The Two-step Liquid Culture: A Novel Procedure for Studying Maturation of Human Normal and Pathological Erythroid Precursors

Transcription

1 The Two-step Liquid Culture: A Novel Procedure for Studying Maturation of Human Normal and Pathological Erythroid Precursors E. Fibach, E. A. Rachmilewitz Department of Hematology, Hadassah University Hospital, Jerusalem, Israel Key Words. Erythropoiesis Erythropoietin Growth factors Hemoglobin In vitro Cell cycle Abstract. We have recently described a novel twophase liquid culture procedure for growing human erythroid cells in vitro. The two phases are 1) an phase, in which the cells are first cultured in the presence of a combination of growth factors excluding EPO; during this phase, early erythroid committed progenitors, burst forming units (BFU-e), proliferate and differentiate into colony forming unit (CFU-e)-like progenitors; 2) a second phase, in which the latter cells are cultured in an EPO-supplemented medium, in which the CFU-e-like progenitors continue to proliferate and mature into orthochromatic normoblasts and then enucleated erythrocytes. This procedure yields large (up to 5 x lo*) and pure (9598%) populations of erythroid cells, which allow detailed study of normal and pathologic erythroid maturation, including 1) the effects of growth factors on proliferation and differentiation at various erythroid developmental stages, 2) intracellular iron metabolism in normal and thalassemic erythroid cells and the role of ferritin as an iron donor for heme synthesis, 3) the expression of surface antigens: transferrin receptor, glycophorin, A, B, H, D and I/i antigens, 4) synthesis of erythroid-specific membrane proteins, 5) the kinetics of globin mrna accumulation during erythroid maturation, 6) the expression of exogenous human f3 globin gene in p-thalassemic cells as a model for gene therapy, and 7) the enhancement of y globin chain synthesis by chemical agents. Introduction The study of the developmental aspects of erythroid cell differentiation and maturation is Correspondence: Dr. E. Fibach, Department of Hematology, Hadassah University Hospital, POB 12000, Jerusalem 91120, Israel. Received December 1, 1992; accepted for publications December 1, OAlphaMed Press /93/$5.00/0 complicated by the fact that, like other hemopoietic cells, the origin of RBC is in precursors in the bone marrow (BM). However, the BM is not readily available, and it consists of a mixture of cells of various types and stages of differentiation, which makes the purification of a subpopulation of cells at a particular stage quite difficult. One approach to overcome this difficulty is to study in vitro established, immortalized, erythroid cell lines. These include cells derived from virus infected erythroleukemic mice [ 11 and cells derived from patients with chronic myeloid leukemia at blast crisis [2]. These cells grow continuously in suspension in liquid culture as proerythroblasts. They maintain their immature phenotype as manifested morphologically, immunologically and biochemically and their leukemogenicity as demonstrated by injection into suitable experimental animals. Various chemical agents, including dimethylsulfoxide, hexamethylene bisacetamide and butyric acid, can induce these cells to undergo further maturation [I, This maturation process has many similarities to normal erythropoiesis, including typical morphological changes, accumulation of globin mrnas, synthesis of globin chains and heme, formation of hemoglobins (Hb), the appearance of RBC-specific membrane antigens and cessation of cell division [5]. Although these established cell lines serve as convenient experimental models, because of their leukemic origin and long history in culture they do not recapitulate all aspects of erythropoiesis. For example, most cell lines are not inducible by erythropoietin (EPO), the normal physiological inducer of erythropoiesis. In addition, most human cell lines do not synthesize adult Hb and do not undergo terminal cell division [6]. An alternative approach for studying erythropoiesis is cloning of BM or peripheral blood (PB) cells in EPO-supplemented semisolid medium and STEM CELLS 1993;11(s~ppl 1):36-41

