1. Art and Diseño

Fetal Liver Generates Low CD4 Hematopoietic Cells in Murine
Stromal Cultures
By Angelo Tocci, Francine Rezzoug, Kamal Wahbi, and Jean-Louis Touraine
We have demonstrated that 0.2% t o 11% of cells from the
fetal liver (FL) reacted specifically with high concentrations
of anti-CD4 monoclonal antibody (MO&). CD4+ cells from
FL were similar in surface phenotype and fluorescence characteristics t o the CD4' population found previously in adult
bone marrow (BM). FL and BM cells were seeded in cultures
that allow differentiation t o primitive precursors. FL cells
released many low CD4' and low Thy+ cells in thesupernatant, while BM cells seeded under the same conditions did
not. We studied the nonadherent cells harvested from 10day FLcultures (greater than 90% low CD4+). In methylcellulose, they were able t o produce more colonies that appear
t o be characteristicof earlier stages
in the hierarchy of hematopoietic precursors (especially erythroid bursts and colonies composedof both myeloid and erythroid elements) in
comparison with CD4- cells from 10-day BM cultures. CD4+
cells harvestedfrom FL cultures initiated secondary cultures
containing both a stromal layer andlarge hematopoietic colonies when replated under conditions similar t o those of
primary cultures. Furthermore, a limited number of CD4+
cells from 10-day FL cultures were able t o repopulatelethally
irradiated mice. Although we cannot formally exclude the
possibility that the low CD4 cells produced in FL cultures
were derived exclusively from the proliferation of the few
CD4 cells found in fresh FL, the dynamic analysisof the development of these cells in culture favors the generation of
this important population from a CD4- subset of hematopoietic stem cells (HSCs). We speculatethat FL contains a prevalent population of very primitive cells not expressing the
CD4 antigen, tentatively called "pre-low CD4 precursors."
These primitive cells can differentiate into low CD4+ cells
that share many characteristics with pluripotent HSCs of
the adult type. These data indicate the possibility of using
hematopoietic progenitors obtained by the expansion/differentiation of fetal stem cells in culture for transplantation
purposes.
0 7 9 9 5 by The American Society of Hematology.
T
the ability to produce CFU-C in methylcellulose, and the
capacity to start miniaturized hematopoietic stromal cultures.
The results were compared with those obtained from BM
cultures established under the same conditions. Moreover,
low CD4 cells harvested from FL cultures were assayed
for the presence of HSCs by injection into marrow-ablated
semiallogeneic hosts.
HE CD4 ANTIGEN is expressed by murine thymocytes,
mature T cells that recognize class I1 major histocompatibility complex proteins,' and thymic precursors that can
give rise to T and B lymphocytes in vivo.' In addition, a
CD4-positive cell subset in bone marrow (BM) is enriched
with myeloid progenitors (colony-forming unit-culture;
CFU-C), multipotential hematopoietic cells (colony-forming
unit-spleen; CFU-S): and pluripotent hematopoietic stem
cells (HSCs), which are responsible for the long-term reconstitution of irradiated
This heterogeneous population
of precursor cells has been termed low CD4 precursors, as
it expresses very low levels of surface antigen, and optimal
immunostaining has been achieved using a high concentration of anti-CD4 monoclonal antibody (MoAb): However,
the exact level of these cells in the hierarchy of hematopietic precursors is not clearly defined, and some studies have
demonstrated that a CD4-depleted cell subset in BM is also
enriched with HSCs.6,' A recent report seems to reconciliate
these conflicting results, demonstrating that c-kit-positive
cell subsets in BM, whether or not they coexpress the CD4
antigen, can repopulate lethally irradiated mice efficiently.*
These results suggest that the antigen is expressed variably
on the heterogeneous population of functionally defined
HSCs.
Fetal liver (FL) is a rich source of primitive HSCs devoid
of mature T cells and can induce hematopoietic recovery
when injected into lethally irradiated mice.9."In humans,
we showed that FL can successfully reconstitute immunodeficient children and unhealthy fetuses transplanted prenatally
in utero.'','2
In the studies reported here, we examined whether the
CD4 antigen is expressed on murine FL hematopoietic cells
under conditions similar to those reported earlier, ie, using
high concentrations of anti-CD4 MoAbs.4 Moreover, FL or
BM cells were seeded into a culture system that allows differentiation to primitive precursors for several weeks." The
nonadherent cells produced in these cultures were harvested
weekly and analyzed for the presence of CD4-positive cells,
Blood, Vol 85, No 6 (March 15), 1995: pp 1463-1471
MATERIALS AND METHODS
Mice. Adult C57BU6 (H-2b),DBA2 (H-2d), and BDFl (H-2")
mice were maintained under standard laboratory conditions in our
facility at the Transplantation and Clinical Immunology Unit (HBpital Ed. Hemot, Lyon, France).
Cell preparation, antibodies, and staining procedures. Cells
were obtained from thymus, spleen, and femoral BM of adult BDFl
mice and from either FL or BM cultures established as described
below. BDFl FL cells were obtained as described previously.'o
Briefly, estrus-synchronized C57BL/6 mice were mated with DBA2
males for 48 hours. At day 14 after the first contact, pregnant mice
were killed, BDFl fetuses were removed, and FLs were dissected
away from surrounding tissue and pooled. A monocellular suspension was prepared using syringes fitted with 25-gauge needles. Cell
counts were performed in a hemocytometer. The viability of cells
was always higher than 95% as assessed by Trypan blue dye exclu-
Fromthe Transplantation and Clinical Immunology Unit, INSERM U 80; and the Department of Hematology, H6pital Ed. Herriot, Lyon, France.
