1795
Journal of Cell Science 109, 1795-1801 (1996)
Printed in Great Britain © The Company of Biologists Limited 1996
JCS9438
IL-3, GM-CSF and CSF-1 modulate c-fms mRNA more rapidly in human early
monocytic progenitors than in mature or transformed monocytic cells
Béatrice Panterne1, Antoinette Hatzfeld1,*, Patricia Sansilvestri1, Angelo Cardoso2, Marie-Noëlle Monier1,
Pascal Batard1 and Jacques Hatzfeld1
1Centre National de la Recherche Scientifique, IFC1, UPR 9044, 7 rue Guy Môquet, 94801 Villejuif Cédex, France
2Division of Hematologic Malignancies, Dana Farber Cancer Institute, 44 Binney Street, Boston, 02115 MA, USA
*Author for correspondence
SUMMARY
We have previously shown that a low concentration of
CSF-1 (1 U/ml) can trigger human immature monocytic
progenitor proliferation in the presence of low concentrations of IL3 (1.7 U/ml). No c-fms down-regulation was
observed during this early cell activation. In contrast, 20
U/ml of CSF-1, active on late monocytic cell growth, downregulated c-fms mRNA expression in immature progenitors
and monocytes derived from bone marrow CD34+ cells in
culture. We have now extended this study to include the
effects of various concentrations of GM-CSF, IL3 and GCSF on c-fms expression. We observed that high doses of
GM-CSF or IL3 down-modulated c-fms mRNA, whereas
low doses of GM-CSF or IL3, which were active on early
monocytic growth, had no such effect. Similar results were
observed at the protein level. In contrast, whatever the concentration, G-CSF had no effect on c-fms mRNA or protein
levels. We further observed that the more immature the cfms expressing progenitors, the faster the down-modulation of this receptor. This was observed within less than 1
hour for immature bone marrow cells, 6 hours for peripheral blood monocytes and even longer for transformed
monocytic cells. These results suggest that oncogene
expression can be regulated much more rapidly in
immature progenitors than was previously observed in
mature cells or transformed cell lines.
INTRODUCTION
down-modulation on murine macrophages (Baccarini et al.,
1992), TGF-β1 down-modulates CSF-1R expression on Lin−
murine bone marrow cells (Jacobsen et al., 1993). Binding
experiments with murine bone marrow cells provide evidence
that CSF-1 and G-CSF receptors are trans-modulated by GMCSF and IL3 (Walker et al., 1985). Gliniak and Rohrschneider
(1990) showed that in the FDC-P1 murine cell line, CSF-1R
mRNA is controlled post-transcriptionally by GM-CSF and
IL3.
The study of messenger RNA expression in rare immature
normal human progenitors is more difficult than in cell lines
or more mature cells which can be obtained in large amounts.
We previously used in situ hybridization (ISH) to show c-fms
mRNA expression in small immature normal human progenitor cells with light chromatin and visible nucleoli (Panterne et
al., 1993). Maximal expression of c-fms mRNA was observed
in these cells after 7 days of culture with 1.7 U/ml IL3. This
expression was down-modulated by 20 U/ml but not by 1 U/ml
CSF-1. We also detected by immunocytochemistry that these
immature cells expressed the CSF-1R. In the present study, we
have applied a semi-quantitative ISH method developed in our
laboratory to compare c-fms trans-modulation in immature
progenitors with that of more mature and more easily obtainable cells such as peripheral blood monocytes or TPA induced
HL60 cell line. We observed CSF-1R trans-modulation by high
Hematopoiesis is a continuous process whereby progenitor
cells develop into mature blood cells under the control of
numerous signals such as growth factors (Metcalf, 1993;
Ogawa, 1993). One of these growth factors, the colony-stimulating-factor-1 (CSF-1), is specific for the development of the
monocytic lineage (Stanley et al., 1978; Tushinski et al., 1982).
It acts through the CSF-1 tyrosine kinase receptor (CSF-1R)
encoded by the proto-oncogene c-fms (Sherr et al., 1985;
Yeung et al., 1987). The CSF-1R is essentially expressed in
murine monocytes, macrophages and their committed precursors (Byrne et al., 1981) and it has been suggested that it is
also expressed in earlier mouse progenitors (Cumano et al.,
1992; Gilmore and Shadduck, 1995). We also observed its
expression in immature human progenitors which were able to
develop into the monocytic lineage (Panterne et al., 1993).
CSF-1R expression is altered by CSF-1, TPA, LPS and PKC
(Downing et al., 1989; Guilbert and Stanley, 1984; Panterne et
al., 1993; Sariban et al., 1989). CSF-1R can also be controlled
by growth factors other than CSF-1. This effect, called ‘transmodulation’ (Nicola, 1987), may be important in understanding how cytokines maintain hematopoietic homeostasis. For
example, Epo induces CSF-1R down-modulation (Van Zant
and Chen, 1983) and IFN-γ enhances LPS-induced CSF-1R
Key words: c-fms, Down-modulation, Cytokine, Hematopoietic
progenitor, Monocyte
1796 B. Panterne and others
concentrations of GM-CSF or IL3 but not by G-CSF. Interestingly, we observed a much faster effect on CSF-1R mRNA
expression by CSF-1, GM-CSF or IL3 in the most immature
cells than in mature or transformed monocytic cells.
