erythroid, myeloid and megakaryocytic progenitors (StemGem 1d, StemBio Research, CNRS, Villejuif, France). Cultures were performed in duplicate, and incubated in a fully humidified atmosphere with 5% CO2 in air for 9–12 d for late BFU-E and CFU-E and 14–18 d for early BFU-E and CFU-GM. Colonies were counted under an inverted microscope.

Immunophenotyping assays

105 cells were washed in PBS-BSA, and pellets were incubated with fluorescein isothiocyanate (FITC)- or phycoerythrin (PE)-conjugated antibodies for 20 min at 4°C. For double staining two antibodies conjugated to different fluorochromes were combined. Cells were washed to remove excess antibodies and resuspended in PBS-BSA for flow cytometry analysis. Antibodies to CD34, CD38, CD36, CD41, CD14, CD71 and Glycophorin A (GpA) (Immunotech, Marseille, France) were used. To assess the specificity of antibody binding, IgG1 isotype controls were used. Antibody-bound cells were counted on a Coulter EPICS XL (Coultronics-France, Margency, France).

Biochemical analysis of CD36+ cells

(i) Cell preparation. After CD36 purification, cells were secondarily cultured for 48–72 h in the presence or absence of Epo in order to obtain sufficient numbers of cells for biochemical studies. To analyse the phosphorylation pattern of the Epo receptor (EpoR) as well as JAK2 and STAT5 activation in response to Epo stimulation, overnight cytokine deprivation was performed: cells were washed twice and resuspended at a concentration of 106/ml in cytokine-free culture medium. On account of the numbers of cells available in each culture condition, Epo (10 U/ml) was then added to 5 × 106 cells (when secondarily cultured without Epo) or to 20 × 106 cells (when secondarily cultured with 2 U/ml Epo) for 10 min. As a control, we used other sets of 5 × 106 or 20 × 106 cells incubated in the same conditions, but without Epo stimulation. Cells were solubilized in 1% Nonidet P-40.

(ii) Antibodies. Immunoprecipitations (IP) were performed with: an antibody prepared in our laboratory and directed against the human EpoR intracellular domain, as previously described ( Verdier et al, 1997 ); an anti-JAK2 antibody, purchased from UBI (Upstate Biotechnology); a mixture of anti-STAT5A and anti-STAT5B antibodies produced in our laboratory. ECL (Amersham Ltd, Les Ullis, France) was used for Western blot visualization.

(iii) Bandshift experiments. STAT5 bandshift experiments were performed as previously described ( Chrétien et al, 1996 ). Briefly, the STAT5 binding site of the bovine β-casein promoter was used as a probe. 2 μl of nuclear extracts were mixed with 20 μl of binding buffer containing 60 000 cpm end-labelled probe. The mixture was incubated for 30 min at 4°C and complexes were separated on 6% non-denaturing polyacrylamide gels, then analysed by autoradiography.

Haemoglobin analysis

On day 7 of secondary culture, cells were suspended in distilled water and disrupted by sonication. DNA was precipitated by adding spermine to a 2 m M final concentration ( Davis et al, 1986 ). Samples were incubated on ice for 15 min and spun in a microcentrifuge at 14 000 rpm for 10 min. Supernatants were collected and haemoglobin composition was determined. Isoelectric focusing (IEF) analysis on agarose gels was performed using ampholines (pH 6–8 and pH 7–9) (Amersham Pharmacia Biotech, Uppsala, Sweden) ( Arad et al, 1981 ) and haemoglobin fractions were revealed by the specific O-dianisidine staining (Helena Laboratories, Beaumont, Texas, U.S.A.). Haemoglobin fractions were also resolved and quantified by ion exchange-high performance liquid chromatography (IE-HPLC) on a Variant system (BioRad Laboratories, California, U.S.A.) using the A1c program provided by the manufacturer. Globin chain composition was determined by reverse phase–high performance liquid chromatography (RPH-HPLC) on a LC10 chromatograph (Shimadzu, Tokyo, Japan) using a Vydac C4 column (The Sep/a/ra/tions Group, Hesperia, California, U.S.A.) according to the method of Shelton et al (1984 ).


