In order to evaluate the hypothetical activity of foetal hepatic factors on putative yolk-sac haemopoietic stem cells we used the Double Diffusion Chamber (DDC) technique. The DDC were made of a regulator compartment, where foetal hepatic tissue was introduced and a test compartment where visceral yolk-sac cells were cultured. In this system a hepatic signal induced the yolk-sac stem cells to differentiate along the granulocytic pathway but did not stimulate yolk-sac CFUs growth. Contrary to CFUs originating from foetal liver or adult bone marrow, yolk-sac CFUs do not increase numerically in diffusion chamber culture.

The current view of vertebrate haematopoietic ontogenesis holds that a succession of pluripotent stem-cell migrations originate in the yolk-sac blood islands and invade the hepatic rudiment first and then the spleen and bone marrow (Moore & Owen, 1967a,b; Moore & Metcalf, 1970). However, this view has been challenged by other authors. Experimental evidence in birds (Dieterlen-Lièvre, 1975; Dieterlen-Lièvre, Beaupain & Martin, 1976; Toivanen, Eskola & Toivanen, 1976; Beaupain, Martin & Dieterlen-Lièvre, 1979; Martin et al., 1979) in amphibians (Hollyfield, 1966; Volpe & Turpen, 1977) and in mice (Marks & Rifkind, 1972a,b; Cudennec, Thiery & Le Douarin, 1981) questions the dependence of the intraembryonic haematopoiesis on the migratory stream of haemopoietic stem cells which would develop ab initio within the yolk-sac mesoderm.

We have previously studied the proliferation and the differentiation of visceral and parietal yolk-sac cells from 9-day-old mouse embryos by the diffusion chamber (DC) technique (Symann et al., 1978). The yolk-sac blood islands are mainly composed of primitive erythroblasts; macrophages are also occasionally present. Between the 7th and 12th day of gestation there is no evidence of the production of other differentiated haemopoietic cell lines in the yolk-sac environment. In the DC cultures of visceral yolk-sac cells CFUs were undetectable, macrophages were the most numerous cells and granulocytes never exceed 6 % of all the harvested cells. These findings gave results contrary to those obtained in the foetal liver experiments (Symann et al. 1976a,b,c). Indeed, foetal liver CFUs proliferate actively in DC. On the 16th gestational day, foetal liver, which is also predominantly an erythroid organ, mainly generates vigorously growing granulocytic cells and, to a lesser extent macrophages, in DC cultures.

The present study was designed to evaluate the hypothetical activity of foetal hepatic factors on putative yolk-sac stem cells. For this purpose, we used the double diffusion chamber (DDC) technique introduced by Pfeffer & Böyurn (1977). Briefly, the DDC were made of a regulator compartment, where a tissue capable of producing a haematopoietic regulation factor was introduced and a test compartment where a target population was cultured.

In DDC, a hepatic signal induced differentiation of the yolk-sac stem cells along the granulocytic pathway. This differentiation did not occur in this signal’s absence.

Animals

In all experiments virgin female C3H mice (Centre d’Animaux de Laboratoire, Heverlee, Belgium) 10 –16 weeks old were used as single diffusion chamber (SDC) and double diffusion chamber (DDC) hosts or as irradiated recipients of spleen colony assay. Yolk-sac donors were C3H mice mated overnight and sacrificed 9 days later. The stage reached by the embryo was carefully checked by counting somite pairs: only the embryos which did not develop beyond the 17-somite stage were retained. Yolk sacs were dissected as previously described (Symann et al., 1978). Foetal liver donors were C3H mice mated overnight and sacrificed 13 days later or WISTAR rats mated overnight and sacrificed 13 days later. All animals were killed by cervical dislocation. All the manipulations were performed on a laminar air flow bench at room temperature.

Single diffusion chamber culture

We used the single diffusion chamber (SDC) technique as described by Benestad (1970). Yolk-sac cells, obtained by passing the yolk sacs through needles of sequentially decreasing diameters, were cultured. After 6 days in culture, the SDC were removed from the host animal and incubated in Hanks Balanced Salt Solution (HBSS) with 0-5 % pronase for 1 h at 25 °C. The cells were harvested and counted in a haemocytometer.

Double diffusion chamber culture

We used the double diffusion chamber (DDC) technique as described by Pfeffer & Böyum (1977). The DDC consisted of a regulator compartment in which five mouse foetal livers from 13-day-old foetuses or five rat foetal fivers from 13-day-old foetuses were introduced and a test compartment with murine yolk-sac cells. After 6 days of culture, the DDC were treated as SDC and the contents of the two different compartments harvested. The cells from each compartment were counted in a haemocytometer, smeared and processed for CFUs assays.

