(Received for publication, January 28, 1997, and in revised form, May 19, 1997)
From the Division of Medical Oncology, University of
Colorado Health Science Center, Denver, Colorado 80262, the
¶ Department of Biochemistry, University of Kentucky, Lexington,
Kentucky 40536-0084, and the
Division of Medical Oncology,
University of Washington and Veterans Administration Medical Center,
Seattle, Washington 98108
Granulocyte-macrophage colony-stimulating factor
(GM-CSF) regulates differentiation, survival, and proliferation of
colony-forming unit-granulocyte-macrophage progenitor cells. The
biologic actions of GM-CSF are mediated by binding to a specific
receptor consisting of two chains designated as and
subunits.
We have demonstrated that the murine FDC-P1-derived cell line WT-19
transfected with the human GM-CSF receptor
and
subunits
(GM-CSFR
and
) can be induced to differentiate by the addition of
human GM-CSF (hGM-CSF). By expressing a series of GM-CSFR
mutants in
WT19 cells, we have determined the amino acid domains of the GM-CSFR
cytoplasmic domain that regulate cell differentiation, proliferation,
and survival. We found that the membrane proximal proline-rich domain and adjacent 16 residues are essential for both
hGM-CSF-dependent cell proliferation and differentiation.
In contrast, the C-terminal region of the GM-CSFR
cytoplasmic domain
was not necessary for cell differentiation mediated by hGM-CSF, but the
removal of this region severely impaired the ability of hGM-CSF to
support cell survival. While the activation of JAK2, Shc, Erk, and
STAT5 proteins correlated with hGM-CSF-mediated cell growth, cellular
differentiation occurred in the absence of activation of these signal
transduction pathways.
Granulocyte-macrophage colony-stimulating factor (GM-CSF)1 is a 22-kDa glycoprotein, which is secreted by activated T cells, endothelial cells, fibroblasts, mast cells, B cells, and macrophages (1-4). GM-CSF plays an important role in promoting differentiation, survival, and proliferation of colony-forming unit-granulocyte-macrophage progenitor cells as well as enhancing the function of mature neutrophils, monocytes, and eosinophils (5, 6) and stimulating burst promoting activity for burst-forming units, erythroid (7, 8). GM-CSF causes a major cytoskeletal reorganization in plasma cells and hairy cells, resulting in the inhibition of motility and loss of adhesion to cellular and matrix ligands (9).
The biologic actions of GM-CSF are mediated by binding to a specific
receptor consisting of and
subunits, both of which are members
of the type-I cytokine receptor family (10, 11). The
subunit binds
GM-CSF with low affinity (10). A soluble form of human GM-CSF receptor
subunit (GM-CSFR
) has also been identified, whose function
in vivo is unclear (12, 13). While the
subunit does not
bind GM-CSF by itself, it forms a high affinity receptor in combination
with the
subunit (11). The
chain is called the common
chain
(
c) because it is shared by interleukin 3 (IL3) and interleukin 5 (IL5) receptors (14, 15). Although the cytoplasmic domain of GM-CSFR
is only 54 amino acids, we and others have demonstrated that the
GM-CSFR
cytoplasmic domain is necessary for GM-CSF-induced cell
proliferation (16-18). Several studies showed that the cytoplasmic
domain of the human GM-CSF receptor
chain (GM-CSFR
) is also
essential for the mitogenic signal (18, 19). However, because of the lack of adequate biologic model cell systems, the role of GM-CSFR
and
subunits in GM-CSF-induced cell differentiation has not been
clearly demonstrated.