  1. Top of page
  2. Abstract
  6. Acknowledgements
  7. References

In vitro amplification of CD34+-derived cells

Stem cell immunomagnetic separation from CB or PBMC led to an average recovery of 86% CD34+ cells. These cells were cultured for 7 d in serum-free conditions with a cytokine mixture composed of IL-3 + IL-6 + SCF, as described in Materials and Methods. At that time a 34-fold mean cell amplification was observed. It must be noted that a proportion of 11.6% CD34+ cells was observed at day 7 of culture, corresponding to a 4–5-fold amplification of these cells in absolute number. Among amplified cells, 46% of CB-derived cells and 65% of PBMC-derived cells were CD36 positive. As presented in Table IA, relative proportions of clonogenic BFU-E and CFU-GM on day 7 of CD34+ cell culture were 53/47 and 60/40 for CB-derived progenitors and PBMC-derived progenitors, respectively.

Table 1. Table I. (A) Composition in progenitors as observed in semi-solid cultures (performed at day-7 liquid culture of CD34+ cells). (B) Phenotypic characteristics of cells eluted from CD36 column for both CB- and adult PBMC-derived cells. Results are expressed as means ±SD of 15 CB and six PBMC experiments. CB, cord blood cells; PBMC, adult peripheral blood mononuclear cells.Thumbnail image of

Separation of CD36+ cells

CD36+ cells were purified by an immunomagnetic procedure from the liquid cultures of CD34+ cells (see Materials and Methods). The optimal time-course of the separation procedure was first determined: in several experiments we separated CD36+ cells on either day 4 or day 10 of CD34+ liquid cultures. In both cases enrichment in erythroid progenitors was poor (data not shown). Separation on days 6 and 7 displayed the best results. The mean total cell recovery after CD36 column elution was 28.7%. For both CB- and PBMC-derived cells, analysis of surface markers of the cells eluted from CD36 columns showed that these cells were 92.5% CD36+, 96% CD71high and were negative for the lineage-specific differentiation markers CD41 (megakaryocytes/platelets) and CD14 (monocytes) ( Figs 1A–1C and Table IB). Therefore, considering the distribution of the CD36 antigen through the haemopoietic cell populations ( Van Schravendijk et al, 1992 ), the cells eluted from CD36 columns in our experiments most likely corresponded to progenitor cells of the erythroid lineage. This was confirmed by the fact that, in semisolid culture conditions, 96.6% of clonogenic cells were small day-10 BFU-Es and CFU-Es, whereas only 3.4% of the cells were CFU-GMs ( 1 Table IA). Moreover, these cells were negative for GpA which is a marker of mature red cell precursors ( 1 Table IB).


Figure 1. -PE control.

Download figure to PowerPoint

On the whole, CD36+ cell recovery from separation columns reached about 10 times the input of CD34+ cells in cultures at day 0.

A total of 15 and six independent separation and amplification experiments have been performed with cord blood and PBMC samples, respectively. Results obtained with both the CB- and PBMC-derived cells were very similar. The main difference resided in the percentage of clonogenic cells throughout the cultures ( 2 Table II): CB-derived cells always displayed higher clonogenic potential than adult PBMC-derived cells. It must be pointed out that most of these clones were erythroid after CD36 cell separation.

Table 2. Table II. Cloning efficiencies of CB- and PBMC-derived cells at various times of the cultures. CB, cord blood cells; PBMC, adult peripheral blood mononuclear cells.Thumbnail image of

Therefore our method leads to the isolation of populations of cells which only contain progenitors of the erythroid cell lineage.

