Development 102, 279-285 (1988)Printed in Great Britain © The Company of Biologists Limited 1988

279

The wall of the chick embryo aorta harbours M-CFC, G-CFC, GM-CFC

and BFU-E

FRANCOISE CORMIER and FRANCOISE DIETERLEN-LIEVRE

Institut d'Embryologie du CNRS el du College de France, 49 bis, avenue de la Belle-Gabrielle, 94736 - Nogent-sur-Mame - France

Summary

In the 3- to 4-day avian embryo, after the first wave ofhaemopoiesis which derives in the yolk sac fromhaemopoietic stem cells formed in situ, haemopoieticcells emerge in an intraembryonic site, the wall of theaorta. In this paper, we demonstrate that this siteharbours M-CFC, G-CFC, GM-CFC and late andearly BFU-E. In serum-free medium, the growth of M-CFC and GM-CFC was strictly dependent on CSFpresent in fibroblast-conditioned medium (FCM). Thegrowth of G-CFC was improved when FCM wasreplaced by a minute quantity of chicken and fetal calfserum. Like erythroid progenitors from bone mar-row, BFU-E detected here required anaemic chickenserum to differentiate into haemoglobinized cells. Thefrequency of the different types of haemopoietic pro-genitors in the aortic population was very high: 80 M-

CFC, 25 G-CFC, 4 GM-CFC and 70 BFU-E for 12 500aorta cells, i.e. two to eight times more frequent thanin the bone marrow population, depending on the typeof progenitors.

Abbreviations: ACS, anaemic chicken serum; ANAE,alpha-naphthyl acetate esterase; BFU-E, burst-formingunit, erythroid; CAE, naphthol AS-D chloro-acetateesterase; CFU-E, colony-forming unit, erythroid; CFU-GEMM, colony-forming unit, granulocytic, erythroid,monocytic, megakaryocytic; CSF, colony-stimulatingfactor; FCM, fibroblast-conditioned medium; G-CFC,granulocytic, colony-forming cell; GM-CFC, granulocyticmonocytic, colony-forming cell; M-CFC, monocytic,colony-forming cell.

Key words: chick embryo, haemopoietic progenitors,colony-forming cells, aorta, erythroid progenitors.

Introduction

During embryonic development of amniotes, earlyhaemopoiesis in the yolk sac derives from stem cellswhich segregate in situ (Romanoff, 1960). By con-trast, the rudiments of the definitive haemopoieticorgans become colonized by extrinsic stem cells (seefor review Metcalf & Moore, 1971; Le Douarin &Jotereau, 1973; Le Douarin et al. 1975; Houssaint,1981). In the avian embryo, an intraembryonic originof these cells was clearly demonstrated in chimaerascomposed of a quail embryo and a chick yolk sac(Dieterlen-Lievre, 1975).

It seems important to pinpoint the origin of theseintraembryonic stem cells, which play a role ofparamount importance. Histological studies and invivo experiments have already disclosed that theaortic wall of the early avian embryo (day 3-4 ofincubation) is a site from which haemopoietic cells

emerge (Dieterlen-Lievre & Martin, 1981; Dieterlen-Lievre, 1984). We now attempt to assess precisely thepotentialities of precursor cells present in this regionby assaying their colony-forming capacities in theclonal culture system in vitro. Our aim is to establishwhether some cells in the wall of the aorta areendowed with properties of stem cells, that is whetherthey are multipotential and selfreplicating.

In recent investigations using agar medium con-taining chicken serum, we have detected progenitorsin the wall of the aorta capable of forming macro-phage colonies (Cormier et al. 1986), which might beeither unipotent monocytic progenitors (M-CFC) orpluripotent progenitors unable to express otherpotentialities because of the culture conditions.These progenitors were selectively detected in a cellpopulation prepared from aortic walls; cells fromembryos deprived of their aorta did not yield anycolonies.

