/. Embryol. exp. Morph. Vol. 36, 2, pp. 247-260, 1976
Printed in Great Britain
247
Properties and development of erythropoietic
stem cells in the chick embryo
By JACQUES SAMARUT 1 AND VICTOR NIGON 1
From the Departement de Biologie generate et appliquee,
Universite Claude-Bernard, Lyon I, France
SUMMARY
1. When injected into irradiated chickens, haemopoietic stem cells give rise to well-defined
erythrocytic colonies in the host marrow. Such stem cells (CFU-M = Colony Forming Unit
in Marrow) have been found in different tissue of the chick embryo (yolk sac, blood, marrow).
Analysis of the properties of CFU-M reveals that they represent two classes of stem cells:
pluripotent stem cells mainly in adult marrow and erythrocytic-committed stem cells present
in yolk sac.
2. Yolk sac contains the main pool of CFU-M during the major part of embryonic life.
In the blood of 6-day-old embryo, there are three or four times more CFU-Ms than in the
yolk sac; they are no longer detected in the blood after the 16th day of incubation. During
development of the marrow, stem cells are actively differentiating and their total number
remains the same from 16 days to hatching.
INTRODUCTION
The origin of stem cells in the different haemopoietic organs of the vertebrate
embryo is still rather controversial.
According to Moore & Owen (1965, 1967), the stem cells of the chick embryo
yolk sac colonize every embryonic haemopoietic tissue through the blood
circulation. The same hypothesis has been proposed by Moore & Metcalf
(1970) for the development of haemopoietic system in the mouse embryo.
However, Dieterlen-Lievre (1975) proved that spleen and thymus of quail
embryo grafted on to chick yolk sacs show no colonization by chick stem cells,
and Rifkind, Chui & Epler (1969) suggest that in the embryonic liver of the
mouse, haemopoietic stem cells originate from the hepatic mesenchyme. We
have attempted to discover to what extent development of vitelline stem cells
could effectively satisfy the conditions required to prove Moore & Owen's
theory.
We have demonstrated that the injection of haemopoietic tissues into
irradiated chicken induced the appearance of erythrocytic colonies in the
marrow of the grafted chicken. The linearity between the number of injected
1
Authors' address: Departement de Biologie generate et appliquee, Universite ClaudeBernard, Lyon I, 43, boulevard du 11 novembre 1918, 69621 Villeurbanne. France.
248
J. SAMARUT AND V. NIGON
cells and the number of colonies in one tibia, as well as the synchronous development of cells in each colony prove the clonal origin of these colonies. The cells
which give rise to these clones have been called CFU-M (Colony Forming Unit
in Marrow) (Samarut & Nigon, 1975). The methodology used for detection
precludes, among CFU-M, any distinction between pluripotent stem cells and
erythroid-committed cells if the latter are able to multiply extensively.
We analysed the production of CFU-Ms in grafted chickens by measuring the
number of erythroid cells arising from CFU-Ms implanted into the marrow.
These observations showed that CFU-Ms from adult marrow and embryonic
blood exhibit similar kinetic parameter and give rise to stable haemopoietic
populations. Conversely, CFU-Ms from yolk sac generate clones in which the
rate of differentiation is higher than the rate of multiplication (Samarut & Nigon,
1975). In addition, synthesis of foetal haemoglobin (Hb F) in the grafted
chickens appears three to four times higher when the injected stem cells come
from yolk sac than if they originate from adult marrow (Godet, Samarut &
Nigon, 1974). These observations led us to propose the existence in the chick
embryo of two kinds of haemopoietic stem cells, respectively in the yolk sac and
in the circulating blood.
In the present work, we attempt to state precisely some of their properties.
Are they cells with unlimited proliferative abilities or with limited self renewal ?
Study of the numerical evolution of some of them during development leads us
to propose a hypothetical model about the relationship between erythropoietic
forms during chicken embryogenesis.
MATERIAL AND METHODS
Most of the techniques have been previously described (Samarut & Nigon,
1975) and will only be briefly reported for some minor changes.