MATERIALS AND METHODS
Peripheral blood monocytes, after detachment, were incubated with
450 U/ml of GM-CSF or 20 U/ml of CSF-1. HL60 cells were induced
by 10 ng/ml TPA (LC Service, USA) for 24 to 48 hours and then
treated with the same concentration of cytokines as peripheral blood
monocytes. As peripheral blood monocytes and HL60 cells required
longer times than CD34+ cells to respond to cytokines, they were
harvested at belated times i.e. 0, 1, 3, 6, 12 and 24 hours after cytokine
addition.
Cytokines and CSF-1 antiserum
Recombinant IL3, GM-CSF, G-CSF and CSF-1 were supplied by the
Genetics Institute (Cambridge, MA, USA). Specific activities were
2.3×106 U/mg, 9.3×106 U/mg, 7×104 U/mg and 6×105 U/mg, respectively. Recombinant erythropoietin (Epo) was obtained from Amgen
Biological (Thousand Oaks, CA, USA) and its specific activity was
5×105 U/mg. Rabbit antiserum against human CSF-1 was from the
Genetics Institute (Cambridge, MA, USA); a 1/50 dilution neutralizes
5 U/ml of CSF-1.
In situ hybridization
This procedure was based on the methods described by Haase et al.
(1984) and Angerer et al. (1987) and is detailed elsewhere (Panterne
et al., 1993). Briefly, it is as follows:
Bone marrow, peripheral blood and cell line
Rib, sternum or iliac crest bone marrow was obtained from normal
transplant donors in agreement with the Institution’s ethical guidelines. Peripheral blood was obtained from normal transfusion donors.
Cells of the HL60 pro-myelocytic cell line (Collins et al., 1977) were
compared with normal cells.
Preparation of the probes
Both v-fms and human c-fms probes were used in these experiments
and results were identical. They were a gift from Drs C. J. Sherr and
M. Roussel (Donner et al., 1982; Roussel et al., 1983) and were
purified as described elsewhere (Panterne et al., 1993). The DNA and
RNA labelled probes were obtained by two different methods. The
DNA labelled probes (v-fms, c-fms and λ DNA used as a negative
control) were synthesized by the multiprime DNA labelling system
with [35S]dCTP from Amersham. The RNA labelled probes (v-fms,
c-fms and the transferrin receptor, a generous gift from Dr Claudio
Schneider, EMBL, Heidelberg (used as a positive control) were synthesized by the Riboprobe II Core System (Promega) with [35S]UTP
(Amersham). The specific activity of the labelled probes was usually
2-4×108 cpm/µg for DNA probes and 1.3×109 cpm/µg for RNA
probes.
CD34+ cell purification
CD34+ cells were obtained as previously described (Cardoso et al.,
1995; Hatzfeld et al., 1994). Briefly, after Ficoll separation, the cell
suspension was enriched for CD34+ cells on soybean agglutinin
(SBA) in a SBA CELLector flask to remove mature cells, and then
on a ICH3 CD34 antibody-covered CELLector flask (RPRGENCELL, Santa Clara, CA) to select CD34+ cells. The manufacturer’s procedure (Lebkowski et al., 1992) was modified (Cardoso et
al., 1995). Firstly, we eliminated the non immunoadsorbed cells from
the CD34 CELLector with more washes (up to 10) to completely
remove non CD34+ cells and the late progenitor CD34+low cells.
Secondly, we replaced the shaking procedure to recover the CD34+high
cells by a two hour (or overnight) incubation in IMDM with 10% fetal
calf serum (FCS) which allows an easy detachment and increased
recovery of all CD34+high cells. With this procedure, we routinely
obtain 95±3% CD34+ cells with more than 72% recovery of CFUGEMM. This population was then used for ISH or immunocytochemistry.
Peripheral blood monocyte purification
Peripheral blood monocytes were obtained after separation on a Ficoll
gradient and adherence on a plastic culture flask (Falcon) for 2 hours
in IMDM supplemented with 10% FCS at 37°C. After several washes,
monocytes were detached by incubation with PBS containing 2 mM
EDTA for at least 15 minutes at 37°C. This population was used for
ISH.
Down-modulation conditions
Five thousand CD34+ cells/ml were plated in 35 mm Petri dishes in
IMDM containing 0.9% (w/v) methylcellulose, 10 µg/ml transferrin,
1 mg/ml deionized bovine serum albumin, 4×10−5 β-mercaptoethanol,
2 mM glutamine, 30% FCS, and only 1.7 U/ml IL3 and 1 U/ml Epo
as cytokines. Cells were incubated for 7 days in a humidified atmosphere (5% CO2 in air at 37°C). At day 7, medium (0.5 ml/ Petri dish)
was added on top of the methylcellulose with the following growth
factors: IL3 alone (1.7 or 100 U/ml), IL3 (1.7 U/ml) with GM-CSF
(9 or 450 U/ml) or G-CSF (3.3 or 90 U/ml). It should be specified that
the low doses of cytokines were growth-effective in culture assay
(Cardoso et al., 1993; Zhou et al., 1988). The total cell population was
harvested at 0, 30, 60, and 180 minutes after growth factor addition
and was cytocentrifuged for further study.
