 |
INTRODUCTION |
Proinflammatory cytokines are the principal intercellular
mediators of the tissue reaction to trauma and infection (1). Members
of the interleukin 6 hematopoietic cytokine family, that include
IL-6,1 leukemia inhibitor
factor (LIF), and oncostatin M (OSM), play a particularly prominent
role in orchestrating initiation and progression of the inflammatory
response and controlling homostatic processes. These roles were
identified by the effects of transgenic cytokine expression, the
knockout of cytokine genes in mice (2-5), or by in vivo
treatments with pharmacological doses of cytokines or activity
neutralizing antibodies (6-11). OSM is produced by activated monocytes
and lymphocytes (12, 13) (e.g. at sites of inflammation) and
acts locally on stromal cells. Stromal cells in turn respond
prominently by enhanced production of IL-6 and LIF (14). IL-6 and LIF
enter into circulation and participate in the recruitment of the
systemic inflammatory response that includes the acute phase reaction
of the liver (1, 15).
Each of the IL-6 cytokines is recognized by a specific ligand binding
subunit, i.e. IL-6 by IL-6 receptor 
(IL-6R), LIF by
LIFR (note, this subunit has also been called LIFR
; Ref. 16), and
OSM (human) by gp130. These cytokine receptor complexes cooperate with
a second, signal-transducing subunit to form a signaling-competent
receptor unit (abbreviated as follows: IL-6R consisting of
IL-6R-gp130; LIFR consisting of LIFR-gp130; and OSMR (also
defined as type II OSMR) consisting of gp130-OSMR (17, 18)). In
human cells, hOSM also engages a type I OSMR consisting of
gp130-LIFR that is equivalent to LIFR (19). Ligand-induced receptor
subunit interaction results in trans- and autophosphorylation of Janus
protein tyrosine kinases (JAKs) that are associated with the
cytoplasmic domains of the receptor subunits. Tyrosine phosphorylation of the receptor subunits, including gp130, creates docking sites for
signal-communicating STAT3 and linker proteins which, upon their
phosphorylation, propagate the signal to other pathways (MAPK and
phosphatidylinositol 3-kinase) (reviewed in Refs. 17, 18, 20,
and 21). The combined effects of the receptor-dependent processes and the complement of existing transcription factors determine the mode of gene regulation that, in the case of hepatocytes treated with IL-6 cytokines, results in enhanced transcription of genes
encoding type 2 APP, including: fibrinogen, haptoglobin (HP), and
thiostatin (TST), and 2-macroglobulin
(2-MG) (15, 22).
The patterns of APP gene expression in rodent and human hepatic cells
suggest that quantitative differences in regulatory activity exist
among IL-6 family members, where OSM shows the strongest activity
(23-26). These differences in cell responses have been interpreted to
be caused by differences in receptor levels, ligand interaction with
subunits, or the ability of heteromeric or homomeric complexes of the
three signaling subunits (gp130, LIFR, and OSMR) to engage
subunit-specific signal-communicating pathways (16, 17, 27-30). Based
on structural and functional analyses of receptor subunits in
transfected cells (27, 29), we hypothesize that the different receptor
subunits contribute specific signaling reactions that define the cell
responses to the different IL-6 cytokines. Hence, we predict that the
OSMR subunits are particularly effective in signal transduction,
accounting for the high level of OSM effects.
It has not previously been possible to precisely compare the regulatory
action of the key IL-6 cytokine receptors, IL6R, OSMR, and LIFR, and to
specifically define the cell response as a function of OSMR
expression since appropriate cell lines were not available. However,
several properties of the recently cloned mouse OSMR (31, 32) now
make it possible to define OSMR-specific action. The amino acid
sequence of mOSMR shows 55% identity to the hOSMR sequence.
Furthermore, mOSMR appears to form a type II OSMR complex only with
rodent OSM (31-33). This differs from the type I and II OSMR engaged
by hOSM in human cells. To assess the function of
OSMR-dependent receptor systems and distinguish between
the activity of type I and II OSMR, we reconstituted OSMR in a clonal line of rat H-35 hepatoma (22). These cells have active IL-6R, IL-11R,
and LIFR systems, but are deficient in the expression of OSMR. The
present studies demonstrate that 1) OSMR subunit assists in
recruitment of a broader spectrum of signaling pathways than achieved
by gp130 activation alone, 2) the particularly prominent activation of
STAT5 and ERK pathways by OSMR accounts in part for the OSM specific
gene regulatory pattern, and 3) the OSM response is influenced by
coactivation of the LIFR.
| |
MATERIALS AND METHODS |
Tissue Culture Cells--
Human HepG2 (clone 86-6; Ref. 34), rat
H-35 (clone T-7-18; Ref. 22), and H-35 cells stably expressing
G-CSFR-gp130 (35), were cultured in Dulbecco's modified Eagle's
medium containing 10% fetal calf serum and antibiotics. Cells used for
analyzing signaling were maintained for 18-24 h in serum-free
Dulbecco's modified Eagle's medium. Treatments were carried out in
serum-free medium containing cytokines at concentrations ranging from
0.01 to 1000 ng/ml (standard treatments used 100 ng/ml) of recombinant human IL-6 and LIF (Genetics Institute), human granulocyte-colony stimulating factor (G-CSF), hOSM (Immunex Corp), mOSM (prepared in
COS-1 cells as described in Ref. 32), or 1 µM
dexamethasone (Sigma). MEK-1 activity was inhibited by 5-25
µM U0126 (Promega, Madison, WI).
Plasmid Constructs--
The cDNA to mOSMR in pME18S (32)
and hOSMR in pDC409 (33) were modified by the addition of FLAG epitope
(DYKDDDK) to the C terminus. The OSMR-FLAG constructs were
transferred as NotI fragments into the retroviral vector
MINV (36) and the vector-derived viruses were used to transduce H-35
cells (35, 37). Two days after viral infection, the transduced cells
were selected in medium containing 2 mg of G418/ml. From the primary cultures of transduced cells (which normally contain 100 to 10000 separate clones), 24 to 48 clones were picked, expanded, and subcloned. Subclonal lines were characterized for relative expression of OSMR
by FLAG immunoblotting, ligand-induced receptor tyrosine phosphorylation, and induction of APPs. The remaining cells on the
primary culture plates were maintained and classified as a pool of
stably transduced H-35 cells (termed hOSMR-H-35 cell or
mOSMR-H-35 cells). The following expression vectors have been described previously: human LIFR with deleted cytoplasmic domain (LIFR
cyto) in pDC302 (27), rat STAT5B and STAT5B4OC (lacking 41 residue of C-terminal transactivation domain) in pSV-Sport 1 (38),
and the reporter-CAT gene constructs containing the indicated APP
promoters, p2-MG (2700)-CAT (39), pHP (4200)-CAT (40),
and pTST(1516)-CAT (41) and the STAT3- and STAT5-sensitive construct
p(8XHRRE)-CAT (42).
