(Received for publication, January 25, 1996, and in revised form, October 25, 1996)
From the Department of Biochemistry and Molecular
Biology, University of Georgia, Athens, Georgia 30602, the
¶ Department of Biochemistry and Molecular Biology, University of
Kansas Medical Center, Kansas City, Kansas 66160, and
Athena
Neurosciences, Inc., South San Francisco, California 94080
Both astrocytes in the central nervous system and fibroblasts in somatic tissues are not only the major sources of extracellular matrix components but also of matrix metalloproteinases (MMPs), a family of enzymes directly involved in extracellular matrix breakdown. We have analyzed the regulation of the expression of MMPs and TIMPs (tissue inhibitors of metalloproteinases) in human primary astrocytes stimulated with oncostatin M (OSM) and other extracellular mediators in comparison with normal human dermal fibroblasts. It was found that OSM induced/enhanced transcription of MMP-1 (interstitial collagenase) and MMP-3 (stromelysin 1) in astrocytes, and MMP-1, MMP-9 (gelatinase B), and TIMP-1 in fibroblasts. Analysis of the signal transduction leading to activation of the MMP-1 gene revealed the presence of an OSM-responsive element (OMRE) encompassing the AP-1 binding site and the signal transducer and activator of transcription (STAT) binding element, which mediate activation by OSM. OMRE is also present in the TIMP-1 gene promoter and, although there are some differences in these two motifs, both appear to be targets for the simultaneous action of OSM-induced nuclear effectors. The induced enhancement of transcription by synergistically acting AP-1 and STAT binding elements in response to OSM is Raf-dependent. Cross-talk between the mitogen-activated protein kinase and JAK-STAT pathways is required to achieve maximal induction of the OMRE-driven transcription by OSM.
Astrocytes and fibroblasts each produce a complex mixture of extracellular matrix components (1), turnover of which is an integral part of many physiological processes such as development, cell migration, angiogenesis, tumor invasion (2), and neurodegenerative diseases (3, 4). The transcriptional regulation of matrix metalloproteinases (MMPs)1 has been observed in these processes (5, 6), and after extracellular activation the activity of these enzymes can be controlled by specific TIMPs (5, 7, 8).
The expression of MMPs is greatly modulated by cytokines and growth factors (reviewed in Refs. 6, 7, 8), which induce cellular responses by activating intracellular signaling cascades including MAP kinase (9, 10) or JAK-STAT (11) pathways via specific cell surface receptors. Activated MAP kinases enter a nucleus and activate (or inactivate) transcription factors by phosphorylation of serine or threonine residues. Genetic and biochemical studies have revealed a series of MAP kinase nuclear substrates including those directly involved in MMPs transcriptional regulation such as the gene products of fos and jun oncogenes which compose the AP-1 transcription factor and ets family members (10, 12, 13, 14, 15). The coordinated up-regulation of MMP-1 and TIMP-1 genes by growth factors, as well as by several nuclear and non-nuclear oncogenes (6, 16), can be explained by the combined actions of transcription factors operating through composite Ets/AP-1 motifs of the type first identified as ras-responsive elements in some mammalian promoters, including MMP-1 and TIMP-1 (16, 17).
Oncostatin M (OSM) is a multifunctional cytokine that affects the growth and differentiation of a variety of cell lines and is produced by activated monocytes and T lymphocytes (18, 19, 20). Recently, OSM was demonstrated to be an immediate early gene induced by multiple cytokines through the JAK-Stat5 pathway (21). This factor exerts a negative effect on the growth of some tumor cell lines and also acts as a positive growth regulator of normal fibroblasts, AIDS-related Kaposi's sarcoma and a TF-1 cell line (22, 23, 24). In addition, OSM induces differentiation of the M1 murine myeloid leukemia cell line (25), modulates the function of endothelial cells (26), up-regulates the expression of the TIMP-1 in human fibroblasts (27), and stimulates acute phase protein expression in hepatocytes (28).