Preparation of cells
Cells were cytocentrifuged on a Cytospin (Shandon Elliott; 7 minutes,
500 rpm, 5×103 cells/Denhardt treated slide; Haase et al., 1984), airdried for 5 minutes, fixed and preserved in 70% ethanol at 4°C.
In situ hybridization
For DNA multiprime labelled probes, the procedure was as follows:
the hybridization solution was mixed with denatured 35S-labelled cfms probe with or without a 100-fold excess of unlabelled probe, or
35S-labelled λ DNA at a concentration so that 12 µl of each mixture
contained 106 cpm which was then layered onto cell coated slide. The
slides were incubated at room temperature for 16-18 hours in a humidified atmosphere. At the end of the hybridization reaction, the slides
were washed in various solutions of increasing stringency. For the
RNA probes, the procedure was as follows: after removal of ethanol,
slides were immersed in water for 5 minutes, then in PBS and postfixed in 4% paraformaldehyde for 20 minutes at 4°C, rinsed twice
with PBS, once with water, then acetylated for 10 minutes in 100 mM
triethanolamine with 0.25% anhydrid acetic, rinsed again twice in
water, dehydrated and air dried. Hybridization solution (12 µl) containing 106 cpm of probes was applied per slide. The hybridization
was performed in a humid chamber at 50°C for 16 to 18 hours in the
dark. Slides were washed in buffer (50% formamide, 2× SSC, 10 mM
DTT), treated with RNase A (100 µg/ml) and RNase T1 (10 U/ml) in
2× SSC (Bernaudin et al., 1988) and rinsed in 2× SSC. In both procedures, slides were then dehydrated and coated for autoradiography
with Kodak NTB2 emulsion. After 7 to 10 days of exposure, the slides
were developed and stained with Giemsa.
Morphological evaluation and grain counting
Both cell size and numbers of grains per cell were automatically
analysed using autoradiography software on a SAMBA image
analyser (Alcatel TITN Answare, Meylan, France). The specificity of
the c-fms probe was checked on c-fms expressing cell lines (human
CSF-1R transfected NIH 3T3 cell line (Roussel et al., 1988) and TPA
induced HL60 cell line) and non expressing cells (WT NIH 3T3 and
non induced HL60 cells). For cells which did not express c-fms (WT
Fast c-fms modulation in early monocytic progenitors 1797
NIH 3T3, non induced HL60), the average number of grains/cell was
less than 10 (4.7±5.5) after hybridization with the antisense probe and
it was the same for cells expressing c-fms (transfected NIH 3T3, TPA
induced HL60 cells and blood monocytes) when hybridized with the
sense probe. Therefore, cells hybridized with the antisense c-fms
probe were scored as positive when they contained more than 10
grains per cell.
Bone marrow monocytes were easily distinguished from immature
progenitors by morphological criteria. Monocytes were cells of 15 to
20 µm in diameter, containing an indented nucleus showing a coarser
chromatin pattern without nucleoli. The nuclear/cytoplasmic ratio was
less than 40%. Immature cells were smaller, of 10 to 12 µm in
diameter and presenting a high nuclear/cytoplasm ratio (>60%). The
nucleus was regular and exhibited a homogenous and light chromatin
pattern with visible nucleoli.
Immunocytochemistry
Rat monoclonal anti-FMS antibody (clone 3-4A4-E4) was supplied
by Oncogene Science (Uniondale, NY), mouse monoclonal anti-TrfR antibody by Immunotech (Marseille, France) and rat serum used as
control was from Dako (Glostrup, Denmark). The experiments were
performed using the Universal Dako LSAB Kit following the manufacturer’s instructions. After 7 days of culture, cells were cytocentrifuged on slides, air dried for at least 1 hour and then fixed for 10
minutes in acetone. Endogenous biotin was blocked first with 0.1%
avidin and then with 0.01% biotin in 5 mM TBS, pH 7.6, prior to the
first step of the staining protocol. Slides were then incubated with the
primary antibody for 30 minutes at room temperature, rinsed,
incubated with the secondary biotinylated antibody, rinsed and
incubated with streptavidin conjugated to the alkaline phosphatase.
After application of this enzyme substrate, positive cells were stained
red. Cells were identified as described in the previous paragraph.
The CSF-1R transfected NIH 3T3 cell line was used as positive
control: an average of 90% of transfected NIH 3T3 cells expressed cfms product (data not shown).
Statistical analysis
The Student’s t-test was used to determine the significance of differences between values obtained for several (n) experiments. The probabilities are detailed in the results.