2 Fibach/Rachmilewitz 37 scoring the erythroid colonies that develop [7]. Since each colony is the outcome of the proliferation and differentiation of one particular progenitor, the quantity and quality of the progenitors in the hemopoietic tissue can be evaluated. Based on the time of maturation and the proliferative potential (size of the colonies), progenitors at different developmental stages-the early burst forming units (BFU-e) and the late colony forming units (CFU-e)--can be distinguished [8]. The fact that the cloned cells are immobilized in the viscous medium results in several disadvantages: 1) The yield is low (<lo5 celldm1 culture), making it technically dificult to carry out quantitative analysis of growth kinetics as well as biochemical, molecular and immunological characterization of the developing cells; 2) It is a one-step continuous culture, thus making it difficult to determine the effects of various factors (i.e., cytokines) at different maturation stages, since it is difficult to add or subtract components to/ from the culture; 3) In addition, since the cells are immobilized, cell-cell interactions are limited. These disadvantages can be overcome using a liquid suspension culture, recently described by us, that supports the growth and maturation of human erythroid progenitors [9]. In this procedure, the culture is divided into two phases. The frst is an EPO-independent phase, in which the cells are first cultured in the presence of a combination of growth factors excluding EPO. During this phase, early erythroid committed progenitors, BFU-e, proliferate and differentiate into CFU-e-like progenitors. In the second phase, the latter cells are cultured in an EPO-supplemented medium, in which the CFU-e-like progenitors continue to proliferate and mature into orthochromatic normoblasts and then enucleated erythrocytes. This procedure yields large (up to 5 x lo8) and pure (95-98%) populations of erythroid cells, which allow detailed study of normal and pathologic erythroid maturation and analysis of the effects, of various pharmacological agents in cultures of cells derived from relevant patients. Materials and Methods The two-phase liquid culture procedure has been previously described in detail 191. PB mononuclear cells (PBMNC) obtained from normal volunteers or patients with hemoglobinopathies were isolated by centrifugation on a gradient of Ficoll-Hypaque and seeded at a density of 5 x lo6 celyml in alpha-minimal essential medium (WM) supplemented with 10% fetal calf serum (FCS) (both from GIBCO, Grand Island, NY), 1 pg/ml cyclosporin A (Sandoz, Basel, Switzerland) and 10% conditioned medium collected from cultures of the 5637 bladder-carcinoma cell line. The cultures were incubated at 37 C in an atmosphere of 5% CO, in air with extra humidity. Following 5-7 day incubation in this phase I culture, the nonadherent cells were harvested, washed and recultured in fresh medium composed of amem, 30% FCS, 1% deionized bovine serum albumin (BSA), 1 x M /3-mercaptoethanol, 1.5 mm glutamine, 1 x M dexamethasone, and 1 U/ml human recombinant EPO (Cilag AG, Schauffhausen, Switzerland). After a 4-day incubation in phase 11, lymphocytes were removed as follows: The cells were harvested and spun down, the medium saved, and the cells layered on a Percoll (Sigma, St. Louis, MO) solution (density = ghl) and centrifuged at 1,000 g for 20 min at room temperature. The upper layer containing the proerythroblasts was collected and resuspended in the saved medium, and incubation continued. Results and Discussion PB cells have been used in this procedure because of the availability of PB from normal individuals and patients and because of the homogeneity of the circulating erythroid progenitor population, namely the early BFU-e, as opposed to the BM, which contains progenitors at various developmental stages [lo]. Depending on the number of cells required for the specific study, either ml blood or the bufq coat fraction derived from a whole blood unit may be used. The latter source is readily available as it is routinely separated by blood banks during the procedure of preparation of packed RBC and platelet-rich plasma. Other sources may be the buffy coat fractions obtained following phlebotomy or leukophoresis. The MNC can be easily separated from the buffy coat fraction by adding a Ficoll-Hypaque solution directly into the transfer bag, spinning and collecting the interphase layer. Monocytes can be depleted by adherence, but, since they do not interfere with the development of erythroid progenitors, they may not removed prior to culturing. Lymphocytes, which

