(Received for publication, January 6, 1997, and in revised form, March 14, 1997)
From the Division of Endocrinology, University of Arkansas for Medical Sciences Center of Osteoporosis and Metabolic Bone Diseases, and GRECC, Veterans Administration Medical Center, University of Arkansas for Medical Sciences, Little Rock, Arkansas 72205
Glycoprotein 130 (gp130), a shared component of
all the receptors for the interleukin-6 cytokine family, transduces
cytokine signals in part by activating latent cytoplasmic signal
transducers and activators of transcription (STATs). STATs subsequently
translocate into the nucleus and stimulate gene expression. In the
studies reported here, the 5-flanking region of the human gp130 gene was isolated and the transcription initiation sites were mapped. To
demonstrate that the isolated DNA fragment contained a functional promoter, a plasmid construct containing 2433 base pairs of the gp130
5
-flanking region, inserted upstream from the firefly luciferase gene,
was transiently transfected into HepG2 hepatoma cells. The construct
exhibited constitutive promoter activity. In addition, a 5-h treatment
with interleukin-6 or oncostatin M stimulated the activity of this
promoter severalfold. Localization of the cytokine response element by
5
-deletion analysis and site-directed mutagenesis revealed a
cis-acting binding site for activated STAT complexes. Furthermore, DNA
binding analysis demonstrated that this element binds activated STAT1
and STAT3 homo- and heterodimers. This STAT-binding element was
sufficient to confer cytokine stimulation to a minimal herpesvirus
thymidine kinase promoter. These results establish that the DNA
fragment we have isolated contains the human gp130 promoter and that
interleukin-6 type cytokines may influence the activity of this
promoter via activated STATs.
Members of the interleukin-6 (IL-6)1
cytokine family, which includes IL-6, interleukin-11 (IL-11),
oncostatin M (OSM), ciliary neurotrophic factor, leukemia inhibitory
factor, and Cardiotropin-1, have pleiotropic but functionally redundant
effects on a wide variety of mammalian cells (1, 2). This redundancy
can now be explained, at least in part, by the discovery that the
receptors for each of these cytokines share a common signal transducing component known as gp130 (1, 3-5). Ligand binding to the specificity determining subunit of these receptors ( subunits) causes tyrosine phosphorylation of gp130 (
subunit) by members of the JAK family of
tyrosine kinases (6). This event results in tyrosine phosphorylation of
several downstream signaling molecules including members of the STAT
family of transcription factors (7, 8). Phosphorylated STATs in turn
undergo homo- and heterodimerization and translocate to the nucleus
where they activate transcription of cytokine responsive genes (9).
In mammals, the STAT family includes at least six members which are differentially activated by a variety of growth factors and cytokines (10). Ligand-activated gp130-JAK complexes predominantly phosphorylate STAT3 (also known as acute-phase response factor or APRF) and to a lesser extent STAT1 (11-13). The cis-acting DNA sequences recognized by STAT complexes, termed STAT-binding elements (SBEs) have the general structure TT(N)5AA (9, 14). The sequence, and in some cases the size, of the spacer region between the palindromic TT-AA motif confers specificity for different STATs (11, 14). An alternative sequence (CTGGGA), originally termed acute-phase response element (APRE), is required for IL-6 induction of many acute-phase response genes (3) and is found in the spacer region of some palindromic SBEs or in incomplete palindromic forms which may also bind STATs (11, 15).
Intrigued by the evidence that IL-6 increases gp130 mRNA in
vitro in human monocytes, epithelial cells, or hepatoma cells (16-18); that administration of IL-6 to mice produces a striking up-regulation of gp130 mRNA levels in several tissues (19); and
that several systemic hormones regulate gp130 expression (20-22), we
have cloned and characterized the 5-flanking region of the human gp130
gene. Using chimeric gp130 promoter/luciferase reporter constructs, we
demonstrate that this region functions as a constitutively active
transcriptional promoter and that its activity is stimulated by
IL-6-type cytokines. In addition, we have localized a cis-acting sequence element responsible for induction by these cytokines and show
that it conforms to the consensus binding site for activated STAT
complexes. The STAT complexes which bind the gp130 cytokine response
element are identified as STAT1 and STAT3 homo and heterodimers. Finally, we show that this SBE is sufficient to confer cytokine inducibility on a heterologous promoter to a level comparable to
previously characterized SBEs.