RESULTS
GM-CSF and IL3 but not G-CSF down-regulate c-fms
mRNA expression in cultured immature progenitors
and monocytes
To understand better the regulation of c-fms mRNA that we
had previously detected in normal early monocytic progenitors
(Panterne et al., 1993), we studied whether cytokines other
than CSF-1, such as GM-CSF, IL3 or G-CSF, could promote
c-fms mRNA trans-modulation in these cells. After 7 days of
culture of bone marrow CD34+ cells with 1.7 U/ml IL3 and 1
U/ml Epo alone, considered as t=0, high or low doses of GMCSF, IL3 or G-CSF were added and the kinetics of expression
of c-fms mRNA was followed by ISH on the whole cell population. Fig. 1 shows that in the presence of 450 U/ml GMCSF (dark columns), we observed a trans-modulation of c-fms
mRNA expression within 30 minutes in bone marrow progenitors and monocytes derived from cultured CD34+ and discriminated by morphological criteria as described in Material
and Methods. For immature cells, a labelling reduction of 78%
was observed and this value did not increase thereafter. For
bone marrow monocytes, after 30 minutes, a smaller reduction
Fig. 1. Down-modulation of c-fms expression in CD34+ derived bone
marrow immature progenitors and monocytes in response to 9 U/ml
(stippled columns) or 450 U/ml (filled columns) of GM-CSF. After
incubation with GM-CSF, cells were harvested at the indicated time
and treated for ISH with the labelled 35Sc-fms probe. The number of
c-fms positive cells is expressed as a percentage of the number of
positive cells counted in the absence of GM-CSF at day 7 defined as
time 0 (open columns). Student’s t-test shows that the percentage of
positive cells with a high dose of GM-CSF was significantly reduced,
as compared to the percentage of positive cells treated with a low
dose or untreated cells (for immature cells P≤0.003 and P≤0.001,
respectively; for monocytes P≤0.02 and P≤0.05, respectively).
of 54% was observed and this percentage remained constant
for 3 hours. On the contrary, when no cytokines (white
columns) or when low but growth-efficient concentrations
(grey columns) of GM-CSF (9 U/ml) were added, no transmodulation was observed.
To see whether the effect of GM-CSF on c-fms mRNA
expression was mediated by a possible induction of CSF-1
secretion, we treated the cells with a low or a high dose of GMCSF in the presence of an anti-CSF-1 serum. In culture, this
antiserum decreased by 75% the number of monocytic colonies
obtained with CSF-1 (1 U/ml) and IL3 (1.7 U/ml) (data not
shown). The anti-CSF-1 serum did not abolish the down-modulation of the c-fms mRNA observed in the presence of 450
U/ml GM-CSF (data not shown).
A similar down-modulation was observed with IL3. Fig. 2
shows that by 30 minutes there was a 72% reduction in the
number of immature cells expressing c-fms mRNA in the
presence of 100 U/ml IL3 (dark columns). For monocytes, a
75% decrease in the number of labelled cells was observed
during the same time. No significant alteration was obtained
with 1.7 U/ml IL3 (white columns).
In contrast to GM-CSF and IL3 data, G-CSF did not alter
the level of c-fms mRNA expression in any cell type, whatever
the dose used, as shown in Fig. 3. The percentage of labelled
immature cells or monocytes in the presence of 3.3 or 90 U/ml
G-CSF was similar to that observed with 1.7 U/ml IL3 alone.
Whatever the growth factor conditions used and whatever
the incubation times, the Trf-R mRNA expression did not
1798 B. Panterne and others
Fig. 2. Down-modulation of c-fms expression in CD34+ derived bone
marrow immature progenitors and monocytes in response to 1.7
U/ml (open columns) or 100 U/ml (filled columns) of IL3. The
number of positive cells is expressed as a percentage of the number
of positive cells in the presence of 1.7 U/ml of IL3 at day 7 (time 0).
The percentage of positive cells was significantly different with the 2
IL3 concentrations (P≤0.005 for immature cells and P≤0.001 for
monocytes).
Fig. 4. c-fms mRNA expression in (A) bone marrow monocytes and
(B) immature progenitors, after 7 days of culture in the presence of
IL3 and after ISH with a 35Sc-fms labelled probe. (C) Cells treated
with a λ DNA probe as a negative control. (D) Cells treated with a
35S-Trf-R probe; about 75 to 80% cells were positive. Bars, 20 µm.
Table 1. Down-modulation of CSF-1R expression detected
by anti-CSF-1R antibody in CD34+ derived bone marrow
immature progenitors and monocytes in response to a 3
hour treatment with high or low concentrations of GM-CSF
Culture condition
IL3 1.7 U/ml
IL3 + 9 U/ml GM-CSF
IL3 + 450 U/ml GM-CSF
% Monocytes
expressing CSF-1R
46.1±4.4
49.2±4.7
14.2±2.9*
% Immature progenitors
expressing CSF-1R
27.9±5.8
30.8±3.7
10.7±1.8**
Bone marrow CD34+ cells were cultured for 7 days in the presence of IL3
(1.7 U/ml) and Epo (1 U/ml), followed by IL3 added alone or together with 9
or 450 U/ml GM-CSF for 3 hours. Cells were then treated by
immunocytochemistry as described in Materials and Methods. The number of
cells expressing CSF-1R after treatment with 450 U/ml of GM-CSF was
significantly lower than after treament with 9 U/ml of GM-CSF for the
monocytes as well as for immature cells (*degrees of freedom=3, P=0.0006
and **degrees of freedom=3, P=0.0012, respectively). No significant
difference was observed between cells with IL3 alone or IL3 with 9 U/ml of
GM-CSF.