3 38 Erythropoiesis in Liquid Culture Fig. 1. Erythroid cells in phase I1 of the liquid culture. A = Cultures 10 days after addition of EPO, demonstrating proerythroblasts. B = Cultures 15 days after addition of EPO, demonstrating orthochromatic normoblasts and an enucleated erythrocyte. do interfere with erythroid development (manuscript in preparation), may be removed following treatment with anti-lymphocytic antibodies by complement-mediated cytolysis [91, panning or using magnetic beads. When large volumes of blood are processed, the use of antibodies is quite expensive; instead, the lymphotoxic drug cyclosporin A is included in the culture medium. This drug interferes with lymphocyte activation and lymphokine production [ 111, but at the concentrations used (1 pg/ml) has no effect on erythroid cell development. PBMNC are then cultured for one week in liquid medium supplemented with various hemopoietic growth factors excluding EPO. Prior to the second phase, cells are washed thoroughly to remove the growth factors and then recultured under conditions conducive to the development of erythroid cells, including the presence of EPO and incubation under low oxygen pressure. This results in the exclusive development of erythroid cells. During the first five days of the second phase, the total cell number decreased, but the proportion of erythroid cells increased. The decrease in nonerythroid, mainly lymphocyte cells is gradual, and by day 17 (day 12 of phase 11), when erythroid cells reach maximum number, they represent only a minor (<5%) contamination. Yet, in order to collect pure populations of erythroid precursors at early maturation stages, when lymphocytes constitute a considerable proportion of the population, the much larger proerythroblasts may be separated by sedimentation on a gradient of Percoll. Hb-containing cells, as determined by benzidine cytochemical staining [ 121, start to appear three to five days following exposure to EPO. A maximum number and percentage of benzidinereactive cells is achieved after days, when cell concentration reaches 3-5 x 106/ml. At this stage, 90% of the cells can be identified morphologically as orthochromatic normoblasts and the rest as enucleated RBC (Fig. 1). The two-step liquid culture has the following advantages compared with the semisolid culture [ 131: A. It supports almost exclusively the growth of erythroid cells, providing uniform erythroid cell populations, while the EPO-supplemented semisolid culture supports the growth of myeloid as well as erythroid colonies, and pure populations of erythroid cells can be obtained only by picking up individual colonies. B. The yield of erythroid cells per MNC cultured is significantly higher in the liquid culture. This is mainly due to the longer period of proliferation, resulting in more cell divisions in liquid culture. Proliferation stops in semisolid culture after 14 days, while in liquid culture it continues until day 17. Assuming that the same number of BFU-e are capable of developing in both culture systems, these results imply that in liquid culture each BFU-e yields more progeny. This suggests that colonies grown in semisolid culture do not rep resent the full growth potential of the progenitors. C. Even more impressive is the difference in terms of erythroid cell concentration per ml culture, which is about tenfold higher in the liquid culture. Colony formation (number and size) in semisolid cultures is cell-dose dependent at low seeding concentrations, but it reaches plateau and even decreases at high seeding cell concentration due to overcrowding effect. Seeding 5 x los PBMNC/ml semisolid culture is optimal for colony growth. In liquid culture, initial cell concentration can be ten times higher without reducing the final cell yield, which could reach up to 7.5 x lo6 erythroid cells/ml culture. D. Moreover, the thickness of the semisolid layer that pennits cell growth is limited, only cm, as compared with 2 cm or more of liquid medium. Therefore, in liquid culture less surface area (culture vessels) is required to set up the same volume of culture. For example, one usually gets 4.2 x lo* cells from one unit (400 ml) of blood in one 150 cm2 flask containing ml culture. In comparison, to get the same culture volume using methylcellulose medium, 800 dishes (35 mm in diameter) would be required.