The present study defines the role of GM-CSFR in GM-CSF-mediated
differentiation by studying WT19 cells, an FDC-P1-derived cell line
that uniformly differentiates toward the monocytic lineage in response
to murine GM-CSF (mGM-CSF), but grows and does not differentiate in the
presence of murine IL3 (mIL3) (20, 21). We find that when the wild type
human GM-CSFR
and
subunits are both transfected into WT19
cells, these cells respond to the addition of human GM-CSF (hGM-CSF) by
undergoing differentiation. To identify the residues of GM-CSFR
cytoplasmic domain necessary for the induction of cell differentiation,
WT19 cell lines were established which express mutated cytoplasmic
domains of the
subunit along with the wild type
subunit. The
ability of GM-CSF to support cell survival of WT19 correlated with the
tyrosine phosphorylation of Jak2, STAT5, Shc, and extracellular
signal-regulated kinases (ERKs). However, the induction of
differentiation in the cells containing the 18-amino acid deletion of
the C-terminal region occurred without the detectable tyrosine
phosphorylation of these four signaling molecules. Our results suggest
that cell survival and differentiation are controlled by different
signal transduction pathways regulated by varying portions of the
GM-CSFR
.
WT19 is a cell line established from a mouse factor-dependent myeloid cell line, FDC-P1 (a generous gift from Dr. Larry Rohrschneider, Fred Hutchinson Cancer Research Center, Seattle, WA). The cell line was cultured in RPMI 1640 medium (Life Technologies, Inc.) supplemented with 10% fetal bovine serum (FBS), and 10% WEHI-3B conditioned medium containing mIL3.
ReagentsRecombinant hGM-CSF was purchased from Immunex Corp. (Seattle, WA). Recombinant mGM-CSF and mIL3 were obtained from Genzyme (Cambridge, MA).
Site-directed Mutagenesis and Construction of Expression PlasmidsThe human GM-CSFR cDNA was removed from the
plasmid pKH97 (a gift from Dr. A. Miyajima, DNAX Research Institute,
Palo Alto, CA), and the 2.9-kilobase pair fragment was ligated into
pCEP4 (Invitrogen), which contains a hygromycin selection marker giving the plasmid pCEP4-GM-CSFR
.
The 1.3-kilobase pair human GM-CSF receptor chain cDNA was
removed from pCDM8 vector (12) and ligated into the pcDNA3 vector
(Invitrogen) giving the plasmid pCMV-GM-CSFR
.
To construct GM-CSFR mutants, GM-CSFR
cDNA was cloned into
M13 mp19. Site-directed mutagenesis was carried out using a kit from
Amersham, using oligonucleotides: GGGAACAGCCGTCATCACCTAAGGAAC (ter1),
CTTTCCCTTCTCATCAGGTGAATTCCTC (ter3), CCTCATGGTTATCCCTAAGGAACCTT (del1),
TCCCTTCCTCTGGATTCAGTTTGTCT (del2), GTCTTTGATCTGTATCCTAAGGAACC (del3), TGGAACTGGACCGAACAGC (P357G), TCTGTTCGTCAGTGTGAAGATCAGAGC (P358G), CTTTGATCTGACCAACTGGCGG (P360G). The mutated cDNAs
were isolated and ligated into pcDNA3 vector. The structure of the constructs was confirmed by restriction enzyme mapping and DNA sequence
analysis.
pCEP4-GM-CSFR was
introduced into WT19 cells by electroporation at 260 V, 975 microfarads
using a Bio-Rad Gene Pulser, and transfectants isolated using
hygromycin (0.4 mg/ml). A clone termed WT19
1 expressing GM-CSFR
was used for the transfection of pCMV-GM-CSFR
wild-type or mutant
subunits. Resistant clones containing the
subunit were then
isolated using G418 selection (0.4 mg/ml). Resulting clones were
screened by flow cytometry using anti-human GM-CSFR
monoclonal
antibody, and three to five positive clones from each construct were
expanded for further studies.
5,000 cells were incubated in 100 µl of RPMI 1640 containing 10% FBS and various concentrations of hGM-CSF for 14 h at 37 °C in a humidified 5% CO2 atmosphere. 20 µl of freshly prepared combined 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt/phenazine methosulfate (MTS/PMS) solution (Promega, Madison, WI) was added to each sample. After an additional 4 h of incubation at 37 °C, the conversion of MTS into the aqueous soluble formazan was measured at an absorbance of 490 nm.