280 F. Cormier and F. Dieterlen-Lievre

In the experiments reported here, chicken serumhas been replaced in the culture media by definednutritional elements. Adding different cell-con-ditioned media then made it possible to control thesupply of colony-stimulating factors, which arerequired by the different types of haemopoieticprogenitors (Iscove, 1985). This has permitted thedetection of granulocytic and granulomonocyticprogenitors. Furthermore, the addition of ACS(Samarut, 1978) has led to the development oferythropoietic progenitors. We report characteristicsof, and quantitative data about, these different typesof progenitors and compare their concentration anddevelopmental schedule to those of progenitors in thebone marrow of young chicks.

Materials and methods

Cell suspensionsAorta cells from 4-day outbred White Leghorn chickenembryos were prepared as described previously (Cormier etal. 1986) except that we used about 50 embryos for eachexperimental series. Bone marrow cells were preparedfrom chickens younger than 1 month. Briefly, the bonemarrow was flushed from one tibia with alpha-medium(Gibco). It was dissociated by pipetting in a syringe with a22-gauge needle. Small clusters were eliminated by filteringthe cell suspensions over a nylon tissue. The single cellsuspensions were centrifuged for 15min at 1000 revs min"1

(150g). The pellet was resuspended in alpha-medium andthe cells were counted using the trypan blue exclusion test.

Sources of growth factorsFCM, used as a source of CSF, was prepared as describedby Dodge & Moscovici (1973) and subsequently concen-trated ten times on an ultrafiltration membrane PM 10(Amicon).

ACS, used as a source of erythropoietin (Samarut, 1978),was kindly provided by Dr J. Samarut (Villeurbanne,France).

Plasma clot culturesThe cultures were performed in plasma clot, which permitsbetter histological staining than agar, making the distinc-tion between macrophages and granulocytes easier.

Cells (6250 to 50 000 ml"' depending on the experiments)were seeded in a volume of 0-5 ml, either in the serum-freemedium or in alpha-medium containing chicken serum(Gibco; 10 %, v/v) for the assay of erythroid progenitors,each being supplemented with citrated bovine plasma(Gibco; 10 %, v/v) and clotted by the addition of thrombin(Produits Roche, France; 1 i.u.ml"1). Sera, tryptose phos-phate broth (Gibco), FCM or ACS were added as indi-cated.

The serum-free medium was alpha-medium containinglOmgml"1 bovine serum albumin (Boehringer-Mann-heim), 30/Jgml"1 of iron-saturated transferrin (Boehr-inger-Mannheim), 25jigmF' of soy-bean lipids (Boehr-inger-Mannheim), 5-6^gml~' of linoleic acid (Sigma),

7-8 jig ml ' of cholesterol (Prolabo) (these two last com-ponents were prepared together as described by Stewart etal. (1984)), 2xKT4M-hemin (Serva), 10~4M-alpha-thio-glycerol (Calbiochem), 1 % BME vitamins solution(Gibco), 2mM-glutamin, lOOi.u. ml"1 of penicillin andlOOjUgml"1 of streptomycin.

Cultures were performed in duplicate and incubated at40°C for 3 or 6 days in a humidified atmosphere of 5 % CO2

in air.

Harvesting and staining the plasma clot culturesPlasma clots were transferred onto a microscopic slide,dehydrated and air dried according to Lanotte's techniquefor collagen cultures (1984). Cultures were routinelystained by the May-Griinwald-Giemsa technique forsmears. The pseudoperoxidase reaction of haemoglobinwas revealed by diaminobenzidine staining according toMcLeodera/. (1974) after fixation in glutaraldehyde (5 % in0-01 M-phosphate buffer) and air drying.

Cytochemical stainings to detect ANAE, CAE, acidphosphatase or alkaline phosphatase were performed usingthe Sigma kits, 90-A1, 90-C2, 387-A or 86-R, respectively.