Fig. 3. c-fms expression in CD34+ derived bone marrow immature
progenitors and monocytes in response to 3.3 U/ml (stippled
columns) or 90 U/ml (filled columns) of G-CSF. The number of
positive cells is expressed as a percentage of the number of positive
cells counted in the absence of G-CSF (open columns). There was no
significant difference between the percentages of positive cells in the
different groups.
change. About 70 to 80% of cells were consistently labelled
with a Trf-R antisense probe.
Fig. 4 shows immature cells and monocytes treated by ISH.
Cells were hybridized with the c-fms labelled probe (Fig.
4A,B), with a λ probe as a negative control (Fig. 4C), or with
the Trf-R labelled probe as a positive control (Fig. 4D). An
additional negative control was performed with non-induced
HL60 cells. These cells when hybridized with the c-fms
antisense labelled probe were negative (data not shown).
We confirmed these results on mRNA expression at the
protein level by immunocytochemistry. After 7 days of culture
in the presence of IL3 and Epo alone, as shown in Table 1,
about 28% of the immature cells and 46% of the monocytes
expressed CSF-1R. Three hours after the addition of a low dose
of GM-CSF (9 U/ml), no significant difference was observed:
about 30% of immature cells and 49% of monocytes were
positive after incubation with an antiCSF1-R antibody. On the
Fast c-fms modulation in early monocytic progenitors 1799
contrary, in the presence of 450 U/ml GM-CSF, only 11% of
immature cells and 14% of monocytes still expressed CSF-1R.
Similar results were obtained with high and low doses of IL3
but no alteration was observed with G-CSF (results not
shown). Whatever the experimental conditions, the expression
of the Trf-R at the surface level was constant: about 75% to
80% of cells expressed this receptor.
c-fms mRNA modulation is more rapid in normal
bone marrow immature cells than in peripheral
blood monocytes and TPA-induced HL60 cells
In our previous study, we have shown that the percentage of
bone marrow cells expressing c-fms mRNA in culture
decreased within 30 minutes after the addition of 20 U/ml of
CSF-1 (Panterne et al., 1993). Such a short time for c-fms
mRNA down-regulation has never been observed before by
northern blot analysis on peripheral blood monocytes (Sariban
et al., 1989) or cell lines (Gliniak and Rohrschneider, 1990).
As detailed by image analysis after ISH, Fig. 5 shows the
frequency of peripheral blood monocytes expressing various
amount of c-fms mRNA as visualized by the amount of autoradiographic grains. We can observe at time 0 that 42% of the
cells contained significantly more than 10 grains, i.e. more than
the background level which was equivalent to 4.7±5.5
grains/cell (Fig. 5A). In Fig. 5B, we confirm that after 1 hour
there was no significant down-regulation of c-fms expression
in peripheral blood monocytes. On the contrary, in Fig. 5C, 6
hours after 20 U/ml CSF-1 were added, less than 5% of these
cells still contained more than 10 grains. Fig. 6 indicates
further that this down-regulation was largely completed
between 3 and 6 hours. Therefore, these results show that ISH
Fig. 5. Frequency of peripheral blood monocytes expressing various
amounts of mRNA as detected by image analysis of autoradiographic
grains. (A) Time 0, (B) 1 hour, (C) 6 hours, after the addition of 20
U/ml of CSF-1 and hybridization with a c-fms antisense riboprobe.
(D) At time 0 after hybridization with the c-fms sense riboprobe.
provided similar results to those obtained by northern blot
analysis. Moreover, ISH allows a more detailed analysis at the
level of single cells or cell subpopulations.
Fig. 6 summarizes the kinetics of c-fms down-regulation
observed with bone marrow monocytes and early progenitors,
peripheral blood monocytes and TPA induced HL60 cells in
the presence of 20 U/ml CSF-1. We observed a down-regula-
Fig. 6. Kinetics of c-fms expression in human immature bone
marrow cells (s), bone marrow monocytes (h), peripheral blood
monocytes (j) and TPA treated HL60 cell line (n) at different times
after addition of 20 U/ml CSF-1. Percentages of labelled cells were
determined after grain counts following ISH with a 35S-c-fms
antisense riboprobe. Percentages were calculated by considering the
value at time 0 as 100%.
Fig. 7. Kinetics of c-fms expression in human immature bone
marrow cells (s), bone marrow monocytes (h), peripheral blood
monocytes (j) and TPA treated HL60 cell line (n) at different times
after addition of 450 U/ml GM-CSF. Percentages of labelled cells
were determined after grain counts following ISH with a 35S-c-fms
antisense riboprobe. Percentages were calculated by considering the
value at time 0 as 100%.