AntibodiesPolyclonal anti-human GM-CSFR anti-sera were
prepared using a glutathione S-transferase fusion protein
containing amino acids 47-93 of GM-CSFR
(16). The
anti-phosphotyrosine monoclonal antibody (4G10) was purchased from
Upstate Biotechnology Inc. (Lake Placid, NY). The anti-Shc polyclonal
antibody and the anti-STAT1 polyclonal antibody were obtained from
Signal Transduction Laboratories (Lexington, KY). The anti-STAT3 and
STAT5 polyclonal antibodies were purchased from Santa Cruz
Biotechnology, Inc. (Santa Cruz, CA). The STAT5 antibody (sc-835) is
specific for both STAT5a and STAT5b. The anti-human GM-CSFR
monoclonal antibody was a generous gift from Dr. A. F. Lopez (Instiute
of Medical and Veterinary Science, Adelaide, Australia). The anti-Mac1
and Mac3 rat monoclonal antibodies were purchased from PharMingen (San
Diego, CA). The anti-F4/80 rat monoclonal antibody was purified from
the culture supernatant of HB198 rat hybridoma cell line obtained from
ATCC.
Cells were lysed in PLC buffer (50 mM HEPES, pH 7.0, 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1.5 mM MgCl2, 1 mM EGTA, 100 mM NaF, 10 mM NaPPi, 1 mM Na3VO4, 1 mM phenylmethanesulfonyl fluoride, 10 mg/ml aprotinin, 10 mg/ml leupeptin) at 108 cells/ml. The lysates were separated on SDS-polyacrylamide gels, electrotransferred to Immobilon polyvinylidene difluoride membranes (Millipore). The membranes were blocked for 2 h in 2% bovine serum albumin-TBST (20 mM Tris-HCl, pH 7.6, 0.15 M sodium chloride, 0.1% Tween 20), incubated with primary antibodies in TBST for 1 h, washed three times with TBST, and incubated for 1 h with horseradish peroxidase-conjugated anti-mouse or rabbit immunoglobulin (Amersham) diluted 1:10,000 in TBST. After three washes in TBST, the blot was developed with the enhanced chemiluminescence system (Amersham) according to the manufacturer's instructions.
ImmunoprecipitationThe cell lysates were incubated with the indicated antibody for 2 h at 4 °C, followed by protein A-Sepharose beads (Pharmacia Biotech Inc.) for an additional 1 h. The beads were washed three times in PLC lysis buffer, and then suspended in SDS-sample buffer, heated at 95 °C for 5 min. The eluted proteins were applied to an SDS-polyacrylamide gel and proteins detected by Western blotting.
Northern Blot AnalysisTotal cellular RNAs were extracted using Trizol reagent (Life Technologies, Inc.) according to the manufacturer's instructions. 15-µg aliquots of the total RNAs were fractioned on 1% agarose gel containing 20 mM MOPS, 5 mM sodium acetate, 1 mM EDTA, pH 7.0, and 6% (v/v) formaldehyde, and transferred to a nylon membrane. After UV-cross-linking, membranes were soaked in prehybridization solution (6 × SSC, 5 × Denhardt's reagent, 0.5% SDS, 100 mg/ml denatured salmon sperm DNA, and 50% formamide) for 3 h at 42 °C followed by incubation with 32P-labeled probe in hybridization solution (6 × SSC, 0.5% SDS, 100 mg/ml denatured salmon sperm DNA, and 50% formamide) for 14 h at 42 °C. The membranes were washed in 2 × SSC, 0.1% SDS for 10 min twice at room temperature; in 0.1 × SSC, 0.1% SDS for 10 min twice at 50 °C; and then exposed to Kodak XAR films.
hGM-CSF Binding AssayhGM-CSF binding assay was performed
as described previously (22). Briefly, cells were incubated in RPMI
1640, 10% FCS, 20 mM Hepes for 4 h at 37 °C, and
incubated in binding buffer (RPMI 1640 + 2% bovine serum albumin, 20 mM Hepes) containing varying concentrations of
125I-hGM-CSF (NEN Life Science Products) for 30 min at
37 °C. To measure nonspecific binding, 100-fold excess of cold
hGM-CSF was added. The cells were centrifuged through a phthalate oil
layer (3:2, dioctyl phthalate/di-n-butyl phthalate). The
radioactivity of the cell pellet was counted in a counter. The
binding data were subjected to Scatchard analysis.
