ABSTRACT
Levels of haem synthesis achieved by foetal liver erythroblasts responding to erythropoietin in vitro are similar in dissociated cell cultures and in cultures of organized tissues. Erythroid colony-forming cells reach maximum numbers on the sixteenth day of gestation. Their presence in foetal liver is associated with the period of most rapid production of erythrocytes, and with in vitro sensitivity to erythropoietin measured as enhanced haem synthesis. It is concluded that at least a proportion of erythroid colony-forming cells in the foetal liver are dependent on erythropoietin in situ and that these cells are separated from the earliest recognizable pro-erythroblast by 1 –2 cell divisions. Populations of granulocyte-macrophage colony-forming cells change independently of erythroid colony-forming cell numbers.
INTRODUCTION
The development of techniques which allow prenatal haemopoietic tissue to respond to erythropoietin in vitro by enhanced haem synthesis (Cole & Paul, 1966; Cole, Hunter & Paul, 1968; Bateman & Cole, 1971) or globin synthesis has facilitated investigation of the molecular events underlying differentiation of erythroid cells and the initiation of haemoglobin synthesis (Djaldetti, Preisler, Marks & Rifkind, 1972; Terada et al. 1972; Harrison, Conkie & Paul, 1973). The cell kinetics of prenatal haemopoietic tissues are now well documented (Paul, Conkie & Freshney, 1969; Tarbutt & Cole, 1970; Wheldon et al. 1974). Extension of the spleen-colony assay for haemopoietic stem cells (Till & McCulloch, 1961) to prenatal mouse haemopoietic organs has led to detailed descriptions of the changes in numbers of this cell type during gestation (Silini, Pozzi & Pons, 1967; Moore & Metcalf, 1970). Recent advances permit the in vitro clonal analysis of progenitor cells of both the granulocyte/macrophage (Moore & Metcalf, 1970; Moore & Williams, 1973) and erythroid populations (Stephenson, Axelrad, McLeod & Shreeve, 1971). Haematological studies on foetuses indicate that during the initial phase of hepatic erythropoiesis foetal tissues are likely to be relatively poorly oxygenated, with consequently high levels of erythropoietin or similar erythropoietic stimulatory factor in the foetal circulation. During later hepatic erythropoiesis, as a result of a very rapid increase in red cell number, a more stable equilibrium is reached and erythrocyte production in the liver slows. This is probably a consequence of both reduced erythropoietin levels, and the operation of directly acting negative feedback controls on erythroid proliferation or differentiation (Cole, 1975).
The present report is concerned with interrelationships between erythropoietin sensitivity of mouse foetal liver cells in vitro, when explanted as cell suspensions and as organized tissues, and the numbers and behaviour of erythroid colonyforming cells (CFUe) and granulocyte-macrophage colony-forming cells (CFUc) in foetal liver, spleen and bone marrow.
MATERIALS AND METHODS
Animals
Foetal material was obtained from timed natural matings of inbred strain-129 mice. The morning on which mating plugs were found began day 0.
Culture methods
(a) Dissociated cell cultures
Foetal livers were disaggregated to single-cell suspensions by pipetting in culture medium, without the use of enzymes, and cultured at 1 –2 × 106 cells/ml in Waymouth’s medium supplemented with 7·5% foetal calf serum and 2·5% mouse serum. Cultures were maintained at 37 °C and pH 7·4 in 5% CO2 in air.
(b) Liver fragments
Isolated livers were cut into fragments of approximately 2 mg, excess liquid removed, and their wet weight determined in pre-weighed tubes containing culture medium. Each fragment was cultured in 3 ml of medium as above, in sealed 15 ml culture tubes on an inclined roller-tube apparatus.