1800 B. Panterne and others
tion of c-fms mRNA expression in peripheral blood monocytes
(j) within 3 to 6 hours of CSF-1 addition whereas this
phenomenon was already completed within 1 hour in bone
marrow cells (s and u). No down-regulation of c-fms mRNA
was observed in HL60 cells (n) with 20 U/ml CSF-1. Quite
similar kinetics were observed when 450 U/ml GM-CSF
instead of CSF-1 was used to down modulate c-fms mRNA.
Fig. 7 shows that in bone marrow cells a decrease in c-fms
mRNA was already observed within 1 hour after GM-CSF
addition. On the contrary, in the same conditions, c-fms mRNA
decreased in peripheral blood monocytes between 6 and 12
hours after GM-CSF addition and this GM-CSF concentration
did not modify significantly the percentage of HL60 cells
expressing c-fms mRNA.
DISCUSSION
In this study, we observed that IL-3 and GM-CSF, but not GCSF, down-modulate c-fms mRNA and CSF-1R in normal
human bone marrow progenitors and monocytes which were
both derived from CD34+ cells in culture. We further observed
that c-fms mRNA expression decreased significantly more
rapidly in the more primitive monocytic cells as compared to
peripheral blood monocytes or established cell lines.
One interest of this study is that we are dealing with subpopulations of early bone marrow monocytes or progenitors,
some of which still express CD34 antigen (data not shown).
Bone marrow monocytes are more immature and can sustain
more cell divisions than peripheral blood monocytes.
Immature monocytic cells cannot be routinely obtained in
large amounts for northern blot analysis. This is why we have
only been able to compare northern blot and ISH methods with
peripheral blood monocytes as they can be obtained easily in
large amounts. Because ISH methods allow analysis both at a
single cell and molecular level, it is quite suitable for the study
of rare hematopoietic subpopulations which are distinguishable
among a heterogenous population. The use of riboprobes
which provide very low autoradiographic background and the
use of image analysis has rendered this method semi-quantitative and reliable. With our riboprobe ISH method, we can
observe as few as 10 c-fms mRNA copies, whereas with cytofluorometry or immunocytochemistry, we can rarely detect less
than 200 to 1,000 CSF-1 receptors per cell. This methodological point is important in view of the fact that immature cells
express low numbers of receptors, in between 100 and 1,000
per cell. The fact that we observe only 46% of bone marrow
monocytes that express CSF-1R at the membrane level, in
culture, may therefore be an underestimation.
The ISH method is validated here by the fact that in peripheral blood monocytes we obtained similar kinetics of c-fms
mRNA down-regulation as observed by northern blot analysis.
Sariban et al. (1989) found c-fms down-modulation in peripheral blood monocytes with a messenger decrease of 35% and
62%, 2 and 6 hours, respectively, after CSF-1 treatment. Two
different methods of ISH using either DNA or RNA probes
provided the same results in our hands. Therefore this suggests
that the extremely fast reduction in c-fms mRNA level
observed in human bone marrow progenitors and monocytes
by ISH is not due to a technical artefact.
The down-modulation of c-fms mRNA in blood monocytes
by CSF-1 appeared to be transient (Fig. 6). This observation is
in agreement with the results of Sariban et al. (1989) who
showed that c-fms mRNA was up-regulated at 24 hours after
the down-regulation by CSF-1. We have observed a quite
similar transient effect with c-kit modulation in CD34+ cells
(Sansilvestri et al., 1995).
We know that GM-CSF induces CSF-1 secretion in
monocytes (Rambaldi et al., 1987). However, using the antiCSF-1 blocking antibody, we demonstrated that the rapid
trans-modulation of c-fms mRNA by GM-CSF was not due to
a possible production of CSF-1 which could down-regulate its
own receptor. On the other hand, Vellenga et al. (1988) have
shown that IL3 induces CSF-1 mRNA in human monocytes
but it would be unlikely that both CSF-1 secretion and c-fms
down-regulation could occur in 30 minutes.
The down-modulation of c-fms has been observed in many
other hematopoietic cell types (Gliniak and Rohrschneider,
1990; Van Zant and Chen, 1983; Walker et al., 1985; Weber
et al., 1989) but here, for the first time, by a detailed study in
rare normal immature human bone marrow progenitors,
derived from CD34+ cells in culture, we describe a very rapid
process of mRNA disappearance. In the HL60 cell line,
Weber et al. (1989) have observed a post-transcriptional cfms mRNA stabilization by a labile protein. In the FDCP-1
murine cell line, a post-transcriptional c-fms mRNA regulation was also observed, due to a GM-CSF stimulated ribonucleasic activity (Gliniak et al., 1992). This suggests that
various pathways could exist for the regulation of c-fms
expression. Molecular micro-methods will have to be
developed in order to study the mechanisms which could
occur in rare immature progenitors.
Our study is reminiscent of Metcalf’s observation in culture
assay where variable doses of GM-CSF were added to micromanipulated daughter cells: with a low dose of GM-CSF,
daughter cells could give rise to monocytic colonies but with
a high dose, granulocytic colonies appeared (Metcalf, 1980).
It would be interesting to study how local concentrations of
these cytokines in the bone marrow microenvironment could
control the development of CFU-GM.
