Erythropoietic activity of foetal rat femoral marrow was examined during the last four days of intra-uterine life. Insignificant at day 18, it develops slowly thereafter until birth. In the non-suckled neonate (not older than two hours), it appears notably enhanced. In order to test the potential of the foetal marrow to develop precocious or increased erythropoiesis, the activity of the erythropoietic organ predominant at this time, the liver, was altered by modifying the level of circulating corticosteroids which govern its function. Maturation and involution of the hepatic erythron were prevented by corticosteroid deprivation of the foetus (maternal adrenalectomy and foetal hypophysectomy). Precocious maturation and exhaustion of the hepatic erythron was induced by submitting foetuses to corticosteroids excess from day 14. Both corticosteroid deprivation and excess increase the erythropoietic activity of the femoral marrow. This activity can reach and even exceed by day 20 of intrauterine life that in neonatal marrow. Foetal hepatic erythron misfunction can therefore initiate and stimulate bone marrow erythropoiesis. The study of circulating red blood cells demonstrates that: (1) anaemia initiates medullary erythropoietic activity; (2) this anaemia is largely corrected by the bone marrow. The regulatory mechanism is presumably erythropoietin mediated.

It is well known that in most mammalian species, the foetal liver assumes a major role in haematopoiesis, especially erythropoiesis (see for instance Ackerman, Grasso, & Knouff, 1961; Grasso, Swift & Ackerman, 1962; Rifkind, Chui & Epler, 1969). In the rat foetus, the cellularity of the hepatic erythron decreases sharply after day 18 of gestation (Nagel, 1968). This decrease is governed by the circulating corticosteroids: in foetuses largely deprived of these hormones by maternal adrenalectomy and foetal hypophysectomy, further evolution of liver erythropoietic tissue is prevented, but can be restored by hydrocortisone administration (Nagel & Jacquot, 1969). Conversely, stressing the pregnant rat by laparotomy results in an accelerated evolution and a precocious regression of the foetal hepatic erythron, which is evident 24 h after the stress, whereas previous maternal adrenalectomy suppresses the effects of the laparotomy (Nagel & Jacquot, 1968). The foetal hepatic erythron responds to corticosteroids as early as day 14 of gestation (Billat, Nagel, Nagel & Jacquot, 1980). It is therefore experimentally possible to prematurely deprive the foetus of a large part of its hepatic erythropoietic tissue by repeated stress inflicted on the pregnant mother during and after day 14.

Little information on foetal marrow erythropoiesis in the rat is available. Lucarelli et al. (1968) assert that there is no detectable erythropoietic activity in the rat foetal femur at day 18 of gestation; this activity appears at birth and increases strikingly towards the seventh day of post-natal life. Therefore, we studied the evolution of haematopoietic activity in the foetal rat femur from day 18 of gestation until birth. Then, we looked for an eventual modification of femoral erythropoietic activity in foetuses: (1) where normal regression of the hepatic erythron was prevented by maternal adrenalectomy and foetal hypo-physectomy; (2) where evolution of the hepatic erythron was precociously initiated and accelerated by repeated maternal laparotomies at days 14, 16 and 18 of gestation.

In parallel, the cellular content of the blood was investigated in each experimental situation.

Animals

Wistar rats (CF strain of the C.N.R.S.) were used. They were housed in a constant temperature room with 12 h day/12 h night. They had free access to water and food (UAR rat commercial food). Coitus was assessed by the presence of spermatozoa in the morning vaginal smear. In this strain of rats, delivery generally occurs during the night between days 21 and 22 of pregnancy or at day 22 in the morning.

Surgery

Surgery was performed under ether anaesthesia. Laparotomy consists of an incision of skin, muscular wall and peritoneum followed by suture, without touching the viscera. Adrenalectomy of the mother was performed classically, by the dorsal approach, on day 14 of gestation. Foetal hypophysectomy was performed by decapitation according to the technique of Jost (1951) on day 18 of gestation.

Cortisol was administered through the maternal uterine wall under the skin of the foetuses (0·6 mg of hydrocortisone acetate Roussel).