CFUs assay

Pluripotent stem cells were assayed according to the Till & McCulloch (1961) method. 0 ·25 ml of cell suspension in HBSS was injected via the tail vein into three to nine lethally irradiated (950R from a caesium source) mice. In these experiments irradiated mice received, respectively, 1 × 106 visceral yolk-sac cells and 1 × 106 DDC test compartment harvested cells. CFUs assays have been performed at least twice with each kind of cell suspension.

Cytology

Smears of yolk-sac cells and foetal liver cells prior to and after diffusion chamber culture were made using a cytocentrifuge (Shandon). The slides were air dried, fixed with absolute methanol and then stained with May-Grünwald-Giemsa or with ortho-tolidine and counterstained with Giemsa (a modification of the Sato method). In addition, other slides were stained by the sodium alpha-naphtylphosphate according to the Gomori (Gabe, 1968) method to identify alkaline phosphatases. Differential counts were based on the evaluation of at least 300 cells per smear. The cells were classified as proliferative granulocytes (myeloblasts, promyelocytes and myelocytes), non-proliferative granulocytes (metamyelocytes, bands and segmented granulocytes) or macrophages. Differentiated erythroid cells were classified together as erythroblast.

Statistics

All the results are expressed as weighted mean and standard error of the mean which were calculated as described by Armitage (1971).

1. Yolk-sac cells proliferation and differentiation in SDC and DDC

The yolk-sac cells were cultured in DC in four different ways: 300000 murine vitelline cells were introduced into SDC, or into the test compartment of a DDC where the regulator compartment was either empty or contained five 13-day mouse or rat foetal livers. Murine yolk-sac cells were cultured in the presence of rat foetal liver tissue to demonstrate that liver cells could not cross the Millipore filter. Indeed, contrary to mouse neutrophils, rat neutrophils contain alkaline phosphatases. Thus, staining of the cell contents of each DDC compartment by sodium alpha-naphtylphosphate allowed us to determine the cells’ real source. This checked our DDC system’s cell tightness.

The presence of foetal hepatic tissue in the DDC regulator compartment did not significantly (P = 0 ·05) influence the total cell production by the visceral yolk sac present in the test compartment (Fig. 1). The small decrease of DDC cell yield could not thus be attributed to the presence of the foetal hepatic tissue in the regulator compartment since it was identical in the DDC with an empty regulator compartment.

Fig. 1.

Total production of 9-day visceral yolk-sac cells in SDC and DDC culture after 6 days. Each point is the weighted mean ± S.E.M. of pooled data from three replicate experiments. The number of chamber is shown. SDC(YS): culture of yolk-sac cells in SDC. DDC(YS/-): culture of yolk-sac cells in DDC where the regulator compartment was empty. DDC(YS/FLM): culture of yolk-sac cells in DDC where the regulator compartment contained five mouse 13-day foetal liver. DDC(YS/ FLR): culture of yolk-sac cells in DDC where the regulator compartment contained five rat 13-day foetal liver.

Fig. 1.

Total production of 9-day visceral yolk-sac cells in SDC and DDC culture after 6 days. Each point is the weighted mean ± S.E.M. of pooled data from three replicate experiments. The number of chamber is shown. SDC(YS): culture of yolk-sac cells in SDC. DDC(YS/-): culture of yolk-sac cells in DDC where the regulator compartment was empty. DDC(YS/FLM): culture of yolk-sac cells in DDC where the regulator compartment contained five mouse 13-day foetal liver. DDC(YS/ FLR): culture of yolk-sac cells in DDC where the regulator compartment contained five rat 13-day foetal liver.

When foetal liver was present in the regulator compartment, granulocytic proliferation was significantly increased at the expense of macrophage production. Figure 2 illustrates macrophage proliferation. The presence of foetal hepatic tissue reduced visceral yolk-sac cells macrophage production when compared to results obtained from SDC (P< 0 ·05).

Fig. 2.

Macrophages production of 9-day visceral yolk sac in SDC and DDC culture after 6 days. Each point is the weighted mean ± S.E.M. of pooled data from three replicate experiments. The number of chamber is shown. Abbreviations as in Figure 1.

Fig. 2.

Macrophages production of 9-day visceral yolk sac in SDC and DDC culture after 6 days. Each point is the weighted mean ± S.E.M. of pooled data from three replicate experiments. The number of chamber is shown. Abbreviations as in Figure 1.