We are indebted to Dr Mary Osborne for her critical reading of the
manuscript. This work was supported by Centre National de la
Recherche Scientifique (CNRS), Direction des Recherches Etudes et
Techniques (DRET), Association pour la Recherche sur le Cancer
(ARC), Institut National de la Santé et de la Recherche MédicaleCNAMTS and Ministère de la Recherche et de l’Enseignement
Supérieur. B.P. has been a recipient of fellowships from DRET and
from Fondation Singer-Polignac. PB has been a recipient of a fellowship from DRET.
REFERENCES
Angerer, L. M., Cox, K. H. and Angerer, R. C. (1987). Demonstration of
tissue-specific gene expression by in situ hybridization. Meth. Enzymol. 152,
649-661.
Baccarini, M., Dello Sbarba, P., Buscher, D., Bartocci, A. and Stanley, E.
R. (1992). IFN-γ/lipopolysaccharide activation of macrophages is associated
with protein kinase C-dependent down-modulation of the colony-stimulating
factor-1 receptor. J. Immunol. 149, 2656-2661.
Bernaudin, J.-F., Yamauchi, K., Wewers, M. D., Tocci, M. J., Ferrans, V. J.
and Crystal, R. G. (1988). Demonstration by in situ hybridization of
dissimilar IL-1β gene expression in human alveolar macrophages and blood
monocytes in response to lipopolysaccharide. J. Immunol. 140, 3822-3829.
Fast c-fms modulation in early monocytic progenitors 1801
Byrne, P. V., Guilbert, L. J. and Stanley, E. R. (1981). Distribution of cells
bearing receptors for a colony-stimulating factor (CSF-1) in murine tissues.
J. Cell Biol. 91, 848-853.
Cardoso, A., Li, M., Batard, P., Hatzfeld, A., Brown, E., Levesque, J.,
Sookdeo, H., Panterne, B., Sansilvestri, P., Clark, S. and Hatzfeld, J.
(1993). Release from quiescence of CD34+CD38− human umbilical cord
blood cells reveals their potentiality to engraft adults. Proc. Nat. Acad. Sci
USA 90, 8707-8711.
Cardoso, A. A., Watt, S., Batard, P., Li, M. L., Hatzfeld, A., Genevier, H.
and Hatzfeld, J. (1995). An improved panning technique for the selection of
CD34-positive human bone marrow hematopoietic cells with high recovery
of early progenitors. Exp. Hematol. 23, 407-412.
Collins, S. J., Gallo, R. C. and Gallagher, R. E. (1977). Continuous growth
and differentiation of human myeloid leukaemic cells in suspension culture.
Nature 270, 347-348.
Cumano, A., Paige, C. J., Iscove, N. N. and Brady, G. (1992). Bipotential
precursors of B cells and macrophages in murine fetal liver. Nature 356, 612615.
Donner, L., Fedele, L. A., Garon, C. F., Anderson, S. J. and Sherr, C. J.
(1982). Mc Donough feline sarcoma virus: characterization of the
molecularly cloned provirus and its feline oncogene (v-fms). J. Virol. 41,
489-500.
Downing, J. R., Roussel, M. F. and Sherr, C. J. (1989). Ligand and protein
kinase C downmodulate the colony-stimulating factor 1 receptor by
independent mechanisms. Mol. Cell. Biol. 9, 2890-2896.
Gilmore, G. L. and Shadduck, R. K. (1995). Inhibition of day-12 spleen
colony-forming units by a monoclonal antibody to the murine
macrophage/monocyte colony-stimulating factor receptor. Blood 85, 27312734.
Gliniak, B. C. and Rohrschneider, L. R. (1990). Expression of the M-CSF
receptor is controlled posttranscriptionally by the dominant actions of GMCSF or multi-CSF. Cell 63, 1073-1083.
Gliniak, B. C., Park, L. S. and Rohrschneider, L. R. (1992). A GM-colonystimulating factor (CSF) activated ribonuclease system transregulates MCSF receptor expression in the murine FDC-P1/MAC myeloid cell line. Mol.
Biol. Cell. 3, 535-544.
Guilbert, L. J. and Stanley, E. R. (1984). Modulation of receptors for the
colony-stimulating factor, CSF-1, by bacterial lipopolysaccharide and CSF1. J. Immunol. Meth. 73, 17-28.
Haase, A., Brahic, M., Stowring, L. and Blum, H. (1984). Detection of viral
nucleic acids by in situ hybridization. Meth. Virol. 7, 189-226.
Hatzfeld, J., Batard, P., Cardoso, A. A., Li, M. L. and Hatzfeld, A. (1994).
Purification and in vitro assay of early human hematopoietic progenitors. In
Culture of Hematopoietic Cells (ed. R. I. Freshney, I. B. Pragnell and M. G.
Freshney), pp. 205-221. John Wiley & Sons, Inc., New York.
Jacobsen, S. E. W., Ruscetti, F. W., Roberts, A. B. and Keller, J. R. (1993).