Results

Development of monocytic and granulomonocyticcolonies

As demonstrated previously (Cormier et al. 1986), inthe presence of chicken serum, FCM stimulated thegrowth of monocytic colonies from aorta cells. Addedto the serum-free medium, FCM induced the devel-opment of different types of colonies. In 3-daycultures, most were macrophage colonies and veryfew were granulocytic clusters (Table 1). Mixed col-onies containing granulocytes and macrophages alsodeveloped (Fig. 1A,B). Without FCM, the serum-Fig. 1. (A,B) 3-day granulomonocytic colony developedin serum-free medium with FCM (3%, v/v). May-Grunwald-Giemsa staining. Mixed colonies display atypical morphology, with a central core of tightlyassociated granulocytes surrounded by dispersedmacrophages (A, bar, 200 jJ-m). The distinction betweenmacrophages and granulocytes is clearcut; the former arelarge cells filled with vacuoles, while the latter containstainable granules (B, bar, 20um). (C) 3-day granulocyticcluster developed in 'control' culture. May-Griinwald-Giemsa staining. Granulocytes are smaller than those inthe mixed colony (B), indicating increasing maturity.They derive from a unipotent progenitor, later thanGM-CFC. They are probably immature heterophils.(D) 6-day erythroid colony developed in the presence ofanaemic chicken serum. Diaminobenzidine staining.(E-J) Enzyme histochemistry. Macrophages positive forANAE (E), CAE (F) and acid phosphatase (G).Granulocytes positive for ANAE (H), CAE (I) and acidphosphatase (J). ANAE is revealed as fine blackgranules, CAE as fine magenta granules and acidphosphatase as large magenta granules. C-J: samemagnification as Fig. IB.

H

Embryonic haemopoietic progenitors in chick 281

free medium supported exclusively the growth of raregranulocytic clusters.

Macrophages showed granulations for ANAE,CAE and acid phosphatase as demonstrated by cyto-chemical staining (Fig. 1E-G). They did not containgranules with alkaline phosphatase activity.

In the range studied, the number of macrophagecolonies was linearly dependent on the number ofcells seeded (Fig. 2A). The dose-response curve ofthe number of macrophage colonies to the FCMconcentration displayed a typical sigmoidal pattern(Fig. 3A).

Development of granulocytic coloniesFibroblast-conditioned medium contained chickserum (5%, v/v), fetal calf serum (5%, v/v) andtryptose phosphate broth (10%, v/v). In order todiscern the activity supported by the secretion prod-ucts of fibroblasts from the one provided by thesecomponents, we have performed cultures in thepresence of chick serum, fetal calf serum and tryptosephosphate broth in concentrations similar to thosereached when FCM is added, that is to say 1-5%chick serum, 1-5 % fetal calf serum and 3 % tryptosephosphate broth. In these so-called 'control' cultures,numerous clusters or little bursts of granulocytesdeveloped (Fig. 1C). Their number was greater thanin serum-free medium, supplemented or not withFCM (Table 1). It was linearly dependent on thenumber of cells seeded (Fig. 2B). A few clustersof macrophages also grew (Table 1). Granulocytescontained ANAE, CAE and acid phosphatase(Fig. 1H-J).

Development of erythroid coloniesIn order to detect erythropoietic progenitors, ACS,which permits the development of BFU-E andCFU-E from bone marrow cells (Samarut & Bouab-delli, 1980), was added to cultures containing normalchicken serum (10 %, v/v). Aorta cells seeded in suchcultures yielded erythroid colonies, which were eitherisolated colonies or bursts. Because of the high

Table 1. Progenitors of the granulomonocytic lineagedeveloped from 12 500 aorta cells

serum-freemedium

+ FCM(3%, v/v)

"control'medium

M-CFCG-CFCGM-CFC

08-6 ±1-4

0

82-3 ±112 ±0-5

3-8 ±0-7

2-8 ± 124 ± 4-6

0

Serum-free medium: see Materials and Methods.'Control' medium: serum-free medium supplemented with

chick scrum (1-5%, v/v), fetal calf serum (1-5%, v/v) andtryptose phosphate broth (3 %, v/v).