Together with leukemia inhibitory factor, IL-6, and IL-11, OSM is a
member of a family of pleiotropic cytokines that share (20, 29, 30) a
common signal transducer, gp130, which is the component of a
multiple subunit cell surface receptor (31). More recently, the
JAK-STAT signaling pathway has been shown to play an integral role in
intracellular signaling by the cytokine receptors (32). Signal
transducers and activators of transcription (STATs) are a unique family
of transcription factors activated by many cytokine receptors (11). The
ligand-receptor interaction brings the receptor-associated JAK kinases
into apposition, enabling them to recruit a latent cytoplasmic STAT
member family. The SH2 domains of STAT proteins interact with sites of
receptor tyrosine phosphorylation and become accessible for the
activated JAKs. Following activation by tyrosine phosphorylation, the
STAT proteins form homo- or heterodimers that are nucleus translocated
where they bind to specific regulatory elements in target genes
(reviewed in Refs. 11 and 33).
Although astrocytes appear to perform a function in the central nervous system similar to that of fibroblasts in somatic tissue, only limited information is available about the regulation of MMP expression in these cells (54, 74). Analysis of the expression of MMPs and TIMPs in human primary astrocytes stimulated with OSM and other extracellular mediators in comparison to human fibroblasts were, therefore, performed in order to determine whether this cytokine modulates MMP gene expression in these cell lines. The mechanisms by which OSM activates transcription of MMP-1 (interstitial collagenase) were analyzed, and an OSM-responsive element (OMRE) in which the AP-1 binding site and the SBE element cooperate together to achieve maximal inducibility by OSM has been characterized.
Human recombinant OSM and human recombinant IL-6
were generously donated by Immunex Comp., Seattle, WA. Human
recombinant interferon and PMA were purchased from Calbiochem, La
Jolla, CA. Polyclonal anti-Jak1, anti-Jak2, and anti-Tyk2 antibodies were purchased from Upstate Biotechnology, Inc. (Lake Placid, NY).
Rabbit polyclonal anti-Stat1 and anti-Stat3 antibodies were from Santa
Cruz Biotechnology, Inc. (Santa Cruz, CA). Anti-phosphotyrosine monoclonal antibodies were from Transduction Laboratories. Human cDNA MMP-3 and TIMP-1 probes were prepared by polymerase chain reaction. Human MMP-1, MMP-2, MMP-9, and TIMP-2 cDNA probes were supplied by Dr. B. L. Marner (Washington University Medical Center). A
cDNA probe for human TIMP-3 was a gift from Dr. S. Apte (Harvard Medical School). Raf mutant expression vectors, pRSV-raf-C4 and pRSV-raf-C4PM17 (Ref. 34; no. 195) were provided by Dr. Ulf R. Rapp
(University of Wurzburg).
In order to create a
pGLMP1-Prom reporter vector, the human MMP-1 promoter (12) (525/+15
fragment) was generated by polymerase chain reaction with a human
genomic DNA template and inserted into pGL2-Basic vector (Promega) in
BglII site. In pGLMP1-Prom(muSBE), a substitutive mutation
was introduced by employing the Deng and Nickoloff method (35), using
as a mutagenic primer
5
-CACCTCTGGCTTTCT
AGGGCAAGGACTC-3
.
The oligonucleotides containing the wild-type OMRE of the human MMP-1 (12) and TIMP-1 (36) promoter fragments, with or without mutation in either the SBE site or AP-1 site (see Table I), were inserted as five repeats in a head-to-tail orientation in an XhoI site of pGL2-Promoter reporter vector (Promega). Orientation, copy number, and sequences were verified by DNA sequencing.
|
Subconfluent monolayer cultures of normal human dermal fibroblasts (NHDF; Clonetics Co., San Diego, CA) were maintained in fibroblast growth medium. Experiments with NHDF cells were performed between 3-5 passages in fibroblast growth medium supplemented with 10% fetal bovine serum (FBS), unless otherwise noted. Human normal astrocytes were cultured in minimal essential medium containing 10% FBS, 1% glucose, 1 mM sodium pyruvate, for up to 4-5 passages before they were used for experiments.
Transient TransfectionPlasmid DNA (3 µg of total DNA)
was transferred to exponentially growing cells in Dulbecco's modified
Eagle's medium supplemented with 10% FBS on six-well tissue culture
plates using the calcium phosphate-mediated transfection protocol (75).
The cells were harvested and assayed for luciferase activity according
to the Promega protocol. Transfection efficiency was assessed by
cotransfection of a -galactosidase expression plasmid pSV-
-Gal
(Promega Corp., Madison, WI); luciferase activity was normalized to
-galactosidase activity. Each transfection was performed in
duplicate and repeated three times using multiple DNA preparations.
Yield of transfection was determined by in situ
-galactosidase assays (37) and ranged from 2 to 4% or from 25 to
35%, in NHDF or human astrocytes, respectively.