TGF-β is a bidirectional modulator of cytokine receptor expression on
murine bone marrrow cells. Differential effects of TGF-β1 and TGF-β3. J.
Immunol. 151, 4534-4544.
Lebkowski, J. S., Schain, L. R., Okrongly, D., Levinsky, R., Harvey, M. J.
and Okarma, T. B. (1992). Rapid isolation of human CD34 hematopoietic
stem cell-purging of human tumor cells. Transplantation 53, 1011-1019.
Metcalf, D. (1980). Clonal analysis of proliferation and differentiation of
paired daughter cells: action of granulocyte-macrophage colony-stimulating
factor on granulocyte-macrophage precursors. Proc. Nat. Acad. Sci. USA 77,
5327-5330.
Metcalf, D. (1993). Hematopoietic regulators: Redundancy or subtlety? Blood
82, 3515-3523.
Nicola, N. A. (1987). Why do hemopoietic growth factor receptors interact with
each other? Immunol. Today 8, 134-140.
Ogawa, M. (1993). Differentiation and proliferation of hematopoietic stem
cells. Blood 81, 2844-2853.
Panterne, B., Zhou, Y. Q., Hatzfeld, J., Li, M. L., Lévesque, J.-P., Clark, S.
C. and Hatzfeld, A. (1993). CSF-1 control of c-fms mRNA expression in
normal human bone marrow early progenitors. J. Cell. Physiol. 155, 282289.
Rambaldi, A., Young, D. C. and Griffin, J. D. (1987). Expression of the MCSF (CSF-1) gene by human monocytes. Blood 69, 1409-1413.
Roussel, M. F., Sherr, C. J., Barker, P. E. and Ruddle, F. H. (1983).
Molecular cloning of the c-fms locus and its assignment to human
chromosome 5. J. Virol. 48, 770-773.
Roussel, M., Downing, J., Rettenmier, C. and Sherr, C. (1988). A point
mutation in the extracellular domain of the human CSF-1 receptor (c-fms
proto-oncogene product) activates its transforming potential. Cell 55, 979988.
Sansilvestri, P., Cardoso, A. A., Batard, P., Panterne, B., Hatzfeld, A., Lim,
B., Lévesque, J.-P., Monier, M.-N. and Hatzfeld, J. (1995). Early
CD34high cells can be separated into KIThigh cells in which transforming
growth factor-β (TGF-β) downmodulates c-kit and KITlow cells in which
anti-TGF-β upmodulates c-kit. Blood 86, 1729-1735.
Sariban, E., Imamura, K., Sherman, M., Rothwell, V., Pantazis, P. and
Kufe, D. (1989). Downregulation of c-fms gene expression in human
monocytes treated with phorbol esters and colony-stimulating factor 1. Blood
74, 123-129.
Sherr, C. J., Rettenmier, C. W., Sacca, R., Roussel, M. F., Look, A. T. and
Stanley, E. R. (1985). The c-fms proto-oncogene product is related to the
receptor for the mononuclear phagocyte growth factor CSF-1. Cell 41, 665676.
Stanley, E. R., Chen, D. M. and Lin, H.-S. (1978). Induction of macrophage
production and proliferation by a purified colony stimulating factor. Nature
274, 168-170.
Tushinski, R. J., Oliver, I. T., Guilbert, L. J., Tynan, P. W., Warner, J. R.
and Stanley, E. R. (1982). Survival of mononuclear phagocytes depends on
a lineage-specific growth factor that the differentiated cells selectively
destroy. Cell 28, 71-81.
Van Zant, G. and Chen, B. D.-M. (1983). Erythropoietin causes down
regulation of colony-stimulating factor (CSF-1) receptors on peritoneal
exudate macrophages of the mouse. J. Cell Biol. 97, 1945-1949.
Vellenga, E., Rambaldi, A., Ernst, T., Ostapovicz, D. and Griffin, J. (1988).
Independent regulation of M-CSF and G-CSF gene expression in human
monocytes. Blood 71, 1529-1532.
Walker, F., Nicola, N. A., Metcalf, D. and Burgess, A. W. (1985).
Hierarchical down-modulation of hemopoietic growth factor receptors. Cell
43, 269-276.
Weber, B., Horiguchi, J., Luebbers, R., Sherman, M. and Kufe, D. (1989).
Posttranscriptional stabilization of c-fms mRNA by a labile protein during
human monocytic differentiation. Mol. Cell. Biol. 9, 769-775.
Yeung, Y. G., Jubinski, P. T., Sengupta, A., Yeung, D. C. Y. and Stanley, E.
R. (1987). Purification of the colony-stimulating factor 1 receptor and
demonstration of its tyrosine kinase activity. Proc. Nat. Acad. Sci. USA 84,
1268-1272.
Zhou, Y. Q., Stanley, E. R., Clark, S. C., Hatzfeld, J., Lévesque, J.-P.,
Federici, C., Watt, S. M. and Hatzfeld, A. (1988). Interleukin-3 and
interleukin-1α allow earlier bone marrow progenitors to respond to human
colony-stimulating factor-1. Blood 72, 1870-1874.
(Received 13 October 1995 - Accepted 5 April 1996)