(c) Erythroid colony-forming cells
Foetal livers were disaggregated as above. Foetal spleens were disaggregated by repeated passage through a 0·8 mm diameter syringe needle, in culture medium. Foetal femurs were dissected free of adherent tissue, crushed with a glass rod in a small volume of culture medium and marrow cells suspended by repeated pipetting. Erythroid colony-forming cells were cultured in plasma clots according to the method of Stephenson et al. (1971), in medium NCTC 109 supplemented with 10% foetal calf serum, 1% mouse serum, 0·02mg/ml asparagine, 0·85% bovine embryo extract, 1% bovine serum albumin and 10% citrated bovine plasma. Cultures (1 ml) were initiated with 1–4 × 105 cells in 35 mm diameter ‘Nunc’ plastic Petri dishes and maintained for 3 days in 5% CO2 in air at 37 °C. Colonies were initially characterized by staining, after fixation with glutaraldehyde and drying, with Lepehne’s stain, or in situ, without drying, with 0·2% benzidine in 0·5% acetic acid+ 0·4% 30% H2O2. They were routinely counted without staining at × 100 magnification. Colonies containing more than eight cells and of characteristic erythroid appearance were counted (Gregory, McCulloch & Till, 1973; Iscove, Sieber & Winterhalter, 1974).
(d) Granulocyte-macrophage colony-forming cells
Cell suspensions were obtained as above and cultured in 35 mm ‘Nunc’ plastic Petri dishes in 1-5 ml modified Eagle’s M.E.M. containing 0·3% agar, as described by Metcalf & Moore (1971). Medium conditioned by mouse L cells grown in suspension was used as a source of Colony Stimulating Factor (C.S.F.) at 10% (the concentration which elicited maximum numbers of colonies from foetal liver and adult marrow). Rat-erythrocyte lysate (Bradley, Sumner & McInerney, 1974) was added to this semi-solid agar medium. Cultures were maintained at 37 °C in 5% CO2 in air for 8 days and examined at × 50 magnification for the incidence of colonies of more than 40 cells. Each experiment with foetal material was run in parallel with cultures derived from pooled femoral bone marrow of three or four 25-gm strain-129 male mice. Such cultures invariably gave an incidence of colonies within the range 2·3 × 103–2·5 × 103/106 marrow cells, so the values obtained for foetal tissue are presented without normalizing to a notional bone-marrow standard (Metcalf & Moore, 1971).
BIOCHEMICAL TECHNIQUES
Haem synthesis was estimated by incubating cells or liver fragments in 1–2 μCi/ml 59Fe (ferric chloride) previously equilibrated with mouse transferrin by incubation with 50% mouse serum in culture medium. Haem was extracted from washed cells with acid ethyl methyl ketone and counted in a gas-flow counter.
Human urinary erythropoietin lot EPM-3-TaLSL or lot EPM-10-Ta.LSL, prepared by the Los Angeles Children’s Hospital, was used at 0·2 unit/ml for stimulation of haem synthesis, and at 0·6 unit/ml for estimation of CFUe.
RESULTS
Changes in circulating blood cells and foetal liver
Changes in the peripheral circulation during the hepatic phase of erythropoiesis are shown in Fig. 1. Red-cell concentration rises rapidly until day 16, then remains relatively constant until near birth. Similar changes are seen for the concentration of haemoglobin in the blood. Haemoglobin content per unit of body weight also shows a sharp inflexion on day 16, and falls slightly during the last 2–3 days of gestation, before initiation of red-cell production by the spleen and bone marrow. Changes in the rate of accumulation of total erythrocytes/foetus are less sharp, but again there is an inflexion on day 16.
Developmental changes in the haematological state of strain 129 foetal mice. Erythrocyte concentration in the peripheral circulation (○— ○) is shown in relation to haemoglobin content/unit of body weight (▫— ▫); haemoglobin content/ unit volume of blood (△— △); and total erythrocytes/foetus (●— ●).
Developmental changes in the haematological state of strain 129 foetal mice. Erythrocyte concentration in the peripheral circulation (○— ○) is shown in relation to haemoglobin content/unit of body weight (▫— ▫); haemoglobin content/ unit volume of blood (△— △); and total erythrocytes/foetus (●— ●).
The progression of erythropoiesis in the foetal liver is shown in Fig. 2. The proportion of erythroblasts in the foetal liver rises rapidly to a maximum on day 13, is fairly stable until day 15, but then declines very sharply, so that on day 16 and until birth recognizable erythroblasts represent less than 30% of the total cellular complement of the foetal liver. A similar progression is seen in the relative proportions of early and late erythroblasts.