Evaluation of liver erythropoiesis

The evolution of the liver erythropoietic function was evaluated according to Nagel’s methodology (1968). In brief, livers were homogenized in ice-cold 0-25 M-sucrose (Potter-Elvehjem, standardized number of strokes and of revolutions per minute), the nuclear suspension obtained was enumerated in a haemocytometer and the number of nuclei per gram liver calculated. In previous studies (Nagel, 1968; Jacquot & Nagel, 1976) the number of hepatic cells per liver volume has been shown to be roughly constant at all gestational stages and in all experimental groups; thus, any change in the number of nuclei per g liver reflects a change in the number of nuclei of haematopoietic cells per g liver. Since erythropoietic cells are considerably more numerous than other blood cells in the liver haematopoietic tissue, experimentally induced differences in the number of nuclei in homogenates reflect experimentally induced differences in the importance of the liver erythron.

Cytological determinations on bone marrow

To study medullary erythropoiesis, femoral marrow imprints were used. The bone was rapidly dissected, split longitudinally and stuck on a slide. The imprints were fixed and stained with May-Gruenwald-Giemsa.

Four different cell groups will be considered.

  • ‘Precursor’ cells, i.e. morphologically undifferentiated cells, the progeny of which cannot be predicted on mere cytological observation.

  • Unhaemoglobinized erythroid cells (proerythroblasts + basophilic erythroblasts).

  • Haemoglobinized cells (polychromato and acidophilic erythroblasts).

  • The white cell lineage (myeloid and lymphoid cells, megakaryocytes and mature granulocytes), called for convenience ‘other cells’.

The number of each cellular type per arbitrary surface area of imprint was determined. Medullary activity was assessed by classifying and numbering the different cells present in this arbitrary surface area of the imprint. For each gestational age, control or experimental, at least eight foetuses belonging to four different mothers were used and 3000 to 5000 cells were enumerated and classified. Numbers of cells per surface area are presented as means (±S.E.M.; significance of differences between means was evaluated according to Fisher’s t test. Qualitative composition of imprints is depicted by the percentage of each cellular type versus the whole cellular population of the imprint.

Histological study of bone marrow

The bones, fixed in formol, were treated for paraffin embedding. Sections were stained with haemalum-eosin.

Blood assays

Foetal blood was collected directly from the axillary artery for the determination of: the total number of circulating cells (10 μl), the number of nucleated cells (25μl), the haemoglobin concentration (20 μl), the hematocrit (12μl) and the composition of the cellular population (two smears). The simultaneous determination of these parameters required too much blood to be performed on single foetuses at day 18. Therefore, cellular enumerations and blood smears were carried out on some foetuses, haemoglobin and hematocrit being determined on their littermate brothers. For older stages, all the determinations were performed simultaneously on each foetus.

For each experimental point, 9 to 25 foetuses were studied, belonging to at least four different mothers.

(a) Cellular enumerations

  • Total number of cells (= a) was determined on 10 μl blood samples using the Unopette 585IF (Becton Dickinson) commercial kit for manual methods.

  • Number of nucleated cells (= b) was determined on 25 μl blood samples using the Unopette 5856F (Becton Dickinson) commercial kit for manual methods.

  • Number of anucleated red cells is by definition (a-b).

  • Number of nucleated red cells was determined as follows. Two smears of each blood were fixed and stained with May-Gruenwald-Giemsa. They allowed the determination of the ratio of the nucleated red cells to the total population of nucleated blood cells (more than 1000 nucleated cells were enumerated on smears for each blood sample). The number of circulating nucleated red cells is obtained by multiplying (b) by this ratio.

(b) Haemoglobin concentration

Unopette test 5857F (Becton Dickinson) commercial kit was used on 20 μl blood samples. Cyanmethaemoglobin, oxyhaemoglobin, carboxyhaemoglobin and methaemoglobin are measured. The coloured reaction developed after addition of the Uno-heme reagent was read at 530 nm and the haemoglobin concentration was deduced from a standard calibration curve.

(c) Hematocrit

A standardized micromethod was used for hematocrit determination: 12 μl blood in 15μl Drummond microcaps, 16500 g centrifugation for 15 sec in a Janetzki TH 11 centrifuge.

(d) Presentation of the results

Experimental values are presented in the tables, as means ± S.E.M.: anucleated red cells per μl blood, nucleated red cells per μl blood (assuming, for simplification, that the ratio of the nucleated red cells to the whole nucleated population is determined with a negligible error), haemoglobin content (g/100 ml blood) and hematocrit (volume of the red cells as percentage of the total volume of the blood). Fisher’s t significance test was used for differences between means.