Secondary amplification of CD36+ erythroid progenitors

In order to obtain larger numbers of purified erythroid progenitors, CD36+ cells eluted from the separation columns were cultured again in the same conditions as before separation: cells were therefore cultured in serum-free IMDM with the cytokine cocktail IL-3, IL-6, SCF and with or without Epo (2 U/ml). In the presence of Epo, dramatic cell proliferation took place and the culture doubling time was about 18–20 h. Total cell amplification of CB-derived cells after primary culture and 3 d of secondary culture is presented in 3 Table IIIA. This considerable cell expansion could be observed for up to 7 d of secondary culture. After that time most cells were differentiated or apoptotic. Due to the high variability in the number of cells obtained from cytapheresis samples significant averages for PBMC amplification cannot be presented. However, cell amplification rates were quite similar for both CB- and PBMC-derived cultures.

Table 3. Table III. (A) Recovery and evolution of total number of CB-derived cells after primary cultures (day 7), and after 3 d of secondary cultures (day 7+3) without or with Epo. The evolution of total fold amplification of cells is also indicated. (B) Phenotype of day 7+3 cultured cells in the absence or presence of Epo. * Mean cell recovery after CD36 column: 28.8 ± 8.2%.Thumbnail image of

Phenotypically, cells eluted from CD36 columns were 95–98% CD36+ and CD71high but only 3% were GpA+ (Fig 1D). The GpA marker progressively appeared after 48–72 h of secondary culture with Epo (Fig 1E and Table IIIB) (IgG1 isotype controls are presented in Figs 1F–1H). In the absence of Epo, cell cultures also proliferated, but the doubling time was much slower (about 36 h), leading to a mean 3–4-fold cell amplification after 3 d of secondary culture compared to 15-fold in the presence of Epo ( Table IIIA). Furthermore, no GpA marker appeared in the absence of Epo and only small CFU-Es remained, confirming that cells were blocked at a late CFU-E stage in these Epo-free and serum-free culture conditions. Total numbers of cells available after 3 d of secondary cultures corresponded to a 32-fold and 152-fold increase in the CD34+ cell input on day 0, when secondary cultures were performed in the absence or presence of Epo, respectively ( 3 Table IIIA).

Morphologically, the majority of cells were composed of immature blasts until day 3 of secondary culture, in the presence or absence of Epo (Fig 2A). From day 4 with Epo, most cells were identifiable as erythroblasts ( Figs 2B and 2C) and haemoglobinized red cells could be observed after day 4. From day 6, 100% of cells were erythroblasts and terminal red cell differentiation was observed with a small contingent of enucleated cells (Fig 2D). Levels of cell maturation were rather homogenous during secondary cultures, although some variations could appear at late stages of culture (Fig 2C). Clear haemoglobinization was evidenced after day 4 by benzidine staining (data not shown).


Figure 2. Fig 2. Aspects of CD36+ cells at various times of secondary culture in the presence of Epo: (A) eluted from CD36 column; (B) after 3 d of secondary culture (day 7+3); (C) after 5 d of secondary culture (day 7+5); (D) after 7 d of secondary culture (day 7+7).

Download figure to PowerPoint

Biochemical characteristics of amplified erythroid cells

In order to check for the behaviour of in vitro CD36+-derived erythroid cells in response to Epo, we examined phosphorylation patterns of EpoR, JAK2 and STAT5 after Epo stimulation. Three independent experiments were performed. Serum- and growth factor-deprived erythroid progenitors were stimulated for 10 min with 10 U/ml Epo. Cell lystates were immunoprecipitated sequentially with antibodies directed against EpoR, JAK2 and STAT5, and analysed by Western blotting with anti-phosphotyrosine antibodies. As shown in Fig 3A, Epo induced tyrosine phosphorylation of the EpoR as well as of two associated proteins of 120 kD and 145 kD (Fig 3A1), respectively. The 120 kD protein was JAK2, as confirmed by the anti-JAK2 immunoprecipitation (IP) (Fig 3A2), and the 145 kD protein most probably corresponded to SHIP, as previously reported ( Verdier et al, 1997 ). STAT5 was also phosphorylated after Epo stimulation (Fig 3A3). Furthermore, we looked for the induction of STAT5 DNA binding activity by Epo. To that end, nuclear extracts prepared from starved cells and stimulated for 10 and 30 min with Epo were tested in an electrophoresis mobility shift assay (EMSA) ( Chrétien et al, 1996 ; Gobert et al, 1996 ). No electrophoretic shift of the radiolabelled β-casein probe was observed in unstimulated cells. In contrast, a DNA-protein binding complex was induced by Epo (Fig 3B).