Developmental changes in the foetal liver. Total erythroblasts are shown as a percentage of total liver cells (○— ○) together with the changing proportions of early erythroblasts (pro- and basophilic) (△— △) and late erythroblasts (poly- and orthochromatic) (▫— ▫). The arrows indicate the total number of liver cells characteristic of each day of gestation.
Developmental changes in the foetal liver. Total erythroblasts are shown as a percentage of total liver cells (○— ○) together with the changing proportions of early erythroblasts (pro- and basophilic) (△— △) and late erythroblasts (poly- and orthochromatic) (▫— ▫). The arrows indicate the total number of liver cells characteristic of each day of gestation.
Changes in the sensitivity of liver erythroblasts to erythropoietin in vitro
The effects of erythropoietin on haem synthesis in cultured foetal erythroblasts were examined in two types of culture system, dissociated cell suspensions (Fig. 3) and fragments of foetal livers maintained with their tissue structure intact (Fig. 4). Dissociated foetal liver-cell cultures responded to erythropoietin by increased haem synthesis up to day 16 of gestation. During the period of most rapid haem synthesis in the presence of erythropoietin (24–29 h after explantation) the rate of haem synthesis clearly exceeded that of freshly explanted cells and was two to three times that found when erythropoietin was not present in the culture medium.
Erythropoietin responsiveness of dissociated foetal liver cells in vitro. Data expressed/liver complement of erythroblasts for each day of gestation. Haem synthesis in the first 4 h of culture (○— ○) is compared with that during 25–29 h of culture in the presence (▫— ▫) and absence (△— △) of erythropoietin.
Erythropoietin responsiveness of dissociated foetal liver cells in vitro. Data expressed/liver complement of erythroblasts for each day of gestation. Haem synthesis in the first 4 h of culture (○— ○) is compared with that during 25–29 h of culture in the presence (▫— ▫) and absence (△— △) of erythropoietin.
Erythropoietin responsiveness of foetal liver tissue in vitro. Data expressed/ liver for each day of gestation, cultured in the presence (▫— ▫) and absence (○— ○) of erythropoietin (Ep).
The erythropoietin responsiveness of intact foetal liver tissue was comparable to that observed in dissociated cell cultures prepared from foetuses of similar gestational age (Fig. 4). Similar levels of haem synthesis per liver were obtained in both types of culture. These results indicate that the maintenance of normal cell-cell interactions is not essential for erythropoietin responsiveness or for normal levels of haem synthesis during short-term culture. Changes in the relative proportions of early and late erythroblasts in dissociated foetal livercell cultures are shown in Fig. 5. There is some initial increase in the number of late erythroblasts even in the absence of erythropoietin, but this is only maintained if erythropoietin is present. Early erythroblasts disappear from the culture unless erythropoietin is present. Such findings are consistent with those of others, e.g. Harrison et al. 1973, using various strains of mice.
Changes in the numbers of erythroblasts in dissociated foetal liver cell cultures. Early (▫— ▫) and late (▪— ▪) erythroblasts cultured in the presence of erythropoietin, and early (○— ○) and late (●— ●) erythroblasts without erythropoietin.
Erythroid colony-forming cells in foetal liver
Cells able to form colonies of eight or more haemoglobinized cells after 72 h culture in semi-solid medium occur in the foetal liver until day 17–18 of gestation (Figs. 6, 7). The maximum number of such cells is found on day 16, when they represent 1·5–2% of total liver cells. The population doubling time between days 11 and 16 of gestation is approximately 22 h. Some erythroid colonies can develop even in the absence of added erythropoietin, but from day 14 to day 16 these were clearly exceeded by erythropoietin-dependent colonyforming cells. Day-17 foetal livers contained only colonies which developed without added erythropoietin and day-18 foetal livers contained no erythroid colony-forming cells at all. From day 11 to day 15 of gestation the proportion of colony-forming cells dependent on additional erythropoietin increased, but then declined sharply. In contrast, cells able to form colonies independently of added erythropoietin declined as gestation advanced. The ratio of total erythroid colony-forming cells to erythroblasts (Fig. 8) was 4–5/100 on day 11, and fell to about half this value until day 15, before reaching a second peak on day 16. The ratio of erythropoietin-dependent colony-forming cells to erythroblasts was 1/100 until day 13, doubling until day 16, and then declining rapidly.
