/ . Embryol. exp. Morph. Vol. 50, pp. 57-74, 1979 5 7

Printed in Great Britain © Company of Biologists Limited 1979

The ontogeny of erythropoiesis inthe mouse detected by the erythroid

colony-forming technique

I. Hepatic and maternal erythropoiesis

By IVAN N. RICH1 AND BERNHARD KUBANEKFrom the Department of Internal Medicine and Pediatrics,

Division of Haematology, University of Ulm, West Germany

SUMMARYEmploying the erythroid colony-forming technique, it is shown that throughout hepatic

erythropoiesis in the mouse, the CFU-E population remains sensitive to erythropoietin.Maximum stimulation was achieved during this period using an erythropoietin concentrationof 0075 units/ml. The peak in the CFU-E concentration occurs between the 11th and 12thday while absolute values show a maximum on the 14th day of gestation. These results arediscussed in terms of changing cell populations, both of erythropoietic precursors and hepato-cytes from which it is concluded that at no time during foetal erythropoiesis does the CFU-Epopulation change or become unresponsive to erythropoietin. The BFU-E population followsclosely that of the CFU-E, but declines about 24 h earlier on the 16th day of gestation.

The effect of the foetus on the mother was also studied during the second half of pregnancy.During this period of natural perturbation both femoral and, in particular, splenic erythro-poiesis are increased. However, during this time an erythropoietin concentration of 0-3units/ml was required to maximally stimulate the CFU-E population derived from thesetissues. The fact that both adult and foetal erythroid tissue maintain a rather constantrequirement for erythropoietin for their growth in vitro, indicates that it is an intrinsic propertyof the cells. It is concluded that increased maternal erythropoiesis is due to an increasedoxygen requirement causing hypoxia due to the growing foetus.

INTRODUCTION

Pregnancy in the mouse lasts between 20 and 21 days. During this time,erythropoiesis in both the growing foetus and the pregnant mother undergoeschanges not only in the site of erythropoiesis but also in its homeostatic regula-tion. Foetal haemopoiesis appears between the 7th and 8th day of gestationwhen erythropoiesis is initiated in the yolk sac (Snell & Stevans, 1966; Metcalf &Moore, 1970). Immature erythroblasts enter the foetal circulation from the yolksac at about the 9th day and continue to differentiate into nucleated erythrocytes(Craig & Russell, 1964; de la Chapelle, Fantoni & Marks, 1969; Marks &

1 Author's address: Zentrum fur Klinische Grundlagenforschung, University of Ulm,Parkstrasse 11, D-7900 Ulm/Donau, West Germany (BRD).

58 I. N. RICH AND B. KUBANEK

Rifkind, 1972) which can be detected in the circulation until about the 14th to15th day of gestation (Fantoni et al. 1969; Kubanek, unpublished results).Between the 10th and 11th day, erythropoiesis is initiated in the foetal liver bymigration, seeding and proliferation of pluripotential stem cells from the yolksac (Johnson & Moore, 1975). Hepatic erythropoiesis continues until just afterbirth.

During development of the foetus, the pregnant mother responds by increasingerythropoiesis primarily in the spleen which is erythropoietically dormant understeady-state conditions, the increase reaching a maximum between the 12thand 15th day of pregnancy (Fowler & Nash, 1968; Fruhman, 1968). Fruhman(1968) postulated that the observed increase in erythropoiesis especially in thespleen was the result of the rapidly growing foetus and associated tissues,causing a greater demand for oxygen by the mother. Regulation of erythropoiesisby the mother has been postulated to be independent of that occurring in thefoetus since by hypertransfusion or starving the mother, thus causing a decreasein maternal erythropoiesis, foetal erythropoiesis carried on at a regular rateof red cell production; that is, change in maternal erythropoietin (Ep) levelshas no effect on foetal erythropoiesis (Jacobson, Marks & Gaston, 1959;Lucarelli et al, 1968).

However, the effect of exogenously added erythropoietin to organ andsuspension cultures of mouse yolk sac and foetal liver cells respectively wasshown by Cole & Paul (1966) to have profound effects. Whereas explanted yolksacs did not respond to erythropoietin, as measured by the incorporation of59Fe into haem, foetal liver cells were stimulated but only until a decrease wasobserved. It was postulated that after the 14J-day, foetal liver erythropoieticcells lost their responsiveness to erythropoietin (Cole & Paul, 1966).