(1) Preparation of the injected cellular suspensions
A slight modification has been made to the preparation of the vitelline cells
from 15-day embryos and older ones. After the first washing, the suspensions
are kept motionless for 15 min at 4 °C allowing sedimentation of a part of the
contaminating yolk. The floating suspension is collected, washed three times
with phosphate-buffered saline (PBS) without Ca 2+ or Mg 2+ .
The contamination of the vitelline cell suspensions by the blood cells was
estimated by counting mature erythrocytes. The observed contaminations at the
different incubation stages are reported in Table 1.
(2) Measurement of CFU-M frequencies
Three-week-old Leghorn chickens received a first irradiation of 760 R
(roentgens) followed by a second irradiation of 990 R after one day. The
cellular preparations are injected in the wing vein less than 4 h after the second
Erythropoietic stem cells in the chick embryo
249
Table 1. Contamination of vitelline preparations by cells from the
circulating blood
It is assumed that the more mature erythrocytic cells in yolk-sac preparations
belong entirely to the fractions of circulating blood. For each embryonic stage, if
a is the frequency of these cells in the blood and b their frequency in the preparations,
contamination is expressed, in percentage, by the proportion lOOb/a.
Embryonic age (days)
Percentage, in the
preparations, of vitelline
cells originated in the
circulating blood
6
9
11
13
15
16
18
20
64
41
34
22
21
20
30
60
irradiation. A few days after the graft, the tibia is carefully opened, the marrow
gently removed and, after staining with benzidin reaction, the colonies appearing on the surface are counted (Samarut & Nigon, 1975).
(3) Haematological parameters
Erythropoiesis in the circulating blood
Erythropoietic activity is estimated in the circulating blood by the frequency
of polychromatophilic erythroblasts which are homologous to mammalian
reticulocytes. Their rate is measured by counting their numbers in at least 500
cells on smears stained with May Griinwald Giemsa.
Enumeration of the cells in tibial marrow
Chickens are killed by decapitation. One or two tibias are taken from chickens
or embryos and the marrow is washed out with Hanks balanced salt solution
(HBSS). The cell suspension is homogenized in a known volume of HBSS by
successive passages through a 10 ml syringe without needle, then counted in a
Malassez haematimeter. The total number of cells is obtained by multiplying the
cellular concentration by the collected volume of suspension.
Enumeration of the cells in the yolk sac
For each embryonic stage, the vitelline membranes from each of five embryos
are harvested in 5 ml PBS, containing 0-25 % trypsin. Suspensions are incubated
for 30 min at 38 °C with continuous agitation, then homogenized by successive
aspirations in a Pasteur pipette and finally washed twice in PBS by centrifugations for 15 min at 300 g. The pellets are harvested in a known volume of PBS,
suspended with a Pasteur pipette and cells counted as for the marrow suspensions.
250
J. SAMARUT AND V. NIGON
Cellular composition of the injected suspensions
When only one tibia is used for cell enumeration the other one is taken to
obtain smears from several levels along the length of the marrow. For
vitelline cells and if both tibias are used for cell enumeration, slides are
made from cell suspensions after lOmin centrifugation at 300 g. The pellets
are suspended in a few drops of decomplemented chick serum, the suspensions
homogenized with a Pasteur pipette and a drop is spread on a glass slide. The
smears are stained with May Griinwald Giemsa and the cellular composition is
estimated by counting about 1000 cells. The different cell classes are identified
according to Lucas & Jamroz (1961).
RESULTS
1. Potentialities of erythropoietic stem cells CFU-M
In order to state precisely the potentialities of CFU-M we analysed the
kinetics of erythropoietic recovery in grafted chickens.
1.1. Evolution of the number ofmedullar colonies after stem cell grafting
The counting of colonies is impossible before the 6th day post-grafting, as the
colonies are too small and there is too little contrast between them and the
faintly-stained background.
Colonies which are observed on the 10th day exhibit a size two to three times
that of the 6-day-old colonies. From the 10th day on, minute colonies (less than
0-1 mm of diameter) arise both in grafted and non-grafted irradiated chickens.