The number (○— ○) and percentage (●— ●) of all erythroid colony-forming cells in foetal liver, and number (▫— ▫) and percentage (▪— ▪) of cells able to form colonies without exposure to erythropoietin in vitro.
The absolute number (○— ○) and percentage (● —●) of erythroid colony-forming cells dependent on continued exposure to erythropoietin in vitro.
The relationship between erythroid colony-forming cells and erythroblasts in foetal livers; total colony-forming cells (○— ○), those independent of erythropoietin in vitro (▫— ▫), and erythropoietin-dependent colony-forming cells (● —●).
Erythroid colony-forming cells in foetal spleen and bone marrow
Foetal spleen contains erythroid colony-forming cells by days 16–17 of gestation (Table 1) and these increase up to birth with a population doubling time of approximately 20 h, similar to that found in the foetal liver. The femoral bone marrow also contains erythroid colony-forming cells by days 16–17 of gestation, and these increase at a similar rate to those in the spleen in the 2–3 days immediately preceding birth.
Granulocyte-macrophage colony-forming cells in the foetal liver
The number of granulocyte-macrophage colony-forming cells in the foetal liver reaches a maximum during days 15–16 of gestation and then remains fairly stable until birth. The proportion of colony-forming cells in the liver is maintained at about 0·4% of total liver cells (Fig. 9) but is more variable when expressed as a proportion of non-erythroid liver cells, reaching a peak (about 2%) on day 13, but declining in the 4–5 days preceding birth.
Granulocyte-macrophage colony-forming cells in foetal spleen and bone marrow
The foetal spleen already contains a significant number of granulocytemacrophage colony-forming cells by days 16–17 of gestation, and on days 18 and 19 the numbers remain fairly constant, at about 300 CFUc/spleen. Such cells are rare in day 16–17 femoral bone marrow but increase during days 18–19 of gestation (Table 1).
DISCUSSION
The granulocyte-macrophage colonies which develop in semi-solid media in the presence of colony-stimulating factor are derived from single cells that make up a population intermediate between the multi-potent haemopoietic stem cells, and the histologically recognizable granulocyte precursors. Such populations, definable in terms of their kinetic properties and their dependence on specific growth-regulating factors for further differentiation, may conveniently be termed the ‘progenitor cell compartments’, and are specific for each line of haemopoietic differentiation (Metcalf & Moore, 1971). Granulocyte-macrophage progenitor cells, capable of forming colonies in vitro, have been detected in prenatal mouse yolk sac, liver, spleen and bone marrow (Moore & Metcalf, 1970; Moore & Williams, 1973), although these organs lack marked granulocyte differentiation in utero. The characteristics of cells which give rise to clonal erythroid colonies in semi-solid media are less well defined. Transfusion-induced polycythaemia, accompanied by reduced plasma erythropoietin levels, reduces the numbers of erythroid colony-forming cells in postnatal spleen and bone marrow, so that at least some of these erythroid precursors must themselves be dependent on erythropoietin for differentiation or survival in vivo (Gregory et al. 1973).
There is no general agreement on the minimum number of cells constituting an ‘in vitro erythroid colony’. In the present study we have adopted the minimum number, eight, used by other workers (Gregory et al. 1973) when undertaking kinetic studies of erythroid colony-forming cells in postnatal haemopoietic tissue. Recent studies (Cooper et al. 1974) using methyl cellulose as a support medium have shown that both the number of erythroid precursor cells triggered to form colonies, and the number of haemoglobinized cells within each colony, are dependent on the concentration of erythropoietin. Using conditions comparable to those of our present report, 60%–85% of colonies derived from early foetal livers were reported to contain more than eight cells, and 30%–60% of colonies contained more than 16 cells.
Prenatal haemopoietic tissues appear to contain two types of erythroid colony-forming cell, one dependent on the continued presence of relatively high levels of erythropoietin during culture, the other, able to form colonies without addition of further erythropoietin. The second type may represent relatively further differentiated cells, already triggered by erythropoietin in the foetal circulation. Alternatively, there may exist some potential colony-forming cells that are highly sensitive to erythropoietin and can therefore react to the low levels present in the serum and plasma in the culture medium. Neither the absolute nor the relative numbers of erythroid colony-forming cells in developing foetal haemopoietic tissue can be easily correlated with changes in numbers or proportions of recognizable erythroblasts. In particular, the highest proportion of erythroid colony-forming cells occurs during days 14–15, 2 days after the peak in early erythroblasts. It is therefore unlikely that the colony-forming cell is identical to a histologically recognizable early erythroblast.