In experiments employing the erythroid colony-forming technique (Stephen-son, Axelrad, McLeod & Shreeve, 1971; Iscove, Siber & Winterhalter, 1974),it was shown that under the same conditions, 14-day foetal liver CFU-Eresponded maximally to 0-075 u./ml of erythropoietin, while adult bone marrowCFU-E required 0-4 u./ml for maximum stimulation (Rich & Kubanek, 1976).This difference pointed to the interesting possibility that during the ontogeny ofthe mouse the erythropoietin sensitivity might change in response to changingdemands and/or sites of erythropoiesis. A detailed investigation was thereforeundertaken in order to study the response of early erythropoietic precursor cellsto erythropoietin in the foetus and simultaneously in naturally perturbated andstimulated maternal erythropoiesis during pregnancy.

METHODS AND MATERIALSAnimals

Foetuses were obtained by placing 20 female CBA/Ca mice 8-12 weeks oldin cages for 3 weeks in order to produce an anovulatory cycle by'overcrowding'.Ovulation was then induced by placing two males and two females in a cage

Hepatic and maternal erythropoiesis 59

separated by a partition for 24 h. After this time the partition was removed,the morning after (a period of 12 h) being designated day 0 of gestation.

Preparation of suspensions

All mice were killed by cervical dislocation. Foetuses were removed asep-tically from pregnant mice and placed into cold Hanks's balanced salt solution.After dissecting out the foetal livers, they were put into a 1 ml syringe containingalpha medium (Stanners, Eliceiri & Green, 1971) without ribosides or deoxy-ribosides, but containing 20 mM L-glutamine, 5 % foetal calf serum and 100 mgeach of penicillin and streptomycin (Flow Laboratories, Bonn, West Germany,supplied alpha medium and L-glutamine). The pooled livers were passed throughneedles of decreasing diameter into plastic tubes (Falcon Plastics, BectonDickinson, West Germany) containing 2 ml of alpha medium. After a short timeon ice to allow debris to settle, the exact volume was measured and foetal liversuspensions made up to a specific volume. For 11- and 12-day foetal livers,five to ten organs were suspended in not more than 2 ml of medium. For 13- and14-day foetal livers, two organs were suspended in 1 ml of medium, whileorgans from later days of gestation were suspended as one organ/ml of medium.

Femora from the mothers were cut to a length of 9 mm from the distal end.The proximal end was fitted into a 22-gauge needle connected to a syringecontaining 1 ml of alpha medium and the marrow flushed through the bonethree times (Fruhman, 1964) into a plastic tube containing cold alpha medium.After all the marrows had been flushed out, the crude suspension was allowedto settle for a short time and the exact volume measured by withdrawing thesuspension through a 25-gauge needle. Adult bone marrow suspensions wereusually made up so that one organ was suspended in I ml of medium.

Spleens also obtained from the mothers were first homogenized in a loose-fitting homogenizer and after a few minutes on ice to allow large particles tosettle, the suspension was decanted into a plastic tube and small particles thenallowed to settle. The exact volume was measured and the suspensions weremade up so that a maximum of three spleens were suspended in 5 ml of medium.

Suspensions prepared in the above manner were all single cell suspensionsas seen in the haemocytometer. However, nucleated cells were regularly countedusing a Coulter Counter Model B with a lower threshold of 16 and an upperthreshold of 108 at 1/4 amplification and 1/2 aperture current.

Erythroid colony-forming technique

The methyl cellulose modification (Tscove et al. 1974) of the erythroid colony-forming technique was employed using a standardized procedure previouslydescribed by Rich & Kubanek (1976). In essence, a total volume of 2-5 ml wasmade up consisting of alpha medium, 30 % foetal calf serum, erythropoietinStep HI (Connaught Laboratories, Canada; Lot 3005-1 containing 300 unitsin 91 mg) dissolved in alpha medium, alpha-thioglycerol (end concentration,

60 I. N. RICH AND B. KUBANEK

1 x 10~4M) diluted in alpha medium, cell suspension previously diluted to therequired concentration and 0-8% of a 2% methyl cellulose (Serva, premiumgrade, 4000 cps) stock solution prepared in alpha medium. The componentswere mixed in plastic tubes using a Vortex mixer and 1 ml was dispensed intoeach of two 35 mm Petri dishes (Greiner Plastics, West Germany).

Due to the fact that to obtain BFU-E colonies, more than ten times as mucherythropoietin is required than for CFU-E colonies, the above method wasscaled down so that only one quarter of the above quantities were used. Usingthis method, 0-2 ml were plated in multiwell tissue culture plates (FalconPlastics, Becton Dickinson, West Germany), Like CFU-E, BFU-E multiwellplates were incubated at 37 °C in 5 % CO2 and an approximately 98 % humidifiedatmosphere.