Mean corpuscular haemoglobin (MCH) was calculated as:
formula
Mean corpuscular volume (MCV) was calculated as:
formula
As both MCH and MCV were obtained, for each experimental situation, from the mean values of haemoglobin content, hematocrit and cell numbers, they are reported without statistical information.

(1) Control foetuses (normal mothers)

(a) Marrow haemopoietic activity (Table 1)

The total number of cells per arbitrary surface area of imprint increases regularly from day 18 of gestation until the first hours following birth. The cells belonging to the white myeloid lineage constitute the major part of the marrow population. The number of ‘precursors’ per surface area remains roughly constant at days 19, 20 and 21. The number of immature and mature erythroid cells per surface area increases between days 18 and 21 of gestation but their proportion remains quite low when compared to that of other cell lines (only 4% of marrow cells at day 21 are differentiated erythroid cells).

Table 1

Marrow haematopoietic activity in control foetuses from normal or adrenalectomized mothers and in decapitated foetuses from adrenalectomized mothers

Marrow haematopoietic activity in control foetuses from normal or adrenalectomized mothers and in decapitated foetuses from adrenalectomized mothers
Marrow haematopoietic activity in control foetuses from normal or adrenalectomized mothers and in decapitated foetuses from adrenalectomized mothers

At birth the number of precursors per surface area sharply decreases. In contrast, the number per surface area of the erythroid cells and, more particularly, of the haemoglobinized ones, increases (in the newborn, differentiated erythroid cells represent 9·2% of the whole population). Cells of the white lineage continue to increase after birth. Amongst these cells, the evolution of granulocytes is spectacular: they represent 0·03, 1·l, 6·9, 15·4 and 29·5% of the total population of the imprints at 18, 19,20, 21 foetal days and at birth respectively (not shown on Table 1).

(b) Blood data (Table 2)

The number of circulating anucleated red cells increases between days 18 and 21 of gestation. During the first hours following birth, this increase is very rapid. On the contrary, the number of nucleated red cells falls sharply between days 18 and 19, and then decreases more slowly during the last two days of gestation and after birth. At day 18, the percentage of red cells from yolk-sac origin as determined on blood smears is very high: 49·8% of total nucleated red cells. This percentage considerably decreases thereafter so that at birth, cells from yolk-sac origin have almost disappeared. The percentage of acidophilic erythroblasts increases between day 18 and birth, and that of younger erythroblasts (polychromatophilic and basophilic), roughly stable until day 20, falls after this stage (Fig. 1).

Table 2

Blood data in control foetuses from normal or adrenalectomized mothers and in decapitated foetuses from adrenalectomized mothers

Blood data in control foetuses from normal or adrenalectomized mothers and in decapitated foetuses from adrenalectomized mothers
Blood data in control foetuses from normal or adrenalectomized mothers and in decapitated foetuses from adrenalectomized mothers
Fig. 1

Evolution of nucleated red cells from yolk-sac origin (●), acidophilic erythroblasts (▾) and younger (polychromatophilic and basophilic) erythroblasts (○) as percentages of total nucleated red cells.

Fig. 1

Evolution of nucleated red cells from yolk-sac origin (●), acidophilic erythroblasts (▾) and younger (polychromatophilic and basophilic) erythroblasts (○) as percentages of total nucleated red cells.

Haemoglobin concentration slightly increases between day 18 and birth and MCH decreases slightly.

Between day 18 and birth, hematocrit values stay quite stable. MCV diminishes between day 18 and birth, especially after day 21.

(II) Corticosteroid deprivation (decapitated foetuses)

Maternal adrenalectomy (on day 14) and foetal hypophysectomy (on day 18) reduce the level of foetal circulating corticoids (cf. for instance Arishima et al. 1977).

(a) Marrow erythropoietic activity (Table 1)

At all foetal ages there is no significant difference between intact foetuses from normal and from adrenalectomized mothers. At day 19, the values obtained from decapitated foetuses do not differ significantly from those of intact foetuses. In contrast, at days 20 and 21, in corticosteroid-deprived foetuses, a striking increase in the number of unhaemoglobinized and haemoglobinized cells per surface area occurs. These differences between decapitated foetuses and controls are statistically significant (P < 0·01). The same conclusions can be drawn when the results are expressed as percentages of the whole marrow population.

(b) Blood data (Table 2)

The number of anucleated red cells in intact foetuses from adrenalectomized mothers is not significantly different (P > 0·05) from that found in foetuses from normal mothers at any gestational age.