4 Fibach/Rachmilewitz 39 In addition to the better yield, purity and convenience of the liquid culture procedure, it affords several additional advantages in studying erythroid development, including synchronization in terms of maturation stages and separation of the EPO-independent and EPO-dependent phases, permitting the study of the effect of various factors in each phase specifically and manipulation of culture conditions and components in various stages without terminating the culture. Development of Erythroid Progenitors in Phase I The developmental stages and the role of growth factors in the primary phase of the culture were analyzed by a combination of liquid and semisolid cultures. An aliquot of PB cells was cloned directly in semisolid medium supplemented with EPO. In parallel, another aliquot was first precultured in liquid medium and then cloned in EPO-containing semisolid medium (indirect cloning). The extent of proliferation in the liquid culture was calculated by comparing the number of colonies that developed following direct and indirect cloning. The results indicated up to a tenfold increase in the number of progenitors, as judged by their ability to form colonies, during the sevenday incubation in phase I. Colonies which developed following direct cloning were the classical bursts, consisting of thousands of cells distributed in several subcolonies and reaching full hemoglobinization and terminal differentiation after about two weeks. Following indirect cloning, colonies reached full maturation after one week and were much smaller, similar to CFU-e-derived colonies (Fig. 1). These results suggest that, during the first phase, the BFU-e, present in the PB, proliferated as well as differentiated into CFU-e [9]. This primary phase was considered EPOindependent. To rule out the possibility that trace amounts of EPO present in the serum used for the culture have an effect, the proliferation of the erythroid progenitors was determined either in the absence of EPO, after addition of anti-epo antibodies at titer enough to neutralize EPO in serum, or after addition of 0.5 U/ml EPO. The results indicated no effect of EPO during the first week in liquid culture (Fig. 2). The role of other growth factors at this stage was determined in medium supplemented with only 2% serum to avoid serum-derived endogenous growth factors. The results indicated an effect of granuloc yte-macrophage colony-stimulating factor T ~ D Effect of PO on Erythroid Progenitor Proliferotion ' - 4 t 3 20 T O L / Fig. 2. The effect of EPO on erythroid progenitor proliferation. PBMNC were cultured in liquid medium either in the absence of EPO, after addition of anti- EPO-antibodies, or after addition of 0.5 U/ml EPO. On the indicated days, aliquots were harvested, washed and cloned in EPO (2 U/ml)-supplemented methylcellulose medium. Colonies were scored two weeks after the onset of the experiment. The results are presented as proliferation index calculated as the ratio of the number of colonies that developed following direct and indirect cloning into EPO-supplemented methylcellulose medium. The Effect of Cyloklnei on Erythroid Progenltore CM Cone. (W) lo00 Cytohlne Corn. (Wml) Fig. 3. The effect of cytokines on erythroid progenitor proliferation. PBMNC were cultured in liquid medium in the presence of different concentrations (%V/V) 5637 conditioned medium (CM), or the indicated concentrations of either GM-CSF, G-CSF, IL-3 or IL-6. After one week incubation, the cells were cloned as described in the legend to Fig.2. I

5 40 Erythropoiesis in Liquid Culture (GM-CSF) and interleukin (IL)-3, but no effect of granulocyte CSF (G-CSF) or IL-6 during the first phase (Fig.3). Other cytokines, such as the c-kit ligand, and combinations are presently under study. Development of Erythroid Precursors in Phase II Experiments performed to study the effect of EPO and other cytokines in the second phase of the culture indicated that EPO was required continuously throughout the second phase, mainly for stimulation of cell proliferation. During the first week of phase 11, this EPO-dependency was also associated with apoptosis (programmed cell death) as could be demonstrated by