They probably represent endogenous stem cells, whose development was impaired by irradiation, but have recovered.
Figure 1 shows the changes in the number of medullar erythrocytic colonies
between the 6th and the 12th day after injection of marrow or vitelline cells
from 11-day-old embryo. After bone marrow grafting the number of colonies
in one tibia stays constant until the 12th day. On the contrary, 80 % of the
colonies arising from vitelline CFU-M disappear between the 6th and the 10th
day.
1.2. Frequency ofpolychromatophilic erythroblasts in the blood of grafted chickens
Figure 2 represents the changes in the frequency of circulating polychromatophilic erythroblasts in irradiated controls and chickens grafted with cells
harvested either from adult marrow or from a 11-day yolk sac. Vitelline graft
results in an earlier appearance of polychromatophilic erythroblasts. The
curves show a first maximum on the 6th day and a second one on the 10th day.
although these results must be cautiously interpreted we would like to advance
the hypothesis that production from vitelline CFU-M represents the resultant
of the respective products of a rapid developing population with a short life
Erythropoietic stem cells in the chick embryo
10
Days post-grafting
251
12
Fig. 1. Evolution of the number of medullar colonies in grafted chickens. Threeweek-old chickens are irradiated and injected with 107 cells from adult marrow or
2-107 cells from 11-day-old yolk sac. Chickens are killed at various times after
grafting and medullar colonies are numbered in each of the tibias of five chickens.
Values represent percentage of the number of colonies on the 6th day. Vertical bars
on each point are standard errors. A - - A, Vitelline cells graft; •
• . adult
marrow graft.
time and of a population whose characteristics are similar to those of adult
marrow. These two hypothetical populations have been drawn with dotted line
on Fig. 2.
1.3. Erythropoietic potentialities of medullar colonies in grafted chickens
In order to test the presence of CFU-M in medullar clones, the marrow of
primary recipients Rl5 grafted with adult marrow or with vitelline cells from
11-day-old embryo, has been harvested on the 6th day post-grafting and injected
into secondary irradiated recipients R2. The results are shown in Table 2.
Marrow of chicken grafted with adult marrow contains, on the 6th day, a great
number of CFU-Ms, whereas marrow of chicken injected with vitelline cells are
very poor in stem cells.
252
J. SAMARUT AND V. NIGON
Days post-grafting
Fig. 2. Evolution of the frequency of polychromatophilic erythroblasts in the blood
of irradiated controls and grafted chickens. Irradiated grafted chickens are injected
with 4-108 cells from adult marrow or yolk sac of 11-day embryo. Each point is the
mean with its standard error calculated on at least five determinations. The dotted
lines represent the two erythropoietic populations arising from vitelline stem cells.
See text for explanation. • - - • , Vitelline cell graft; #
• , adult marrow
graft; O
O, irradiated controls.
Table 2. Potentialities ofmedullar colonies
Irradiated chickens Ri are injected with 107 to 108 cells harvested in adult marrow
or yolk sac of 11-day-old embryo. On the 6th day post-grafting, the marrow of these
primary recipients Rx is harvested and injected to secondary recipients R2. Number
of colonies per tibia is determined in R2 chickens on the 6th day after retransplantation. Each value represents the mean with its standard error estimated on both
tibias of at least five chickens. Confluent colonies means more than 50 colonies per
tibia.
Nature of cells
grafted to Rx
recipients
Adult marrow
Vitelline cells from
11-day-old embryo
Number of medullar
colonies observed in
one tibia on the 6th
day and injected to
R2 recipients
Number of medullar
colonies per tibia in
R2 recipients on the
6th day after
retransplantation
10-1 ±1-8
7-1 ±1-9
Confluent
0
Erythropoietic stem cells in the chick embryo
10 -
253
14-6 + 4-3 on the 8th day
4
i
i
8 !
6 -
i
i
i
i
i
i
i
i
i
i
•/A
4 -
V
2 -
i
6
i
10
Embryo
i
i
14
18
f 1
Hatching
i
i
5
9 Days
Chicken
Fig. 3. Evolution of the cellular contents of haemopoietic organs. Values presented
for total blood cells are those estimated by Romanoff (1960). Total numbers of cells
in the yolk sac have been corrected after deduction of the contaminating blood cells.