In contrast, the number of nucleated red cells presents a significant difference (P < 0·01) between foetuses from adrenalectomized mothers and their normal controls at days 19, 20 and 21.

Haemoglobin concentration, although slightly higher in foetuses from adrenalectomized mothers, is not really significantly different from that of normal controls and MCH is roughly identical at every gestational age. Hematocrit and MCV present no difference at any gestational age between foetuses from adrenalectomized mothers and normal controls.

In decapitated foetuses, the number of anucleated red cells at day 19 is significantly lower (P < 0·01) than in intact foetuses from normal or adrenalectomized mothers. On the contrary at days 20 and 21, this number is higher (P < 0·01). The number of nucleated red cells is more elevated (P < 0·01) in decapitated foetuses, except for day 21.

Haemoglobin concentration is significantly lowered in decapitated foetuses at day 19 (P < 0·01); this difference disappears thereafter. MCH in decapitated foetuses is below the control values at all gestational stages. Hematocrit at day 19, is significantly lower in decapitated foetuses than in their controls but is no longer different thereafter. MCV in decapitated foetuses is always below the controls.

(III) Corticosteroid excess; repeated maternal stress

Repeated laparotomies were performed on pregnant rats from day 14. At this early stage the foetal liver erythron is already responsive to corticosteroids (Billat et al. 1980). Mothers were stressed at days 14 and 16 (sampling at day 18), 14–16 and 18 (sampling at days 19 or 20), 14–16–18 and 20 (sampling at day 21). The results are compared to those of foetuses from normal mothers.

(a) Marrow erythropoietic activity (Table 3)

In 18-day-old foetuses from mothers laparotomized at days 14 and 16, the number of erythroid cells per surface area of marrow imprint is not significantly different from that observed in foetuses from normal mothers at the same gestational age (P > 0·05). On the contrary, this value is statistically much higher (P < 0·01) in 19, 20 and 21-day-old foetuses borne by stressed mothers, than in foetuses from normal mothers at the same gestational ages. At day 19, the increase of unhaemoglobinized cells, and, at day 20, that of haemoglobinized erythroblasts, are particularly noticeable.

Table 3

Marrow haematopoietic activity in control foetuses from normal mothers and in foetuses from mothers submitted to repeated stress

Marrow haematopoietic activity in control foetuses from normal mothers and in foetuses from mothers submitted to repeated stress
Marrow haematopoietic activity in control foetuses from normal mothers and in foetuses from mothers submitted to repeated stress

When percentages referring to the whole marrow population are considered there is, again, after day 18 a difference between experimental foetuses and their controls, especially at day 19 for unhaemoglobinized erythroid cells and at day 20 for haemoglobinized erythroblasts.

It is noteworthy that 21-day-old foetuses from four-fold-stressed mothers do not differ significantly from 20-day-old foetuses from three-fold-laparotomized mothers (P > 0·05). In both cases, cell numbers and percentages are higher than in 21-day-old control foetuses borne by normal mothers.

(b) Blood data (Table 4)

The number of anucleated red cells is significantly lowered at days 18 and 19 in experimental foetuses (P < 0·01); this difference is no longer observed thereafter. The numbers of nucleated red cells, slightly higher in experimental foetuses at days 18 and 19 (0·05 > P > 0·01), are thereafter analogous to control values.

Table 4

Blood data in control foetuses from normal mothers and in foetuses from mothers submitted to repeated stress

Blood data in control foetuses from normal mothers and in foetuses from mothers submitted to repeated stress
Blood data in control foetuses from normal mothers and in foetuses from mothers submitted to repeated stress

Haemoglobin concentration is unchanged in experimental foetuses, and MCH, slightly increased at days 18 and 19, is the same as in controls at days 20 and 21. Hematocrit, lower at day 18 in experimental foetuses (0·5 > P > 0·01) is not different from controls at other gestational stages, and MCV, increased at day 19 in experimental foetuses, is not different from controls at other ages.

(IV) Control experiments

  • Foetuses (adrenalectomized mothers) were decapitated at 18 days and simultaneously injected with cortisol. Blood measurements were performed at 19 days. As shown in Fig. 2, the number (per μl) of anucleated red cells was higher than in non-injected decapitated foetuses (P < 0·01) and equal to the values observed in intact foetuses from normal or adrenalectomized mothers.
    Fig. 2

    Cortisol injection at day 18 to decapitated rat foetuses: effect on the number of anucleated red cells at day 19.