A human fibroblast genomic library in the
Lambda FIX II vector (Stratagene) was screened with a probe consisting
of the 5-terminal 550 bp of the human gp130 cDNA (4) according to
established methods (23). DNA was purified from positive plaques and
mapped by partial restriction enzyme digestion combined with Southern blot analysis (24). A 2.7-kb EcoRI fragment which hybridized to the probe was isolated from a clone containing a 12.5-kb insert and
was subcloned into pBluescript II KS+ (Stratagene) to yield the plasmid
pAE3. Both strands of the 2.7-kb fragment were completely sequenced by
the chain termination method (25) using sequence-derived primers
(GenBank accession number U70617[GenBank]).
Primer extension
was performed using an oligonucleotide primer complementary to bases
+116 to +136 of the first exon shown in Fig. 1B. The
5-32P-labeled primer was annealed to 20 µg of
poly(A)+ RNA from the ARD cell line in 1 × first
strand buffer (50 mM Tris-HCl (pH 8.3), 75 mM
KCl, 3 mM MgCl2) for 1 h at 60 °C and then placed on ice. Dithiothreitol and deoxynucleoside triphosphates were added to a final concentration of 10 and 500 mM,
respectively, followed by 200 units of Superscript II reverse
transcriptase (Life Technologies, Inc.). The reaction was incubated at
45 °C for 2 h followed by the addition of ribonuclease A to a
final concentration of 0.25 mg/ml and a 10-min incubation at 37 °C. The reaction mixture was then precipitated with ethanol, resuspended in
95% formamide, heat denatured, and fractionated on a 5%
polyacrylamide, 8 M urea sequencing gel.
The S1 nuclease assay was performed using a 32P-labeled, single-stranded DNA probe (23) generated by extension of the 116-136 oligonucleotide, described above, after it was annealed to single-stranded DNA from the pAE3 plasmid. The probe was annealed to 50 µg of total RNA from HepG2, ARD, or Escherichia coli cells in 30 µl of hybridization buffer (40 mM 1,4-PIPES, pH 6.4, 1 mM EDTA, 400 mM NaCl, and 80% formamide) overnight at 30 °C. S1 nuclease digestion was carried out by the addition of 300 µl of S1 mapping buffer (280 mM NaCl, 50 mM sodium acetate, pH 4.5, 4.5 mM ZnCl2, 20 µg/ml single-stranded DNA, and 1000 units/ml S1 nuclease) and incubation at 42 °C for 1 h. The reaction was terminated with 80 µl of stop buffer (4 M ammonium acetate, 50 mM EDTA, and 50 µg/ml E. coli tRNA) and the products were ethanol precipitated and fractionated as described for the primer extension products.
Cell Culture and Transient TransfectionsHepG2 hepatoma
cells and HeLa epithelial carcinoma cells were obtained from the
American Type Culture Collection (ATCC) and maintained in phenol
red-free minimum essential medium (Life Technologies, Inc.)
supplemented with 10% (v/v) fetal bovine serum (Sigma). A human
myeloma cell line (ARD) was kindly provided by Dr. Bart Barlogie
(University of Arkansas for Medical Sciences) and was cultured in RPMI
1640 medium (Life Technologies, Inc.) containing 10% fetal bovine
serum. A murine preadipocyte cell line, +/+LDA.11, was cultured in
McCoy's 5A medium (Sigma) containing 10% fetal bovine serum. All
cytokines were obtained from R & D Systems and were used at the
following final concentrations: IL-6, 20 ng/ml; OSM, 20 ng/ml; IL-6sR,
40 ng/ml; IFN-, 5 ng/ml.
Transient transfections of all cell types were carried out in 12-well
tissue culture plates (2.2-cm diameter wells) using LipofectAMINE (Life
Technologies, Inc.) as described by the manufacturer. Briefly, the day
before transfection, cells were seeded at 2-4 × 104
cells/well in medium containing 10% fetal bovine serum. The next day,
the cells were washed once with serum-free medium and each well was
incubated with serum-free medium containing 200 ng of chimeric
luciferase reporter plasmid, 200 ng of control plasmid (pSVbetagalactosidase Vector, Promega), and LipofectAMINE for 5 h.