Figure 3. ). The experiments presented were performed using 20×106 cells.

Download figure to PowerPoint

Similar results were obtained, whatever the secondary culture conditions (presence or absence of Epo) (not shown), demonstrating that erythroid progenitors, even at an early stage of differentiation (GpA-negative cells), were able to fully transduce Epo signals.

Haemoglobin analysis

Since no residual red blood cells were ever observed in the cultures after CD36 purification columns, the haemoglobin analysed in cells from the secondary cultures was synthesized by cells derived from the purified erythroid progenitors.

Analysis of the haemoglobin content of cells was performed by three independent methods on day 7 of secondary culture (day 7 + 7) in the presence of Epo. IEF analysis revealed the presence of adult haemoglobin (Hb A) and fetal haemoglobin (Hb F) for both the cord blood and adult samples. However, Hb A was the major component in the adult sample, whereas Hb F and Hb A bands were of similar intensities in the cord blood sample. IE-HPLC analysis gave identical results and in addition provided a quantitative estimate of the two haemoglobin fractions ( 4 Table IV). Finally, determination of the globin chain composition by RP-HPLC not only confirmed the nature of the analysed components, but also their respective amounts.

Table 4. Table IV. Analysis of the haemoglobin content of adult or cord blood cells at day 7 of secondary culture.Thumbnail image of


  1. Top of page
  2. Abstract
  6. Acknowledgements
  7. References

In the present work we have described a method which enabled the purification and in vitro expansion of large numbers of erythroid progenitors from human cord blood or adult peripheral blood CD34+ cells. Importantly, primary cultures were performed in serum-free and Epo-free medium and thus allowed the generation of a population of erythroid progenitors. These cells were mostly composed of late day-10 BFU-Es and CFU-Es which behaved normally in the presence of Epo stimulation.

It must be emphasized that the proportions of clonogenic erythroid cells after CD36 column separation of day-7 culture cells from either CB or PBMC were 16% and 7.8%, respectively ( Table II). These relatively low cloning efficiencies could be related to the presence of a majority of mature erythroid cells in the cultures. However, considering (i) that day-0 CD34+ immature cells were just 17–27% clonogenic (amongst which 8–13% were BFU-E/CFU-E) ( 2 Table II), (ii) the phenotype of the cells eluted from CD36 columns (GpA negative) ( 1 Table IB), (iii) the kinetics of secondary cultures ( Table IIIA), and (iv) the fact that primary Epo-free cultures did not produce differentiated erythroid precursors, we can assume that most of the CD36+ cells isolated belong to the progenitor (BFU-E/CFU-E) cell compartment. This highlights the fact that only some of these cells could usually be cloned in the semi-solid conditions used.

The method described here represents a useful alternative to some other methods previously reported ( Fibach et al, 1989 ; Wickrema et al, 1992 ; Giampaolo et al, 1994 ; Muta et al, 1995 ). Actually, the previous techniques combined both complex negative selections and culture stages before obtaining satisfactory red cell enrichments. For example, the purification of day-6 erythroid colony-forming cells (ECFC) used in several studies ( Fibach et al, 1989 ; Sawada et al, 1989 ; Dai et al, 1991 , 1998; Wickrema et al, 1992 ; Giampaolo et al, 1994 ; Muta et al, 1995 ) consisted of a multistep procedure comprising, after Ficoll-Hypaque separation, sheep erythrocyte rosetting of lymphoid T cells, overnight adherence for depletion of monocytes, followed by the use of a mixture of four monoclonal antibodies to remove granulocytes, B lymphocytes, NK cells and residual T cells. Thereafter, cells were cultured in methylcellulose in the presence of 30% FCS and Epo for 6 d, then collected from the culture and submitted again to adhesion and Ficoll-Hypaque separation before secondary culture in the presence of FCS, human AB serum and Epo. Moreover, total numbers of erythroid progenitors were low when compared to the numbers reached by the procedure described here ( 3 Table IIIA).