    Fig. 2

    Cortisol injection at day 18 to decapitated rat foetuses: effect on the number of anucleated red cells at day 19.

  • Foetuses, decapitated at 18 days (adrenalectomized mothers), were cortisol-injected at 20 days and sampled at 21 days. As shown in Table 5, they presented a significant (P < 0·01, versus non-injected decapitated foetuses) decrease in the number of unhaemoglobinized erythroblasts per surface area of marrow imprint, and a concomitant increase in the number of more mature erythroblasts. The number of their circulating red cells (anucleated and nucleated) was significantly increased (P < 0·01) (Table 5).

    Table 5

    Effects of hydrocortisone administration at day 20 to decapitated foetuses on femoral marrow erythropoietic activity and number of anucleated and nucleated red blood cells at day 21

    Effects of hydrocortisone administration at day 20 to decapitated foetuses on femoral marrow erythropoietic activity and number of anucleated and nucleated red blood cells at day 21
    Effects of hydrocortisone administration at day 20 to decapitated foetuses on femoral marrow erythropoietic activity and number of anucleated and nucleated red blood cells at day 21

  • Liver cellularity (number of nuclei per volume unit of liver) was evaluated, between 18 and 21 days, in intact foetuses from normal and adrenalectomized mothers (Fig. 3), in decapitated foetuses from adrenalectomized mothers (Fig. 4) and in foetuses borne by stressed mothers (Fig. 5). The results previously described (Nagel & Jacquot, 1968; Nagel & Jacquot, 1969; Jacquot & Nagel 1976) are confirmed. Corticosteroid deprivation prevents the decrease of liver cellularity normally observed as pregnancy progresses; corticosteroid excess, on the contrary, induces an early decrease. Cortisol administration at day 20 to foetuses decapitated at day 18 induces a rapid decrease of liver cellularity (Fig. 4).
    Fig. 3

    Evolution of the hepatic haematopoietic tissue in foetuses from normal (●) and adrenalectomized (▾) mothers.

    Fig. 3

    Evolution of the hepatic haematopoietic tissue in foetuses from normal (●) and adrenalectomized (▾) mothers.

    Fig. 4

    Evolution of the hepatic haematopoietic tissue in corticosteroid deprived foetuses (decapitated: ○). Effect of cortisol administration at day 20 (●).

    Fig. 4

    Evolution of the hepatic haematopoietic tissue in corticosteroid deprived foetuses (decapitated: ○). Effect of cortisol administration at day 20 (●).

    Fig. 5

    Anticipated and accelerated evolution of the hepatic haematopoietic tissue in foetuses submitted to corticosteroid excess (▵) by maternal repeated stress as compared to normal controls (●).

    Fig. 5

    Anticipated and accelerated evolution of the hepatic haematopoietic tissue in foetuses submitted to corticosteroid excess (▵) by maternal repeated stress as compared to normal controls (●).

In the foetal rat the adrenals become significantly functional at day 17-18 (Holt & Oliver, 1968; Kamoun, 1970; Cohen, 1973; Milković, Milković & Paunovic, 1973; Dupouy, Coffigny & Magre, 1975; Martin, Cake, Hartmann & Cook, 1977). In intact foetuses borne by adrenalectomized mothers, they are particularly active (Milkovic, Paunovic, Kniewald & Milkovic, 1973 a; Cohen & Brault, 1974), but remain largely inactive in decapitated foetuses (Arishima et al. 1977; Cohen & Brault, 1974; Dupouy et al. 1975; Milković et al. 1973a).

(I) Normal evolution of prenatal erythropoiesis

At the end of intrauterine life in the rat, the number of anucleated red cells increases with age while MCV decreases; this is quite obvious during the first hours after birth. The anucleated red cells produced in the last two days of gestation and after birth are thereafter smaller than the ones produced earlier.

The functioning of the hepatic erythron has direct repercussions on the composition of foetal blood: around day 18, a new population of red cells, differing in size and morphology from those of yolk-sac origin, appears in the blood (Nagel, 1972). Similar conclusions were reported in the foetal mouse (Craig & Russel, 1964: Barker, 1968; Brotherton, Chui, Gauldie & Patterson, 1979). Accordingly, we observed, particularly between days 18 and 19, a rapid decrease in both the number and the percentage of nucleated red cells of yolk-sac origin.