Before lysis, cells were incubated for the indicated time periods in medium supplemented with hGM-CSF (10 ng/ml) followed by a wash in PBS. The cells were then incubated in cell lysis buffer (10 mM EDTA, 50 mM Tris-HCl, pH 8, 0.5% SDS, 0.5 mg/ml Proteinase K) for 14 h at 50 °C. After an additional 3-h incubation with the addition of 0.25 mg/ml RNase, the genomic DNA was extracted with phenol-chloroform and precipitated with ethanol. DNA fragments were visualized after 1.8% agarose gel electrophoresis by ethidium bromide staining.
WT19 cells (20) growing in mIL-3 demonstrated a myeloblastic morphology including rounded nuclei, fine chromatin, and thin and basophilic cytoplasm. In response to mGM-CSF, the cells demonstrated monocytic characteristics: an indented nucleus with fine-stranded appearance, increased cytoplasm containing a variable number of vacuoles, and larger total cell size. Cells treated with mGM-CSF also became positive for nonspecific esterase and acid phosphatase (Table I). To quantitate the number of cells undergoing differentiation after the addition of mGM-CSF or mIL3, surface marker changes were evaluated by FACS analysis. WT19 cells growing in mIL3 showed weak surface expression of F4/80 and Mac3 (Fig. 1A), both of which are monocytic specific markers (23-25). When the WT19 cells were incubated with mGM-CSF, the expression of both these markers was significantly increased, suggesting that the cells differentiated toward monocytic lineage (Fig. 1A). As demonstrated by FACS analysis, cell size and granularity also increased as evidenced by an increase in both forward scattergram (FSC) and side scattergram (SSC) (Fig. 1A). These characteristics were stable for at least 14 days (Fig. 1A). Differentiated cells continued to divide, and the cell number increased. Increases in both F4/80 expression and cell granularity were evident within 1-2 days after the addition of mGM-CSF (Fig. 1B). Washing out the mGM-CSF and replacing it with mIL3 caused the F4/80 expression to decrease to background levels within 3 days (Fig. 1C), suggesting that mGM-CSF-induced monocyte/macrophage differentiation of WT19 cells is a reversible phenomenon. In addition, mGM-CSF induced cell differentiation of WT19 in the presence of mIL3, suggesting mGM-CSF-mediated differentiation signal is dominant over mIL3 (data not shown).
|
hGM-CSF-induced Differentiation of WT19 Cells Transfected with Human GM-CSF Receptors
To examine the ability of hGM-CSF to
mediate the differentiation of WT19 cells, cells were transfected with
an expression plasmid encoding human GM-CSFR subunit containing a
hygromycin resistance selectable marker. A clone expressing high levels
of GM-CSFR
, WT19
clone 1, was then transfected with the
G418-selectable expression plasmid encoding GM-CSFR
wild type
subunit, and the cells were further selected in G418. Treatment of
these doubly transfected cells with hGM-CSF induced differentiation of
WT19 cells as measured by changes in F4/80 and Mac3 surface markers (Fig. 2). This hGM-CSF-induce differentiation was also
found to be reversible upon removal of human hGM-CSF (data not
shown).
Expression of Human GM-CSFR
We have
demonstrated that the cytoplasmic domain of the subunit regulates
growth of factor-dependent hematopoietic cells (16, 17).
Specific residues of the
subunit are highly conserved among growth
factor receptors (Fig. 3B). As hGM-CSF is
capable of inducing the differentiation of WT19 cells, it is possible to evaluate the role of the cytoplasmic domain in GM-CSF-mediated cell
growth, survival, and differentiation. A series of expression plasmids
encoding deletion and substitution mutants of GM-CSFR
(Fig.