Aggregates of cells counted between 36 and 48 h of incubation were consideredCFU-colonies if the aggregates contained eight or more tightly packed cellswhich would quickly stain positive for haemoglobin using benzidine solution(Cooper et al. 1974). Usually, however, CFU-E colonies were counted withoutstaining or fixing with glutaraldehyde (Cooper et al. 1974). In consideringBFU-E colonies counted after 10 days of incubation, the following basiccriteria were used: (a) aggregates consisted of at least one core of cells, usuallyred in colour, (b) the aggregates contained more than 200 cells, (c) aggregates ofcells similar in form to CFU-E colonies on the periphery which stain in apositive manner with benzidine (see Fig. 1). In this way it was easy to distinguishBFU-E colonies from small granulocytic/macrophage colonies consisting ofmuch larger cells formed under the same conditions and probably due to thepresence of colony stimulating factor (CSA) in the erythropoietin preparationand/or the foetal calf serum used. It was found, however, that using 11-dayfoetal liver cells, no aggregates of cells similar to CFU-E colonies were observedin the periphery. These colonies were extremely compact and, in comparisonto background colonies, were much larger. These were therefore considered 11-day foetal liver BFU-E colonies (Fig. 2).

RESULTS

Using the culture conditions described previously, erythropoietin dose-response curves were performed using 0-5 x 105 foetal liver cells/ml.

Figures 3 and 4 show the number of CFU-E/105 cells plated as a function oferythropoietin concentration for the 11th, 12th, 13th, 14th, 15th and 16th daysof hepatic erythropoiesis. The dose response curves depicted in Fig. 3 wereperformed at a later date and using a different batch of foetal calf serum thanthose shown in Fig. 4. However, since the linear regression parameters for 13-and 14-day foetal liver cells are similar in both figures, a continuity in theresults obtained can be assumed.

The high incidence of colony growth obtained from early foetal liver

Hepatic and maternal erythropoiesis 61

Fig. 1. Thirteen-day foetal liver BFU-E as seen in the plate after 8 days of incubationwith 2 x 105 cells/ml and 40 u./ml erythropoietin. The colony consists of a compactcore with many CFU-E like colonies in the periphery (magnification x 80).

EMB 50

62 I. N. RICH AND B. KUBANEK

v mm

Jt-

JFig. 2. Eleven-day foetal liver BFU-E after 8 days of incubation with 2 x 105 cells/mland 40 u./ml erythropoietin. This colony consists of a very tight ball of cells with noCFU-E-like colonies on the periphery (magnification x 50).

4-5 -]

4 0 -

3-5 -

13

o 30 -

2 2 - 5 -x

| 2-0 -I"o

T 15 -u.

10 -

0-5 -

Hepatic and maternal erythropoiesis

11-day

63

12-day

13-day

14-day

000156 000313 000625 00125 0025

Erythropoietin dose (u./ml)

I 0075 I005 01

0-2

Fig. 3. Dose-response curves for CFU-E foetal liver growth as a function of logerythropoietin concentration. All foetal liver cells suspensions plated at 0-5 x 105/ml.11-day (r = 083; P < 005); 12-day (r = 0-92; P < 0001); 13-day (r = 0-96;P < 0001); 14-day (r = 0-96; P < 0001).

cells has allowed colony counts to be obtained using erythropoietin concentra-tions from 0-00156 u./ml to 0-075 u./ml. Spontaneous colony formation wasobserved amounting to about 20 CFU-E colonies/105 for 11- and 12-day foetalliver cells and between 10 and 12 CFU-E colonies/105 for 13- and 14-day livercells.

With the exception of 11-day foetal liver cells which show a maximum at0-025 u./ml in the dose-response curve, a maximum erythropoietin stimulationis found at 0-075 u./ml for all later days of hepatic erythropoiesis. Instead ofa definite plateau being obtained, an almost instant decrease is seen when higherconcentrations of erythropoietin (Step III) are employed. This effect is almostcertainly due to the presence of unspecific toxic substances in the erythropoietinpreparation (Stephenson & Axelrad, 1971), since by further purification a plateauoccurs (Tscove et al. 1974; Iscove & Sieber, 1975). Using a more purified erythro-poietin preparation derived from human urine a dose-response on 14-day foetalliver cells was performed using the same foetal calf serum batch, as that for11-day foetal liver (Fig. 3). It is apparent (Fig. 5) that even though the concentra-tion of CFU-E is increased over the whole erythropoietin dose range, thedose producing 50% stimulation of CFU-E is the same as that using the

5-2

64 I. N. RICH AND B. KUBANEK

1-8 -|

1-6

1-4

* 1-7 -O ' -

o 10 -

I 0-8 -_oou

UJ 0-6 -

0-4-

0-2 -

0

\13-day

000156 I 000625 | 0025 I 0-075 I000313 00125 005 01

Erythropoietin dose (u./ml)

0-20-3

Fig. 4. Dose-response curves for CFU-E foetal liver growth as a function of logerythropoietin concentration. All foetal liver cell suspensions plated at 0-5 x 105/ml.13-day (r = 0-97; P < 0001); 14-day (r = 0-97; P < 0001); 15-day (r = 0-84;P < 0001); 16-day (r = 0-88; P < 0001).

commercially obtained Step III preparation, namely 0-014 u./ml. In contrast,however, to the latter preparation, the purified substance produces aplateau ranging from 0-075 u./ml to 4-00 u./ml, the highest concentrationemployed.