Total number of cells per yolk sac ( •
• x 10~8); per marrow of one tibia
7
(O — Ox 10~ ) and in the whole blood circulation ( • - - • x 10~9). Each point
represents the mean and standard error for five determinations.
2. Development of embryonic erythropoietic sites
2.1. Evolution of total cellular populations (Fig. 3 and Table 3)
Cell numbers obtained from yolk sac increase about 25 times between the
6th and the 16th incubation day. After the 16th day, these numbers decrease
rapidly. Between the 16th and the 20th embryonic day the medullary cell
population remains steady. It starts to increase at the time of hatching. Tibial
cell number grows five-fold during the first post-hatching week. On the 16th
embryonic day, the marrow is populated only by erythrocytic cells. Later the
frequency of granulocytic cells, mostly at young myelocytic stages, increases
reaching a maximum at hatching. After hatching, marrow recovers erythropoietic dominance.
After the 6th incubation day, basophilic erythroblasts are no longer found in
the circulating blood population, which mostly comprises polychromatic
erythroblasts. From hatching onwards, circulating blood contains almost exclusively mature erythrocytes.
EMB 36
254
J. SAMARUT AND V. NIGON
Table 3. Cell composition of yolk sac and marrow
In yolk sac, a high proportion of cells are erythroblasts at late maturation stages
and epithelial cells. These two cellular types are frequently damaged in the smears
and could not be distinguished with security; therefore they have not been counted
separately. For yolk sac values, contaminating blood cells have been substracted.
Values are expressed as percentages.
Stage of
Marrow
embryonic (E)
and postProerythrohatching (P)
Erythro- Granublasts
development
locytic
+ basophilic cytic
(days)
cells
erythroblasts cells
Yolk sac
^
Other
cell
types
E6
E9
Eu
E13
E«
El6
Eis
E20
Pi
p8
2-8
4-8
2-4
80
180
75
61
51
52
75
160
37-6
44-2
360
16-8
90
1-4
4-8
120
8-2
(
ProerythroGranublasts
+ basophilic locytic
erythroblasts cells
330
60
11-2
16-7
13-4
8-3
20
0
30
1-2
1-9
2-5
31
1-3
50
Other
cell
types
640
92-8
86-9
80-8
83-5
90-4
930
2.2. Evolution of CFU-M number
The proportion of tibial colony number to number of grafted cells gives the
apparent frequency of CFU-M in the injected preparations. The real CFU-M
frequency differs from the apparent one by a/factor which could be determined
after Siminovitch, McCulloch & Till (1963).
We make the following assumptions: (a) /factor stays constant for each
injected tissue, whatever the age of the donor from which it is harvested;
(b) grafting efficiency is similar for embryonic blood CFU-Ms and for medullar
CFU-Ms. This assumption is shown to be rated in previously described observations which demonstrate similar kinetic parameters for CFU-Ms of the
embryonic blood and for those of the adult marrow when injected into irradiated
chickens (Samarut & Nigon, 1975).
2.2.1. Frequency of CFU-M. Apparent frequencies of CFU-Ms in haemopoietic tissues during development are plotted in Fig. 4. The corresponding
estimates of real frequencies are given in Table 4.
Frequency of CFU-Ms in yolk sac increases between the 6th and the 11th
embryonic day. Afterwards, it remains constant till hatching. Between the 16th
day and hatching, CFU-M-frequency in the marrow decreases about three-fold;
it increases after hatching. CFU-M-frequency in embryonic marrow is higher
than in yolk sac. In the adult fowl, frequencies vary according to the age of the
animals.
Erythropoietic stem cells in the chick embryo
14
.Embryo
5
Chicken
255
150 Days
75
Adult
Fig. 4. Apparent frequencies of CFU-M in haemopoietic tissues. The apparent
frequency of CFU-M is determined by the number of tibial colonies per 107 cells
injected into 3-week-old irradiated chickens. For yolk sac, numbers of grafted cells
have been corrected by substracting blood cell contamination. Apparent frequency
of CFU-M in circulating blood ( • - - * ) ; in yolk sac ( •
• ) and in marrow
(O — O). Vertical bars represent standard error. Each measurement has been
made on at least five determinations.