Our data also differ from previous human erythroid cell purification procedures performed in the constant presence of Epo and which lead to the expansion of more mature, mostly GpA-positive cells ( De Wolf et al, 1994 ; Mrug et al, 1997 ; Cippolleschi et al, 1997 ; Malik et al, 1998 ; Oda et al, 1998 ). This point was particularly documented by Oda et al (1998 ), who performed biochemical assays with peripheral blood-derived erythroid cells obtained from day-8 liquid cultures in the presence of both serum and Epo. The cells obtained were 71% GpA+, thereby demonstrating late erythroid differentiation, whereas our method leads to the purification and expansion of a homogenous population of immature erythroid progenitors.

Furthermore, we observed that CD34+ cells originating from both cord blood and adult peripheral blood collected by cytapheresis after stem cell mobilization displayed similar kinetics of culture evolution and cell expansion. This suggests that day-7 CD36+ cell purification from CD34+ cells cultured in IL-3 + IL-6 + SCF is an efficient general method for obtaining high concentrations of human immature erythroid progenitors.

Dramatic further cell expansion, particularly in the presence of Epo ( 3 Table IIIA), provided sufficient amounts of material for biochemical and immunological studies of erythroid progenitors. Indeed, the combination of the data presented in 1 Tables IA and IIIA indicates that, for example, an input of 106 CD34+ cells on day 0 leads to the generation of about 107 erythroid progenitors after CD36 column on day 7, and to 1.5 × 108 maturing erythroid precursor cells on day 7 + 3 of secondary culture in the presence of Epo.

With the above cell numbers, we demonstrated that serum- and cytokine-deprived CD36-positive GpA-negative or -positive erythroid cells displayed Epo responsiveness after 10 min of Epo stimulation. This resulted in the tyrosine phosphorylation of Epo-R, Jak2, SHIP and STAT5. These data confirmed those of Oda et al (1998 ), performed on more mature red cell precursors.

We did not detect any tyrosine phosphorylation of STAT3 in these human progenitors (data not shown). In addition, we recently demonstrated that GAB-1 (Grb2-associated Binder 1), a molecular adaptor whose structure is close to IRS-1/2 is tyrosine phosphorylated in response to Epo in these normal human erythroid progenitors ( Lecoq-Lafon et al 1999 ). This result further assesses the usefulness of our ex vivo purification and expansion technique.

The analysis of the haemoglobin content of erythrocytic cells obtained after 7 d of secondary culture revealed that adult-derived cells mostly contained adult haemoglobin, whereas cord blood-derived cells displayed equal proportions of fetal and adult haemoglobin ( Table IV). At that time the cell population in culture was composed of a majority of mature erythroblasts and some enucleated red cells (Fig 2D). A recent report by Malik et al (1998 ) also described the in vitro generation of mature human red blood cells by using a cytokine cocktail comprising IL-3 + GM-CSF + Epo. In that report terminal differentiation and enucleation of erythrocytes seemed to depend on the presence of macrophages in the cultures. It must be emphasized that, in our model, enucleation occurred in the absence of any myelomonocytic cells since on day 7 of secondary cultures 100% of the cells belonged to the erythroyctic lineage and no macrophages were observed.

Taken together, our data indicate that the present work allows in vitro expansion of large numbers of erythroid progenitors at mature BFU-E or CFU-E stages. These cell populations exquisitely represent the main Epo targets and are able to fully differentiate in the presence of Epo. Therefore they are adequate reagents for the study of signal transduction events which follow Epo stimulation and thus represent a useful and abundant source of normal cells for in vitro studies of erythropoiesis. Similar studies can be performed with cells from patients suffering from erythroid pathologies and therefore enable more specific analysis of the pathogenesis of acquired red cell lineage disorders.