3A) were created; ter1 mutant lacks the entire cytoplasmic domain except for the membrane-proximal 5 amino acids; del1 has an
internal deletion of 15 amino acid residues corresponding to the
proline-rich box 1 region, which is well conserved among the type I
cytokine receptor family; del2 has an internal 16-amino acid deletion
adjacent to the box 1 region; del3 has an 8-amino acid deletion within
the box 1 region removing the proline-rich domain (PPVP) (Fig.
3B); ter3 has a deletion of the C-terminal 18 amino acid
residues; and three individual amino acid substitutions have mutations
in the well conserved proline-rich domain: proline 357, 358, or 360 (Fig. 3A).
These GM-CSFR expression plasmids were transfected into WT19
clone 1, and the G418-resistant clones were isolated. The levels of
expression of these mutants were analyzed by staining with a specific
anti-GM-CSFR
monoclonal antibody followed by flow cytometric
analysis with a FACScan (Fig. 4). All these mutants were
confirmed to express
c equally well by immunoblotting using specific
anti-GM-CSFR
antibody (data not shown). Each transfectant was
examined for hGM-CSF binding using 125I-labeled hGM-CSF,
and the results of high affinity binding profiles were shown in Table
II. All mutant transfectants including ter1 clones bind
hGM-CSF with high affinity and Kd values suggesting
that there are equivalent numbers of receptors with similar affinity.
These results suggested that the mutations introduced in the
cytoplasmic domain of GM-CSFR
did not affect the interactions of
GM-CSFR
with the ligand or GM-CSFR
.
|
It is well established that tyrosine phosphorylation of an
array of cytoplasmic proteins is critical for cytokine signal
transduction. Accordingly, we analyzed the spectrum of substrates
tyrosine-phosphorylated by addition of hGM-CSF to WT19 cells expressing
various GM-CSFR mutants.
Several hematopoietic cytokines including GM-CSF induce p52 Shc
tyrosine phosphorylation, which correlates with their ability to
activate Ras (26-30). GM-CSFR transfectants were assayed for hGM-CSF-induced Shc tyrosine phosphorylation by anti-Shc
immunoprecipitation followed by anti-phosphotyrosine immunoblotting. As
shown in Fig. 5A, increased p52 Shc tyrosine
phosphorylation was detected only in wild type and P357G GM-CSFR
transfectants but not in the other mutants. Activated Ras through a
cascade of protein kinases stimulates phosphorylation of ERKs (31).
hGM-CSF induced phosphorylation of both p44ERK1 and p42ERK2 in wild
type and P357G cells, but not in any other transfectants (data not
shown).
GM-CSF addition to cells activates JAK2, which leads to the tyrosine
phosphorylation and activation of STAT5 (32-34). Tyrosine phosphorylation of JAK2 and STAT5 was induced by GM-CSF in wild type
and P357G mutants (Fig. 5, B and C). On shorter
exposure, tyrosine phosphorylation of two STAT5 isoforms (STAT5 a and
b) was observed. A slight decrease in JAK2 and STAT5 phosphorylation seen in Fig. 5 in the P357G transfectants was not constantly
reproducible. In the other GM-CSFR mutant cell lines, GM-CSF did not
induce detectable JAK2 or STAT5 tyrosine phosphorylation. Activation of
STAT5 in wild-type and P357G GM-CSFR
transfectants was also detected
by gel shift assay using
-interferon-activated site of the IRF-1
promoter and anti-STAT5 antibody (data not shown). These data indicated
that the same regions of GM-CSFR
that are essential for
GM-CSF-induced Shc-ERK phosphorylation are also essential for the
induction of JAK2 and STAT5 tyrosine phosphorylation.
GM-CSF
has been shown to induce rapid expression of a number of
protooncogenes, including c-fos, c-jun, and
c-myc (29, 35). In wild-type GM-CSFR and the P357G
transfectants, expression of c-fos, c-jun, and
c-myc mRNAs was rapidly induced upon hGM-CSF stimulation
(Fig. 6). In contrast, the expression of
c-fos and c-jun mRNA was not induced in the
other
subunit mutants except ter3. In ter3 mutant receptor cell
lines, hGM-CSF was capable of inducing c-jun but not
c-fos mRNA. In contrast, induction of c-myc
mRNA expression was repeatedly observed in both the wild type and
all of the mutant clones, indicating that hGM-CSF is able to induce
c-myc mRNA in the absence of GM-CSFR
cytoplasmic domain and that all of the mutant receptors are capable of
signaling.