Each number is the mean result of seven experiments ± S.D.

Table 2. Effect of anaemic chicken serum on thenumber of erythroid colonies developed in 3-day and

6-day cultures of 12 500 aorta cells

-ACS + ACS*

3-day culture6-day culture

19-4 ±3-73-6 ±0-95

65-6 ±8-569-3 ±7

Each number is the mean result of five experiments ± S.D.• ,10%, v/v.

level of chicken serum, numerous macrophage colon-ies also developed. In 3-day cultures, most of theerythroid colonies were composed of basophilicerythroblasts. Some also contained a few benzidine-positive erythroblasts. In 6-day cultures, the numberof erythroid colonies was similar to that in 3-daycultures. They were larger and composed of haemo-globinized (benzidine-positive) cells (Fig. ID). In 3-day cultures where ACS was omitted, erythroidcolonies also developed but to a lesser extent(Table 2). They were small bursts composed of baso-philic erythroblasts. In 6-day ACS-free cultures, allcolonies had disappeared (Table 2).

In the range studied, the relationship between thenumber of erythroid colonies developed in 6-daycultures with ACS and the number of cells seededfollowed a linear curve (Fig. 2B). The dependence oferythroid colony formation on ACS concentration isshown in Fig. 3B.

Development of colonies from bone marrow cellsThe haemopoietic activity of bone marrow cells fromyoung chickens was chosen as a reference to evaluatethe cloning efficiency of aorta cells. Bone marrowcells were cultured in the different conditions used foraorta cells.

In serum-free medium supplemented with FCM,about 40 macrophage colonies developed from 12500bone marrow cells (Fig. 2A). In the 'control' cultures,about 10 granulocytic colonies developed from 12500cells (Fig. 2B). In 6-day cultures containing ACS,about 25 erythroid colonies developed from 25 000bone marrow cells (Fig. 2B). No colonies grew with-out ACS.

Discussion

In the chick and quail embryo, at 6-8 days ofincubation, before the definitive haemopoietic organsbecome functional, haemopoiesis occurs in diffusefoci located in the dorsal mesentery ventral to theaorta (Miller, 1913; Romanoff, I960). Erythroid andgranuloid precursor cells, which differentiate there,emerge in situ (Dieterlen-Lievre & Martin, 1981).

282 F. Cormier and F. Dieterlen-Lievre

12500Number of cells plated

25000

100-

o

ez

125001Number of cells plated

25000

Fig. 2. Concentrations of progenitors compared inembryonic aorta (solid symbols) and bone marrow cell(open symbols) populations. Each point represents themean value of n experiments ± standard deviation.(A) Relationship between the number of macrophagecolonies and the number of cells seeded in serum-freemedium containing FCM (3 %, v/v). • , cells from thewall of the aorta (n = 5); O, bone marrow cells (n = 5).(B) Relationships between the number of granulocyticand erythroid colonies and the number of cells seeded.Granulocytic colonies developed from aorta cells( • , n = 5) and from bone marrow cells ( • , n = 3) seededin 'control' cultures. Erythroid colonies developed fromaorta cells (A, n = 2) and from bone marrow cells(A, n = 2) seeded in cultures containing chick serum(10 %, v/v) and ACS (10 %, v/v).

These foci also harbour cells capable of undergoinglymphoid differentiation (Eskola, 1977; Lassila et al.1979, 1980). In 3- to 4-day embryos, the involvementof the region of the aorta in haemopoietic processesof the mesentery is indicated by morphologicalaspects and experimental data (Dieterlen-Lievre,1984).