DNA gel electrophoresis as well as by flow cytometry. However, cell survival and maturation could proceed thereafter in the absence of EPO, indicating that normal erythroid progenitors undergo an irreversible commitment step beyond which the presence of EPO is not required. Other growth factors had no effect on erythroid development at this stage. The cell cycle status of erythroid precursors in phase II was studied by a double labeling flow cytometric technique [ 141. In the presence of optimal concentrations of EPO (2 U/ml), a high percentage (>40%) of cells was found in the S phase of the cell cycle until day 10, indicating a rapid proliferation rate. Then, as a result of maturation, the proportion of cells in S gradually decreased, reaching less than 2% by day 21. At this time, the culture consisted of >95% Hb-containing, nonproliferating, orthochromatic normoblasts. This normoblast population demonstrated a bimodal distribution with respect to the cell cycle status; while the majority of the cells had a diploid (2C) DNA content, i.e., cells in G, (or Go) phase, a sizeable fraction were tetraploid (4C) corresponding to cells in G,. In contrast, in cultures stimulated with physiological concentrations of EPO (50 mu/ml), all the terminally differentiated cells were arrested at the G, phase. These results support the notion that EPO is an essential growth-promoting factor for erythroid precursors, but suggest that supra-physiological concentrations, such as are present in vivo in severe anemia (e.g., aplastic anemia) or after EPO administration, may be associated with development of normoblasts with abnormal DNA content. The development of the ABH antigens was studied by flow cytometry of cells labeled with specific monoclonal antibodies or lectins [15]. No expression of these antigens was detected in phase I but could be observed two days after exposure to EPO. These results indicated that ABH antigens are absent or unexposed on erythroid progenitors; their development depends on exposure to EPO, and they first appear at early stages of maturation (pronormoblasts) prior to the accumulation of Hb in the developing erythroid cells. The two-step liquid culture was recently utilized to study synthesis of membrane components and to determine the mechanisms underlying band 3 and spectrin deficiencies in subsets of patients with hereditary spherocytosis [ 161 and pyropoikilocytosis [ 171, respectively. We have also studied the detailed kinetics of globin gene expression at the transcriptional level, in adult and newborn erythroid cells in phase IT cultures [18]. The mrnas for a, p, and y globins could be first detected on day 4 following exposure to EPO. In cultures derived from normal adult PB, the mrna levels for y was proportionally high during early stages, but as the cells matured mrna for c1 and p increased much more rapidly than y mrna, resulting in a decrease in the proportion of 'y. In contrast, cultures derived from cord blood demonstrated high expression of all three globin genes. Analysis of the erythroid-specific transcription factor GATA- 1 by gel-shift assay, indicated that the peak of globin mrna transcription was preceded by a exponential increase in an EPOstimulated burst of GATA-1 activity. These results suggest that the level of GATA-1 regulates the extent of transcription of the erythroid-specific genes, including that of the globins (manuscript in preparation). We are currently utilizing the two-step liquid culture for studying drugs that can enhance the production of fetal Hb in cells derived from patients with sickle cell anemia and p-thalassemia. Our results with hydroxyurea, a drug currently in clinical trials, indicated that the cultures recapitulate the hematological effects demonstrated in patients undergoing treatment with the drug [19]. In addition, these cultures were found to be insttumental in screening compounds for their fetal Hb stimulation potential [20] and for studying their mechanism of action. Moreover, the cultures also provide a valuable predictive assay system for evaluation of the response of individual patients with hemoglobinopathies to such agents. Conclusions The two-phase liquid culture procedure is use-