The medium was replaced with serum-containing medium and the cells were
allowed to recover for 24 h before treatment with cytokines or
hormones. The only exception was the experiment described in Fig. 3 in
which the cells were cultured for 48 h following transfection.
Lysate preparation and luciferase activity assays were performed using
a kit (Promega) according to the manufacturer's instructions. Light
intensity was measured with a Turner luminometer. The colorometric
-galactosidase assay was performed using standard protocols (23) and
luciferase activity was divided by the
-galactosidase activity to
normalize for transfection efficiency.
DNA Constructions
All of the chimeric luciferase reporter
constructs in this study were prepared using the pGL3-Basic vector
(Promega). A region spanning 2433 to +64 of gp130 5
-flanking region
was prepared by partial digestion of the 2.7-kb EcoRI
fragment of pAE3 with PvuII and purification of the 2.5-kb
fragment. The 2.5-kb fragment was blunt-ended with the Klenow fragment
of E. coli DNA polymerase I (Klenow) and ligated into the
SmaI site of pGL3-Basic to produce constructs with the
insert in either the sense (p-2433) or antisense (p-2433-AS)
orientation. Each of the 5
-deletion constructs was prepared from
p-2433 utilizing existing restriction enzyme sites (indicated in Fig.
5A). Constructs in which the APRE sequence (mt1) or the SBE
sequence (mt2) were eliminated from the p-381 5
-deletion construct
were prepared using the Chameleon site-directed mutagenesis kit
(Stratagene) and the mutations were confirmed by sequencing (25). The
herpesvirus thymidine kinase (tk) promoter/luciferase construct
(ptk-LUC) was prepared by insertion of the
BamHI/BglII fragment of pBLCAT2 (26), which spans
positions
105 to +51 of the thymidine kinase gene, into the
BglII site of pGL3-Basic. The heterologous constructs
depicted in Fig. 7 were prepared by insertion of the following
double-stranded oligonucleotides (only the top strands are shown) into
the SmaI site of ptk-LUC: gp130, GATCGCGTTACGGGAATCG; SIE,
GATCGATTGACGGGAACT; rat
2-macroglobulin, GATCCTTCTGGGAATTC. Plasmids with single or double insertions were identified by restriction enzyme analysis and confirmed by
sequencing.
Nuclear Extracts and DNA Mobility Shift Assays
Nuclear
extracts were prepared from HepG2 cells using the method
described by Sadowski and Gilman (27). Cells grown to
approximately 70% confluence in 10-cm dishes were treated with
cytokines for 15 min. After treatment, the cells were washed twice with
ice-cold phosphate-buffered saline and once with hypotonic buffer (20 mM HEPES, pH 7.9, 20 mM sodium fluoride, 1 mM Na3VO4, 1 mM
Na4P2O7, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, and 1 µg/ml each of
leupeptin, aprotinin, and pepstatin). The cells on each plate were
lysed with 0.3 ml of ice-cold lysis buffer (hypotonic buffer containing
0.2% Nonidet P-40), scraped into microcentrifuge tubes, vortexed for
5 s, and centrifuged for 30 s at 15,000 × g.
Nuclear pellets were resuspended in 0.2 ml of high salt buffer
(hypotonic buffer with NaCl and glycerol added to 420 mM
and 20%, respectively) and incubated on ice for 30 min with occasional
mixing. Nuclear debris were pelleted at 15,000 × g for
20 min at 4 °C, and the supernatants were quick-frozen on dry ice
and stored at 80 °C.
Probes for the DNA mobility shift assay were prepared by filling in
GATC overhangs of double-stranded oligonucleotides (described in the
DNA construction section) using Klenow and [-32P]dCTP.
DNA binding reactions were performed by preincubating 15 µg of
nuclear extract with 1 µg of poly(dI-dC)·poly(dI-dC) in 1 × binding buffer (10 mM Hepes, pH 7.9, 50 mM
NaCl, 1 mM dithiothreitol, and 5% glycerol) for 15 min on
ice followed by the addition of approximately 5 fmol of probe (20,000 cpm) and an additional 15-min incubation at room temperature. The
reactions were resolved on 5% polyacrylamide gels (39:1
acrylamide:bis) containing 1% glycerol in 0.5 × TBE (Tris
borate/EDTA) buffer. For supershift assays, binding reactions were
preincubated with either anti-STAT1 or anti-STAT3 antibodies (both from
Santa Cruz Biotechnology) or non-immune rabbit IgG for 15 min at room
temperature.