Transient Transfection--
HepG2 cells were transfected by the
calcium-phosphate method (43) with a total of 20 µg of DNA/ml,
including 0.25 µg of pEGFP(N1) (Upstate Biotechnology Inc.) as
internal transfection marker, 15 µg of CAT-reporter construct, 0.5-2
µg/ml expression vector for receptor, and the balance with empty
vector. H-35 cell lines were transfected with FuGene6 (Roche Molecular
Biochemicals) according to the manufacturer's recommendation using a
ratio of 6 µl of FuGene6 to 4 µg of DNA. Cell cultures were
subdivided immediately after transfection, and 24 h later these
cultures were treated for 24 h with serum-free medium containing
the inducing factors. To inhibit gp130 function in HepG2 cells during
cytokine treatment, the cells were maintained in the presence of
activity neutralyzing monoclonal antibodies against human gp130
("144," Ref. 44). Prior to cell extraction, expression of green
fluorescent protein (GFP) was visualized under a Nikon inverted
fluorescence microscope. A digitized image of GFP positive culture
at × 40 magnification was taken by a SPOT camera (Diagnostic
Instruments, Inc.) and GFP fluorescence signal above background
quantitated in the NIH Image program version 1.62. The integrated net
pixel values relative to a constant view area were used as a measure for transfection efficiency. CAT activity in serially diluted cell
extract was determined, normalized to the GFP signal for each culture,
and expressed in relative CAT activity. To obtain enriched cultures of
H-35 cells transfected with expression vectors for STAT5 isoforms,
24 h after transfection, GFP negative (served as controls) and GFP
positive cells were selected by sterile fluorescence-activated cell
sorting as described (45). This approach yielded 0.5 to 1 × 106 GFP positive cells representing a 50% recovery of the
original 5% transfected cells in the culture. The cells were plated
into 24-well culture plates (2.5 × 105 cells per
well). Following 24 h recovery, the cells were treated for an
additional 24 h with cytokines in the presence of dexamethasone. Conditioned medium was used to measure the amounts of plasma proteins. Cell extracts were analyzed by Western blotting for the expression of
STAT5 proteins.
Immunoprecipitation and Western Blotting--
Cell monolayers
were washed twice with ice-cold phosphate-buffered saline and lysed in
RIPA buffer (0.5 ml/10-cm2 cell monolayers). Lysate
containing ~2 mg of protein were diluted to 1 ml of RIPA buffer,
precleared, and incubated for 2 h with specific antibodies to
gp130, LIFR, SHC (Santa Cruz Biotechnology), SHP-2 (Up-State
Biotechnology Inc.), FLAG (M2 antibody; Eastman Kodak Co.),
phosphotyrosine (Py20; Transduction Laboratories, combined with
4G10; Upstate Biotechnology, Inc.), extracellular domain of hOSMR
(Immunex; Ref. 46) for 2 h at 4 °C. Immunocomplexes were
recovered by incubation with 50 µl (50% slurry in RIPA buffer) of
protein G-conjugated Sepharose (Amersham Pharmacia Biotech) for 1 h at 4 °C with agitation. Beads were washed three times with RIPA
buffer and boiled in SDS sample loading buffer. The immunoprecipitates
or aliquots of whole cell lysates were separated on a 6-10%
SDS-polyacrylamide gel and proteins transferred to protean membranes
(Schleicher & Schuell). Wherever possible, separation of replicate
sample aliquots was preferred over sequential probing of the same
membranes in order to circumvent antigen loss. Where separate antigens
of different sizes were analyzed, the membrane was cut into horizontal
sections containing the specific size categories of antigens. The
membranes were reacted with primary antibodies and secondary antibodies
(horseradish peroxidase-conjugated anti-mouse or anti-rabbit antibodies
(Cappel, West Chester, PA) in TBST (20 mM Tris-HCl, pH
7.5, 150 mM NaCl, 0.1% Tween 20) containing 5% milk or
3% albumin. Results were visualized by enhanced chemiluminescence reaction (ECL) according to the manufacturer's directions (Amersham Pharmacia Biotech). To perform additional immunoreactions, membranes were treated for 30 min with 62.5 mM Tris-HCl, pH 6.8, 2%
SDS, and 100 mM 2-mercaptoethanol.
Northern Blot Hybridization--
Total cellular RNA was
extracted by the Trizol method (Life Technology, Grand Island, NY).
Aliquots of 5-20 µg were separated on 1.5%
formaldehyde-agarose gel, transferred onto nylon membrane (Schleicher & Schuell), and reacted with 32P-labeled cDNA probes for
rat 2-MG, HP, or TST. Ethidium bromide staining pattern
of separated RNA was used as a marker for sample loading.
Plasma Protein Analysis--
Synthesis and secretion of APP into
the culture medium of cytokine-treated H-35 cells were quantitated by
immunoelectrophoresis, using equal aliquots of cell-free conditioned
medium (47). The area under the precipitation peaks (proportional to
amount of antigen) was integrated, using the NIH image program and
expressed in arbitrary immunoelectrophoretic units normalized to the
amount of cell protein.
Thymidine Incorporation--
H-35 cells were subcultured in
96-well plates (2.5 × 104 cells/well). After 24 h, the cells were treated first for 8 h with serum-free medium,
then switched to fresh serum-free medium with or without cytokines (8 wells per treatment). Following 16 h incubation, 0.4 µCi
of [3H]thymidine was added to each well and incubation
continued for an additional 8 h. The cells were washed, released
by trypsin, and collected onto filter paper with a cell harvester.
Incorporation of 3H was measured by scintillation counting
(Wallac, Gaithersburg, MD). Statistical evaluation of the data was
performed by Student's t test.
| |
RESULTS |
Mouse mOSMR Signals Through Human gp130--
To assess whether
the cloned mOSMR formed a signaling-competent complex with gp130 and
whether this complex is functional, the expression vector for mOSMR,
together with the OSM-responsive HRRE-CAT reporter gene construct, was
transiently transfected into HepG2 cells (Fig.
1). HepG2 cells have an endogenous OSMR system, comprised of type I and II OSMR, that effectively mediated a
prominent CAT gene induction in response to human, but not mOSM (Fig.
1, left panel). Expression of transfected mOSMR by itself did not detectably modify the activity of the other IL-6 cytokine receptors, but introduced an mOSM-specific gene induction comparable to
the endogenous hOSMR (Fig. 1, second panel). That mOSM did not promote interaction with LIFR was confirmed by overexpression of
a membrane form of hLIFR that lacks the cytoplasmic domain (LIFRcyto). This form acted as a competitive inhibitor of both full-length LIFR and hOSMR of HepG2 cells by generating an
hOSM-dependent, but signaling-incompetent complexes with
gp130. LIFRcyto eliminated the cell response to hOSM but did not
affect mOSM action (Fig. 1, third panel). However, a
functional, ligand-dependent cooperation of transfected
mOSMR with the endogenous gp130 was identified by its inhibition
with anti-gp130 antibodies (Fig. 1, right panel). These
results established functionality of the mOSMR construct and
indicated that human gp130 was capable of cooperating similarly with
mouse and human OSM and OSMR in forming a signal-transducing complex.