The Cytoplasmic Domain of GM-CSFR
We next examined
hGM-CSF-induced cell proliferation of WT19 transfectants expressing chain mutants. As shown in Fig. 7, both the wild type
GM-CSFR
transfectant and the P357G mutant proliferated upon addition
of hGM-CSF to the medium. ter1, del1, del2, del3, P358G, and P360G did
not show any proliferative response to hGM-CSF, suggesting that some
residues of the
chain cytoplasmic domain is indispensable for
hGM-CSF-mediated growth signal transduction. These studies confirm our
earlier findings about the role of GM-CSFR
cytoplasmic domain in
promoting growth of BaF/3 cells (16).
Treatment of ter3 clones that lack the C-terminal 18 amino acid
residues of chain with hGM-CSF did not lead to an increase in the
cell numbers. Instead, the cells died, but more slowly than ter1 clones
(Fig. 7). In MTS cell proliferation assays, ter3 clones clearly showed
hGM-CSF-mediated cell proliferation signal, although it was weaker than
wild type or P357G clones (Fig. 8). Cell death in
factor-dependent cells is known to occur through apoptotic
mechanisms. To examine if apoptosis occurred in ter3 cell number in the
presence of hGM-CSF, genomic DNA was isolated from
chain
transfectants incubated with hGM-CSF and DNA fragmentation was analyzed
by agarose gel electrophoresis (Fig. 9). ter3 showed detectable DNA fragmentation characteristic of apoptosis by 9 h
and showed pronounced apoptosis by 24 h after withdrawal of mIL3
and the addition of hGM-CSF. These results suggest that ter3 mutants
that are able to transduce a cell proliferation signal by GM-CSF have
severely impaired anti-apoptotic signaling.
hGM-CSF-induced Monocytic Differentiation of WT19 Cells Expressing hGM-CSF Receptors
Next we analyzed hGM-CSF-induced
differentiation of WT19 cells expressing human GM-CSFR. All the
subunit transfectants examined retained the ability to differentiate
when mGM-CSF was added to the medium (data not shown). Because several
transfectants died within 24 h after the withdrawal of mIL3, cell
lines were treated with hGM-CSF in the presence of mIL3. After
incubation with hGM-CSF for 3 days, the cells were examined for
monocytic differentiation by morphology, F4/80, and Mac3 surface
expression. As shown in Fig. 10, after 3 days of
incubation with GM-CSF, cells transfected with either the wild type
GM-CSFR
or P357G showed characteristic morphology of monocytic
lineage: larger cell sizes, fine-stranded nuclear appearance, and a
variable number of cytoplasmic vacuoles. Both the morphologic and cell
surfaces changes induced by hGM-CSF were identical to those induced by
mGM-CSF treatment. They also showed increased surface expression of
F4/80 and Mac3 (Fig. 2). None of the clones containing ter1, del1,
del2, del3, P358G, or P360G was able to differentiate when incubated
with hGM-CSF (Figs. 2 and 10). In contrast, all ter3 clones that were derived (five independent clones) differentiated as well as wild type
clones in response to hGM-CSF (Figs. 2 and 10).
The GM-CSF receptor signals by ligand-mediated heterodimerization
of GM-CSFR and GM-CSFR
. Although the cytoplasmic domain of the
GM-CSFR
is only 54 amino acids, this short region of the receptor is
necessary for GM-CSF-induced cell proliferation (16-18). In the
present study, we have compared the role of GM-CSFR
cytoplasmic domain in GM-CSF-mediated cell proliferation, survival, and
differentiation. This analysis was made possible by our use of the WT19
cell line, which grows but does not differentiate in mIL-3. In
comparison, mGM-CSF induces differentiation, but does not stop the
growth of these cells. Our studies suggest that the mGM-CSF-mediated differentiation process is reversible upon removal of mGM-CSF. The WT19
cells are equally capable of responding to both murine and human GM-CSF
when transfected with the wild type human GM-CSF receptor subunits.