A human fibroblast
genomic DNA library was screened with a probe consisting of the extreme
5-end of the cloned human gp130 cDNA (4). Several clones were
isolated and the hybridizing region from each was subcloned and
partially sequenced. One of the clones contained a region which was
identical to the first 152 bp of the reported gp130 cDNA sequence
(Fig. 1, A and B). This clone was
selected for further analysis and a partial restriction enzyme map of
the 12.5-kb insert is shown in Fig. 1A. One of the remaining
clones possessed approximately 3 kb of continuous homology to the gp130
cDNA but contained several mismatches, small insertions and
deletions (not shown). Since the homologous region in this clone was
uninterrupted by introns, it was concluded that this fragment
represented a processed gp130 pseudogene. This conclusion is consistent
with the previous observation that two distinct loci, one each on
chromosomes 5 and 17, hybridized to gp130 probes (28) and that a
pseudogene-like sequence was polymerase chain reaction amplified from
chromosome 17 DNA (29).
A 2.7-kb EcoRI fragment containing the 152-bp identity with
the gp130 cDNA was subcloned and completely sequenced. To identify the transcription start sites, primer extension analysis was performed using a primer complementary to bases 116- 136 of the exon 1 sequence (from Fig. 1B) and poly(A)+ RNA from ARD cells.
ARD cells were chosen for these experiments because they produce
relatively large amounts of gp130 mRNA compared with other cell
lines examined (not shown). The primer extension products indicated
that there are three groups of start sites, designated A, B, and C,
which begin 17 bp upstream from the 5-end of the known cDNA
sequence (Fig. 2, left panel). S1 nuclease
protection assays performed on total RNA from HepG2 and ARD cells
revealed protected fragments corresponding to the same three sets of
start sites as the primer extension analysis ( Fig. 2, right
panel).
Sequence analysis of the region upstream of the transcription start
sites revealed a G + C-rich region that did not contain a TATA box
(Fig. 1B). However, a potential Sp1-binding site was present
at 47 bp and two CCAAT motifs were located at
135 and
167 bp.
Besides several additional Sp1 sites, the first 600 bp of 5
-flanking
region also contained several sequences homologous to IL-6 type
cytokine response elements including a palindromic SBE (TTACGGGAA), a
non-palindromic SBE or APRE (CTGGGA), and two adjacent
NFIL6/C-EBP
-binding sites (3, 30).
To confirm that the isolated DNA fragment contained a
functional promoter, we prepared reporter gene constructs by inserting the 5-flanking region (spanning bases
2433 to +64 relative to the
most 3
-transcription start site) upstream from the firefly luciferase
gene in the pGL3-Basic vector. When a construct with this fragment in
the sense orientation (p-2433) was transiently transfected into HepG2
cells, a significant level of transcriptional activity was observed
(Fig. 3). The parental vector or the parental vector
containing the fragment in the antisense orientation produced only
trace amounts of transcriptional activity. Similar results were
obtained when this set of constructs was transfected into HeLa or
+/+LDA.11 cells (not shown). These results establish that the
5
-flanking region of gp130 contains a promoter that is constituitively active in epithelial and fibroblastoid cell types.
Transcriptional regulation of the gp130 promoter was subsequently
investigated in HepG2 cells transiently transfected with the p-2433
construct. Consistent with earlier evidence suggesting that gp130
production is regulated, 5-h treatment with IL-6 in combination with
the IL-6-soluble receptor (IL-6sR) or OSM stimulated the activity of
the promoter approximately 4-fold, while IL-6 alone produced
approximately a 2-fold increase (Fig. 4). Essentially the same results were obtained when the treatment time was extended to
24 h, with the only exception that the effects of IL-6 + IL-6sR and OSM were slightly lower and that for IL-6 was higher compared with
the 5-h time point. In line with the evidence that interferon (IFN)-
and IL-6 activate overlapping sets of transcription factors and genes
(31), 5-h treatment with IFN-
stimulated the reporter construct to a
level similar to that produced by IL-6. However, after 24 h of
treatment, the response to IFN-
was significantly decreased.