View larger version (93K):
[in this window]
[in a new window]
|
Fig. 1.
Signaling by mOSMR
in combination with human gp130. HepG2 cells were
transfected with p8XHRRE-CAT (15 µg/ml) alone or together with
expression vector for mOSMR (0.5 µg/ml) with or without addition
of the expression vector for human LIFRcyto (2 µg/ml).
Subcultures were treated with the cytokines (100 ng/ml) and anti-gp130
antibodies (10 µg/ml) for 24 h as indicated. CAT activities were
determined (lower panel) and expressed relative to the
untreated control (upper panel).
|
|
OSMR Introduces a Type II OSMR Activity in H-35 Cells--
To
characterize the regulatory activity of OSMR on endogenous genes as
function of OSMR expression, we employed H-35 cells. These cells
showed an approximately equal expression of gp130 and LIFR and
responded to IL-6 or LIF by a similar level of tyrosine phosphorylation
of gp130 (Fig. 2). As described below,
the cytokine treatment also resulted in a comparable activation of
STAT3 (Fig. 4A) and induction of type 2 APPs (Fig.
7B). In H-35 cells, expression of OSMR mRNA was
undetectable by hybridization analysis (data not shown) and mOSM
treatment did not produce an appreciable regulatory effect. The
LIF-like response to hOSM treatment had been attributed to the
hOSM-mediated engagement of LIFR into a type I OSMR complex, an activity not exerted by mOSM (Fig. 2A). An
unexpected finding, and convenient tool for subsequent functional
analysis of receptor subunit-associated proteins, was that under the
conditions selected for immunoprecipitation, the hOSM-directed
interaction of LIFR and gp130 was less stable than that directed by
LIF. This is demonstrated by the absence of gp130 from the
LIFR-containing complex immunoprecipitated from hOSM-treated cells
(Fig. 2A). Identical immunoprecipitation of the LIF-induced
complex yielded equal immunoblot signals for the
tyrosine-phosphorylated LIFR and gp130 subunits.
View larger version (65K):
[in this window]
[in a new window]
|
Fig. 2.
Expression and activities of IL-6 cytokine
receptors in H-35 cells. A, confluent monolayers of
parental H-35 cells in 6-cm diameter dishes were treated for 15 min
with the indicated cytokines (100 ng/ml). One-half of the cell extract
was reacted with anti-gp130 and the other half with anti-LIFR. The
immunoprecipitates (I.P.) were analyzed by Western blotting
(W.B.) for proteins recognized by antiphosphotyrosine
(PY), anti-LIFR, and anti-gp130. Two protein forms are
detected for gp130 and LIFR. The larger size form in each pattern
comigrates with the fully processed, cell surface-exposed gp130 or
LIFR, which are also the only form subjected to tyrosine
phosphorylation. The smaller proteins represent not yet fully processed
intracellular cytokine receptor forms. B, subcultures of
parental H-35 cells and pool cultures of H-35 cells transduced with
hOSMR or mOSMR were treated for 15 min with the cytokines and
then lysed. Cell lysates were reacted with anti-FLAG or anti-LIFR
and the immunoprecipitated proteins were analyzed on a single gel by
Western blotting, first for tyrosine phosphorylated proteins
(upper panel), then for anti-FLAG reactive proteins
(lower panel). Immunoprecipitation of LIF-activated LIFR was
used to demonstrate comparable LIF response of the cultures,
reproducibility of immunoprecipitation technique, and consistent
recovery of co-activated LIFR and gp130, and to gain a marker for
the electrophoretic mobilities of these subunits. Note, that a single
size form of FLAG-tagged OSMR proteins is observed.
|
|
Based on the data from HepG2 cell transfection (Fig. 1), it is expected
that expression of human or mouse OSMR would introduce in H-35
cells, an OSM-directed gene regulatory activity that is similar to, or
exceeds that of IL-6. Additionally, the expression of hOSMR should
introduce type II OSMR signaling in parallel to hOSM-coordinated
endogenous type I OSMR. Stable expression of human and mouse OSMR
was established in H-35 cells by retroviral transduction. Initial
experiments verified that native OSMR, and OSMR carrying a
C-terminal FLAG epitope, were equivalently active in signal
communication. Hence, to facilitate structure/functional analyses of
the OSMR-containing complexes, most studies were carried out with
cells expressing the FLAG-modified receptor proteins.
Several independently generated pools of H-35 cells stably transduced
with hOSMR or mOSMR consistently showed a single form of
anti-FLAG reactive OSMR protein with the predicted molecular mass of 160 to 180 kDa (Fig. 2B, Refs. 31-33 and 48).
Northern blot (not shown) and immunoblot analyses indicated that in
each receptor-transduced culture, the relative expression of mRNA
and proteins was closely correlated. The comparison of hybridization signals on Northern blots also indicated that the average level of
mOSMR mRNA was in the range of that of NIH 3T3 cells, and the
level of hOSMR mRNA and protein was similar to that of HepG2 cells.
Due to technical limitations of the binding assay for OSM (29, 33), we
could not obtain a convincing quantification of OSMR-dependent, high affinity binding sites for human or
mouse OSM introduced by transduction in H-35 cells. Therefore, the cell surface-assessable and functionally relevant OSMR proteins were determined by ligand-dependent tyrosine phosphorylation of
the subunit proteins (33, 48). LIF-induced modification of endogenous LIFR and gp130 served as internal references (Fig. 2B).
Treatment of the cells for 15 min with OSM initiated tyrosine
phosphorylation of the receptor subunits in the range of that observed
for endogenous LIFR (Fig. 2B) and gp130 (Fig.
3A, right panel). The cells
showed a strictly species-restricted recruitment of OSMR subunit by OSM (Fig. 2B). The same analysis also confirmed that hOSM,
but not mOSM, engaged endogenous LIFR in forming a signaling
competent type I OSMR (Fig. 3A, right panel). Surprisingly,
when using conditions for co-immunoprecipitation that recovered the
LIF-stabilized complex of LIFR and gp130, no significant
co-immunoprecipitation of gp130 with hOSMR (Fig. 3A, left
panel), and only a very minor amount of gp130 with mOSMR (Fig.
3B), was observed. Similarly, an antibody that recognized an
epitope in the extracellular, hematopoietic domain of hOSMR (33),
rather than the C-terminal FLAG epitope, failed to co-immunoprecipitate
gp130 as part of the ligand-activated complex (Fig. 3A, left
panel). These results suggested that human and mouse OSM mediated
a weaker interaction of receptor subunits than LIF, and that our
immunoprecipitation conditions were sufficiently stringent to
distinguish between LIF and OSM coordinated receptor complexes.