The progression of hepatic erythropoiesis is shown in Figs. 6 and 7. Figure 6illustrates that four distinct phases of growth occur. The first, between the Uthand 13th day of gestation, appears to be one of very rapid growth with theorgan increasing in cellularity by a factor of over 25, that is, an approximatedoubling time of 12 h. Between the 13th and 15th day a doubling time of about24 h occurs followed by a levelling off by the 17th day. Finally, there is a slightdecrease between the 17th and 19th day of gestation.

The absolute or total of CFU-E/organ (Fig. 6) increases up to the 14th dayof gestation, followed by a decrease to the 17th day, but never quite reacheszero, even on the 19th day of gestation. The increase in cellularity seen in thethe foetal liver, together with the changing cell populations, results in thedecrease seen after the 12th day in the CFU-E concentration. (Fig. 6).

In contrast to the CFU-E population, the BFU-E population shows a constantmaximum eryihropoietin stimulating dose of 4-0 u./ml from the 11th to the

Hepatic and maternal erythropoiesis 65

120 -i

110 -

100 -

90 -

80 -

70 -

60 -

50 -

4 0 -

30 -

20 -"

10 -

0 -. • > • • • •

000156 I 000625 I 0025 I 0075 1000313 00125 005 01

Hrythropoietin dose (u./ml.)

0-2 I 04 10-6 I03 0-5 0-8

Fig. 5. Dose-response curve for CFU-E foetal liver growth as a function of logerythropoietin concentration of Step HI and a more purified preparation fromhuman urine. 14-day-old foetal liver cell suspensions were plated at O-5xlO5/ml.• • , Purified Ep; • • , Step III Ep.

16th day of gestation, which is the same dose as required for maximum growthof BFU-E from adult bone marrow. Furthermore, a similar pattern in the fre-quency and absolute number of BFU-E during hepatic erythropoiesis is observedwhen compared with CFU-E. The BFU-E concentration values for 11- and12-day foetal liver are not significantly different, but it would appear by com-arison of Fig. 7 with Fig. 6 that the peak in the BFU-E concentration wouldperhaps occur between the 10th and 12th day of gestation. It is however clearthat a decline to almost zero levels is seen in the BFU-E concentration by the16th day of gestation, approximately 24 h before the decline of the CFU-Epopulation.

The apparent dilution effect of the CFU-E concentration as the cellularity ofthe foetal liver increases would also be applicable to the BFU-E population. It isof interest to note here that comparison of concentration and absolute numbersof CFU-E and BFU-E indicate that the BFU-E population represents about1% of the CFU-E population throughout the 11- to 16-day hepatic period,assuming the same plating efficiency of both cell populations.

The effect of the foetus on the mother is shown in Figs. 8 and 9 for bonemarrow and spleen. The upper panels of these diagrams illustrate the changein cellularity of the bone marrow and spleen over the same period (11-19 days),

66 I. N. RICH AND B. KUBANEK

1 0 8 - ,

107 -

106 -

1Of'

10 11 12 13 14 15 16I

17 181

19

(h)

10I

17

I

18I

19I

13 14 15 16

Days of gestation

Fig. 6. (a) Total number of cells/foetal liver, plotted on a log scale ( x 106). (b) Numberof CFU-E/105 foetal liver cells, (c) Number of CFU-E/organ. All parameters as afunction of gestational age (vertical bars: mean±standard deviation).

as that described for hepatic erythropoiesis in the foetus. The number of cells/organ is elevated in both bone marrow and spleen above the normal (control)adult animals. The spleen, in particular, doubles in cellularity (over 200 x 106

compared to 100 xlO6 cells/organ) from the 11th to about the 15th day ofpregnancy. After the 15th day, a gradual decrease to near-normal levels isobserved.

Hepatic and maternal erythropoiesis 6760 -i

ecus

b

Ioni

cs/

co

cc

50

40

30

20

10

(fl)

I I I I I I10 11 12 13 14 15 16

1800 -i

1600 -

1400 -

1200 -

10U0 -

800 -

600 -

400 -

200 -

I I I I I I10 11 12 13 14 15 16

Days of gestation

Fig. 7. (a) Number of BFU-E/105 foetal liver, (b) Number of BFU-E/organ. AHparameters as a function of gestational age (vertical bars: mean±standarddeviation).