0-375 0-75 1-5 3 6Fibroblast-conditioned

medium concentration (%, v/v)

1-25 2-5 5 10Anaemic chicken serum concentration (%, v/v)

Fig. 3. Requirements of aorta monocytic and erythroidprogenitors for growth factors. Each point represents themean value of n experiments ± standard deviation.(A) Dose-effect curve of FCM on the development ofmacrophage colonies from 12500 cells seeded in serum-free cultures (n = 3). (B) Dose-effect curve of ACS onthe development of erythroid colonies from 12 500 cellsseeded in plasma clot culture containing chick serum(10%, v/v) (n = 2).

Our earlier experiments demonstrating the pres-ence of monocytic progenitors in the wall of theembryonic aorta (Cormier et al. 1986) were per-formed in a medium containing chicken serum, whichhad been found to be an absolute requirement for thecloning of avian haemopoietic cells (Moscovici &Moscovici, 1973). However, as already reported forbone marrow cells (Dodge & Moscovici, 1973; Dodgeet al. 1982), serum-containing medium was not per-missive for the expression of all developmental poten-tialities of cells from the aorta. It exclusively sup-ported proliferation and differentiation in themonocytic lineage.

Embryonic haemopoietic progenitors in chick 283

Recently, Dodge & Sharma (1985) reported thatthe nutritional requirements of the avian granulocyticand monocytic progenitors are similar to those of themurine haemopoietic progenitors (Iscove, 1983). Theserum-free medium used in the present experimentsis similar to the one described for murine earlyerythroid progenitor cultures (Cormier et al. 1984,1985; Stewart et al. 1984; Eliason & Odartchenko,1985). By comparison with the medium used byDodge & Sharma (1985), the presence of hemin is themain difference. We have found that hemin enhancedmaturation of granulocytes in clones developed fromchicken bone marrow cells (data not shown).

By culturing cells from the wall of the aorta inserum-free medium supplemented with FCM for 3days, it has been possible to obtain monocytic,granulocytic and mixed colonies, derived, respect-ively, from M-CFC, G-CFC and, finally, earlierprogenitors, GM-CFC. M-CFC and GM-CFCrequired CSF produced by fibroblasts, since they didnot grow in 'control' cultures deprived of FCM. Onthe other hand, the growth of G-CFC was consider-ably improved when FCM was replaced by minutequantities of serum. In comparison with the culturemedium used previously (Cormier etal. 1986), replac-ing most chicken serum by nutritional elementschanged the pattern of development in favour ofgranulocytic progenitors. This has also been de-scribed with bone marrow cells (Dodge & Sharma,1985). These authors demonstrated that the mono-cytic activity of chicken serum is not supported by agrowth factor but by a 'Monocytic DifferentiationFactor' which might either induce the conversion ofearly granulocytic precursors into monocytic cells orcommit pluripotent progenitors towards monocyticdifferentiation. Such a factor probably also exerts itsactivity on progenitors present in the wall of theaorta.

Both macrophages and granulocytes differentiatedin our cultures displayed acid phosphatase activity,but no alkaline phosphatase activity, a pattern ofenzymatic activity previously described in ultrastruc-tural studies of avian cells (Daimon '& Caxton-Martins, 1977). To our knowledge, there are no datain the literature about ANAE and CAE in avianleukocytes. Chicken macrophages and granulocytesin our culture expressed these two enzymes, contraryto mouse and human macrophages and granulocytes,which can be distinguished on the basis of theseenzymatic activities (Willcox et al. 1976; Konwalinkaet al. 1980; Lanotte, 1984). Important biochemicaldifferences between mammals and avian granulocyteshave already been reported, in particular the lack ofalkaline phosphatase and myeloperoxidase in thegranules of avian myelocytes (Brune & Spitznagel,1973; Daimon & Caxton-Martins, 1977).