Localization of the Cytokine Response Elements in the gp130 5
To identify the cis-acting sequence elements
responsible for the induction of the gp130 promoter by IL-6 type
cytokines, sequential 5-deletion constructs of the plasmid p-2433 were
prepared (depicted in Fig. 5A). Transfection
of these constructs into HepG2 cells and treatment with either IL-6 or
OSM for 5 h revealed that deletion beyond
191 bp abolished the
response to either cytokine (Fig. 5A). Besides failing to
exhibit cytokine responsiveness, the
191 construct also displayed a
significant decrease in basal transcriptional activity. Taken together,
these results indicate that the region between
381 and
191 bp
contained sequences responsible for the majority of the cytokine
responsiveness as well as sequences contributing to the basal
transcriptional activity of the gp130 promoter.
The region between 381 and
191 bp contains two potential cytokine
response elements identified by comparison to consensus sequences: a
palindromic SBE (TTACGGGAA) and a non-palindromic SBE or APRE (CTGGGA).
Based on this and evidence that IL-6-type cytokines utilize STAT
proteins in their signal transduction pathway, we reasoned that one or
both of these potential SBEs might be responsible for the cytokine
induction of the gp130 promoter. To address this issue, we prepared
reporter constructs in which either of these potential response
elements were eliminated from the
381 promoter fragment by
site-directed mutagenesis. The constructs were then transiently
transfected into HepG2 cells and their activity was examined following
5 h of treatment with IL-6 or OSM. The wild-type
381 promoter
fragment responded to IL-6 and OSM as expected from the 5
-deletion
analysis and this response was unaffected in the construct containing
the mutated APRE consensus sequence (mt1) (Fig. 5B).
However, the construct containing the mutated palindromic SBE consensus
sequence was unresponsive to either cytokine (mt2 in Fig.
5B). Thus, the SBE-like sequence located between
381 and
191 bp is required for stimulation of the gp130 promoter by IL-6 and
OSM.
The sequence between the
TT-AA motif of the gp130 cytokine response element is identical to the
IL-6-response element from the human 2-macroglobulin
promoter (32) and to the c-fos SBE known as the
sis-inducible element (SIE) (33). The SIE binds STAT1 and
STAT3 homo- and heterodimers (12, 34) and is required in
vivo for the induction of the c-fos gene by a variety
of stimuli (35). A comparison of the gp130 SBE-like sequence with SBEs from various cytokine or growth factor-inducible promoters is shown in
Table I. Both a mutant version of the SIE, denoted
SIEm67, and a SBE from the rat
2-macroglobulin promoter,
which contains the APRE consensus (CTGGGA), have been used as classical
IL-6 response elements (12, 31, 34, 36). In our studies we employed the
latter for comparison purposes.
|
Previous studies have shown that in HepG2 cells, IL-6 and OSM induce
DNA binding homodimers of STAT1 and STAT3 as well as heterodimers of
STAT1 and STAT3; while IFN- induces predominantly STAT1 homodimers
(12, 13, 36). In electrophoretic mobility shift assays with the SIEm67
probe, these complexes migrate as a set of three bands with the slowest
corresponding to homodimers of STAT3, the middle to heterodimers of
STAT1 and STAT3, and the fastest to homodimers of STAT1.
To determine if the gp130 cytokine response element is in fact bound by
activated STAT complexes, we performed electrophoretic mobility shift
assays using an oligonucleotide probe corresponding to the gp130
cytokine response element and nuclear extracts from cytokine-treated
HepG2 cells. IL-6 type cytokines induced the formation of the three
characteristic DNA binding complexes designated A, B, and C, while
IFN- induced the formation of a single complex that co-migrated with
complex C (Fig. 6A). Complexes of the same mobility and intensity were seen when the same extracts were incubated with a probe corresponding to the wild-type SIE from the
c-fos promoter (compare lanes 2-5 with
lanes 7-10). The specificity of the protein binding to the
gp130 probe was demonstrated by competition with a 100-fold molar
excess of either the wild-type gp130 or the SIE oligonucleotide but not
a mutated gp130 oligonucleotide (Fig. 6B). The identity of
the binding proteins in the OSM-treated extract was established by
demonstrating that an anti-STAT1 antibody inhibited the formation of
both the fastest and the intermediate migrating complexes, whereas an
anti-STAT3 antibody supershifted the slowest and the intermediate
complexes (Fig. 6C). As expected, the anti-STAT1 antibody
inhibited the formation of the fastest migrating complex from the
IFN-
-treated extract whereas anti-STAT3 had no effect. These results
demonstrate that the cytokine response element in the gp130 promoter is
bound by activated STAT1 and STAT3 proteins and that the gp130 SBE
binds these proteins with an affinity similar to that of the
c-fos SBE.