Submitted July 7, 1994; accepted October 7, 1994.
Address reprint requests to Jean-Louis Touraine, MD, and Angelo
Tocci, MD, Transplantation and Clinical Immunology Unit, INSERM
U 80, H6pital Ed. Herriot, 5. Place d'Arsonva1, 69437 Lyon, Cedex
03, France.
The publication costs of this article were defrayed in part by page
charge payment. This article must therefore be hereby marked
"advertisement" in accomlance with 18 U.S.C. section 1734 solely to
indicate this fact.
0 1995 by The American Society of Hematology.
ooos-~~~~~~~~~os-o~~o~~.oo/o
1463
1464
sion test. The following MoAbs were used: fluorescein isothiocyanate (F1TC)-conjugated YTS 191.1 (directed against a subset of T
lymphocytes; anti-CD4), Ml/70. 15 (directed against myelomonocytic cells; anti-Macl), 345-2C11 (directed against T cells; antiCD3t), 34-2-12 (directed against major histocompatibility complex
class I antigen; anti-H-2Dd), RA3-6B2 (directed against B cells;
anti-B220). phycoerythrin (PE)-conjugated 5a-8 (directed against T
cells and a subset of primitive HSCs; anti-Thy 1.2) purchased from
Caltag Laboratories (San Francisco, CA), RM 4-4 (directed against
a subset of T cells; anti-CM) purchased from Pharmingen (San
Diego, CA), and biotinilated 53.7.3 (directed against CD4-positive
T-helper lymphocytesL3and a subset of B cells; anti-Lyt 1.2) purchased from Becton Dickinson & CO (Mountain View, CA).To
define background fluorescence, FITC- or PE-conjugated control
isotypes (or streptavidin) were used. Before staining, red blood cells
were lysed (when required), and the cell suspension was washed
once in phosphate-buffered saline (Sigma Chemical, St Louis, MO)
supplemented with 2% fetal calfserum
(Organics, Strasbourg,
France). Staining was performed by reacting freshly explanted (or
cultured) FL cells, BM cells, thymocytes, or splenocytes with antiCD4, anti-Lyt 1.2, anti-Thy 1.2, anti-CDSe, or anti-Mac1 MoAbs
(0.025 to25 pg/lOh cells). Single- or dual-color fluorescence was
analyzed on a FACScan flow cytometer; data from 5,000 to 10,000
cells were collected and analyzed using a Lysis program (BectonDickinson & CO).Appropriate gates were set up to exclude cellular
debris and aggregates. Analysis was performed under similar conditions in comparative studies.
FL and BM culture techniques. Cultures were set up asdescribed
previously,'" according to amethod derived from Dexter's technique
of BM c ~ l t u r e .The
' ~ modifications introduced to Dexter's technique
were as follows. The culture medium consisted of Iscove's modified
Eagle's medium (IMEM; Boehringer Mannheim Biochemicals,
Mannheim, Germany) containing 20% horse serum (Boehringer
Mannheim Biochemicals), 2 mmol/L glutamine (Bio Merieux, Lyon,
France), 0.4% penicillin-streptomycin (Bio Merieux), 30 &mL
transfemne (Boehringer Mannheim Biochemicals), IO-' molL hydrocortisone (Sigma Chemical), and IO-' m o a P,-mercaptoethanol
(Sigma Chemical). A single inoculum of FL or BM was cultured in
complete medium, and cultures were not recharged with fresh FL
or BM. The cultured cells were refed twice a week by removal of
half of the supernatant volume (including cells in suspension), and
replacement was performed using fresh medium containing 15%
horse serum and 5% fetal calf serum. These cultures will be referred
to as primary cultures. Many previous experiments have demonstrated that this one-step culture technique allows either FL or BM
cells to set down their own stlomal layer and sustain hematopoiesis
for several months."' Fat cells, cobblestone areas,I4 and colonies of
more superficial cells laying onthe stromal layer are seen under
inverted microscope visualization; however, fat cells are less numerous in FL than in BM cultures.
Beginning 10 days after seeding, the cultures were refed twice a
week, andthe cells removed were washed once before any other
test was performed. The cells were adjusted to the appropriate cell
concentration for staining with MoAbs, for replating experiments in
methylcellulose andin miniaturized cultures (established as described below), and for morphologic and molecular biology analysis.
The viability of cells harvested from cultures was always higher
than 95% as assessed by Trypan blue dye exclusion test.
An aliquot of nonadherent cells harvested from primary cultures
was replated weekly in methylcellulose as described previously'" to
assess the presence of colonies (aggregates of greater than 40 cells).
Conditioned medium from newborn mouse hearts wasusedas a
source of colony-stimulating factor. Colonies were counted after 14
days of culture. Smaller, translucent-type aggregates were scored
as myeloid colonies [colony forming unit-granulocyte-macrophage
TOCCI ET AL
(CFU-GM)]; larger, red-type colonies were scored as erythroid
bursts (BFU-E): and very large colonies composed of both translucent and red-type elements were scored as mixed-cell colonies (CFUmix).