View larger version (33K):
[in this window]
[in a new window]
|
Fig. 3.
Detection of OSMRI and OSMRII activity in
transduced cells. Pool cultures of H-35 cells stably transduced
with hOSMR (A, left panel) and mOSMR (A,
right panel, and B) were treated for 15 min as
indicated at the bottom, and then lysed. Equal amounts of
cell extract were subjected to immunoprecipitation (I.P.)
using antibodies to the receptor subunits as indicated above
or left of the blots. Recovered proteins were analyzed by
immunoblotting as indicated. Note that treatment with hOSM, but not
mOSM, yields an activation of LIFR. However, this subunit activated
by OSM does not appreciably co-immunoprecipitate gp130, despite the
high level of tyrosine phosphorylated of the latter. In B,
the electrophoretic separation of the proteins of the >100 kDa range
was extended. The positions of the receptor subunit proteins and
molecular size markers are marked.
|
|
The Immediate Signaling Reactions by OSMR Differ from IL-6R and
LIFR--
The gain of specific signaling activity as a function of
transduced OSMR was taken as evidence for the functional
contribution by this receptor subunit (Fig.
4). The comparison of
OSMR-dependent signaling, mediated by the resident IL-6R
and LIFR, allowed for assessment of the signaling specificity. In all
transduced H-35 cell cultures, the relative activity of the IL-6 and
LIF signals were essentially identical to those established in parental
cells (Fig. 4A). Similarly, the hOSM effects in
mOSMR-transduced cells (Fig. 4B) were indistinguishable
from hOSM effects in parental cells.
View larger version (72K):
[in this window]
[in a new window]
|
Fig. 4.
Signal transduction by IL-6 cytokine
receptors. Subcultures of parental H-35 (A),
mOSMR-H-35 cells (B and D, left three
panels) and hOSMR-H-35 cells (C and D,
right panel) were either treated with 100 ng/ml cytokines
for 0 to 120 min (A and C) or with increasing
concentrations of the cytokines for 15 min (D). Whole cell
extracts were Western blotted for the proteins indicated to the
right of each panel. Representative extracts in A
and D were used to demonstrate that cellular amounts of
STATs and ERK remained constant during the treatments. The film
exposures to the ECL reaction in B and C occurred
simultaneously for all corresponding panels, and the films showing
optimal quantitative signal differences among the samples (avoiding
overexposure of most reactive bands) were used to generate the
composite picture. In D, the blots for
phosphorylated ERK were exposed longer to ECL reaction for
demonstrating ERK activation.
|
|
As expected from the comparable level of receptor protein
phosphorylation (Fig. 3), IL-6, mOSM, hOSM, and LIF were similarly effective in activating STAT3 (Fig. 4B). Stimulation reached
maximal level after 5 min treatment, slowly declined to a nadir at
~60 min, and rose minimally by 120 min. Specific to OSMR-mediated response was the prominent activation of STAT5 and ERK, both with a
peak level at 15 min, and thus delayed relative to STAT3. The differential kinetics of signaling toward STAT3 and was also identified by the analyses of cytoplasmic and nuclear fractions of
cytokine-treated cells (data not shown), confirming the sequential
activation and nuclear translocation of tyrosine-phosphorylated STAT3
and STAT5.
LIF, and hOSM through the type I OSMR complex, activated STAT5 and ERK
stronger than IL-6, but less than OSM through type II OSMR. No
appreciable reaction to mOSM treatment was detected in
hOSMR-expressing cells (data not shown), indicating that hOSMR could not serve as a functional partner in a complex with mOSM and rat
gp130. Treatment of hOSMR cells with hOSM produced a signaling
response (Fig. 4C) that clearly exceeded that elicited by
LIF. Of note was that in independently derived H-35 cultures expressing
hOSMR, a temporarily prolonged activation of ERK was detected at 30 min and at later time points, exceeded the effects of mOSM in mOSMR
expressing cells. This finding suggested that at this level of
signaling action of type I and II OSMR was additive. Serial dilution of
the cytokines established that maximal cell response was attained at
approximately 10 ng/ml (Fig. 4D).
Intracellular Signaling by mOSMR Involves Additional
Pathways--
To define the type II OSMR-specific signaling reactions
independent of type I OSMR signaling, we focused on the analysis of mOSMR -transduced H-35 cells in the following experiments. Since continuous culturing of the heterogeneous populations of transduced H-35 cells invariably led to phenotypic drifting of the cultures, a
clonal line was used to further characterize mOSMR-specific action.
From a panel of 24 primary clones, which displayed severalfold difference in receptor protein expression and signal activity, we
selected one line (27-7) that expressed mOSMR mRNA and
FLAG-tagged subunit protein at the average level of the initial pool of
transduced cells.
The signal events that were initiated by the OSMR subunits and then
extended to associated proteins were assessed by the time course of
protein phosphorylation (Fig. 5). The
major changes detectable by anti-phosphotyrosine in cell lysate
included proteins co-migrating with mOSMR and gp130 (Fig.
5A). The anti-FLAG-reactive proteins demonstrated that
mOSMR underwent a ligand-induced, 2-3-fold reduction within 30 min
to 2 h of treatment. At 30 min following treatment, a transient
appearance of two anti-FLAG reactive proteins at ~45 kDa position was
noted. These smaller proteins have been tentatively identified as
intermediate, membrane-associated breakdown products of mOSMR,
resulting from proteolytic release of the extracellular domain within
the endocytic compartment. Since no appreciable changes in mRNAs
for mOSMR could be detected in OSM-treated cells (data not shown), a
modification of receptor levels by a post-transcriptional process such
as enhanced turnover was suggested.
View larger version (62K):
[in this window]
[in a new window]
|
Fig. 5.
Time course of mOSMR activation and
recruitment of signal transduction pathways. Subcultures of the
clonal line of mOSMR-expressing H-35 cells (clone 27-7) in 100-cm
diameter dishes were treated with 100 ng/ml mOSM for the indicated
lengths of time. A, equal amounts of whole cell lysates were
analyzed by immunoblotting for overall changes in the composition of
tyrosine phosphorylated (upper panel) and anti-FLAG
reactive proteins (lower panel). In B, extracts
were subjected to immunoprecipitation with anti-phosphotyrosine
(I.P.:PY) or anti-IRS-2 (I.P.:IRS-2). Cell
lysates or immunoprecipitated proteins were analyzed by Western
blotting (W.B.) for the specific forms of receptor subunits
and signal-transducing proteins as indicated on the left. The
electrophoretic positions of the proteins are marked on the
right.
|
|
The clonal cell line reproduced the two distinct kinetic profiles of
mOSM-dependent signaling reactions (Fig. 5B).