Finally, to determine if the gp130 SBE was sufficient to confer
cytokine inducibility on a heterologous promoter, we inserted an
oligonucleotide corresponding to a single copy of the gp130 SBE
upstream from an herpes simplex virus-thymidine kinase
promoter/luciferase reporter gene. To compare the stimulatory ability
of the gp130 SBE to previously characterized SBEs, we also prepared
similar constructs containing single copies of either the
c-fos SBE (SIE) or the rat 2-macroglobulin
SBE. The gp130 SBE was sufficient to confer IL-6 and OSM inducibility
to the promoter to an extent equal to or greater than that of either
the c-fos or rat
2-macroglobulin SBE (Fig.
7). Multimerization of the SBEs increased responsiveness to IL-6 and OSM, and conferred IFN-
induction as well.
In the work described in this report, we have isolated and
characterized the human gp130 5-flanking region. We determined that
this DNA fragment is similar to the 5
-flanking regions of genes for
other members of the cytokine receptor family in that it lacks a
recognizable TATA box and is rich in G + C residues (37). These
properties along with multiple transcription start sites are also found
in many constitutively expressed genes (38). In addition, we found that
IL-6-type cytokines stimulate the activity of this promoter. Although
several potential sites for IL-6-inducible transcription factors were
identified within the first 600 bp upstream from the transcription
start sites, deletion and mutagenesis analysis revealed that, of these,
only the consensus palindromic SBE was required for IL-6 stimulation
under the conditions used in our experiments.
The observation that the gp130 promoter is stimulated by IL-6 and OSM
suggests that the previously reported increases in gp130 mRNA in
response to these cytokines might be the result of transcriptional regulation. Consistent with this suggestion, the 2-4-fold stimulation of the isolated promoter, in our studies, corresponded closely with the
earlier described magnitude of gp130 mRNA induction by IL-6 in
HepG2 cells (18). OSM exhibited reproducibly greater stimulation of the
gp130 promoter, as compared with IL-6. This phenomenon may be due to a
greater number of functional receptors at the cell surface and/or more
effective activation of STAT complexes by the former cytokine. Several
genes are induced by both IL-6 and IFN-, in many cases through the
same SBEs (31). Nonetheless, we found that IFN-
, although it
stimulated promoter activity following a 5-h treatment, had no effect
on the gp130 promoter in HepG2 cells at 24 h. The lack of a
sustained effect of IFN-
is consistent with the results of studies
by Schooltink et al. (18) demonstrating lack of an effect of
this agent on gp130 expression in HepG2 cells.
Lamb and co-workers (13) have suggested that promoters containing a
single SBE, like gp130, are stimulated preferentially by STAT1/STAT3
heterodimers as compared with homodimers of either STAT1 or STAT3. This
suggestion is in agreement with our finding of a preferential
stimulation of the gp130 promoter by IL-6 or OSM as compared with
IFN-, and can explain the apparent discrepancy between the amount of
STAT-DNA binding activity and transactivation effect produced by these
cytokines, which has also been demonstrated in other systems (14, 39).
Recent experiments in STAT1-deficient mice demonstrated that STAT1 is
not required for IL-6 induction of several IL-6-responsive genes (40).
Therefore it remains possible that the STAT3 homodimer, and potentially
other STAT proteins (41), may be responsible for the stimulation of the gp130 promoter by IL-6 type cytokines.