An aliquot of nonadherent cells was harvested from 10-day FL
or BM cultures, washed once, and transferred at a concentration of
IO' cells per milliliter to a 96-well, flat bottom, tissue culture plate
(Microtest 111; Falcon, Becton Dickinson & CO, Lincoln Park, NJ)
containing 0.2 mL primary culture medium." These cultures will be
referred to as miniaturized cultures. Beginning at 10 days, miniaturized cultures were refed as described for primary cultures. They
were checked weekly under an inverted microscope for development
of a stromal layer, its morphology, and the presence of hematopoietic
foci. Hematopoietic foci were defined by the presence of colonies
of pleomorphic cells or typical cobblestone areas within the adherent
layer." In some experiments, all cells were harvested from miniaturized cultures by gentle pipetting and were seeded into a 96-wel1, flat
bottom, tissue culture plate containing 0.1 mL methylcellulose culture medium described above.
Morphology. Cells harvested from primary or miniaturized FL
or BM cultures were cytocentrifuged onto a glass slide and stained
using May-Griinwald-Giemsa dye.
Repopulationexperiments.
Lethally irradiated (9 Gy) C57BL/6
(H-2h) mice were injected with either 1 X 10' cells from C57BL/6
(H-2h) BM or simultaneously with 2 X 10' cells from BDFl
(H-2h'd) 10-day cultured FL and 1 X 10' cells from C57BL/6 (H2h)BM to provide radioprotection. Many previous experiments have
demonstrated that cultured cells transplanted alone do not protect
mice from lethal irradiation.'" BDFl cells from cultured FL contained greater than 90% low CD4-positive cells. Reconstitution was
assessed by staining peripheral blood lymphocytes with an antibody
specific for the H-2Dd determinant present on BDFl cultured cells.
H-2Dd-positive cells were gated, and reconstitution of the various
hematopoietic lineages was determined by analyzing Thy 1.2-,
B220-, and Macl-positive cells present in the gate.
Detection of CD4 mRNA in fresh, ie, uncultured, and cultured FL
and BM cellsbyreversetranscriptase-polymerasechainreaction
(RT-PCR)analysis. Total cellular RNA was isolated from samples
of thymic, BM, or FL cells, and cells were harvested from FL and
BM primary cultures by the guanidine thiocyanate
For
(40 U RNAsin; 100
cDNA preparation, 5 p g ofRNAwasused
mmol/L Tris, pH 8.3: 140 mmol/L KC1; 10 mmol/L MgC12,pH 8.3;
28 mmol/L p2 mercaptoethanol; 1 mmol/L dNTPs, 10 U of Moloney
murine leukemia virus reverse transcriptase, for 60 minutes at 42°C).
This cDNA was used astemplate for PCR amplification, using primers CD4-85374 (5'-GGAGTCCATCTTGACCTT-3')
and CD485373 (5"GAAGTGAACCTGGTGGTG-3') deduced from the
cloned CD4-cDNA, under conditions described by Perkin-Elmer
Cetus (Norwalk, CT). As control, previous digestion was performed
using ribonuclease before the amplification. PCR products were visualized by ethidium bromide staining on 2% agarose gels."~"'
Statisticalanalysis.
The statistical analysis of parameters was
performed using the Student's t-test and was accomplished using a
Statworks program (Statworks, Meylan, France) on a Macintosh
computer (Apple, Cupertino, CA). Data were considered to be significantly different when P < .OS.
RESULTS
D e t e c t i o n of CD4 on fresh FL or BM cells. An MoAb
concentration of 0.25 pgl106 cells was established previously
to be optimal for detection of CD4 on thymocytes: the PEconjugated anti-CD4 RM 4-4 MoAb stained 84% (n = 2),
and
the
FITC-conjugated anti-CD4 YTS 191.1 MoAb
stained 87% (n = 2) of BDFl thymocytes. A 10-fold higher
EMATOPOIETIC
1465
CD4'
CULTURE
LIVER
IN FETAL
FETAL
LIVER
BONE MARROW
FL2
Fig 1. Detection of CD4 in fresh, ie, uncultured, FL, BM, and thymus cells. Background fluorescence determined by PE-conjugated
isotype control(black areas), percentage of low-fluorescentFL or BM
cells (bold line1 and thymus cells (peak at right) determined by PEconjugated anti-CD4 (RM 4-41 MoAb. Ata MoAb concentration of2.5
pg/106 cells, percentages were as follows: FL, 3% (A); BM, 10% (61;
and Thymus, 98%(A,B). At a MoAb concentration of 25.0 pg/108
cells, percentages were as follows: FL, 11% (Cl; BM, 22% (Dl;
and
thymus, 99% (C,D).