Around birth, two other erythropoietic organs may be implicated: the spleen and the bone marrow. The role of the spleen has not been considered in the present paper: preliminary results suggest that the supply of erythrocytes by the spleen is not important before birth in the rat (cf. also Lucarelli et al. 1968).

Under normal conditions, the bone marrow of the foetal rat contains only few differentiated erythroid elements and the increase of its cell population during the last three days of foetal life concerns essentially other lineages. At day 21, only 4% of the whole population are differentiated erythroid cells, and at birth, 9·2%. This last value is in good agreement with the one published by Lucarelli et al. (1968) for the neonatal rat.

(II) Experimentally induced modifications of prenatal erythropoiesis

The corticosteroid environment of the foetus was modified in two opposite ways: deprivation (maternal adrenalectomy and foetal decapitation) or early excess of these hormones (repeated maternal stress). Paradoxically both modifications produced a similar effect: an increase of the erythropoietic tissue in the bone marrow. The first conclusion which can be drawn is that the foetal bone marrow is able, as early as day 20 of gestation, to adjust its erythropoietic activity. The similarity of the effects produced by two apparently opposed procedures suggests that, in fact, these procedures lead to a common physiological situation, as far as foetal erythropoiesis is concerned.

In corticosteroid-deprived foetuses, the stability of the hepatic haematopoietic tissue was established by Nagel & Jacquot (1969) and Jacquot & Nagel (1976) (and here, Fig. 4). Such a ‘frozen’ situation strongly suggests a block in the delivery of circulating red cells by the liver, at a time when the foetal growth is rapid. According to this hypothesis, the precocious increase of the erythroid line in the bone marrow observed in decapitated foetuses occurs in compensation for precocious cessation of liver erythron functioning.

In fact, at day 19, an anaemia is detectable in decapitated foetuses: the number of anucleated red cells and haemoglobin concentration are lowered. By day 20, the number of anucleated red cells is more than restored. Accordingly, haemoglobin concentration increases between days 19 and 20, but MCH falls. These observations suggest that these new red cells are supplied by the stimulated bone marrow and not by the liver. These red blood cells from medullary origin appear to be smaller than those from hepatic origin.

The anaemia observed at day 19 in decapitated foetuses is really due to a hepatic misfunction and not to a surgical haemorrhage for it is lacking in foetuses decapitated at day 18 and simultaneously injected with cortisol.

Cortisol administration to decapitated foetuses at day 20 restores the efficiency of their liver erythropoiesis. Twenth-four hours later, part of their liver erythroid line has been discharged as suggested by the decrease in nuclear count of liver homogenates and by the fact that the number of erythroid blood cells is significantly higher than in non-treated decapitated controls. This suggests that both anucleated and nucleated red cells are delivered by the liver.

In these cortisol-injected decapitated foetuses, the suddenly resumed supply of blood cells from the hepatic erythron appears to block erythropoiesis in the marrow (this medullary erythropoiesis is initially stimulated by the failure of hepatic erythropoiesis). The number (by imprint surface area) and the percentage of unhaemoglobinized erythroid cells, which were both increasing rapidly decrease sharply after cortisol injection, while the more mature haemoglobinized cells accumulate in the marrow, and are not liberated as circulated reticulocytes.

Repeated maternal laparotomies at 14, 16 and 18 days reduce the number of nuclei in liver homogenates from 18- and 19-day-old foetuses. Here again, it is tempting to speculate that the liver, prematurely drained of the major part of its erythropoietic cells is unable to deliver enough red cells to the blood.

In agreement with this hypothesis, red blood cells, and especially anucleated ones, are less numerous at 18 and 19 days (in fact this situation may well be present already at 17 days, stress having been inflicted at 14 and 16 days). Afterwards, the numbers of anucleated red cells reach the control values, and an increased number of reticulocytes is observed at days 19 and 20. It is therefore logical to think that, here again, alterations in red blood cells number due to hepatic deficiency brought about a medullary response. Nevertheless, foetuses from stressed mothers differ in their behaviour from decapitated foetuses: their liver erythron, although already depleted by the two first maternal laparotomies at days 14 and 16, is still able to react to a third one at day 18. As a consequence, the decrease in the number of their circulating red cells, although already present by day 18, is not so drastic (haemoglobin concentration is not affected) and their medullary response, although occurring earlier (at day 19), is less pronounced.