By using a mutant GM-CSFR lacking most of the cytoplasmic domain, we
have shown that the cytoplasmic domain of GM-CSFR
is essential for
both hGM-CSF-dependent cell differentiation and proliferation. This mutant was still able to interact with the
-chain to form a high affinity receptor complex (Table II),
suggesting that the cytoplasmic domain of GM-CSFR
is not
necessary for receptor dimerization. The GM-CSFR
cytoplasmic domain was necessary for the phosphorylation of signaling
molecules, JAK2, STAT5, Shc, and ERKs and the induction of
c-fos and c-jun mRNA expression. Weak
activation of STAT1 and STAT3 by GM-CSF have recently been reported in
polymorphonuclear leukocytes (36). However, we could not detect the
activation of these STATs in any of the transfectants in response to
GM-CSF (data not shown). Deletion of the intracytoplasmic domain did
not abolish c-myc mRNA induction after GM-CSF
stimulation. Similar results were observed by us using BaF/3 cells
(16).
Using other cell systems and varying approaches other laboratories have
suggested that the internal portion of the GM-CSFR may not have a
major role in regulating receptor function. For example, using a
chimera of the extracellular domain of the erythropoietin receptor and
the intracellular domain of the murine IL-3 receptor
chain (AIC2A),
the addition of erythropoietin to the receptor was able to stimulate
cell growth (37). Another study demonstrated that a chimera comprising
the extracellular region of GM-CSFR
and the intracellular domain of
the h
c can also transduce signals (38). These data suggest that
dimerization of the
-chain is important. However, they do not
necessarily exclude the possibility that the GM-CSFR
is an important
dimerization partner, and there are no physiologic data demonstrating
that two
chains dimerize to initiate signaling in normal cells.
Both the
and
subunits have proline-rich regions close to the
plasma membrane, and both of these regions are important for receptor
function. In addition, we have shown that deletion of the
internal
segment blocks both growth and differentiation.
By using deletion mutants of GM-CSF, we demonstrate that the
membrane proximal proline-rich region and the adjacent 15 amino acids
of the
subunit are indispensable for both cell proliferation and
differentiation (Figs. 2, 7, and 10). The proline-rich region of
GM-CSFR
contains a Pro-X-Pro sequence that exists in the
membrane-proximal box1 region of many other members of cytokine
receptors (Fig. 3B). Mutation of this domain in the IL6
receptor gp130 protein (39) and in the granulocyte colony-stimulating
factor receptor (40) eliminated receptor activity. We have here shown
that similar mutations (P358G, P360G) also result in a receptor that is
unable to mediate proliferation, differentiation, or other signaling events. Proline 357 could be part of a
Pro-X-X-Pro motif, such as has also been found in
cytokine receptors and SH3-binding proteins (41). This proline appears
to be dispensable, however, since the P357G mutant was able to fully
support proliferation and differentiation.
The region downstream from the proline-rich domain was also
indispensable for hGM-CSF-dependent transduction of cell
growth, survival, and differentiation signals. These 15 amino acid
residues, which include aspartic acid 368, are conserved in IL5
receptor , prolactin receptor, growth hormone receptor, and IL2
receptor
-chain. Our studies show that tyrosine phosphorylation of
JAK2 and STAT5 is inhibited by the deletion of this region,
demonstrating that the proline-rich domain alone is not sufficient for
the GM-CSF-induced activation of the JAK2 signal transduction pathway.
Similar results have been obtained in other systems. Deletion of 6 amino acids of the region downstream of proline-rich domain of IL5
receptor
, including the conserved aspartic acid, abolished
IL5-induced JAK2 activation (42), while mutation of the region
immediately downstream of box1 region in the erythropoietin (43) and
IL6 receptors (39) blocked the activation of JAK2.