concentration (ie, 2.5 pg/lOh cells) of the MoAbs anti-YTS
191.1and anti-RM 4-4 stained 3.4% 5 2.3% (n = 5) and
5.0% 5 4.6% (n = 4) of BM, respectively, and 0.2% 2 0.1 %
(n = 4) and 2.5% 5 0.8% (n = 4) of FL cells, respectively,
depending on the MoAb used. Following a protocol reported
previously: we used
100-times more MoAb (ie, 25 &IO"
cells) than that optimal for staining the thymocytes. Under
these conditions, up to 33% of BM cells were stained above
background, but the percentage was highly variable, from
I % to 33%, with a mean of 8.6% and 17.1% with the two
MoAbs (Fig I and Table l ) . The anti-Lyt 1.2 53.7.3 MoAb
was used as control (as Lyt 1 is expressed on CD4-positive
T cells'3) at the same high concentration, and it stained no
more
than
5.3% BM
of
cells (Table 1). At these very
high
concentrations of MoAb, most thymocytes were CD4-positive (compare also Wineman et al'). The presence of a heterogeneous population expressing intermediate levels of the
CD4 antigen in the thymus could explain the difference of
antigen-positive cells detected with MoAb concentrations of
0.25 or 25 pg/lOh cells. However, nonspecific staining of
thymocytes could not be excluded. Therefore, we reasoned
that the possibility of nonspecific staining could be ruled out
by using splenocytes that are expected to contain fewer cells
expressing intermediate levels of the CD4 antigen compared
with thymocytes. The percentage of stained splenocytes was
about 15.4% (n = 2) when a MoAb concentration of 2.5
pg/lO" cells wasusedand15.7%
(n = 2) when a MoAb
concentration of 25 &IOh cells was used. The presence of
(few) early progenitors in the spleen may account for the
slight difference observed." Dual-color staining with antiLyt1.2and anti-CD4 MoAbs showed that most low CD4
cells from the BM did not stain positively withLyt1.2
MoAb, while most CD4-positive splenocytes were costained
with anti-Lyt 1.2 MoAb, although a small percentage was
not (data not shown). The analysis of forward-light scatter
characteristics showed that low CD4 cells from BM include
two distinct populations: a high forward light-scatter population and a low forward light-scatter population. These cells
appear several-fold less fluorescent when compared with thymocytes stained in parallel (Fig 2 and Table l ) . We assumed
that this low CD4-Lyt 1.2-negative population corresponded to that described previously as being HSC-conmining4 To determine whether fresh FL cells expressed the
same staining pattern as BM, we incubated fresh FL cells
with the anti-CD4 and the anti-Lyt 1.2MoAbsunder the
same conditions. Analyzing a population similar to that of
BM by setting appropriate gates and using similar cytofluorimetric parameters for the acquisition of the data, the antiCD4 MoAbs of various origins stained about 0.2% to I 1%
of FL cells even whenusedatthe
highest concentration
(Figs 1 and 2). whereas the anti-Lyt 1.2 MoAb, at the same
high concentration, stained only 0.7% to 2.6% (Table l ) .
Detection qf CD4 anrigen on cultured FL nnd BM cells.
Usingthe same protocol as for thymocytes, splenocytes,
Table 1. Detection of CD4 on BDFl FL, BM, or Thymus Cells
Reagent
Concentration
(pg/106 cells)
Positive Cells 1% ?SD)
Reagent
2.5
Control isotype (PE)
25.0
Control isotype
0.1 (PE1
2.5
Anti-CD4(PE)
4-4
RM
25.0
Anti-CD4 RM
(PE)
4-4
8.1
25.0
Anti-CD5 53.7.3 (biotin)
4.7
ND
FL
0.10.1
2 0.1
(n = 4)
0.1 2 0.1
(n = 5)
2.5 25.0
0.8
(n = 4)
2 4.4
(n = 5)
0.7
2.6
BM
0.1
2 0.1
(n = 4)
0.12 0.1
(n = 5)
295.7
4.6
(n = 4)
17.1 297.2
11
(n = 5)
5.3
ND
Geometric
Mean
THY
2 0.1
(n = 5)
2 0.1
(n = 51
244.8
2.8
(n = 5)
2 56.3
1.8
(n = 5)
98.1
96.8
Fluorescence I-SD)
FL
BM
THY
-
-
-
-
-
-
241.0
1.1
(n = 3)
249.7
1.9
(n = 3)
2 132.1
11.9
(n = 3)
2 11.8
139.4
( n = 5)
2 5.8
(n = 3)
2 8.3
(n = 51
ND
Data were obtained using increasing concentrations of the indicated PE-conjugated MoAbs. Background staining was determined using the
appropriate PE-conjugated control isotypes.
Abbreviations: THY, thymus; ND, not determined.
TOCCI ET AL
1466
i
)
Anti-CD4 YTS 191.1 (FITC)
Anti-CD4 RM 4.4 (PE)
Fig 2. Detection of CD4 on BDFl BM (0,n = 5). FL (0;n = 5). or
thymus (m, n = 5)cells. Barsrepresent the percentage (range) of low
CD4 cells detected by using a concentration of the indicated MoAb
of 25 pg/106 cells. Median values (horizontal segment) and standard
deviations (vertical bars) are also indicated.
fresh BM, andfresh FL cells, increasing concentrations (0.25
to 25 pg/lO" cells) of the anti-CD4 MoAb were usedto
stain nonadherent cells harvested from FL and BM cultures.
Although the general pattern of staining was consistently
observed in our culture system, the percentage of CD4-positive cells varied between experiments. After IO days of culture, as many as greater than 90%of nonadherent cells from
FL cultures were found to be positive when an anti-CD4
MoAb concentration of 25 pg/lO" cells was used (in two of
three experiments). In contrast, the nonadherent cells recovered from BM cultures and incubated with MoAbs under the
same conditions contained virtually no CD4-positive cells (in
two of three experiments; Fig 3). Nonadherent cells from
IO-day FL cultures expressed very low percentages of CD4
and Lyt I .2 antigen when the MoAbs were used at a concentration of 0.25 pg/lOh cells, whereas as many as 26% of cells
stained positively for the Thy 1.2 antigen at the highest
concentration used and expressed low levels of the antigen
(low Thy; Fig 4). At subsequent weeks of culture, the percentage of CD4-positive cells decreased progressively (Fig
3). Staining of nonadherent cells from BM cultures showed
negligible percentages of CD4- (Fig 3), Lyt 1.2-, and Thypositive cells. Both FL and BM cells contained low to no
FETAL LIVER
detectable percentages of MacI- and CD3-positive cells
(even when high concentrations of MoAb were used; data
not shown). When nonadherent cells from FL cultures were
analyzed after 24 hours and 48 hours of culture, no staining
was detected (in this particular experiment, 1 I .8% fresh, ie,
before seeding in culture, FL cells were low CD4). At the
endpoint of cultures, low CD4-positive cells were still found
in both BM and FL stromal layers. The results obtained at
the cell surface using MoAbs were confirmed at the RNA
level, using the RT-PCR technique (Fig S).