Blood CFU-Ms are no longer detected after the 16th embryonic day. Before
this stage their frequency remains nearly constant.
2.2.2. Total CFU-M number in haemopoietic tissues (Fig. 5). Total number of
vitelline CFU-Ms increases between the 6th and the 15th incubation day. Until
the 11th day, the population grows nearly as an exponential with 20-h doubling
time. After the 15th day the number of CFU-Ms decreases.
The population of medullar CFU-Ms remains stable during the last days of
embryonic development. From hatching onwards, the number of these cells in a
tibia increases approximately exponentially with a 55-h doubling time.
In the 6-day-old embryo the total number of CFU-Ms in circulating blood is
about three times higher than in yolk sac (Table 4). Blood CFU-Ms-number
reaches a maximum (8000) on the 13th day. Between the 6th and 13th incuba17-2
256
J. SAMARUT AND V. NIGON
Table 4. Estimates of real frequencies and total numbers of CFU-M
Frequencies as numbers of CFU-M for 107 cells are estimated from apparent
frequencies presented in Fig. 4. Values are estimated by the ratio
apparent frequency
7
'
Values for / a r e respectively 0023 for CFU-M of yolk sac and 0049 for those of
marrow and blood circulation (Samarut & Nigon, 1975). Total numbers of CFU-M
are determined by multiplying real frequencies by total numbers of harvested cells.
Stage of
embryonic (E)
or-posthatching (P)
development
(days)
Real frequency of CFU-M
(for 107 cells)
f
Marrow Yolk sac
E6
E9
Eu
61
87
E13
Ei6
Ei8
E20
Pi
p8
Adult 2-5 months
5 months
Total number of CFU-M
629
531
367
198
298
475
269
235
204
226
157
218
248
Blood
One tibia
29
20
29
22
8
4
2
Blood
Yolk sac circulation
180
457
362
405
595
4350
1830
8000
10400
17500
11800
8050
2280
640
2120
5490
8080
4130
2350
1470
tion day, the growth of the blood CFU-M population is nearly exponential with
20-h doubling time. Afterwards the population decreases until it contains only
1000 CFU-Ms in the 18-day-old embryo.
DISCUSSION
(1) Nature of CFU-M
Kinetic and cytological studies of erythrocytic colonies appearing in the
marrow of irradiated and grafted chickens led us to suggest that the colonies
represent clones (Samarut & Nigon, 1975). This original study showed different
values for the developmental parameters of the clones according to their origin.
From a certain time on, the rate of differentiation appears to be higher than the
rate of multiplication in clones arising from vitelline CFU-Ms suggesting that
these clones would rapidly vanish.
The results of this present work confirm this hypothesis: (a) on the 10th day
post-grafting most of the colonies obtained from vitelline CFU-Ms have disappeared although the colonies from adult marrow persist; (b) vitelline clones
contain very few CFU-Ms compared with clones from adult marrow.
Erythropoietic stem cells in the chick embryo
10
14
Embryo
5
257
9 Days
Chicken
Fig. 5. Total number of CFU-M in haemopoietic tissues. Numbers have been taken
from Table 4. Total number of CFU-M per yolk sac ( •
• ) ; per tibia (O - - - O)
or in the whole blood circulation (*
•).
These properties suggest the existence of two kinds of CFU-M with different
potentialities:
CFU-M I, in adult marrow, which exhibits properties of pluripotent stem
cells. Most of these cells in marrow develop an erythrocytic determination. We
indeed observed that granulocytic proliferation is about 20 times slower than
erythrocytic proliferation in marrow of grafted chickens (Samarut & Nigon,
1975). However, we cannot separate the respective responsibilities of stem cell
determination on one hand, and on the other hand, the multiplicative differential capacities of those stem cells to develop into granulocytes or erythrocytes.