  1. Top of page
  2. Abstract
  6. Acknowledgements
  7. References

We are grateful to Professor Jacques Elion for his help in haemoglobin analysis experiments and to Dr Stephen Ting for reviewing the manuscript.


  1. Top of page
  2. Abstract
  6. Acknowledgements
  7. References
  • 1
    Arad, Y., Mayer, T.K., Arvan, D.A. (1981) Isoelectric focusing of hemoglobins on thin-layer agarose. American Journal of Clinical Pathology, 76, 200 205.
  • 2
    Broudy, V.C. (1997) Stem cell factor and hematopoiesis. Blood, 90, 1345 1364.
  • 3
    Brugger, W., Möcklin, W., Heimfeld, S., Berenson, R.J., Mertelsmann, R., Kanz, L. (1993) Ex vivo expansion of enriched peripheral blood CD34+ progenitor cells by stem cell factor, interleukin-1β (IL-1β), IL-6, IL-3, interferon-γ, and erythropoietin. Blood, 81, 2579 2584.
  • 4
    Chrétien, S., Varlet, P., Verdier, F., Gobert, S., Cartron, J.P., Gisselbrecht, S., Mayeux, P., Lacombe, C. (1996) Erythropoietin-induced erythroid differentiation of the human erythroleukemia cell line TF-1 correlates with impaired STAT5 activation. EMBO Journal, 15, 4174 4181.
  • 5
    Cippolleschi, M.G., D'Ippolito, G., Bernabei, P.A., Caporale, R., Nannini, R., Mariani, M., Fabbiani, M., Rossi-Ferrini, P., Olivotto, M., Sbarba, P.D. (1997) Severe hypoxia enhances the formation of erythroid bursts from human cord blood cells and the maintenance of BFU-E in vitro. Experimental Hematology, 25, 1187 1194.
  • 6
    Dai, C.H., Krantz, S.B., Zsebo, K.M. (1991) Human burst-forming units-erythroid need direct interaction with stem cell factors for further development. Blood, 78, 2493 2797.
  • 7
    Dai, C.H., Price, J.O., Brunner, T., Krantz, S.B. (1998) Fas ligand is present in human erythroid colony forming cells and interacts with Fas induced by interferon γ to produce erythroid cell apoptosis. Blood, 91, 1235 1242.
  • 8
    Davis, L.G., Dibner, M.D., Battey, J.F. (1986) Methods in Molecular Biology, pp. 120 122. Elsevier, New York.
  • 9
    De Wolf, J.T.M., Muller, E.W., Hendriks, D.H., Halie, R.M., Vellenga, E. (1994) Mast cell growth factor modulates CD36 antigen expression on erythroid progenitors from human bone marrow and peripheral blood associated with ongoing differentiation. Blood, 84, 59 64.
  • 10
    Edelman, P., Vinci, G., Villeval, J.L., Vainchenker, W., Henri, A., Miglierina, R., Rouger, P., Reviron, J., Breton-Gorius, J., Sureau, C., Edelman, L. (1986) A monoclonal antibody against an erythrocyte ontogenic antigen identifies fetal and adult erythroid progenitors. Blood, 67, 56 63.
  • 11
    Erslev, A.J. (1983) In: Production of Erythrocytes (ed. by W. J. Williams, E. Beutler, A. J. Erslev and M. A. Lichtman), pp. 365 375. McGraw-Hill, New York.
  • 12
    Fibach, E., Manor, D., Oppenheim, A., Rachmilewitz, E.A. (1989) Proliferation and maturation of human erythroid progenitors in liquid culture. Blood, 73, 100 103.
  • 13
    Giampaolo, A., Sterpetti, P., Bulgarini, D., Samoggia, P., Pelosi, E., Valtieri, M., Peschle, C. (1984) Key functional role and lineage-specific expression of selected HOXB genes in purified hematopoietic progenitor differentiation. Blood, 84, 3637 3647.
  • 14
    Gobert, S., Chrétien, S., Gouilleux, F., Muller, O., Pallard, C., Dusanter-fourt, I., Groner, B., Lacombe, C., Gisselbrecht, S., Mayeux, P. (1996) Identification of tyrosine residues within the intracellular domain of erythropoietin receptor crucial for STAT5 activation. EMBO Journal, 15, 2434 2441.