Tyrosine phosphorylation of mOSMR, gp130, JAK1, SHC, ERK, and STAT5
was temporally coordinated and was maximal at 15 min. Tyrosine
phosphorylation of SHP-2 and STAT3 displayed peak values after a 5-min
treatment. The faster time course of signaling was comparable to that
observed for the signaling in response to IL-6 and LIF (Fig. 4,
A and B; Refs. 37 and 49). The
mOSMR-expressing cells also exhibited a mOSM-dependent
activation of JNK (Fig. 5B) but not p38 MAPK. The
engagement of the stress MAPK pathway by mOSMR signal (but not observed
with IL-6 or LIF treatments) was, however, only a fraction of that
achieved by IL-1 in the same cells (not shown).
The recruitment of insulin receptor substrate (IRS) proteins
represented an additional signaling pathway activated by mOSMR (Fig.
5B) and to a lesser extent by LIFR, but not detectably by IL-6 (data not shown; Refs. 50 and 51). In H-35 cells, IRS-2 was
primarily phosphorylated in response to mOSM and was found in
association with phosphatidylinositol 3-kinase (Fig. 5B,
bottom). Although activation of the IRS/phosphatidylinositol
3-kinase pathway by mOSM was not as effective as by insulin (52),
together with the prominent activation of STATs (a process not
appreciably mediated by insulin), it produced a signal combination
highly characteristic for mOSMR.
Recruitment of SHC Is Specific to OSMR--
Recently, we showed
that in hepatoma cells, the dimeric gp130 engages the ERK pathway
through a mediator role of SHP-2 (49). In mOSMR-expressing H-35
cells, mOSM treatment also led to tyrosine phosphorylation of SHP-2
(Fig. 5B). Since the cytoplasmic domain of mOSMR, in
contrast to that of gp130, does not contain a docking motif for SHP-2,
the recruitment of SHP-2 by OSM treatment had been tentatively
attributed to gp130 within the type II OSMR complex. The kinetics by
which SHP-2 and the downstream ERK were phosphorylated differed
significantly from each other (Fig. 5B), suggesting a separate link from mOSMR to the MAPK pathways that acted in addition that controlled by gp130. SHC appeared as a possible candidate because
it was also enhanced phosphorylated following OSM treatment (Fig.
5B). A similar phosphorylation of SHC by the action of
hOSMR was also observed in hOSMR-transduced H-35 cells and in
OSM-treated HepG2 cells (data not shown).
Immunoprecipitation of SHP-2 was effective in recovering
ligand-activated (tyrosine-phosphorylated) gp130, as well as the tightly interacting LIFR-gp130 complex, but not of mOSMR (Fig. 6A, left panel). In contrast,
immunoprecipitation of SHC specifically yielded the activated mOSMR
(Fig. 6A, right panel). A comparable interaction of SHC with
activated human and mouse OSMR was detected (Fig. 6B).
The receptor subunit-specific recruitment of SHC and SHP-2 was
interpreted as a potential explanation for the prominent ERK, and
conceivably also of the JNK pathway by OSMR. Additional downstream
effects of OSMR signaling were recognized by the presence of
strongly tyrosine-phosphorylated proteins that co-immunoprecipitated with SHP-2 and SHC (Fig. 6A, open arrows). These proteins
exhibited apparent molecular sizes of 120 kDa and may include Gab-1 and the major forms of SIRPs (53, 54).
View larger version (57K):
[in this window]
[in a new window]
|
Fig. 6.
Specific association of receptor subunits
with SHP-2 and SHC. A, mOSMR-H-35 cells (clone 27-7)
in 10-cm diameter dishes were treated with indicated cytokine (100 ng/ml) for 15 min. One-half of each cell lysates was reacted with
anti-SHP-2 and the other half with anti-SHC. The immunoprecipitates
were separated on a single 7.5% SDS-polyacrylamide gel; each group
separated by a lane containing the molecular size markers. Only the
immunoblot result of the anti-phosphotyrosine antibody reaction is
shown, representing a 2-s (lower section) and a 30-s
exposure (upper section) to the ECL reaction. The position
of each of the indicated protein has been determined in additional
immunoblot reactions. Open arrows indicate SHP-2 and
SHC-associated ~120-kDa phosphoproteins which are increase following
cytokine treatment. B, in a separate experiment, the clonal
line 25-4 of hOSMR-H-35 cells and the clonal line 27-7 of
mOSMR-H-35 were treated for 15 min with cytokines. Cell lysates were
subjected to immunoprecipitation of SHC. Immunoprecipitated proteins
were separated on a 6% SDS-polyacrylamide gel electrotophoresis
followed by Western blot analysis for anti-FLAG-reactive proteins.
Positions of the co-immunoprecipitated OSMR subunits are indicated.
hIg, heavy immunoglobulin chain.
|
|
OSMR Signals Have Specific Effects on Proliferation and APP Gene
Expression--
Events downstream of OSMR signaling have been
described for different cell types, including enhanced or suppressed
proliferation, induced apoptosis, or enhanced gene transcription.
Hepatoma cells, like primary cultures of rodent hepatocytes, did not
show any detectable apoptotic reactions, when treated with any of the
IL-6 cytokines at doses as high as 1 µg/ml (data not shown). However, H-35 cells responded to IL-6 cytokines, in particular
mOSMR-expressing H-35 cells to mOSM, by suppressed proliferation as
demonstrated by reduced [3H]thymidine incorporation (Fig.
7A) and lower cell counts
after 3 days of treatment (data not shown). Flow cytometric analysis indicated a proportionally increased fraction of the cells in G1 phase.2
View larger version (37K):
[in this window]
[in a new window]
|
Fig. 7.
Specific cell responses activated by cytokine
receptors. A, the effect of cytokine treatment on
3H incorporation was determined in parental and
mOSMR-H-H35 cells (clone 27-7) as described under "Materials and
Methods." The asterisk (*) indicates values with
p < 0.05 compared with control. B, mOSMR-H-35 cells in
24-well cluster plates were treated for 24 h with serially diluted
cytokines (100, 10 and 1 ng/ml) in the absence or presence of
dexamethasone (DEX). Aliquots of the conditioned medium were
separated on 10% SDS-polyacrylamide gels. Proteins were transferred to
a single membrane and reacted with antibodies against
2-MG, TST, and HP. The immunoblot pattern was developed
for all samples by a single ECL reaction. C, parental and
mOSMR-H-35 cells (clone 27-7) were treated for 24 h into
increasing concentration of cytokines in the presence of dexamethasone.
The relative amounts of the indicated plasma proteins were quantitated
by immunoelectrophoresis (mean ± S.D.; n = 3)
|
|
Simultaneously with growth inhibition IL-6 cytokines induced and
supported elevated APP expression. The relative effects of mOSM, hOSM,
IL-6, and LIF on expression of the representative APPs, TST, HP, and
2-MG, were determined by immunoblotting (Fig. 7B) and immunoelectrophoresis Fig. 7C) of
the secreted proteins, and by RNA hybridization (e.g. Fig.
8A; below). Characteristic patterns of regulation by the individual cytokines, alone or in combination with dexamethasone, emerged that could be confirmed in
independently derived pools as well as in clonal lines of transduced cells.