Monitoring of BM and FL cultures by production qf nucleated cells and clonogenic progenitors. Nucleated cell numbers, CFU-GM, BFU-E, and CFU-mix were assayed weekly
in nonadherent cells harvested from FL or BM cultures (Figs
6 and 7). The cells from FL cultures produced a larger number of colonies and, after I O days of culture, most precursors
contained in the nonadherent cells produced erythroid or
mixed colonies. Later in culture, FL produced prevalently
CFU-GM progenitors. Conversely, the cells from BM cultures produced fewer colonies and, at any time, CFU-GM
were prevalent while mixed colonies were veryrare. The
cells harvested from FL cultures were greater than 90% CD4positive, and those from BM cultures were virtually antigennegative after I O days of culture. (The data of Fig 6 correspond to the first and second experiments shown in Fig 3.)
Miniaturized cultures. Miniaturized cultures were obtained by seeding nonadherent cells harvested from IO-day
FL or BM cultures. Cultures from FL cells could be maintained for several weeks (longer than 8 weeks in some experiments), whereas cultures from BM cells didnot progress
after weeks 2 to 3 (Table 2). Nonadherent cells harvested
from FL cultures (unlike those harvested from BM cultures)
were able to produce a confluent stromal layer in about 3
weeks. This stromal layer was typical in aspect (ie, such as
that observed in primary cultures), although fat cells were
consistently more numerous compared withprimary cultures. Some foci of hematopoietic activity [very large polymorphic cell aggregates (Fig 8A) and rare cobblestone areas]
and many colonies of large translucent cells on the stromal
layer were observed. BM cells produced only some foci of
adherent cells withvery small aggregates of pleomorphic
cells (Fig 8B). In some experiments, nonadherent and adherent cells were harvested from miniaturized FL or BM cultures after 5 weeks of culture and seeded into mcthylcellu-
BONE MARROW
~
Days of culture
1st experiment
2nd experiment
3rd experiment
Fig 3. Comparisonof FL and
BM culture production of low
CD4 cells. The curves represent
percentages of antigen-positive
cells detected in the nonadherent population present in cultures and are expressedas a
function of time (days). Data are
from three different experiments
for either FL or BM.
CD4' HEMATOPOIETICCELLS
1467
IN FETALLIVERCULTURE
Fig 4. Detection ofCD4, Lyt
1.2, and Thy 1.2 in cultured FL.
Nonadherent cells were harvested after 10daysof culture
and reacted with anti-RM4-4,
anti-53.7.3,
and
anti-Sa-8
MoAbs used at the givenconcentrations. Background
fluorescence was determined by PEconjugated isotype control. Data
are from one representative experiment.
L
P
:
Y
p:
L", 1.2 W P t l l O ~au.1
0
- 1
FLl
lose. After 14 days of culture, notypical colonies were
observed from seeded FL cells, while some clusters (aggregates of 5 to I O cells) were observed from seeded BM cells:
in a typical experiment, nine clusters of well-separated, large,
macrophage-like cells were counted (Table 2).
Morphology of cells harvested from FL ond BM cultures.
May-Griinwald-Giemsa staining of the cytocentrifuged cell
suspensions showed that the most prevalent cells in primary
cultures of FL were large and macrophage-like cells; some
of these cells contained multiple nuclei (up to four) and
numerous nucleoli (Fig 9). These cells were nonspecific esterase-positive. In a nonadherent subset from BM primary
cultures, smaller cells were found: binucleated macrophagelike cells were rarely seen, and no plurinucleated cells were
identified. Very few lymphocyte-like cells were seen. Cells
found in BM cultures were more polymorphic than those in
FL cultures. Mitotic figures were repeatedly observed in FL
but rarely seen in BM nonadherent cell populations.
In nonadherent and adherent subsets of BM miniaturized
cultures, only macrophage-like cells were observed, whereas
many undifferentiated cells with a high nuc1ear:cytoplasmic
ratio were harvested from miniaturized FL cultures.
Low CD4 cells generated in FL cultures contain pluripotent HSCs. At 13 weeks after transplantation, all miceinjected with either host-type BM or simultaneously with cells
from FL cultures (greater than 90% low CD4-positive) and
host-type BM survived the lethal effects of irradiation. Mice
were analyzed for the presence of H-2d cells after 5 and 13
weeks from injection. Low numbers of cultured FL cells
were able to repopulate all lineages in lethally irradiated
I
L
THYFL
DISCUSSION
In the present study, we have demonstrated that murine
FL contains a population of low CD4 cells that are similar
in surface phenotype and fluorescence characteristics to the
primitive population found previously in BM."' However,
these cells did not constitute a major population as might be
expected of FL, which is a rich source of early HSCs devoid
of mature cells, and in some experiments the percentage of
antigen-positive cells was as low as less than 1.0% of the
total population. Preliminary experiments onhuman FL
show that, as early as the 16Ihweek of age, FL contains only
about less than 2% of low CD4 cells, a finding that is lower
than the reported 9% incidence in adult human BM.4
A great interexperiment variation was observed in our
study on BM, and this confirms previous studies in which
percentages varying from 3% to 60% of cells were found to
be positive in
In our study on FL, this variation
is
probably caused by the fact that cells are obtained by pooling
FL from fetuses of slightly different conceptional ages, with
a variable distribution of younger and older fetuses in the
pooled suspension.
m
A
L LIVER
BONE MARROW
I
BM
NC
mice; theywere also able to compete with host-type BM
(Table 3).