In the present report, we also demonstrate that thewall of the aorta harbours progenitors belonging tothe erythroid lineage. The size and the developmentalcourse of erythroid colonies developing in the pres-ence of ACS indicate that they derived from earlyprogenitors (BFU-E) rather than late progenitors(CFU-E; Samarut & Nigon, 1976). According to thesequential development process demonstrated formouse erythroid progenitors (Iscove, 1978; Iscove etal. 1982), kinetics of development and growth re-quirement further reveal two categories of progeni-tors, primitive BFU-E, ACS-independent during thefirst 3 days of culture, and late BFU-E requiring ACSfrom the start. The primitive BFU-E from mousebone marrow require burst-promoting activity (BPA)to undergo initial proliferation (Iscove, 1978). In ourcultures, no source of avian BPA, such as chickenspleen-conditioned medium (Samarut & Bouabdelli,1980), was added. Either BFU-E detected in ourcultures did not require this factor, or BPA wassupplied by normal chicken serum.

Until now, in mammals, no experimental resultshave questioned the theory that all haemopoieticstem cells derive from the yolk sac (Moore & Owen,1967). In recent experiments, no clonogenic progeni-tors could be detected among cells dissociated from 8-day mouse embryos separated from their yolk sac(Wong et al. 1986). However, cell budding from theaortic endothelium, similar to that seen in the avianembryo, has been described in mammalian embryos(Jordan, 1917). In the mouse, this process occurs atday 10 of gestation (Smith & Glomski, 1982). It ispossible that Wong et al. (1986) have not detectedclonogenic progenitors because of the very early stageat which their experiments have been performed. Atthat time, cellular interactions critical for the emerg-ence of haemopoietic stem cells may not yet haveoccurred. In this regard, it is interesting that 9-daymouse yolk sac cells become capable of giving rise tospleen colonies in irradiated adults, only if the yolksac undergoes organ culture prior to dissociation andinjection (Perah & Feldman, 1977).

In conclusion, in early chick development, thedorsal aorta with surrounding intraembryonic mesen-chyme harbours a population which is very rich,compared to bone marrow population, in haemopoi-etic progenitors committed towards different lin-eages. In recent experiments, we have also detectedpluripotent cells which could be similar to mouse(Johnson, 1980) and human (Fauser & Messner,1979) CFU-GEMM (data not shown). This study,with our earlier findings (Cormier etal. 1986), arguesfor the important role of the wall of the aorta inintraembryonic haemopoiesis.

284 F. Cormier and F. Dieterlen-Lievre

We thank Dr J. Samarut for kindly providing anaemicchicken serum and Mrs M. Klaine for her skilful technicalassistance.

This work was supported by the Centre National de laRecherche Scientifique.

References

BRUNE, K. & SPITZNAGEL, J. K. (1973). Peroxidaselesschicken leukocytes: Isolation and characterization ofantibacterial granules. J. Infect. Diseases 111, 84-94.

CORMIER, F., BAINES, P., LUCIEN, N. & BOFFA, G. A.(1985). Complete replacement of serum in cultures ofmurine primitive erythroid and multipotentialprogenitor cells: absolute requirement for spleenconditioned medium. Cell Diff. 17, 261-269.

CORMIER, F., DE PAZ, P. & DIETERLEN-LIEVRE, F. (1986).In vitro detection of cells with monocytic potentiality inthe wall of the chick embryo aorta. Devi Biol. 118,167-175.

CORMIER, F., LUCIEN, N. & BOFFA, G. A. (1984).D6veloppement des prog6niteurs 6rythroides prdcoces(BFU-E) de la moelle osseuse de Souris dans un milieude culture d^pourvu de s6rum. Effet de I'h6mine. C. r.hebd. Sianc. Acad. Sci. Paris 299, 143-146.

DAIMON, T. & CAXTON-MARTINS, A. E. (1977). Electronmicroscopic and enzyme cytochemical studies ongranules of mature chicken granular leukocytes. J.Anat. 123, 553-562.

DIETERLEN-LIEVRE, F. (1975). On the origin ofhaemopoietic stem cells in the avian embryo: anexperimental approach. J. Embryol. exp. Morph. 33,607-619.

DIETERLEN-LIEVRE, F. (1984). Emergence ofintraembryonic blood stem cells studied in avianchimeras by means of monoclonal antibodies. Dev.Comp. Immunol. Suppl. 3, 75-80.