View larger version (63K):
[in this window]
[in a new window]
|
Fig. 8.
Influence of OSM-regulated ERK on APP
expression. A, subcultures of mOSMR-H-35 cells (clone
27-7) were treated for 24 h with cytokines (100 ng/ml),
dexamethasone, and with or without UO126 (10 µM). Total
cell RNA was extracted, and aliquots of 20 µg were subjected to
Northern blot analysis for mRNA encoding 2-MG, HP,
and TST. B, cells were treated for 24 h with mOSM or
IL-6 (100 ng/ml), dexamethasone, and increasing doses of UO126. The
effect of MEK-1 inhibitor on 2-MG, and HP was
quantitatively assessed by immunoelectrophoresis. (mean ± S.D.;
n = 3)
|
|
In the presence of dexamethasone (required for expression of
2-MG), mOSM proved to be the most effective inducer of
2-MG (Fig. 7C). At the maximal dose of OSM
tested, the expression of 2-MG was 2-fold higher than
that induced by IL-6. mOSM induced TST at a level close to that of
IL-6, but was least effective on HP (Fig. 7, B and
C). Notable was that TST and HP regulation by OSM, in
contrast to IL-6, did not demonstrate a strong synergism with
dexamethasone (Fig. 7B). By utilizing maximal magnitudes of
induction as defining criteria, the comparison indicated the following
preference of the cytokines to regulate APP: mOSM (or hOSM through type
II OSMR) 2-MG > TST > HP; IL-6: TST > 2-MG > HP; and LIF (or hOSM through type I OSMR):
HP >TST > 2-MG. The relative activities of the
cytokines on APP gene expression were maintained for at least 96 h
of treatment (not shown).
Cytokine dose-response analysis (Fig. 7, B and C)
uncovered a second important difference among the cytokines. Mouse or
human OSM, acting through type I or II OSMR, and IL-6 produced a
similarly sharp decline in APP inducing activity in the range of 10 to
1 ng/ml (35, 49). In contrast, LIF displayed a sustained induction profile with half-maximal stimulation at ~0.1 ng/ml. This distinct dose-response of LIF relative to IL-6 had been reported before for the
induction of fibrinogen in H-35 and HepG2 cells (23, 24, 26) but had
been left unexplained. The fact that hOSM through the same LIFR-gp130
(type I OSMR) complex was not reproducing a LIF-like regulatory pattern
suggested a ligand-dependent signal control that may relate
to the difference in ligand-directed receptor subunit interaction.
Differences in Receptor-activated Signal Proteins Contribute to
Differences in APP Regulation--
To correlate some of the signaling
pathways engaged by mOSMR (Figs. 4 and 5) with the mOSM-specific
effects on APP (Fig. 7), namely the relative high 2-MG
and low TST or HP expression, we focused on the influence of ERK and
STAT5. The effects of the two signaling pathways were assessed either
by inhibiting ERK activation or by overexpressing wild type or the
transdominant negative STAT5B. Treatment of mOSMR-expressing H-35
cells with the MEK-1 inhibitor UO126 effectively abolished activation
of ERK (not shown) and suppressed induction of 2-MG
mRNA by OSM (Fig. 8A) and protein production (Fig.
8B). A similar, but less prominent inhibitory effect of
UO126 was also observed on the action of IL-6 or LIF (or hOSM through
type I OSMR) on 2-MG expression. The same UO126-treated
cells revealed a 10-fold higher HP expression after OSM treatment but
only a 2-fold higher expression after IL-6 treatment. In contrast, the
regulation of TST by each of the cytokine treatments was only minimally
affected by UO126. These results demonstrated that cytokine-regulated
ERK activity exerted a differential regulatory effect on specific APP
genes. The relative ERK activation by different cytokines could, in
part, contribute to the cytokine-specific APP pattern (45).
To probe the regulatory effects of STAT5, we transfected plasmid
constructs containing the promoters of the three rat APP genes fused to
the CAT genes (Fig. 9A). These
constructs reproduced the cytokine-specific CAT expression in
OSMR-H-35 cells (Fig. 9, left panel). Elevating the
concentration of wild type STAT5B by co-transfection further enhanced,
by 2-fold, the mOSM-stimulated expression of the 2-MG
construct but reduced the effect on HP and TST promoters by 50% (Fig.
9, right panel). Overexpression of the dominant-negative
STAT5B40C produced the opposite activity to STAT5B, lowering the OSM
stimulation of 2-MG promoter by approximately 75% and
suppressing induction of TST and HP promoters in the range as noted for
wild type STAT5B (Fig. 9, center panel). Overexpressed STAT5B proteins had only a minor influence on the inducing action of
IL-6 (Fig. 9) that appeared proportional to the recruitment of STAT5 by
IL-6 relative to OSM (Fig. 4B).
View larger version (36K):
[in this window]
[in a new window]
|
Fig. 9.
Contribution of OSMR-regulated STAT5 to APP
expression. A, mOSMR-H-35 cells (clone 27-7) were
transfected with the APP-CAT reporter gene constructs (noted at the
right) and expression vector for wild type STAT5B or
C-terminal truncated STAT5B40C (noted at the top).
Subcultures of transfected cells were treated with cytokines (100 ng/ml) as indicated at the bottom. The relative CAT activity was
determined, and data of three separate analyses are presented as mean + S.D. B and C, mOSMR-H-35 cells were
transfected with expression vectors for GFP (0.5 µg/ml) and wild type
STAT5B of STAT5B40C (3.5 mg/ml). The fluorescence-activated cell
sorter-enriched population of GFP (= internal control)
and GFP+ cells derived from STAT5B transfected cultures and
GFP+ cells from STAT5B40C transfected cultures were
subcultured. B, the control treated cells in each group were
analyzed by Western blotting for the level of STAT5 protein.
C, the production of 2-macroglobulin and
thiostatin in response to 24 h treatment with cytokines was
measured by immunoelectrophoresis.
|
|
To demonstrate that the regulation of endogenous APP genes displays
similar responses to cytokine-activated STAT5 proteins as the
transfected CAT reporter constructs, we analyzed mOSMR-H-35 cells
which had been transfected with expression vectors for the wild type,
or the transdominant-negative STAT5B, together with GFP and then were
selected by FACS for GFP positive population (Fig. 9B). The
OSM and IL-6 stimulated production of 2-MG and TST
showed a STAT5 isoform-dependent regulation that was
similar to that observed for the regulation of the promoter sequences. Expression of 2-MG was appreciably enhanced by OSM in
the presence of wild type STAT5B, whereas expression of TST was reduced
by both the wild type and truncated STAT5B. The relatively high amounts of STAT5 proteins achieved by the technique used in Fig. 9,
B and C, was also in part effective in modulating
the effects of IL-6. Taken together, these results supported the model
that mOSM-activated STAT5B, by interacting with the APRE at 
200 of
the 2-MG promoter (38),3 could exert a positive
transactivation for which the C-terminal domain of STAT5 was required.