The effect of pregnancy on erythropoiesis measured by the CFU-E techniqueis shown in the middle and lower panels of Figs. 8 and 9. The CFU-E concen-tration of the bone marrow decreases from the 11th to the 14th day and thendeclines more sharply to the 19th day. The effect on the absolute CFU-E valuesdoes not vary to any great extent from normal bone marrow. The effect onsplenic erythropoiesis is more pronounced. From the 11th to the 14th day of

68

t>

I. N. RICH AND B. KUBANEK

109 —i Bone marrow (a)

108 -

107 -

500 - i

400 -

o 300 —

J 200 -p& 100 —

100 - ,

a so —

X

col.

J j

D• _

60 -

40 -

20 -

(0

r10 II

\12 13

I14

115

I16 17 18

I19

Days of pregnancy

Fig. 8. The effect of pregnancy on myeloid erythropoiesis in the mouse, {a) Totalnumber of cells/organ, plotted on a log scale, (b) Number of CFU-E/105 cells,(c) Number of CFU-E/organ. All parameters as a function of days of pregnancy.Continuous horizontal line indicates the mean value of normal (control) valuesthrough the period of investigation (vertical bars: mean±standard deviation).

pregnancy, splenic CFU-E concentrations seem to follow those of bone marrowbut remain at much higher levels (between 10- and 40-fold) than normal spleen.A similar pattern of events is seen for the absolute splenic CFU-E whichdecrease to almost normal values by the 19th day of pregnancy.

Figure 10 shows a scatter graph for the erythropoietin dose-response curves

Hepatic and maternal erythropoiesis 69

1 0 9 - ,

io8 -

io7

Spleen (a)

500 - i

» 400 -o

o 300 -

a 200 -

u.u 100 -

(h)

100 - i

80 -

2 60 -

40 -

(c)

HU

-b

20 -

0

101

11

V

1p

1

ni

14

\

\

115

116

117

118

119

Days of pregnancy

Fig. 9. The effect of pregnancy on splenic erythropoiesis in the mouse, (a) Totalnumber of cells/organ, plotted on a log scale, (b) Number of CFU-E/105 cells plated,(c) Number of CFU-E/organ. All parameters as a function of days of gesta-tion. Continuous horizontal line indicates the mean value of normal (control)values throughout the period of investigation (vertical bars: mean ±standarddeviation).

from the 11, 13th and 15th days of pregnancy in the maternal spleen comparedwith that for normal adult spleen. These normalized results are expressed as thepercentage of the CFU-E response from the maximum stimulating Ep dose. Onexamination of the curves it is seen that they are parallel to the normal adult

70 I. N. RICH AND B. KUBANEK

100 n

9 0 -

80 -

70 -

| 60 -

1 50Hc

£.2 40 H

30 -

20 -

10 -

00125 0025 005 0075 01 0-2Erythropoietin dose (u./ml.)

0-3 0-4 0-5

Fig. .10. Scatter graph for splenic CFU-E from pregnant mice as the percentresponse from the maximum stimulating erythropoietin dose as a function oferythropoietin concentration. Shaded area represents normal animals, y = 161-9 +57-0(x); #, l l-day:y = .137-7 +60-8(x); 1,13-day:y = 125-5 +65-6(x); A15-day:y = 137-8 + 59-2 (x).

splenic erythropoietin. response, indicating that neither the cell nor its mechanismof response to erythropoietin has changed. In addition, the erythropoietinconcentration required to produce maximum colony formation remains at0-3 u./ml.

DISCUSSION

During hepatic erythropoiesis, the foetal liver is growing very rapidly witha doubling time of about 12 h between the 11th and 13 day, decreasing to 24 hfrom the 13th to the 15 day followed by an even slower rate until the 17th dayof gestation. The most frequent erythropoietic cell type up to about the 13th dayis the pronormoblast, the number of which steadily declines to the 19th day(Silini, Pozzi & Pons, 1976; Tarbutt & Cole, 1970; Kubanek, Bock, Bock &Heit, 1975); the basophilic erythroblasts show a maximum on the 15th daywhile the polychromatic and orthochromatic erythroblasts peak on the 16th dayof gestation (Tarbutt & Cole, 1970). It has also been shown by Paul, Conkie &Freshney (1969) and Tarbutt & Cole (1970) that during early hepatic ery-thropoiesis, 55-70 % of all cells belong to the erythropoietic series. However,during the last 5-6 days of hemopoiseis in the foetal liver the number of hepa-tocytes also increases (Silini et al. 1967; Paul et al. 1969). The pattern observed

Hepatic and maternal erythropoiesis 71

for the CFU-E during hepatic erythropoiesis would appear to be similar to thatof the pronormoblast.