Figure 3 shows the production of non-proliferative granulocytes after 6 days of culture. 6000 ± 1000 mature granulocytes were collected from SDC, 5000 ± 900 from DDC without foetal hepatic tissue, 19 000 ±3000 from DDC with mouse foetal livers and 12 000 ± 3000 from DDC containing rat foetal tissue. The. presence of mouse or rat foetal liver in the regulator compartment increased the production of mature granulocytes in the test compartment when compared to results obtained from SDC or DDC with an empty regulator compartment (P<0 ·05). This phenomenon was even more striking in the production of proliferative granulocytes. There were 2500 ± 600 granulocytes in SDC, 1900 ± 600 in DDC without foetal hepatic tissue, 11000 ± 3000 in DDC with five mouse foetal livers and 10 800 ± 1800 in DDC with five rat foetal livers after 6 days of culture (Fig. 4). Again the presence of foetal liver in the regulator compartment enhanced the production of proliferative granulocytes in the test compartment (P< 0 ·05 for SDC or DDC(YS/-) versus DDC(YS/FLM) or DDC(YS/FLR)).

Fig. 3.

Production of yolk-sac proliferative granulocytes after 6 days of SDC and DDC culture. Each point is the weighted mean ± S.E.M. of pooled data from three replicate experiments. The number of chamber is shown. Abbreviations as in Figure 1.

Fig. 3.

Production of yolk-sac proliferative granulocytes after 6 days of SDC and DDC culture. Each point is the weighted mean ± S.E.M. of pooled data from three replicate experiments. The number of chamber is shown. Abbreviations as in Figure 1.

Fig. 4.

Production of yolk-sac proliferative granulocytes after 6 days of SDC and DDC culture. Each point is the weighted mean ± S.E.M. of pooled data from three replicate experiments. The number of chamber is shown. Abbreviations as in Figure 1.

Fig. 4.

Production of yolk-sac proliferative granulocytes after 6 days of SDC and DDC culture. Each point is the weighted mean ± S.E.M. of pooled data from three replicate experiments. The number of chamber is shown. Abbreviations as in Figure 1.

The influence of the foetal hepatic environment upon the erythroid population is shown in Fig. 5. No statistic difference between the various DC culture was observed.

Fig. 5.

Production of yolk-sac nucleated erythrocytes in SDC and DDC culture. Each point is the weighted mean ± S.E.M. of pooled data from three replicate experiments. The number of chamber is shown. Abbreviations as in Figure 1.

Fig. 5.

Production of yolk-sac nucleated erythrocytes in SDC and DDC culture. Each point is the weighted mean ± S.E.M. of pooled data from three replicate experiments. The number of chamber is shown. Abbreviations as in Figure 1.

2. CFUs assay

In order to check whether the yolk sac contained pluripotent stem cells and whether the foetal hepatic tissue was able to influence their proliferation, a CFUs assay was performed after dissection and after SDC and DDC cultures in the presence of 13-day foetal livers. CFUs were almost undetectable after yolk-sac dissection or after conventional DC culture if 1 × 106 cells were injected into lethally irradiated mice. A similar result was obtained after DDC culture with foetal liver tissue in the regulator compartment (Table 1).

Table 1.

CFUs assay from 9-day visceral yolk sac before and after DDC cultures in presence of 13-day foetal livers

CFUs assay from 9-day visceral yolk sac before and after DDC cultures in presence of 13-day foetal livers
CFUs assay from 9-day visceral yolk sac before and after DDC cultures in presence of 13-day foetal livers

Our experiments provide evidence that the properties of vitelline haemopoietic stem cells are different from those of foetal liver CFUs or adult bone marrow CFUs. It should be stressed that their detection by the spleen colony assay is questionnable (Table 1). Like Niewish et al. (1970) we did not find any significant level of visceral yolk-sac CFUs in the various experimental DC system situations. According to Moore & Metcalf (1970) and to our previous results (Symann et al., 1978) the incidence of CFUs in the yolk sac is very low. Perah & Feldman (1977) claimed that after 48 h in vitro, undetectable yolk-sac CFUs become activated and may increase as much as 84-fold. The 6-day DC culture did not permit activation of day-9 yolk-sac stem cells nor did a humoral influence from day-13 foetal hepatic tissue in DDC culture (Table 1). Contrary to yolk-sac CFUs, foetal liver CFUs proliferate and increase numerically in DC culture even faster than adult bone-marrow CFUs (Symann et al., 1976a,c; Breivik, Benestad & Böyum, 1971).