DIETERLEN-LIEVRE, F. & MARTIN, C. (1981). Diffuseintraembryonic hemopoiesis in normal and chimericavian development. Devi Biol. 88, 180-191.

DODGE, W. H. & Moscovici, C. (1973). Colonyformation by chicken hematopoietic cells and virus-induced myeloblasts. J. cell. Physiol. 81, 371-386.

DODGE, W. H. & SHARMA, S. (1985). Serum-freeconditions for the growth of avian granulocyte andmonocyte clones and primary leukemic cells induced byAMV, and the apparent conversion of granulocyticprogenitors into monocytic cells by a factor in chickenserum. J. cell. Physiol. 123, 264-268.

DODGE, W., SHARMA, S. & RUDEL, L. (1982). Effect ofhypercholesterolemia on the in vitro production ofgranulocytes and monocytes in the chick. Exp. molec.Pathol. 36, 44-56.

ELIASON, J. F. & ODARTCHENKO, N. (1985). Colonyformation by primitive hemopoietic progenitor cells inserum-free medium. Proc. natn. Acad. Sci. U.S.A. 82,775-779.

ESKOLA, J. (1977). Cell transplantation intoimmunodeficient chicken embryos. Reconstitutingcapacity of cells from the yolk sac at different stages of

development and from the liver, thymus, bursa ofFabricius, spleen and bone marrow of 15-day embryos.Immunology 32, 467-474.

FAUSER, A. A. & MESSNER, H. A. (1979). Identificationof megakaryocytes, macrophages and eosinophils incolonies of human bone marrow containingneutrophilic granulocytes and erythroblasts. Blood 53,1023-1027.

HOUSSAINT, E. (1981). Differentiation of the mousehepatic primordium. II. Extrinsic origin of thehaemopoietic cell line. Cell Diff. 10, 243-252.

ISCOVE, N. N. (1978). Erythropoietin-independentstimulation of early erythropoiesis in adult marrowcultures by conditioned media from lectin-stimulatedspleen cells. In Hematopoietic Cell Differentiation (ed.D. W. Golde, M. J. Cline, D. Metcalf & C. F. Fox),pp. 37-52. New York: Academic Press.

ISCOVE, N. N. (1983). Culture of lymphocytes andhemopoietic cells in serum-free medium. In Methods inMolecular and Cellular Biology (ed. D. Barnes, D.Sirbasku & G. Sato). New York: Alan R. Liss, Inc.

ISCOVE, N. N. (1985). Specificity of hemopoietic growthfactors. In Leukemia, Dahlem Konferenzen, (ed. I. L.Weissman), pp. 5-20. Berlin: Springer-Verlag.

ISCOVE, N. N., ROITSCH, C. A., WILLIAMS, N. &

GUILBERT, L. J. (1982). Molecules stimulating early redcell, granulocyte, macrophage and megakaryocyteprecursors in culture: Similarity in size, hydrophobicityand charge. J. cell. Physiol. suppl. 1, 65-78.

JOHNSON, G. R. (1980). Colony formation in agar byadult bone marrow multipotential hemopoietic cells. J.cell. Physiol. 103, 371-383.

JORDAN, H. E. (1917). Aortic cell clusters in vertebrateembryos. Proc. natn. Acad. Sci. U.S.A. 3, 149-156.

KONWALINKA, G., GLASER, P., ODAVIC, R., BOGUSH, E.,

SCHMALZL, F. & BRAUNSTEINER, H. (1980). A newapproach to the morphological and cytochemicalevaluation of human bone marrow CFUc in a g a r

cultures. Expl Hematol. 8, 434-440.LANOTTE, M. (1984). Terminal differentiation of

hemopoietic cell clones cultured in tridimensionalcollagen matrix: In situ cell morphology and enzymehistochemjstry analysis. Biol. Cell. 50, 107-120.