It is also possible that the glucocorticoids act directly on marrow erythroid cells. The status of effects, tested with in vitro systems, is controversial (Golde, Bersh & Cline, 1976), Singer, Samuels & Adamson (1976), Gidari & Levere (1979), Urabe, Hamilton & Shigeru (1979) and Zalman, Maloney & Pratt (1979)). In our study, erythropoiesis in the marrow is not stimulated at day 21 by the fourth maternal laparotomy, but is considerably increased in corticoid-deprived foetuses. Such results do not sustain the hypothesis of corticosteroids acting directly on the marrow.

The case of intact foetuses borne by adrenalectomized mothers (littermate brothers of decapitated foetuses) deserves some comment. Their hepatic erythron presents an accelerated evolution (Nagel & Jacquot, 1969) and seems to deliver to the blood an increased supply of nucleated and perhaps also anucleated red cells.

Misfunction of the hepatic erythron, whether induced by glucocorticoid deficiency or by glucocorticoid excess, and leading to transitory anaemia, is associated with premature initiation of medullary erythropoiesis. It is tempting to believe that the bone marrow activity is triggered by an erythropoietin-mediated mechanism of adult type, as suggested by Matoth & Zaizov (1971) and reported by Meberg (1980) in the case of foetal hypoxia. Preliminary data indicate that plasma erythropoietin-like activity is higher in both types of experimental foetuses.

This work was supported in part by the DGRST Grant no. 77-7-0673.