When the vitelline cells were submitted to the influence of a hepatic environment by the DDC culture, a hepatic signal from the day-13 foetal liver induced the yolk-sac stem cells to differentiate along the granulocytic pathway but did not stimulate yolk-sac CFUs to increase in number. Indeed, granulocytic production increased more than four-fold when the foetal hepatic tissue was introduced in the DDC regulator compartment (Figs 3, 4). Cudennec et al. (1978) have also documented the influence of hepatic factors upon yolk-sac haemopoiesis and have analysed the in vitro capacity of yolk-sac haemopoietic cells to produce either primitive or definitive erythrocytes. Prior to the colonization of the liver rudiment by haemopoietic cells, yolk sac explanted alone produced solely primitive erythrocytes and only for a short time. When allowed to colonize a liver rudiment, haemopoietic cells from the yolk sac gave rise to definitive erythrocytes. These cells could express the same capacity when stimulated by liver rudiment even if no direct cell-cell contact was established between stimulating tissue and target haemopoietic cells.

Murine foetal liver is predominantly an erythropoietic organ enclosing few granulocytic cells. Foetal haemopoietic stem cells do not seem to express their granulocytic potential in the foetal liver microenvironment. However, the foetal liver cell suspensions as well as the adult bone-marrow cells generate almost exclusively granulocytes and to a lesser extend macrophages in diffusion chamber cultures (Symann et al., 1976b; Vilpo & Vilpo, 1976). This suggests that the induction of the process of differentiation in our experiments with DDC results from a foetal liver signal, while the choice of a specific pathway, the granulocytic expression in this instance, comes from the diffusion chamber’s microenvironment.

To summarize, our results show that yolk-sac CFUs have different properties than CFUs originating from foetal liver or adult bone marrow. In DDC, a hepatic signal induced the yolk-sac stem cells to the granulocytic differentiation.

Par la technique de culture en doubles chambres à diffusion (DDC), nous avons cherché à évaluer l’influence éventuelle de facteurs hépatiques foetaux sur de putatives cellules souches vitellines. La DDC comprend un compartiment régulateur où est introduit du tissu hépatique foetal et un compartiment test où les cellules du sac vitellin viscéral sont cultivées. Dans ce système, un signal hépatique oriente les cellules souches vitellines vers la différenciation granulocytaire mais n’induit pas une augmentation numérique des CFUs vitellines. Contrairement aux CFUs provenant du foie foetal ou de la moelle osseuse adulte, les CFUs du sac vitellin n’augmentent pas en nombre lorsqu’elles sont cultivées en chambres à diffusion.

We are indebted to Drs J. Rodhain and A. Fedida for their help. We gratefully acknowledge Mrs D. Chintinne for her technical assistance and Mrs K. Deleuse for the expert typing of the manuscript.