C
NC
C
I
Days of culture
Fig 5. RT-PCR analysis on2% agarose gel of CDdmRNA obtained
from fresh (ie, uncultured; NCI and cultured IC1 cells from BM, FL, or
thymus (THY). L, controls.
Fig 6. Comparison of FL and BM culture production of nucleated
cells. The curves represent mean numbers of nucleated cells per culture t standard deviation (vertical bars; when not indicated, SD is
less than 1) and are expressed as a function of time (days). Data are
from three different experiments.
1468
TOCCIET AL
BONE MARROW
FETAL LIVER
310
250
E!
S
*
z
‘;j
190
Y
2
v
3
8
130
70
10
L
Days of culture
Fig 7. Comparison of FL and BM culture production
of clonogenic
progenitors. The boxed areas represent mean numbersof clonogenic
progenitors per culture and
are expressed as a function of time
(days). The data are the meanof two different experiments.
translucent colonies (CFU-GM); 0, large, hemoglobinized colonies
(BFU-E) and mixed colonies containing both translucent and hemoglobinized elements (CFU-mix).
Despite the relatively lower number of low CD4 cells in
FL compared with BM, the FL cultures release high numbers
of antigen-positive cells in the supernatant (which becomes
several-fold enriched withlow CD4 cells compared with
fresh, ie, uncultured, cells), whereas BM cultures give rise
to few low CD4 cells. Interestingly, the population produced
in FL cultures also contains a substantial proportion of low
Thy-positive cells after 10 days of culture. The presence of
these markers on cells produced by FL cultures suggests that
most of these cells are earlier in the hierarchy than those
produced in BM cultures.
In our experiments, we did not purify further our greater
than 90% low CD4 cell population, as we reasoned that cell
sorting would have produced a population containing a still
high level of contaminants because of a lack of clear separation of the positive from the negative populations.
It is necessary to exclude the possibility that these cells
could derive from the proliferation of mature T cells (if any)
present in the fresh FL or BM. First, the morphologic study
of this population showed a very limited number of lymphocyte-like cells. Furthermore, it is well known that
Dextertype cultures do not lead to maintenance of mature lymphocytes,’‘ anda short period of culture under conditions similar
to those used in this study has been proposed as a technique
to purge adult BM to avoid graft-versus-host disease in allogeneic transplantation.2’ BM,which logically contains a
higher number of mature lymphocytes in comparison with
FL, produces few to no detectable CD4-positive cells in our
culture system. Moreover, few cells in FL cultures react with
a standard (0.25 pg/106 cells) concentration of anti-CD4 or
with standard to high concentrations of anti-Lyt 1.2 and antiCD3 MoAbs.
It is notable that many macrophage-like, nonspecific esterase-positive, multinucleated cells were identified in the nonadherent population harvested from primary FL cultures,
whereas in BM cultures, binucleated, macrophage-like cells
were rarely seen and no plurinucleated cells were identified.
The multinucleated cells identified in FL cultures are similar
to those described by others in human cord blood:’ ie, macrophage-like, large, multinucleated, and nonspecific esterasepositive cells. These cells have excellent replating capacity,
and it is speculated thattheymaybe
associated withthe
repopulating capacity of cord blood cells. More importantly,
these cells virtually lack the CD14 antigen, which is found
on human monocytes and macrophages. Our results seem to
mimic these previous findings. The production of these cord
blood cells has been referred to as being hyperstimulationmediated by added growth factors, particularly the steel factor, interleukin-3, and the granulocyte-macrophage colonystimulating factor. The exact mechanism of development
of these cells has not been fully elucidated, although their
presence in our FL (but not in BM) cultures might be caused
by a higher production of growth factors by the FL stromal
layer compared with that of the BM. This explanation seems
to be confirmed by recent studies that show higher levels of
Table 2. Generation of Stromal Layer and Hematopoietic Foci in Secondary Miniaturized Cultures Obtained by
Seeding
Nonadherent Cells Harvested From 10-Day FL or BM Primary Cultures
Stromal Layer and Foci of Hematopoietic Cells (wk)
No. of
Cell Type
Experiments
FL
BM
4
3
1
+
+
2
++
+++
++
3
+++
++
4
5
+
+++
+
Colonies in Methylcellulose at wk 5
None
None; some clusters of flattened
macrophage-like cells
Nonadherent cells (2 x lo5)harvested from IO-day FL or BM primarycultures were transferred into a 96-well, flat bottom, tissue culture plate
containing 0.2 mL culture medium. Starting
at 10 days,cultures were refed as described for primarycultures. Miniaturized cultures were checked
weekly under invertedmicroscope for development of the stromallayer, its morphology, and the presence of hematopoietic foci. Hematopoietic
foci were defined by the presence of aggregates of hematopoietic cells on the adherent layer and/or cobblestone areas. In two experiments,
the cells were harvested from a single well of miniaturized cultures at week 5 of culture by gentle pipetting and were seeded into a 96-well,
flat bottom, tissue culture plate containing the methylcellulose culture mediumdescribed in Materials and Methods.
Abbreviations: +, stromal layer and hematopoietic foci present in less than 30% of thewell; ++, stromal layer and hematopoietic foci present
in 30% to 60% of the well; +++, stromal layer and hematopoietic foci present in greater than 60% of the well.