258
J. SAMARUT AND V. NIGON
CFU-M II, in the yolk sac, represents a class of cell which has already got
some erythrocytic differentiation and exhibits limited self-renewal. They may be
considered as erythroid-committed stem cells.
(2) Development of haemopoiesis in yolk sac
After the 11th day of embryonic development yolk sac contains essentially
CFU-M II. Increase of vitelline CFU-M-number between the 11th and 14th day
could be explained in two ways: (a) either CFU-M II shows at this time intermediate properties and is still able to multiply for a few days giving rise to new
CFU-M Us. This ability would be lost when implanted into marrow of irradiated chickens, (b) or yolk-sac colonies are continuously added to by stem cells
arising from an extravitelline site according to the hypothesis of DieterlenLievre (1975).
Decrease of CFU-M population from 16th day on may be attributed to
several factors: (a) cessation of CFU-M-multiplication although its differentiation continues; (b) migration of CFU-Ms from yolk sac into other haemopoietic
organs. However, our observations on the tibia show that the supply of medullar
sites in CFU-M is completed on the 16th incubation day when the decrease of
vitelline CFU-Ms has just started. Therefore, it seems that migration does not
play a decisive role in disappearance of CFU-Ms in the yolk sac.
(3) Stem cells in embryonic blood
Development of CFU-Ms from embryonic blood grafted into irradiated
chickens revealed that blood CFU-Ms exhibit a kinetic pattern identical to that
of CFU-Ms from adult marrow (Samarut & Nigon, 1975). Thus, embryonic
blood contains mainly pluripotent stem cells CFU-M I.
The number of blood stem cells has been estimated by assuming equal grafting
efficiency for embryonic blood and adult marrow stem cells. Experimental
proof of such an hypothesis will be difficult because of the great dilution of stem
cells in embryonic blood. If grafting efficiency is closer to that of vitelline CFUM, the total numbers of blood stem cells plotted in Fig. 5 would have been
underestimated.
The decrease of blood stem cell number from the 12th day onwards coincides
with population growth of medullar sites. If both phenomena are really correlated, it could mean that blood CFU-Ms colonize medullar sites.
(4) Stem cells in embryonic marrow
The stem cell population in embryonic marrow remains nearly constant from
the 16th day on. So we must conclude that colonization of medullar sites occurs
before the 16th day. However, the small volume of marrow has not yet allowed
estimation of stem cell contents during that period.
Two hypotheses can be proposed to explain the decrease in the frequency of
the marrow stem cell during the end of embryogenesis: (a) differentiation rate
Erythropoietic stem cells in the chick embryo
259
of medullar stem cells is greater than their multiplication rate; (b) the development of the marrow starts from two kinds of stem cells, one of which rapidly
disappears.
After hatching, multiplication seems accelerated with respect to differentiation which leads to an increasing number of stem cells.
(5) Cellular traffic between haemopoietic organs and origin of stem cells
On the 6th incubation day there are three to four times more CFU-Ms in the
blood than in the yolk sac which may indicate that, at this stage, colonization
of the yolk sac occurs from pluripotent stem cells present in the blood circulation, in contrast to Moore & Owen (1965,1966) hypothesis. This intepretation is
in keeping with the results of Dieterlen-Lievre (1975).
Observations of Godet (1974) show that embryonic marrow assumes first a
haemoglobin production similar to that displayed by yolk-sac erythroblasts,
whereas, from the 18th incubation day onwards, the parameters of haemoglobin
synthesis change rapidly into those of an adult-type erythropoiesis. This, on the
other hand, would be in line with a correlation between yolk erythropoiesis and
the start of marrow erythropoiesis.
These results lead us to propose a working hypothesis for the relationship
of erythrocytic stem cells during embryonic development of the chick.
(1) Megalocytic erythropoiesis originates in the Wolff islands and occurs in
yolk sac and blood.
(2) Normocytic erythropoiesis develops in yolk sac from colonizing circulating
blood stem cells. Relationship between these later cells and the mother cells of
megalocytic erythropoiesis remains to be defined.