Armitage
,
P.
(
1971
).
Statistical Methods in Medical Research
.
Oxford and Edinburgh
:
Blackwell Scientific Publications
.
Beaupain
,
D.
,
Martin
,
C.
&
Dieterlen-Lièvre
,
F.
(
1979
).
Are developmental hemoglobin changes related to the origin of stem cells and site of erythropoiesis?
Blood
53
,
212
225
.
Benestad
,
H. B.
(
1970
).
Formation of granulocytes and macrophages in diffusion chamber cultures of mouse blood lymphocytes
.
Scand. J. Haemat
.
7
,
279
288
.
Breivik
,
H.
,
Benestad
,
H. B.
&
Böyum
,
A.
(
1971
).
Diffusion chamber and spleen colony assay of murine hematopoietic stem cells
.
J. cell. Physiol
.
78
,
65
72
.
Cudennec
,
C. A.
,
Thiery
,
J. P.
&
Le Douarin
,
N. M.
(
1981
).
In vitro induction of adult erythropoiesis in early mouse yolk sac
.
Proc. natn. Acad. Sci., U.S.A
.
78
,
2412
2416
.
Dieterlen-Lièvre
,
F.
(
1975
).
On the origin of haemopoietic cells in the avian embryo: an experimental approach
.
J. Embryol. exp. Morph
.
33
,
607
619
.
Dieterlen-Lièvre
,
F.
,
Beaupain
,
D.
&
Martin
,
C.
(
1976
).
Origin of the erythropoietic stem cells in avian development: shift from the yolk sac to an intra-embryonic site
.
Annis. Immunol. (Inst. Pasteur)
127c
,
943
946
.
Gabe
,
M.
(
1968
).
Techniques histologiques
.
Paris
:
Masson et Cie
.
Hollyfield
,
J. G.
(
1966
).
The origin of erythroblasts in Rana pipiens Tadpoles
.
Devi Biol
.
14
,
461
480
.
Marks
,
P. A.
&
Rifkind
,
R. A.
(
1972a
).
Protein synthesis: its control in erythropoiesis
.
Science
175
,
955
961
.
Marks
,
P. A.
&
Rifkind
,
R. A.
(
1972b
).
Fetal liver erythropoiesis and yolk sac cells
.
Science
v11
,
187
.
Martin
,
C.
,
Lassila
,
O.
,
Nurmi
,
T.
,
Eskola
,
J.
,
Dieterlen-Lièvre
,
F.
&
Toivanen
,
P.
(
1979
).
Intraembryonic origin of lymphoid stem cells in the chicken: studies with sex chromosome and IgG allotype markers in histocompatible yolk sac embryo chimaeras
.
Scand. J. Immunol
.
10
,
333
338
.
Moore
,
M. A. S.
&
Owen
,
J. J. T.
(
1967a
).
Chromosome marker studies in the irradiated chick embryo
.
Nature, Lond
.
215
,
1081
1082
.
Moore
,
M. A. S.
&
Owen
,
J. J. T.
(
1967b
).
Stem cell migration in developing myeloid and lymphoid systems
.
Lancet
II
,
658
659
.
Moore
,
M. A. S.
&
Metcalf
,
D.
(
1970
).
Ontogeny of the haemopoietic system: yolk sac origin in vivo and in vitro colony forming cells in the developing mouse embryo
.
Brit. J. Haemat
.
18
,
279
296
.
Niewish
,
H.
,
Hajdik
,
L
,
Sultanian
,
L
,
Vogel
,
H.
&
Matioli
,
G.
(
1970
).
Hemopoietic stem cell distribution in tissues of fetal and newborn mice
.
J. cell. Physiol
.
76
,
107
116
.
Perah
,
G.
&
Feldman
,
M.
(
1977
).
In vitro activation of the in vivo colony-forming units of the mouse yolk sac
.
J. cell. Physiol
.
91
,
193
200
.
Pfeffer
,
P.
&
Böyum
,
A.
(
1977
).
Bone marrow cell culturing in double diffusion chamber
.
Scand. J. Haematol
.
18
,
129
136
.
Symann
,
M.
,
Fontebuoni
,
A.
,
Quesenberry
,
P.
,
Howard
,
D.
&
Stohlman
,
F.
Jr
(
1976a
).
Fetal hemopoiesis in diffusion chamber cultures. I. The pattern of the pluripotent stem cell growth
.
Cell. Tissue Kinet
.
9
,
41
49
.
Symann
,
M.
,
Quesenberry
,
P.
,
Fontebuoni
,
A.
&
Stohlman
,
F.
Jr
(
1976b
).
Fetal hemopoiesis in diffusion chamber cultures. II. Cell proliferation and differentiation
.
Nouv. Rev. Franc. Hemat
.
16
,
321
328
.
Symann
,
M.
,
Quesenberry
,
P.
,
Fontebuoni
,
A.
,
Howard
,
D.
,
Ryan
,
M.
&
Stohlman
,
F.
Jr
(
1976C
).
Fetal hemopoiesis in diffusion chamber cultures. III. The effect of neutropenia
.
Blood
48
,
283
291
.
Symann
,
M.
,
Anckaert
,
M. A.
,
Cordier
,
A.
,
Rodhain
,
J.
&
Sokal
,
G.
(
1978
).
Murine yolk sac hematopoiesis studied with the diffusion chamber technique
.
Expl Hemat
.
6
,
749
759
.
Till
,
J. E.
&
McCulloch
,
E. A.
(
1961
).
A direct measurement of the radiation sensitivity of normal mouse bone marrow cells
.
Radiat. Res
.
14
,
213
222
.
Toivanen
,
A.
,
Eskola
,
J.
&
Toivanen
,
P.
(
1976
).
Restorative effects of different embryonic cells transplanted into immunodeficient chick embryo
.
Annis Immunol. (Inst. Pasteur)
127c
,
923
929
.
Vilpo
,
J. A.
&
Vilpo
,
T.
(
1976
).
Growth of haematopoietic cells of mouse fetal liver in diffusion chambers
.
Acta haemat. Basel
55
,
224
229
.
Volpe
,
E. P.
&
Turpen
,
J. B.
(
1977
).
Lymphocyte differentiation and allograft reactivity: experimental studies on the origin of thymic lymphocytes
.
Transplant. Proc
.
9
,
785
788
.