  • 15
    Goldsmith, M.A., Mikami, A., You, Y., Liu, K.D., Thomas, L., Pharr, P., Longmore, G.D. (1998) Absence of cytokine receptor-dependent specificity in red blood cell differentiation in vivo. Proceedings of the National Academy of Sciences of the United States of America, 95, 7006 7011.
  • 16
    Huh, H.Y., Lo, S.K., Yesner, L.M., Silverstein, R.L. (1995) CD36 induction on human monocytes upon adhesion to tumor necrosis factor-activated endothelial cells. Journal of Biological Chemistry, 270, 6267 6271.
  • 17
    Huh, H.Y., Pearce, S.F., Yesner, L.M., Schindler, J.L., Silverstein, R.L. (1996) Regulated expression of CD36 during monocyte-to-macrophage differentiation: potential role of CD36 in foam cell formation. Blood, 87, 2020 2028.
  • 18
    Kieffer, N., Bettaieb, A., Legrand, C., Coulombel, L., Vainchenker, W., Edelman, L., Breton-Gorius, J. (1989) Developmentally regulated expression of a 78 kDa eythroblast membrane glycoprotein immunologically related to the platelet thrombospondin receptor. Biochemical Journal, 262, 835 842.
  • 19
    Lecoq-Lafon, C., Verdier, F., Fichelson, S., Chrétien, S., Gisselbrecht, S., Lacombe, C., Mayeux, P. (1999) Erythropoietin induces the tyrosine phosphorylation of GAB1 and its association with SHC, SHP2, SHIP and phosphatidylinositol 3-kinase. Blood, 93, 2578 2585.
  • 20
    Lessard, J., Baban, S., Sauvageau, G. (1998) Stage-specific expression of polycomb group genes in human bone marrow cells. Blood, 91, 1216 1224.
  • 21
    Malik, P., Fisher, T.C., Barsky, L.L.W., Zeng, L., Izadi, P., Hiti, A.L., Weinberg, K.I., Coates, T.D., Meiselman, H.J., Kohn, D.B. (1998) An in vitro model of human red blood cell production from hematopoietic progenitor cells. Blood, 91, 2664 2671.
  • 22
    Miller, J.L., Njoroge, J.M., Gubin, A.N., Rodgers, G.P. (1999) Prospective identification of erythroid elements in cultured peripheral blood. Experimental Hematology, 27, 624 629.
  • 23
    Mrug, M., Stopka, T., Julian, B.A., Prchal, J.F., Prchal, J.T. (1997) Angiotensin II stimulates proliferation of normal early erythroid progenitors. Journal of Clinical Investigation, 100, 2310 2314.
  • 24
    Muta, K., Krantz, S.B., Bondurant, M.C., Dai, C.H. (1995) Stem cell factor retards differentiation of normal human erythroid progenitor cells while stimulating proliferation. Blood, 86, 572 580.
  • 25
    Nakahata, T. & Okumura, N. (1994) Cell surface antigen expression in human erythroid progenitors: erythroid and megakaryocytic markers. Leukemia and Lymphoma, 13, 401 409.
  • 26
    Oda, A., Sawada, K., Druker, B.J., Ozaki, K., Takano, H., Koizumi, K., Fukada, Y., Handa, M., Koike, T., Ikeda, Y. (1998) Erythropoietin induces tyrosine phosphorylation of Jak2, STAT5A, and STAT5B in primary cultured human erythroid precursors. Blood, 92, 443 451.
  • 27
    Ogawa, M. (1993) Differentiation and proliferation of hematopoietic stem cells. Blood, 81, 2844 2853.
  • 28
    Okumura, N., Tsuji, K., Nakahata, T. (1992) Changes in cell surface antigen expressions during proliferation and differentiation of human erythroid progenitors. Blood, 80, 642 650.
  • 29