(3) Medullar erythropoiesis is initiated simultaneously by erythroidcommitted cells which have undergone predifferentiation in the yolk sac, and by
pluripotent blood stem cells. The erythroid-committed cells are characterized by
quick development and give rise to foetal-type erythropoiesis. The slower
development of the pluripotent-blood stem cells leads them to supplant erythroidcommitted cells of vitelline origin a few days after the beginning of medullar
erythropoiesis. Cessation of foetal erythropoiesis in embryonic marrow can
then be attributed to intrinsic properties of the vitelline-committed cells: their
rapid differentiation should lead to their disappearance.
RESUME
Lorsqu'elles sont injectees a des poussins irradies, les cellules souches hematopoiietiques
donnent naissance a des colonies erythrocytiques bien individualists dans la moelle des
poussins receveurs. Ces cellules souches (CFU-M: Colony Forming Unit in Marrow) sont
trouvees dans differents tissus de l'embryon (sac vitellin, sang, moelle). L'analyse des proprietes des CFU-M montre que ces cellules constituent deux populations de cellules souches:
d'une part, des cellules souches pluripotentes tres abondantes dans la moelle adulte, d'autre
part des cellules ayant deja acquis une predifferenciation erythrocytique et qui sont presentes
dans le sac vitellin.
260
J. SAMARUT AND V. N I G O N
Le sac vitellin constitue le reservoir principal de CFU-M durant la majeure partie du
developpement embryonnaire. Dans le sang de l'embryon de 6 jours, on compte trois a
quatre fois plus de CFU-M que dans le sac vitellin; ces cellules ne sont plus decelees dans le
sang apres le 16eme jour d'incubation. Durant le developpement de la moelle, les cellules
souches se differencient activement et leur nombre reste constant entre le 16eme jour du
developpement embryonnaire et l'eclosion.
This work was supported by a D.G.R.S.T. grant.
REFERENCES
DIETERLEN-LIEVRE, F. (1975). On the origin of haemopoietic stem cells in the avian embryo:
an experimental approach. /. Embryol. exp. Morph. 33, 607-619.
GODET, J. (1974). Hb F synthesis in chicken embryonic and postnatal development. Studies
in various explanted erythropoietic tissues. Devi Biol. 40, 199-207.
GODET, J., SAMARUT, J. & NIGON, V. (1974). Synthese d'Hb F chez des poussins irradies et
greffes a l'aide de tissus hematopoietiques embryonnaires et adultes. C.r. hebd. Seanc Acad.
Sci. Paris 279, 699-701.
LUCAS, A. B. & JAMROZ, C. (1961). Atlas of Avian Hematology. Agriculture Monograph 25.
Washington DC: United States Department of Agriculture.
MOORE, M. A. S. & METCALF, D. (1970). Ontogeny of the haemopoietic system: yolk sac
origin of in vivo and in vitro colony forming cells in the developing mouse embryo. Br. J.
Haemat. 18, 279-296.
MOORE, M. A. S. & OWEN, J. J. T. (1965). Chromosome marker studies on the development
of the haemopoietic system in the chick embryo. Nature, Lond. 208, 956.
MOORE, M. A. S. & OWEN, J. J. T. (1967). Chromosome marker studies in the irradiated
chick embryo. Nature, Lond. 215, 1081-1082.
RIFKIND, A. R., CHUI, D. & EPLER, H. (1969). An ultrastructural study of early morphogenetic events during the establishment of fetal hepatic erythropoiesis. /. Cell Biol., 40,
343-365.
ROMANOFF, A. L. (1960). The Avian Embryo. Structural and Functional Development.
New-York: The Macmillan Company.
SAMARUT, J. & NIGON. V. (1975). Potentialites hematopoietiques de tissus embryonnaires et
adultes analysees par greffe a des poussins irradies. /. Embryol. exp. Morph. 33, 259-278.
SIMINOVITCH, L., MCCULLOCH, E. A. &TILL, J. E. (1963). The distribution of colony-formingcells among spleen colonies. /. CellPhysiol. 62, 327-336.
Received 27 August 1975; revised 24 February 1976)