The pattern of appearance of the BFU-E in the foetal liver follows that ofthe CFU-E fairly closely, although the peak of the BFU-E concentrationcannot be positioned exactly during early hepatic development. The earlierdecline in BFU-E on the 16th day is about 24 h before that seen by the CFU-Epopulation. Furthermore, the BFU-E population represents about 1 % of theCFU-E population during this period suggesting that within the erythropoietichierarchy, the CFU-E is derived from this earlier BFU-E.

Throughout hepatic erythropoiesis, with the exception of the 11th day, themaximal erythropoietin sensitivity of the foetal liver CFU-E remains at 0-075u./ml. This would suggest that changes in the concentration of erythropoietinin the foetal circulation during this period are not the main determining factorfor the rate of red cell production, since the CFU-E are capable of respondingto extremely low doses of erythropoietin. BFU-E derived from foetal liver havea similar high requirement for erythropoietin to grow in vitro as BFU-E derivedfrom adult marrow. This may imply that BFU-E growth is governed by a factorindependent of erythropoietin.

Recents reports by Zucali, McGarry & Mirand (1977c) and Zucali, McDonald,Gruber & Mirand (19776) have indicated that foetal liver cells in culture arecapable of producing an erythropoietic stimulating factor which reaches maxi-mum production between the 14th and 15th day of gestation, after whichproduction decreases. Gruber, Zucali & Mirand (1977) have implicated theKupffer cell or macrophage of the foetal liver capable of storing if not alsoproducing erythropoietin. If this is so, then it may be envisaged that internalhepatic, local concentrations of erythropoietin, however small, would, in fact,be high concentrations for the cells responding to it in the immediate vicinity ofthe erythropoietin-producing and/or storage cells. Whether the minimumerythropoietin concentration in vitro reflects the actual minimum erythropoietinconcentration in vivo can only be assumed. Nevertheless, the fact remains thatthroughout hepatic erythropoiesis, the CFU-E continue to respond to erythro-poietin in vitro, with the decreased absolute response reflecting the decreasedCFU-E concentration.

In 1966 Cole & Paul reported that whereas yolk sac erythroblasts did notincorporate 59Fe into haem in response to exogenous erythropoietin, foetal livercells did, but only until the 14th day of gestation. They postulated that up tothe 14th day erythropoietin production increased; thereafter it was in excess andthe erythropoietic cells which up to this time had incorporated 59Fe, failed to doso. This was interpreted as a loss in potential for erythropoietic stimulation.However, Cole, Regan, White & Cheek (1975) demonstrated that CFU-E couldbe obtained until at least the 16th day of gestation, this being correlated withthe response of foetal liver cells suspensions to erythropoietin as measured bythe rate of haem synthesis. These authors not only proposed that erythroid

72 I. N. RICH AND B. KUBANEK

colony formation was associated with high erythropoietin levels in the circula-tion of the foetus, but that this also corresponded with the greatest demand forerythrocytes. In addition, the rapid decrease in CFU-E after the 16th day wasassociated with a loss in erythropoietin sensitivity. The difference between theseresults and those presented here are difficult to interpret.

Jacobson et al. (1959) and Lucarelli et aJ. (1968), using mice and rats respec-tively, showed that if the mother is subjected to polycythaemic conditions byhypertransfusion or starvation, foetal erythroid production continued in anormal manner. It was therefore postulated that either adult and embryonicerythropoiesis were controlled by different mechanisms or, that embiyonicerythropoiesis, although being erthropoietin dependent, is regulated by very lowlevels of endogenous erythropoietin since such concentrations could be presenteven in the polycythaemic mouse (Bleiberg & Feldman, 1969). As shown inFigs. 3 and 4, foetal liver cells are, in fact, extremely sensitive to low erythro-poietin concentrations. It is therefore concluded that at no time during hepaticerythropoiesis does the CFU-E population cease to respond to erythropoietin.

The observations of Bleiberg & Feldman (1969) imply that mouse adult bonemarrow cells are less sensitive to erythropoietin than foetal liver cells. Rich &Kubanek (1976) have shown that normal adult bone marrow CFU-E requirefive times as much erythropoietin than foetal liver CFU-E to achieve maximumstimulation in vitro. This same requirement for erythropoietin is also observedfor bone marrow or spleen taken at any time during the second half of pregnancyin the mother. Despite increased splenic erythropoiesis and changing erythroidcomposition seen during this naturally perturbated phase, a change in erythro-poietin sensitivity of the CFU-E population is not observed. These findingsindicate that erythropoietin requirement for growth of CFU-E is an intrinsicproperty of the adult or foetal erythroid tissue and not due to a differentcomposition of these tissues.