The level of IL-6-type cytokine stimulation of the gp130 promoter is
slightly lower than that reported for some of the acute-phase response
and immediate early gene promoters, which demonstrate 4-8-fold
stimulation by IL-6 in hepatoma cell lines (15, 42-44). This is the
case even though the gp130 SBE binds activated STAT dimers with a
similar affinity as the c-fos SBE and is able to confer
cytokine stimulation to a heterologous promoter to approximately the
same extent as the c-fos and rat
2-macroglobulin SBEs. A possible explanation for this
difference comes from the observation that many acute-phase response
gene promoters contain multiple SBEs or an SBE in close proximity to
binding sites for other inducible transcription factors. However, we
have identified only a single SBE in the gp130 promoter which appears
to confer the majority of the cytokine responsiveness.
The biologic significance of the in vitro evidence for the
stimulation of the gp130 promoter by IL-6 type cytokines, to the in vivo situation, is at this stage a matter of conjecture.
Nonetheless, it is well established that following cytokine binding,
gp130, as well as the subunit of the IL-6 receptor, is internalized and degraded (45-48). Considering this and the evidence that several cell types express receptors for more than one member of the IL-6 type
cytokine family (49, 50), and that all of them utilize gp130,
maintenance of responsiveness to other members of the family, following
stimulation by one of them, may require gp130 replenishment. The
STAT-mediated increase in gp130 promoter activity may therefore be part
of a mechanism to achieve this replenishment. A similar mechanism may
be operative for other receptors that activate STAT proteins as
evidenced by the recent finding that STAT6 binds to and activates the
IL-4 receptor promoter (51).
Currently, it is unknown whether, in at least some of the target cells
of IL-6-type cytokines, the abundance of gp130 may be limiting so as to
render regulation of gp130 expression consequential (47). Nonetheless,
it has been shown that not only IL-6 type cytokines, but also IL-1,
TNF, and the systemic hormones estrogen, androgen,
1,25-dihydroxyvitamin D3
(1,25-(OH)2D3), all-trans-retinoic acid, and parathyroid hormone regulate gp130 expression in
vitro (17, 20-22). Specifically, in studies reported elsewhere,
we have shown that estrogen and androgen decrease the expression of the
gp130 transcript and its protein by cells of the bone marrow stromal/osteoblastic lineage in vitro (22). Conversely,
ovariectomy in mice causes an increase in the expression of gp130
mRNA and protein in ex vivo bone marrow cell
cultures.2 In contrast to sex steroids, we
found that parathyroid hormone and 1,25-(OH)2D3
stimulate gp130 expression. A stimulatory effect of gp130 by
parathyroid hormone and 1,25-(OH)2D3 has been
also found in studies by Romas et al. (21). Furthermore,
Sidell and co-workers (20) have observed that
all-trans-retinoic acid decreases the cell surface
expression of gp130 in freshly isolated myeloma cells. Moreover, it has
been demonstrated that anti-gp130 antibodies dose dependently inhibit
IL-6-type cytokine-induced osteoclast formation (21) and Kaposi's
sarcoma cell proliferation in vitro (52); and that the
decrease in gp130 induced by all-trans-retinoic acid is
associated with inhibition of IL-6-dependent human myeloma cell growth (20). Finally, in studies reported elsewhere, we have shown
that the inhibitory effect of estrogen on gp130 is accompanied by a
decrease in the amount of activated STAT complexes induced in response
to IL-6 type cytokines (53). These latter observations may bear
relevance to the evidence that bone marrow cells from
estrogen-deficient mice are more sensitive to osteoclastogenic signals
mediated through gp130, than cells from estrogen replete animals
(54).
In conclusion, our results establish that the DNA fragment we have isolated contains the human gp130 promoter and that IL-6 type cytokines can influence the activity of this promoter via activated STATs. In addition, the in vitro demonstration of a STAT mediated up-regulation of gp130 suggests a mechanism whereby activation of gp130 by its ligands may modulate the production of new gp130. Whether modulation of gp130 production by its cognate ligands and/or other agents may serve to replenish receptor consumed upon ligand activation and/or to modulate the sensitivity of cells to the actions of IL-6 type cytokines will of course require future studies, as will the mechanisms by which other cytokines and hormones regulate gp130 production.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) U70617[GenBank].
We thank T. Bellido and R. Jilka for useful discussions, T. Kishimoto for reagents, and N. Stahl and J. Darnell for reviewing the manuscript.