    Oquendo, P., Hundt, E., Lawler, J., Seed, B. (1989) CD36 directly mediates cytoadherence of Plasmodium falciparum parasitized erythrocytes. Cell, 58, 95 101.
  • 30
    Papayannopoulou, T., Brice, M., Blau, C.A. (1993) Kit ligand in synergy with interleukin-3 amplifies the erythropoietin-independent, globin-synthesizing progeny of normal human burst-forming units-erythroid in suspension cultures: physiologic implications. Blood, 81, 299 310.
  • 31
    Rigotti, A., Acton, S.L., Krieger, M. (1995) The class B scavenger receptor SR-BI and CD36 are receptors for anionic phospholipids. Journal of Biological Chemistry, 270, 16221 16224.
  • 32
    Sauvageau, G., Lansdorp, P.M., Eaves, C.J., Hogge, D.E., Dragowska, W.H., Reid, D.S., Largman, C., Lawrence, H.J., Humphries, R.K. (1994) Differential expression of homeobox genes in functionally distinct CD34+ subpopulations of human bone marrow cells. Proceedings of the National Academy of Sciences of the United States of America, 91, 12223 12227.
  • 33
    Sawada, K., Krantz, S.B., Dessypyris, E.N., Koury, S.T., Sawyer, S.T. (1989) Human colony-forming units-erythroid do not require accessory cells but do require direct interaction with insulin-like growth factor 1 and/or insulin for erythroid development. Journal of Clinical Investigation, 83, 1701 1709.
  • 34
    Shelton, J.B., Shelton, J.R., Schroeder, W.A. (1984) High-performance liquid chromatographic separation of globin chains on a large-pore C4 column. Journal of Liquid Chromatography, 7, 1969 1977.
  • 35
    Socolovsky, M., Dusanter-Fourt, I., Lodish, H.F. (1997) The prolactin receptor and severely truncated erythropoietin receptors support differentiation of erythroid progenitors. Journal of Biological Chemistry, 272, 14009 14012.
  • 36
    Sui, X., Tsuji, K., Tajima, S., Tanaka, R., Muraoka, K., Ebihara, Y., Ikebuchi, K., Yasukawa, K., Taga, T., Kishimoto, T., Nakahata, T. (1996) Erythropoietin-independent erythrocyte production: signals through gp130 and c-kit dramatically promote erythropoiesis from human CD34+ cells. Journal of Experimental Medicine, 183, 837 345.
  • 37
    Van Schravendijk, M.R., Handunnetti, S.M., Barnwell, J.W., Howard, R.J. (1992) Normal human erythrocytes express CD36, an adhesion molecule of monocytes, platelets, and endothelial cells. Blood, 80, 2105 2114.
  • 38
    Verdier, F., Chrétien, S., Billat, C., Gisselbrecht, S., Lacombe, C., Mayeux, P. (1997) Erythropoietin induces the tyrosine phosphorylation of insulin receptor substrate-2. Journal of Biological Chemistry, 272, 26173 26178.
  • 39
    Wickrema, A., Krantz, S.B., Winkelmann, J.C., Bondurant, M.C. (1992) Differentiation and erythropoietin receptor gene expression in human erythroid progenitor cells. Blood, 80, 1940 1949.
  • 40
    Wu, H., Klingmüller, U., Acurio, A., Hsiao, J.G., Lodish, H.F. (1997) Functional interaction of erythropoietin and stem cell factor receptors is essential for erythroid colony formation. Proceedings of the National Academy of Sciences of the United States of America, 94, 1806 1810.
  • 41
    Wu, H., Klingmüller, U., Besmer, P., Lodish, H.F. (1995 a) Interaction of the erythropoietin and stem-cell-factor receptors. Nature, 377, 242 245.
  • 42
    Wu, H., Liu, X., Jaenisch, R., Lodish, H.L. (1995 b) Generation of committed erythroid BFU-E and CFU-E progenitors does not require erythropoeitin receptor. Cell, 83, 59 67.