LASSILA, O., ESKOLA, J. & TOIVANEN, P. (1979).Prebursal stem cells in the intraembryonic mesenchymeof the chick embryo at 7 days of incubation. J.Immunol. 123, 2091-2094.

LASSILA, O., ESKOLA, J., TOIVANEN, P. & DIETERLEN-

LIEVRE, F. (1980). Lymphoid stem cells in theintraembryonic mesenchyme of the chicken. Sand. J.Immunol. 11, 445-448.

LE DOUARIN, N. M., HOUSSAINT, E., JOTEREAU, F. V. &

BELO, M. (1975). Origin of hemopoietic stem cells inembryonic bursa of Fabricius and bone marrow studiedthrough interspecific chimeras. Proc. natn. Acad. Sci.U.S.A. 72, 2701-2705.

LE DOUARIN, N. M. & JOTEREAU, F. (1973). Origin andrenewal of lymphocytes in avian embryo thymusesstudied in interspecific combinations. Nature, New Biol.246, 25-27.

Embryonic haemopoietic progenitors in chick 285

MCLEOD, D. L., SHREEVE, M. M. & AXELRAD, A. A.

(1974). Improved plasma culture system for productionof erythrocytic colonies in vitro: Quantitative assaymethod for CFU-E. Blood 44, 517-534.

METCALF, D. & MOORE, M. A. S. (1971). Haemopoieticcells. In Frontiers of Biology (ed. A. Neuberger & E. L.Tatum). Amsterdam: North-Holland PublishingCompany.

MILLER, A. M. (1913). Histogenesis and morphogenesisof the thoracic duct in the chick: Development of bloodcells and their passage to the blood stream via thethoracic duct. Am. J. Anat. 15, 131-198.

MOORE, M. A. S. & OWEN, J. J. T. (1967). Stem cellmigration in developing myeloid and lymphoid systems.Lancet 2, 658-659.

Moscovici, C. & Moscovici, M. G. (1973). Tissue cultureof avian hematopoietic cells. In Methods in CellBiology, vol. 7 (ed. V. Prescott), pp. 313-328. NewYork: Academic Press.

PERAH, G. & FELDMAN, M. (1977). In vitro activation ofthe in vivo colony-forming units of the mouse yolk sac./. cell. Physiol. 91, 193-200.

ROMANOFF, A. L. (1960). The hematopoietic, vascularand lymphatic systems. In The Avian Embryo. Structuraland Functional Development, pp. 569-678. New York:The Macmillan Company.

SAMARUT, J. (1978). Isolation of an erythropoieticstimulating factor from the serum of anemic chicks.Expl Cell Res. 115, 123-126.

SAMARUT, J. & BOUABDELLI, M. (1980). In vitrodevelopment of CFU-E and BFU-E in cultures ofembryonic and post-embyronic chicken hematopoieticcells. J. cell. Physiol. 105, 553-563.

SAMARUT, J. & NIGON, V. (1976). In vitro development ofchicken erythropoietin-sensitive cells. Expl Cell Res.100, 245-248.

SMITH, R. A. & GLOMSKI, C. A. (1982). "Hemogenicendothelium" of the embryonic aorta: does it exist?Dev. comp. Immunol. 6, 359-368.

STEWART, S., ZHU, B. & AXELRAD, A. (1984). A "serum-free" medium for the production of erythropoieticbursts by murine bone marrow cells. Expl Hematol. 12,309-318.

WILLCOX, M. B., GOLDE, D. W. & CLINE, M. J. (1976).Cytochemical reactions of human hematopoietic cells inliquid culture. /. Histochem. Cytochem. 24, 979-983.

WONG, P. M. C , CHUNG, S. W., CHUI, D. H. K. &EAVES, C. J. (1986). Properties of the earliestclonogenic hemopoietic precursors to appear in thedeveloping murine yolk sac. Proc. natn. Acad. Sci.U.S.A. 83, 3851-3854.

(Accepted 5 October 1987)