However, STAT5B also seemed to act as inhibitor of other APP genes,
such as TST and HP. Conceivable modes of action could include
attenuation of transcription by interaction of STAT5 with APP promoter
elements or by sequesteration of other APP-promoter binding
transcription factors. For either function, the presence of the
C-terminal domain of STAT5 was apparently inconsequential.
Recruitment of LIFR Is a Dominant Determinant of OSMR
Action--
LIF dose-response analysis indicated a higher sensitivity
of H-35 cells to LIF (Fig. 7, B and C) suggesting
that LIF, in contrast to hOSM, was more effective in recruiting and
coordinating the functions of LIFR and gp130. A more stable
LIF-dependent complex (Figs. 2 and 3) could be responsible
for this higher sensitivity. Since separate IL-6 cytokine receptor
systems co-exist in hepatic cells and that their signaling depends on a
common gp130, a more effective recruitment and retention of gp130 by
LIF in a LIFR complex could result in a squelching of gp130 activity.
By limiting accessibility of gp130 to other IL-6 cytokine receptor
subunits, LIF would determine the ability of the cells to bind other
IL-6 cytokines. Thus, under conditions in which several IL-6 cytokines simultaneously interact with the cell, LIF would assume a dominant role. This influence should be particularly evident by the level of
expression of those APP genes that are poorly induced by LIF. mOSMR-expressing H-35 cells appeared optimal for testing the hypothesis that there is a hierarchy in receptor engagement and signaling because in these cells, unlike HepG2 cells, there is no
shared usage of LIFR and OSMR by mOSM (Fig. 3A).
Treatment with dose gradients of LIF and mOSM resulted in APP
expression patterns consistent with a dominant regulatory influence of
LIF. One example of TST expression is shown in Fig.
10A. At low cytokine concentrations, the expression of TST was additive following addition of the two cytokines. At high LIF concentrations, the prominent action
of mOSM was reduced by approximately 50%. In contrast, the low HP
inducing activity of mOSM was only minimally effective in reducing the
stimulation of HP expression by LIF (data not shown).
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 10.
LIFR action determines OSM and IL-6
responsiveness. mOSMR-H-35 cells (clone 27-7) (A)
and G-CSFR-gp130-H-35 cells (B) were treated for 24 h
with a medium containing dexamethasone and increasing concentrations of
LIF, alone, or together with increasing concentrations of mOSM
(indicated in ng/ml), or 100 ng/ml IL-6 or G-CSF. The relative change
in the production and secretion of TST was quantitated by
immunoelectrophoresis. For clarity, only the mean values of 3 separate
experiments are shown. Error bars were not included, with
the exception of the LIF dose response in A.
|
|
In agreement with the hypothesis that gp130 is sequestered by
LIF-activated LIFR, induction of TST by IL-6 was also reduced by LIF
(Fig. 10A). To rule out the possibility of an inhibitory activity was specifically induced by LIF that either mediated a reduced
APP induction or neutralized the signaling by other IL-6 cytokines
independently from the competition for gp130, we determined the effects
of LIF treatment on induction of TST in H-35 cell lines stably
expressing the chimeric receptor G-CSFR-gp130 (49). In these cells, the
signaling by gp130 cytoplasmic domain was triggered by G-CSF acting
independently of the endogenous gp130. These cells displayed the
expected low level induction of TST by LIF in the presence of
dexamethasone (Fig. 10B). LIF reduced the action of IL-6 but
not of G-CSF.
Taken together, this study documents that the IL-6 cytokines, despite
their overlap in engaging common signal transduction mechanisms, have
the capability through different relative ratios of signal activation
to mediate cytokine-specific gene regulation. Also, in situations of
multiple cytokine exposure, LIFR assumes a dominant role over the more
effective signaling complex of OSMR or IL-6R.
| |
DISCUSSION |
The central issue addressed in the present study is the extent to
which different members of the IL-6-type cytokines have redundant or
interchangeable functions or how activities of these are specific to
individual members of the family. Focusing on the OSMR, we tested the
hypothesis that signal-transducing subunits of IL-6 cytokine receptors
have distinct signaling capabilities and the combination of subunits
establishes the specificity of cellular response to the cytokines. The
following key features of the cell system used allowed us to identify
signaling specificities: 1) the activities of the different receptor
forms were defined in the same cellular milieu (constant gp130
content); 2) the cell response was quantifiable by the induction of
several major regulated genes; and 3) the contribution of the receptor
subunits to the cell response was assessed as a function of manipulated
receptor expression. The results document that the representative
members IL-6, LIF, and OSM elicit specific regulatory patterns and that the relative quantitative manifestation of cytokine effects is prominently influenced by the ligand-activated receptor type.
The study also led to the question of how multiple cytokines control
cell function. Responsiveness of target cells to IL-6 cytokines has
generally been assessed by measuring the effects of individual
cytokines (25, 26, 55, 56). However, in physiologically relevant
settings, such as in the developing organism, at the site of
hematopoiesis, or during immune or inflammatory reactions in
vivo, a temporal and local coexistence of several members of IL-6
cytokines, as well as a plethora of other bioactive molecules, is
expected (1, 14, 57, 58). The sum of these factors will inevitably
influence the function of the individual factors. This work
demonstrates two such interaction: 1) LIF action has dominant effects
on OSM or IL-6 signaling, possibly due to the more effective
recruitment of gp130 by LIFR that is documented here, and 2) the
action of different IL-6 cytokines is differently affected by dexamethasone.
Control of Signaling as Function of Receptor Subunit
Expression--
If one assumes that gp130 is the only relevant
signal-transducing subunit for all IL-6 cytokines, then a generic IL-6
cytokine profile of cell response should be detected (17). When
recording the effects of IL-6 cytokine receptors on proliferation (as
exhaustedly done in the model of Ba/F3 cells), only a partial view of
the signaling by IL-6 cytokines is obtained. The regulation of multiple APP genes samples a much broader array of receptor-derived signals. The
results clearly demonstrate cytokine receptor-specific effects. The
concept that receptor signaling is directed by modular information contained within the receptor proteins (59) leads us to propose that
the composition of the oligomeric receptor subunits, recruited by the
cytokines through the extracellular receptor domains, are instrumental
in determining the intracellular signal specificity. The homomeric
gp130 that has been predicted to be the core of the functional IL-6
receptor (60) displays structural and biochemical information by its
cytoplasmic domains that is distinct from that of the heteromeric
cytoplasmic domains of OSMR or LIFR with gp130 (19, 31-33, 60,
61). The signal-transducing pathways recruited by OSMR or LIFR
blend with those controlled by the common gp130.
This work suggests a highly effective contribution of OSMR, which
exceeds that of LIFR resulting in a greater activation of ERK,
STAT5, and IRS pathways and, in addition, an activation of JNK (Fig.