EMATOPOIETIC CD4'
IN FETAL
1469 LIVER CULTURE
Fig 8. Hematopoietic foci in FL (A) and BM IBI
miniaturized cultures after 16 days of culture loriginal magnification, x40).
Fig 9. Nonadherent cells found in the supernatant of the FL primary cultures after 10 days of culture. A multinucleatedcell isshown (originalmagnification. x 1.000).
1470
TOCCI ET AL
Table 3. Repopulation of Lethally Irradiated Mice With Low CD4 Cells From FL Cultures
Donor Cells in Blood
(% H-Zd-positive)
TotalPeripheral Blood
Markers*
Lineage
Lymphocytes
(13wkst)
(no.of mice = 5)
(no. of mice = 5)
Cells
+
2 x lo5BDFl cultured FLS lo5 C57BU6 fresh BM
1 x lo5C57BU6 fresh BM (negative control)
Untreated BDFl cells (positive control)
5 wkst
? 3
20 71
93 2 2
13 wkst
t 19
5 5 3
Thy 7.2
32 t 7
43 t 3
B220
35
5
10
36 2 2
Mac-l
3 2 1
-
13 2 7
Lethally irradiated (9 Gy) C57BU6 mice were injected with the indicated numbers of either host-type BM cells (negative control) or simultaneously with BDFl cultured FL and host-type BM cells.
* Denotes the absolute percentage of donor-type (BDF1) FL cells found in peripheral blood.
t Indicates the interval between the intravenous injection and the blood testing.
BDFl cultured cells contained greater than 90% low CD4-positive cells.
*
RNA transcripts of steel factor and interleukin-3 in stromal
clones obtained from murine FL compared with those obtained from BM.24
Our data show that clonogenic precursors earlier in the
hierarchy (CFU-mix and BFU-E) are more numerous in FL
than in BM cultures at day 10 of culture; after this time,
even the FL cultures shift towards a prevalent production
of CFU-GM. In a system with no addition of exogenous
cytokines, these observations confirm the ontogeny-related
differences between fetal and adult HSCs reported prev i o ~ s l yWe
. ~ ~used replating experiments (miniaturized stromal cultures) to assess the presence of very primitive cells
in the nonadherent population harvested from either FL or
BM primary cultures, as described previo~sly.'~
Miniaturized cultures were successfully established from the nonadherent cells harvested from either FL and BM cultures, but
the latter did not progress after weeks 2 to 3. The morphology
of FL miniaturized cultures was typical, with the exception
of a higher number of fat cells, which are rarely observed
inprimary FL cultures but are more frequently found in
primary BM cultures (F. Rezzoug, unpublished observations
from 1988 to 1993; compare also Slaper-Cortenbach et al").
The presence of more fat cells could be caused by the culture
system, as described previo~sly,'~
but the exact mechanism
that would explain this difference is not fully understood. In
some experiments, we replated the nonadherent and adherent
cells harvested from the miniaturized cultures after 5 weeks
in methylcellulose to assess the presence of clonogenic progenitors derived from more immature cells possibly present
in the inoculum. However, the absence of CFU-C raises
doubts on the efficacy of this one-step culture system in
assessing the presence of primitive cells in the inoculum.
However, in vivo, our results show that a limited number of
cells harvested from FL cultures are able to repopulate lethally irradiated animals, that low CD4 cells harvested from
FL cultures contained pluripotent HSCs, as shown by the
repopulation assay in lethally irradiated mice, and that the
number of HSCs in the harvested population should be relatively high, because they were able to compete with either
coinjected host-type BM and (few) residual HSCs inthe host.
Alternatively, cultured cells might possess some advantages
over coinjected fresh BM cells in repopulating mice.
These results indicate that FL and adult BM hematopoietic
cells have different developmental potential in vitro. It is
conceivable that functional differences between primitive
hematopoietic cells of tissues at various stage of development2' could explain these findings: more primitive cells,
which are more highly represented in FL compared with BM,
could have undergone partial differentiation in our culture
system, with production of cells phenotypically similar to
BM primitive precursors. Althoughwe cannot absolutely
exclude the possibility that the low CD4 cells produced in
FL cultures were derived exclusively from the proliferation
of the few CD4 cells found in fresh FL, the dynamic analysis
of the development of these cells in culture is in favor of
the generation of this important population from a CD4negative subset of HSCs. When the nonadherent cells from
FL cultures were tested within the first 2 days for the presence of low CD4 cells, no detectable staining was observed,
suggesting the exhaustion of the preexisting low CD4 cells
(via differentiation or death) and the active production of
new low CD4 cells. Alternatively, internalization of the antigen occurring in the first phase of culture may also explain
these findings. We speculate that FL contains a primitive
population of cells, not expressing the CD4 antigen (prelow CD4 precursors), which are able, in the appropriate
microenvironment, to produce low CD4 cells that share
many characteristics with adult-type HSCs. Conversely, BM,
at a lower stage in the hierarchy, does not sustain the production in vitro of a large number of low CD4 cells under similar
conditions. These differences between adult and fetal tissues
should be taken into account when planning to expand HSCs
in vitro.
ACKNOWLEDGMENT
We thank Prof Glaichenhaus (Institut of Molecular and Cellular
Pharmacology, Nice, France) for the gift of the cDNA insert corresponding to mouse CD4. We acknowledge the invaluable assistance
given by G. Panaye in expert flow cytometry. We thank G. Vivier
for technical assistance and Dr Blanc-Brunat for encouragement.