The pattern shown by the changing splenic cellularity during the second halfof pregnancy in the mother is in good agreement with the change in spleenweight described by Fruhman (1968) when, on the 12th day, both these para-meters have doubled in comparison with normal animals. The CFU-E populationappears to reach a maximum on or about the earliest day measured, namelyday 11 of pregnancy, while the peak in the morphologically identifiable erythro-blasts has been shown to occur on the 12th day (Fowler & Nash, 1968). Thefemoral cellularity is also increased above normal levels during this time. Con-centrations of CFU-E parallel those of the spleen, again with an apparent peakoccuring at about the 11th day of pregnancy. Fruhman (1968) has shown thatthe nucleated erythroblasts peaked at between the 9th and 12th days althoughthe 59Fe incorporation decreased throughout pregnancy, an effect considered tobe due to preferential uptake of the isotope by the growing foetus (Fruhman,1970). The fluctuation in concentration and absolute values of CFU-E betweenthe 11th and 13th day of pregnancy despite the relatively stable cellularity of

Hepatic and maternal erythropoiesis 73

the organs during this period is unclear, although Fowler & Nash (1968) havepointed out that maternal erythropoiesis and litter size are directly related toeach other.

During pregnancy, a disproportional increase in plasma volume and red bloodcells results in a decreased haematocrit causing anaemia. Despite an increasedred cell production, maintenance of erythropoietic homeostasis in this stresssituation appears to be mainly a function of the spleen. Viewed in comparisonwith the total myeloid content of the animal (calculated from one femur rep-resenting about 6 % of the total bone marrow as described by Smith & Claytonin 1970), the total splenic CFU-E content is of the same order of magnitude as,that calculated for the total bone marrow. That is, erythropoiesis has, in effectbeen doubled due to the addition of splenic erythropoiesis between the days11 and 14 of pregnancy.

As to the mechanisms of increased maternal erythropoiesis, Fruhman (1968)suggested that the growing foetus and the associated tissues have an increasingrequirement of oxygen causing hypoxia in the mother leading then to increasedmaternal erythropoiesis. Furthermore, as foetal growth progresses, the oxygenrequirement decreases so causing a decrease in maternal hypoxia leading to adecline in erythropoiesis.

The decrease in CFU-E values observed after the 14th day of pregnancy mayreflect the possibility that hypoxia in the mother is decreasing, which would bein agreement with the proposed regulation of maternal erythropoiesis (Fruhman,1968).

This work was supported by the Deutsche Forschungsgemeinschaft, Sonderforschungs-bereich 112 Project A2, and partly by the Volkswagen Foundation.

REFERENCES

BLEIBERG, I. & FELFMAN, M. (1969. On the regulation of haemopoietic spleen coloniesproduced by embryonic and adult cells. Devi Biol. 19, 566-580.

CHAPELLE, DE LA, A., FANTONI, A. & MARKS, P. A. (1969). Differentiation of mammaliansomatic cells: DNA synthesis and haemoglobin synthesis in foetal mice. Proc. natn. Acad.Sci. U.S.A. 63, 812-810.

COLE, R. J. & PAUL, J. (1966). The effects of erythropoietin on haem-synthesis in mouse yolksac and cultured foetal liver cells. / . Embryol. exp. Morph. 15, 245-260.

COLE, R. J., REGAN, T., WHITE, S. L. & CHEEK, E. M. (1975). The relationship betweenerythropoietin-dependent cellular differentiation and colony-forming ability in prenatalhaemopoietic tissue. / . Embryol. exp. Morph. 34, 575-588.

COOPER, M. C, LEVY, J., CANTOR, L. N., MARKS, P. A. & RIFKIND, R. A. (1974). The effectof erythropoietin in clonal growth of erythroid precursor cells in vitro. Proc. natn. Acad.Sci. U.S.A. 71, 1677-1680.

CRAIG, M. L. & RUSSELL, E. S. (1964). A development change in haemoglobins correlatedwith an embryonic red cell population in the mouse. Devi Biol. 10,191-201.

FANTONI, A., DE LA CHAPELLE, A., CHUI, D., RIFKIND, R. A. & MARKS, P. A. (1969). Controlmechanism of the conversion from synthesis of embryonic to adult haemoglobin. Ann.N.Y. Acad. Sci. 165, 194-204.

FRUHMAN, G. J. (1964). Cell counts of mouse bone marrow. Life Sci. 3, 737-748.

74 I. N. RICH AND B. KUBANEK

FRUHMAN, G. J. (1968). Blood formation in the pregnant mouse. Blood. 31, 242-248.FRUHMAN, G. J. (1970). Splenic erythropoiesis. In Regulation of Haemopoieis (Ed. A. S.