5). The contribution of gp130 would account in part for the qualitative
similarity among IL-6 cytokine responses in the various cells types
analyzed. If just the influence of receptor subunit levels is
considered, the relative amount of gp130 is a major determinant of the
maximal signaling capability of IL-6 cytokines in a given cell. There
is no evidence for a gp130-independent signaling event elicited by
either LIFR or OSMR, although their cytoplasmic domains have
signaling capabilities that have been identified in the context of
chimeric receptors (29, 61, 62). Aside from gp130, the relative amounts
of IL-6R, LIFR, and/or OSMR expressed at the cell surface not
only determine specificity of cytokine recognition, but also influence
the quantitative cell response. Limited availability of these receptor
subunits will invariably result in a reduction, or absence, of a
cytokine response. The extreme case represents cells with only gp130
expression, which would provide low affinity binding sites for OSM but
otherwise be unable to respond to any IL-6 cytokine. However, the
function of gp130 in such a cell type can still be verified by reaction with a complex of soluble IL-6R with bound IL-6 that mimics an IL-6-specific recruitment and activation of the membrane resident gp130
(63, 64).
Graded expression of transduced hOSMR in
Hep3B4 and H-35 cells
indicated a receptor dose-dependent inducibility of APP
genes and that the relative profile of the regulated APPs was not
appreciably different in cells with low versus high amounts
of receptors subunits (data not presented). In our experimental system,
limited amounts of OSMR or LIFR was not an issue; the majority of
the receptor-transduced pools and clones appeared to express the
receptors at a level that saturated gp130 recruitment. Similarly,
further enhancement of LIFR expression achieved by transduction of
hLIFR was unable to improve the cells response to LIF
(27).5 Taken together these
data indicate that, in reconstituted H-35 cells, the signal by IL-6,
LIF, and OSM is limited by gp130 and the cell response measures maximal
signaling capability by each receptor complex. The cytokine dose
response (Fig. 7, B and C) in turn
illustrates the signaling efficiency. We assume that any excess amounts
of LIFR or OSMR expressed by H-35 cells are inconsequential as
far as signaling is concerned. However, excess subunits could enhance
low affinity cytokine binding activity of the cells (19, 32).
Receptor Subunits Define Signaling Specificity--
The marked
difference in APP gene regulation and suppressed proliferation by OSM
and LIF, compared with IL-6, is attributed to signaling by the
cytoplasmic domains of OSMR and LIFR. Although the primary
structures of mouse and human OSMR show 45% sequence difference,
both have comparable signaling effects when the contribution of
co-activated LIFR-gp130 by hOSM is discounted (Fig. 4). The cytoplasmic domains of human and mouse of OSMR contain the same structural motifs associated with signal communication, e.g.
JAK and STAT3 (32, 33). Both subunits are devoid of a SHP-2 docking element, thus explaining the failure to co-immunoprecipitate OSMR with SHP-2 (Fig. 6A). The recovery of SHC with OSMR
suggested that this adaptor protein may contribute to the prominent
MAPK recruitment by OSMR complex (46).2 The coactivated
gp130 may be responsible for recruitment and phosphorylation of SHP-2
(28, 35, 49). A single SHP-2-binding site per OSMR complex, as opposed
to two in the IL-6R complex, could also reduce the probability of the
receptor complex (including their associated kinases) being
dephosphorylated and thereby de-activated by receptor-recruited SHP-2
(35). Although in other cell types SHC has been described in
association with IL-6-induced gp130 signaling (65, 66), its relative
contribution in the hepatoma cells to the IL-6 response appears minor
(49). Recovery of LIFR as part of the ligand-activated LIFR
co-immunoprecipitated with anti-SHP-2 is probably due to gp130 that is
tightly associated with LIFR. LIFR does not contain an effective
SHP-2 docking site in its cytoplasmic domain, although one at tyrosine
115 has been suggested (67). LIFR cannot be recovered as
SHP-2-associated protein when activated by OSM suggesting a lower
affinity interaction with gp130. Similarly, no SHP-2 association with
the cytoplasmic domain of OSMR or LIFR could be detected when
these were part of G-CSF-activated, homomeric
G-CSFR-chimeras.4,5
Inasmuch as the precise signal transduced from OSMR and LIFR
toward the MAPK/ERK pathway and the basis for the kinetic differences in activation remain to be established (68, 69), it is evident (Fig. 4) that both receptor subunits assist in a more active process than the IL-6-stimulated gp130 dimer. Furthermore, the subunits, in
particular OSMR, promote a more effective STAT5 recruitment (Fig. 4,
B and C; Ref. 50). Preliminary characterization
of mutations in the 6 tyrosine residues conserved between human and mouse OSMR could not assign a STAT5-specific docking and activation site.4 An alternative model proposes that JAKs could be
responsible for direct STAT5 activation (70) as well as for the
engagement of IRS (50). Both the prominent ERK and STAT5 activation by OSM has been associated with modulated APP gene expression in H-35
cells (Fig. 8) and primary cultures of mouse and rat hepatocytes (data
not presented). OSMR signaling matches the effect of growth hormone
receptor on STAT5 recruitment in liver cells (71). However, in contrast
to growth hormone, OSM may contribute to the nuclear accumulation of
STAT5 in liver cells during an adjuvant-induced inflammatory response
(38). It remains to be determined whether STAT5 is also responsible for
the particularly abundant hepatic expression of 2-MG
during an inflammatory reaction in rats (39, 72).
Cross-modulation between IL-6 Cytokine Family Receptors as a
Mechanism Controlling Cytokine Responsiveness in Cells--
LIFR is
expressed in many cell types and seems particularly relevant, among
others, in directing neuronal regeneration (73, 74), control of the
hypothalamic-adrenal axis (75), stress response of myocardiocytes (76),
embryonal implantation (77), and differentiation of the renal
mesenchyme to an epithelial structure (78). Surprisingly, the mode of
LIFR action is still largely unknown. In regards to liver cell
regulation, the molecular biology behind the relative low level of
LIFR-directed induction of several APPs is similarly unexplained. A
relevant finding is that cells respond more sensitively to LIF than to
IL-6 and OSM. The same subunits as part of hOSM-activated type I OSMR
do not generate a sensitivity as great as that seen with the
LIF-activated complex (Fig. 7B). The difference may be due
to the more stable receptor complex recruitment by LIF. An efficient
interaction among the components of the LIFR complex is suggested by
the binding of LIF to LIFR through two interaction sites that
involve a coordination of LIF with the first hematopoietin domain and
the immunoglobin domain of LIFR (79, 80). The immunoglobulin domain
is not involved in OSM binding (80), potentially explaining the lower affinity association of type I OSMR subunits. A critical consequence of
the effective interaction of LIF and LIFR on signaling functions is
that LIFR action appears dominant over the more active OSMR or IL-6R
complexes (Fig.
,
, and
TOP
ABSTRACT
INTRODUCTION