6 Fibach/Rachmilewitz 41 ful in elucidating various aspects of normal and pathologic erythropoiesis, including 1) the effects of growth factors on proliferation and differentiation at various erythroid developmental stages, 2) intracellular iron metabolism in normal and thalassemic erythroid cells and the role of ferritin as a iron donor for heme synthesis, 3) the expression of surface antigens: transferrin receptor, glycophorin, A, B, H, D and Jli antigens, 4) synthesis of erythroid-specific membrane proteins, 5) the kinetics of globin mrna accumulation during erythroid maturation, 6) the expression of exogenous human P globin gene in P-thalassemic cells as a model for gene therapy, and 7) the enhancement of y globin chain synthesis by chemical agents. References 1 Friend C, Scher W, Holland J, Sato J. Hemoglobin synthesis in murine virus induced leukemic cells in vitro: stimulation of erythroid differentiation by dimethylsulfoxide. Proc Natl Acad Sci USA 1971;68: Lozzio CB, Lozzio BB. Human chronic myelogenous leukemia cell line with positive Philadelphia chromosome. Blood 1975;45: Fibach E, Reuben RC, Rifkind RA, Marks PA. Effect of hexamethylene hisacetamide on the commitment to differentiation of murine erythroleukemia cells. Cancer Res 1977;37: Leder A, Leder P. Butyric acid, a potent inducer of erythroid differentiation in cultured erythroleukemia cells. Cell 1975;5: Marks PA, Rifkind RA. Erythroleukemic differentiation. Annu Rev Biochem. 1978;47: Rutherford T, Clegg JB, Higgs DR, Jones RW, Thompson J, Weatherall DJ. Embryonic erythroid differentiation in the human leukemia cell line K562. Proc Natl Acad Sci USA 1981;78: Stephenson JR, Axelrad AA, Mcleod DL, Shreeve MM. Induction of colonies of hemoglobin synthesizing cells by erythropoietin. Proc Natl Acad Sci USA 1971;68: Gregory CJ. Erythroid sensitivity as a differentiation marker in the hemopoietic system: studies of three erythropoietic colony responses in culture. J Cell Physiol 1976;89: Fibach E, Manor D, Oppenheim A, Rachmilewitz EA. Proliferation and maturation of human erythroid progenitors in liquid culture. Blood 1989;73: Clarke BJ, Housman D. Characterization of an erythroid precursor cell of high proliferative capacity in normal peripheral blood. Proc Natl Acad Sci USA 1977;74: Bore1 JF, Ryffel B. The mechanism of action of cyclosporin: a continuing puzzle. In: Schnidler R, ed. Cyclosporin in Autoimmune Diseases. Heidelberg: Springer-Verlag. 1985: Orkin SH, Harosi FI, Leder P. Differentiation of erythroleukemic cells and their somatic hybrids. Proc Natl Acad Sci USA 1975;72: Fibach E, Manor D, Treves A, Rachmilewitz EA. Growth of human normal erythroid progenitors in liquid culture: a comparison with colony growth in semisolid culture. Int J Cell Cloning. 1991;9: Fibach E, Rachmilewitz EA. Stimulation of erythroid progenitors by high concentrations of erythropoietin results in normoblasts arrested in G, phase of the cell cycle. Exp Hematol 1993;21: 184. Sharon R, Fibach E. The expression of ABH antigens on human erythroid cells during differentiation. The XX Congress of the International Society of Blood Transfusion, Saad STO, Liu SC, Golan D, Corbett JB, Thattle HS, Derick L, Hanspal M, Jarolim P, Fibach E, Palek J. Mechanism underlying band 3 deficiency in a subset of patients with hereditary spherocytosis (HS). Blood 1991;78(suppl 1): 10. Hanspal M, Hanspal J, Fibach E, Palek J. Spectrin synthesis defect underlies spectrin deficiency in a subset of patients with hereditary pyropoikilocytosis (HPP). Blood 1991;78(suppl 1):lO. Dalyot N, Fibach E, Rachmilewitz EA, Oppenheim A. Adult and neonatal patterns of human globin gene expression are recapitulated in liquid cultures. Exp Hematol 1992;20:1141. Fibach E. Burke LB, Schechter AN, Noguchi CT, Rodgers GP. Hydroxyurea increases fetal hemoglobin in cultured erythroid cells derived from normal individuals and patients with sickle cell anemia or P-thalassemia. Blood 1993; 81: Fibach E, Rodgers GP, Samid D. Sodium phenylacetate (NaPA) increases fetal hemoglobin (HbF) in cultured erythroid cells derived from normal individuals and patients with sickle cell anemia or P-thalassemia. Blood 19928O(Suppl 1): 81a.