Interestingly, our data about proline mutations in the box1 region were
different from the recently published study on the IL-5 receptor
-chain. In the previous report, the existence of any one of the
three proline residues was adequate for IL-5-mediated cell
proliferation signal (42). The difference between these findings could
be secondary to differences in sequence in the
-chains (PPVPQI,
GM-CSF receptor; PPIPAP, IL-5 receptor), or, possibly the difference in
results could be due to the divergence in amino acid residues adjacent
to the proline-rich domain.
The C-terminal deletion of GM-CSFR only partially inhibited the
hGM-CSF-induced cell proliferation (Figs. 7 and 8), but the cells died
within several days due to the extensive apoptosis in the presence of
hGM-CSF (Fig. 9). In the transfectants of this C-terminal deletion
mutant (ter3), hGM-CSF-induced protein-tyrosine phosphorylation was
severely impaired (Fig. 5).
Increased tyrosine phosphorylation of a 52-kDa protein, Shc, in response to GM-CSF stimulation has been reported (27-30, 44). Shc, when tyrosine-phosphorylated, binds to SH2 domains of Grb2, which leads to the recruitment of Sos, a guanine nucleotide exchange factor for Ras, to the plasma membrane (45). Tyrosine phosphorylation of Shc is thought to play an important role in GM-CSF-mediated activation of Ras through this mechanism (27). In ter3 transfectants, tyrosine phosphorylation of Shc was not detectable in response to hGM-CSF stimulation (Fig. 5A). hGM-CSF-induced activation of Ras-ERK pathway appeared to be impaired in these transfectants, since hGM-CSF-mediated ERK phosphorylation and c-fos induction, which are downstream events regulated by Ras activation (46, 47), could not be detected (Fig. 6). Our findings are compatible with a previous report that Ras activation is necessary for anti-apoptotic effect by GM-CSF, but not essential for GM-CSF-mediated DNA synthesis (27).
Our results demonstrate that ter3 transfectants differentiated as well
as wild type transfectants in response to hGM-CSF. The ter3 clones are
capable of stimulating increases in c-myc and to a lesser
extent c-jun, but do not cause the phosphorylation of Jak2,
STAT, and Shc, suggesting that activation of these proteins is not
necessary for differentiation. The findings that ter3 cells die of
apoptosis while they were capable of differentiation in hGM-CSF
suggests that the pathways controlling cell survival and differentiation can be separated and are controlled by different portions of the GM-CSFR. hGM-CSF-mediated tyrosine phosphorylation of JAK2 and STAT5 could not be detected in ter3 clones (Fig. 5, B and C), suggesting that these pathways may be
important for the inhibition of apoptosis. We have recently
demonstrated that the expression of a chimeric protein of CD16 and Jak2
is capable of preventing cell death, implying that part of the function
of the Jak/STAT pathway could be to inhibit apoptosis (51).
In contrast to other signal transduction pathways, both c-myc and c-jun were induced in ter3 transfectants by hGM-CSF to similar levels to that seen in wild type transfectants, suggesting that c-myc and c-jun can be induced without the activation of either ERKs or JAK2. In a recent report, it was shown that the transient expression of the dominant negative form of JAK2 inhibited hGM-CSF-induced transcription of a reporter plasmid containing the c-myc promoter, suggesting that JAK2 is essential for c-myc mRNA induction by hGM-CSF (48). It is possible that inhibition of c-myc promoter was caused by a nonspecific effect of the overexpression of dominant-negative JAK2, although further experiments will be needed to clarify this possibility. c-jun overexpression induces monocytic differentiation of the WEHI-3B (49) and U937 cells (50), while c-fos overexpression did not have similar biologic effects (49). However, it is unlikely that c-jun alone is responsible for hGM-CSF-mediated cell differentiation, as c-jun mRNA expression was equivalently induced by mIL3, which has no effect on the differentiated phenotype of WT19 cells (data not shown).
In summary, specific regions of the intracytoplasmic domain of the subunit play an essential role in hGM-CSF-mediated cell proliferation,
survival, and differentiation, while the signal transduction pathway
which controls c-myc activation is independent of this
subunit. Our results demonstrate that differentiation may occur in the
absence of Shc, ERK, or JAK2 activation, suggesting that there are
specific novel signal transduction pathways, yet to be determined,
which control this process.