REFERENCES
1. Dyalinas DP, Wilde DB, Marrack P, Pierres S, Wail K A , HW-
ran W. Otten G, Loken MR, Pierres M, Kappler J, Fitch W :Charac-
CD4' HEMATOPOIETIC CELLS IN FETALLIVER
CULTURE
terization of the murine antigenic determinant, designated L3T4a.
recognized by monoclonal antibody GK1.5: Expression of L3T4a
by functional T cell clones appears to correlate primarily with class
I1 MHC antigen-reactivity. Immunol Rev 7429, 1983
2.Wu
L, Scollay S, Egerton M, Pearse M, Spangrude GJ,
Shortman K CD4 expressed on earliest T-lineage precursor cells in
the adult murine thymus. Nature 349:71, 1991
3. Fredrickson GG, Basch RS: L3T4 antigen expression by hematopoietic precursor cells. J Exp Med 169:1473, 1989
4. Wineman JP, Gilmore GL, Gritzmacher C, Torbett BE, MiillerSieburg CE: CD4 is expressed on murine pluripotent hematopoietic
stem cells. Blood 80:1717, 1992
5. Onishi M, Nagayoshi K, Kitamura K, Hirai H, Takaku F, Nakauchi H: CD4d""' hematopoietic progenitor cells in murine bone
marrow. Blood 81:3217, 1993
6. Spangrude GJ, Heimfeld S, Weissman L:Purification and
characterization of mouse hematopoietic stem cells. Science 241:58,
1988
7. Szilvassy SJ, Cory S: Phenotypic and functional characterization of competitive long-term repopulating hematopoietic stem cells
enriched from 5-flurouracil-treated murine marrow. Blood 81:2310,
1993
8. Orlic D, Fisher R, Nishikawa S, Nienhuis AW, Bodine DM:
Purification and characterization of heterogeneous pluripotent hematopoietic stem cell population expressing high levels of c-kit receptor. Blood 82:762, 1993
9. Uphoff D E Preclusion of secondary phase of irradiation syndrome by inoculation of fetal hematopoietic tissue following lethal
total-body X irradiation. J Natl Cancer Inst 20:625, 1958
10. Tocci A, Rezzoug F, Aitouche A, Touraine JL: Comparison
of fresh, cryopreserved and cultured fetal liver. Bone Marrow Transplant 13641, 1994
1 I . Touraine JL: Bone marrow and fetal liver transplantation in
immunodeficiencies and inborn errors of metabolism: Lack of significant restriction of T-cell function in long-term chimeras despite
HLA-mismatch. Immunol Rev 71:103, 1983
12. Touraine JL, Raudrant D, Roy0 C: In utero transplantation
of stem cells in bare lymphocyte syndrome. Lancet 17:1382, 1989
13. Ledbetter JA, Herzenberg LA: Xenogeneic monoclonal antibodies to mouse lymphoid differentiation antigens. Immunol Rev
47:63, 1979
14. Dexter TM, Allen TD, Lajtha LG: Conditions controlling the
1471
proliferation of haemopoietic stem cells in vitro. J Cell Physiol
91:335, 1977
15. Mergenthaler HG, Briihl P, Dormer P: Kinetics of myeloid
progenitor cells in human micro long-term bone marrow cultures.
Exp Hematol 16:145, 1988
16. Spooncer E, Lord BI, Dexter "M: Defective ability to selfrenew in vitro of highly purified primitive haematopoietic cells.
Nature 316:62, 1985
17. Rimokh R, Rouault JP, Wahbi K, Gadoux M, Lafage M,
Archimbaud E, Charrin C, Gentilhomme 0, Germain D, Samarut J,
Magaud J P A chromosome 12 coding region is juxtaposed to the
MYC protooncogene locus in a t(8;12)(q24;q22) translocation in a
case of B-cell chronic lymphocytic leukemia. Gene Chromosom
Cancer 3:24, 1991
18. Chirghwin JM, Przybyla AE, McDonald W, Rutter W: Isolation of biologically active ribonucleic acid from sources enriched in
ribonuclease. Biochemistry 185294, 1979
19. Thomas PS: Hybridization of denatured RNA and small DNA
fragments transferred to nitrocellulose. Proc Natl Acad Sci USA
775201, 1981
20. Rimokh R, Magaud JP, Berger F, Samarut J, Coiffier B, Germain D, Mason DY: A translocation involving a specific break-point
(q 35) on chromosome 5 is characteristic of anaplastic large cell
lymphoma. Br J Haematol 71:31, 1989
21. Metcalf D, Moore MAS: Hemopoietic Cells. Amsterdam, The
Netherlands, North-Holland Publishing, 1971
22. Dexter TM, Spooncer E: Loss of immunoreactivity in longterm bone marrow culture. Nature 275:135, 1978
23. Lu L, Xiao M, Shen RN, Grisby S, Broxmeyer HE: Enrichment, characterization, and responsiveness of single primitive
CD34+ + human umbilical cord blood hematopoietic progenitors
with high proliferative and replating potential. Blood 81:41, 1993
24. Gutierrez-Ramos JC, Olsson C , Palacios R: Interleukin (IL1
to IL7) gene expression in fetal liver and bone marrow stromal
clones: Cytokine-mediated positive and negative regulation. Exp
Hematol 20:986, 1992
25. Landsdorp PM, Dragowska W, Mayani H: Ontogeny-related
changes in proliferative potential of human hematopoietic cells. J
Exp Med 178:787, 1993
26. Slaper-Cortenbach I, Ploemacher R, Lowenberg B: Different
stimulative effects of human bone marrow and fetal liver stromal
cells on erythropoiesis in long-term culture. Blood 69:135, 1987