Ackerman
,
G. A.
,
Grasso
,
J. A.
&
Knouff
,
R. A.
(
1961
).
Erythropoiesis in the mammalian embryonic liver as revealed by electron microscopy
.
Lab. Invest
.
10
,
787
796
.
Arishima
,
K.
,
Nakama
,
S.
,
Morikawa
,
Y.
,
Hashimoto
,
Y.
&
Egushi
,
U.
(
1977
).
Maternalfoetal interrelations of plasma corticosterone concentrations at the end of gestation in the rat
.
J. Endocr
.
72
,
239
240
.
Barker
,
J. E.
(
1968
).
Development of the mouse hematopoietic system. I. Types of hemoglobin produced in embryonic yolk sac and liver
.
Devl Biol
.
18
,
14
29
.
Billat
,
C.
,
Nagel
,
M. D.
,
Nagel
,
J.
&
Jacquot
,
R.
(
1980
).
Early reactivity of liver erythropoietic tissue of the rat foetus towards glucocorticoids
.
Biol. Cell
.
38
,
3
, 187-194.
Brotherton
,
T. W.
,
Chui
,
D. H. K.
,
Gauldie
,
J.
&
Patterson
,
M.
(
1979
).
Hemoglobin ontogeny during normal mouse fetal development
.
Proc. natn. Acad. Sci., U.S.A
.
76
,
6
, 2853-2857
Cohen
,
A.
(
1973
).
Plasma corticosterone concentration in the foetal rat
.
Horm. Metab. Res
.
5
,
1
,66.
Cohen
,
A.
&
Brault
,
S.
(
1974
).
Modifications de la concentration en corticosterone des surrénales et du plasma chez le foetus de rat après surrénalectomie maternelle
.
C.r. hebd. Séanc. Acad. Sci. Paris
.
278
,
1755
1758
.
Craig
,
M. L.
&
Russel
,
E. S.
(
1964
).
A developmental change in hemoglobins correlated with an embryonic red cell population in the mouse
.
Devl Biol
.
10
,
191
201
.
Dupouy
,
J. P.
,
Coffigny
,
H.
&
Magre
,
S.
(
1975
).
Maternal and foetal corticosterone levels during late pregnancy in rats
.
J. Endocr
.
65
,
347
352
.
Gidari
,
A. S.
&
Levere
,
R. D.
(
1979
).
Glucocorticoid mediated inhibition of erythroid colony formation by mouse bone marrow cells
.
J. Lab. Clin. Med
.
93
,
5
, 872-878.
Golde
,
D. W.
,
Bersh
,
N.
&
Cline
,
M. J.
(
1976
).
Potentiation of erythropoiesis in vitro by dexamethasone
.
J. Clin. Invest
.
57
,
57
62
.
Grasso
,
J. A.
,
Swift
,
H.
&
Ackerman
,
G. A.
(
1962
).
Observations on the development of erythrocytes in mammalian fetal liver
.
J. Cell. Biol
.
14
,
235
254
.
Holt
,
P. G.
&
Oliver
,
I. T.
(
1968
).
Plasma corticosterone concentrations in the perinatal rat
.
Biochem. J
.
108
,
339
341
.
Jacquot
,
R.
&
Nagel
,
J.
(
1976
).
Glucocorticoids and evolution of haemopoietic tissue in the liver of the rat foetus
.
Ann. Immunol. (nst. Pasteur
.
126 C
,
895
903
.
Jost
,
A.
(
1951
).
La physiologie de l’hypophyse foetale
.
Biol. Med
.
40
,
205
229
.
Kamoun
,
A.
(
1970
).
Activité cortico-surrénale au cours de la gestation, de la lactation et du dévelopment pré et post natal chez le rat
.
J. Physiol
.
Paris
62
,
5
32
.
Lucarelli
,
G.
,
Porcellini
,
A.
,
Carnevali
,
C.
,
Carmena
,
A.
&
Stohlman
,
F.
Jr
. (
1968
).
Fetal and neonatal erythropoiesis
.
Ann. N.Y. Acad. Sci
.
149
,
544
559
.
Martin
,
C. E.
,
Cake
,
M. H.
,
Hartmann
,
P. E.
&
Cook
,
I. F.
(
1977
).
Relationship between foetal corticosteroids, maternal progesterone and parturition in the rat
.
Acta endocrinol
.
84
167
176
.
Matoth
,
Y.
&
Zaizov
,
R.
(
1971
).
Regulation of erythropoiesis in the fetal rat
.
Israel J. Med. Sc
.
7
,
839
843
.
Meberg
,
A.
(
1980
).
Plasma erythropoietin levels in fetal and newborn rats: response to hypoxia
.
Expl Hemat
.
8
,
5
,
615
619
.
Milkoviô
,
K.
,
Paunonió
,
J.
,
Kniewald
,
Z.
&
Milkovic
,
S.
(
1973
).
Maintenance of the plasma corticosterone concentration of adrenalectomized rat by the foetal adrenal glands
.
Endo
.
93
,
115
118
.
Milkovic
,
S.
,
Milkovic
,
K.
&
Paunovió
,
J.
(
1973
).
The initiation of foetal adrenocorticotrophic activity in the rat
.
Endo
,
92
,
380
384
.
Nagel
,
J.
(
1968
).
Le tissu hématopoïetique dans le foie foetal de rat en fin de gestation. L Evolution normale au cours du dernier quart de la gestation
.
Archs Anat. micr. Morph. Exp
.
57
,
99
105
.
Nagel
,
J.
(
1972
).
Le tissu hématopoïetique dans le foie foetal de rat en fin de gestation. IV. Conséquences de son évolution sur la population érythrocytaire circulante
.
Archs Anat. micr. Morph. Exp
.
61
,
301
312
.
Nagel
,
J.
&
Jacquot
,
R.
(
1968
).
Le tissu hématopoïetique dans le foie foetal de rat en fin de gestation. II. Influence de la laparotomie et de la surrénalectomie maternelle
.
Archs Anat. micr. Morph. Exp
.
57
,
99
105
.
Nagel
,
J.
&
Jacquot
,
R.
(
1969
).
Le tissu hématopoïetique dans le foie foetal de rat en fin de gestation. III. Influence de l’hypophysectomie foetale
.
Archs Anat. micr. Morph. Exper
.
58
,
97
104
.
Rifkind
,
R. A.
,
Chui
,
D.
&
Epler
,
H.
(
1969
).
An ultrastructural study of early morphogenetic events during the establishment of fetal hepatic erythropoiesis
.
J. Cell. Biol
.
40
,
343
365
.
Singer
,
J. W.
,
Samuels
,
A. I.
&
Adamson
,
J. W.
(
1976
).
Steroids and hematopoiesis. I. The effect of steroids on in vitro erythroid colony growth: structure/activity relationships
.
J. cell physiol
.
88
,
127
134
.
Urabe
,
A.
,
Hamilton
,
J.
&
Shigeru
,
S.
(
1979
).
Dexamethasone and erythroid colony formation: contrasting effects in mouse and human bone marrow cells in culture
.
Brit. J. Haematol
.
43
,
479
480
.
Zalman
,
F.
,
Maloney
,
M. A.
&
Patt
,
H. M.
(
1979
).
Differential response of early erythropoietic and granulopoietic progenitors to dexamethasone and cortisone
.
J. exp. Med
.
149
,
67
72
.