Kinetic studies of erythropoiesis in prenatal mice (Paul et al. 1969; Tarbutt & Cole, 1970; Wheldon et al. 1974) suggest that there are four to five cell divisions in vivo between entry into the pro-erythroblast compartment and the formation of orthochromatic erythroblasts. If it is assumed that the cloning efficiency of the erythropoietin-dependent colony-forming cell is fairly high, i.e. not less than 20% (as found for the granulocyte-macrophage progenitor by Metcalf & Moore, 1970) it is likely that this cell type is separated from the earliest recognizable pro-erythroblast by one or two divisions, since colony-forming cells are present in foetal liver at a level of 1–2/100 erythroblasts.
The period of rapid rise in absolute numbers of erythroid colony-forming cells in the foetal livers and their increase relative to liver cells is closely correlated with the period during which dissociated erythroblasts respond to erythropoietin by increased haem synthesis in vitro, i.e. up to day 15–16 of gestation. This is also the period during which the amount of haemoglobin, relative to the mass of the embryo and associated parameters (Fig. 1), is increasing most rapidly. A similar trend is followed by those erythroid colonyforming cells requiring erythropoietin in vitro and by those seemingly independent of further stimulation. The period of relative stability of the parameters describing the peripheral circulation, e.g. red cell concentration, which begins on day 16–17 of gestation, is also characterized by loss of erythropoietin sensitivity in explanted foetal liver cells and by a rapid loss of erythroid colonyforming cells. Erythroid colony-forming cells in the foetal liver therefore only appear and increase during a period of great demand for, and rapid production of, erythrocytes, and while the total haemoglobin content of the foetus is low relative to its mass. This suggests that the maintenance of such colony-forming cells in the foetal liver is associated with relatively high erythropoietin levels in the foetal circulation. In prenatal tissue as in postnatal (Gregory et al. 1973), at least a proportion of the in vitro erythroid colony-forming cells appear dependent on erythropoietin for their production, as well as their expression.
Initiation of haemopoiesis in foetal spleen and bone marrow follows immigration of precursor cells, probably from the foetal liver, around day 15–16 of gestation (Petrakis, Pons & Lee, 1969; Metcalf & Moore, 1971). Both types of erythroid colony-forming cells found in foetal liver are already represented in day 16–17 spleen and femoral bone marrow. They increase in number, with a population doubling time similar to that found in foetal liver, over the next 2–3 days, while numbers in the liver decline. Cessation of production of erythroid colony-forming cells by the foetal liver may therefore also involve changes in the liver microenvironment which render it less able to support erythropoiesis (Trentin, 1970). Levels of haemoglobin/unit of body weight also begin to decline on day 17–18, so that the proliferation of erythroid colonyforming cells in both prenatal spleen and bone marrow can be correlated with an increased demand for erythrocytes.
The progenitor cells forming granulocyte and macrophage colonies in vitro were found in foetal liver, spleen and bone marrow, together with erythroidcolony progenitor cells. The sharp reduction in the number of erythroid progenitor cells, in later foetal liver, was not matched by any corresponding increase in granulocyte-macrophage progenitors. This underlines the independence of such haemopoietic cell compartments. The in vitro development of granulocyte-macrophage progenitors from all three prenatal tissues was predominantly towards macrophage expression. Although this tendency was enhanced by use of the rat-erythrocyte lysate (Bradley et al. 1974), which also markedly reduced the proportion of clusters (i.e. colonies with fewer than 40 cells), it is consistent with the relative frequency of macrophages in prenatal mouse haemopoietic tissues and the corresponding lack of granulocyte differentiation.
This work was supported by the Medical Research Council.
Erythropoietin was supplied by the Committee on Erythropoietin of the U.S. National Heart and Lung Institute. It was procured by the Department of Physiology, University of the Northeast, Corrientes, Argentina, and processed by the Haematology Research Laboratories, Children’s Hospital of Los Angeles.