Gordon) pp. 339-368. Apelton-Century-Crofts.FOWLER, J. H. & NASH, D. (1967). Erythropoiesis in the pregnant mouse. Expl Haemat. 12,

72-73.GRUBER, D. F., ZUCALI, J. R. & MIRAND, E. A. (1977). Identification of erythropoietin

producing cells in foetal mouse liver cultures. Expl Haemat. 5, 392-398.ISCOVE, N. N., SIEBER, F. & WINTERHALTER, K. H. (1974). Erythroid colony formation in

cultures of mouse and human bone marrow; analysis of the requirement for erythropoietinby gel filtration and affinity chromatography on agarose-concanavalin A. / . cell. Physiol.83, 309-320.

ISCOVE, N. N. & SIEBER, F. (1975). Erythroid progenitors in mouse bone marrow detectedby macroscopic colony formation in culture. Expl Haemat. 3, 32-43.

JACOBSON, L. O., MARKS, E. K. & GASTON, E. O. (1959). Studies on erythropoiesis. II. Theeffect of transfusion-induced polycythaemia in the mother on the foetus. Blood 14,644-652.

JOHNSON, G. R. & MOORE, M. A. S. (1975). Role of stem cell migration in initiation ofmouse foetal liver haemopoiesis. Nature, Lond. 258, 726-738.

KUBANEK, B., BOCK, E. L., BOCK, O. & HEIT, W. (1975). Growth patterns of foetal haemo-poiesis after injury. In Erythropoiesis. Proceedings of the Fourth International Conferenceon Erythropoiesis (Ed. F. Nakao, J. W. Fisher & F. Takaku) pp. 371-379. University ofTokyo Press.

LUCARELLI, G., PORCELLINI, A., CARNEVALI, C, CARMENA, A. & STOHLMAN, F. Jr. (1968).Foetal and neonatal erythropoiesis. Ann. N.Y. Acad. Sci. 149, 544-559.

MARKS, P. A. & RIFKIN, R. A. (1972). Protein synthesis: its control in erythropoiesis. ScienceMK 175, 955-961.

METCALF, D. & MOORE, M. A. S. (1970). Haemopoietic Cells. Amsterdam: North HollandPublishing Company.

PAUL, J., CONKIE, D. & FRESHNEY, R. I. (1969). Erythropoietic cell population changesduring the hepatic phase of erythropoiesis in the foetal mouse. Cell Tissue Kinet. 2,283-294.

RICH, I. N. & KUBANEK, B. (1976). Erythroid colony formation in foetal liver and adult bonemarrow and spleen from the mouse. Blood 33, 171-180.

SILINI, G., Pozzi, L. V. & PONS, S. (1967). Studies on haemopoietic stem cells of mouse foetalliver. J. Embryol. exp. Morph. 17, 303-318.

SMITH, L. H. & CLAYTON, M. C. (1970). Distribution of injected 59Fe in mice. Expl Haematol.20, 82-86.

SNELL, G. D. & STEVENS, L .C. (1966). Early embryology. In Biology of the Laboratory Mouse(Ed. E. L. Green), pp. 205-345. McGraw-Hill.

STANNERS, C. P., ELICEIRI, G. L. & GREEN, H. (1971). Two types of ribosomes in mouse-hamster hybrid cells. Nature, New Biol. 230, 52-54.

STEPHENSON, J. R. & AXELRAD, A. A. (1971). Separation of erythropoietin-sensitive cellsfrom haemopoietic spleen colony-forming stem cells of mouse foetal liver by unit gravitysedimentation. Blood 37, 417-427.

STEPHENSON, J. R., AXELRAD, A. A., MCLEOD, D. L. & SHREEVE, M. M. (1971). Induction ofcolonies of haemoglobin-synthesizing cells by erythropoietin in vitro. Proc. natn. Acad.Sci. U.S.A. 68, 1542-1546.

TARBUTT, R. G. & COLE, R. J. (1970). Cell population kinetics of erythroid tissue in the liverof foetal mice. J. Embryol. exp. Morph. 24, 429-446.

ZUCALI, J. R., MCGARRY, M. P. & MIRAND, E. A. (1977a). Stimulation of erythropoiesis ina grafted animal by mouse foetal liver culture media. Expl Haemat. 5, 103-108.

ZUCALI, J. R., MCDONALD, T. P., GRUBER, D. F. & MIRAND, E. A. (19776). Erythropoietinthrombopoietin and colony-stimulating factor in foetal mouse liver culture media. ExplHaemat. 5, 385-391.

(Received 3 July 1978, revised 24 October 1978)