Total cellular RNA was extracted from cells by the guanidinium thiocyanate method of Chomczynski et al. (38). Northern blots were made as described previously (39) and analyzed in a PhosphorImager (Molecular Dynamics, Sunnyvale, CA).
Gel Shift and Antibody Supershift AssayWhole cellular extracts (5 µg of total protein) were used for EMSA, which was performed as described previously (40) with 1 ng of 32P-labeled double-stranded oligonucleotides (Table I) as probes. DNA-protein complex formation was carried out in a binding buffer containing 10 mM Hepes, pH 7.8, 50 mM KCl, 1 mM EDTA, 5 mM MgCl2, 10% glycerol.
ImmunoblottingFor immunoprecipitation followed by immunoblotting, precleared cell lysates in IPA buffer (150 mM Tris-HCl at pH 7.4, 1% Nonidet P-40, 0.25% sodium deoxycholate, 150 mM NaCl, 1 mM EGTA, 1 mM pehnylmethylsulfonyl fluoride, 1 mg/ml aprotinin, 1 mg/ml leupeptin, 1 mg/ml pepstatin, 1 mM Na3VO4, 1 mM NaF) were incubated with indicated antibodies for 2 h at 4 °C. Immunocomplexes bound to protein A-Sepharose were collected by centrifugation and washed several times in IPA buffer. Immunoprecipitated proteins were resolved by 8% SDS-PAGE (41) and transferred to nitrocellulose membranes (Schleicher & Schuell). Immunoblots were developed by use of ECL (Amersham Corp.), according to the manufacturer's protocol.
We
compared matrix metalloproteinase gene expression in two "stromal"
human primary cell lines, which share some common functional characteristics despite their different origin. Analysis of
non-activated cells had shown the presence of MMP-1, MMP-2, MMP-3, and
MMP-9 in NHDF cells, but only MMP-2 could be observed in primary
astrocytes at a detectable amount. In both cell lines, the level of the
MMP-2 expression was unchanged after treatment with different
activators (Fig. 1). IL-1 and TNF induced MMP-1 and
MMP-3 in NHDF cells after 7 h (data not shown), but not MMP-9, and
this effect was enhanced after 24 h (Fig. 1). MMP-1 and MMP-9
expression was up-regulated by IL-1 in human astrocytes, whereas MMP-3
was not (Fig. 1).
When NHDF cells were exposed to OSM, an increased level of MMP-1
mRNA (2.5 times) was observed (Fig. 1). The treatment of NHDF cells
with OSM for 24 h in the presence of IL-1, TNF
, or fibroblast growth factor resulted in enhancement of MMP-1 mRNA up
to 12, 24, or 9.5 times, respectively, while mediators without OSM
induced synthesis only 5.8, 8.8, and 4 times (Fig. 1). OSM alone or
together with IL-1
, TNF
, and fibroblast growth factor transiently
enhanced MMP-9 after 7 h in NHDF cells (data not shown). In human
astrocytes, MMP-3 was induced only when cells were challenged with OSM
and IL-1 simultaneously (Fig. 1). The cooperation between signals
induced by OSM in presence of either IL-1 or PMA resulted in 100-fold
or 365-fold, respectively, stimulation of MMP-1 in human astrocytes
(Fig. 1). Northern blot analysis did not detect an increase of
endogenous MMP-1 mRNA in response to OSM alone in these cells,
perhaps, because basal MMP-1 expression was below the level of
detection (Fig. 1). However, the MMP-1 promoter was responsive to OSM
in NHDF cells as well as in human astrocytes (Fig.
2).
The role of OSM in the regulation of the human endogenous matrix
metalloproteinase inhibitors, TIMP-1, TIMP-2 (Fig. 1), and TIMP-3 (data
not shown), was investigated. In human astrocytes, TIMP-1 was strongly
activated by IL-1 and PMA, and only weakly by OSM (Fig. 1). We did
not detect any up-regulation of the other two inhibitors in these
cells, apart from TIMP-3 activation by PMA. In contrast, TIMP-1 was not
induced by IL-1 or by TNF
in NHDF cells, similar to results obtained
with both synovial fibroblasts and endothelium (42), but was increased
significantly when the cells were exposed to either OSM or PMA.
In order to determine the mechanism by which
transcription of MMP-1 can be regulated by OSM, we addressed the
question of which regulatory element(s) in the MMP-1 gene confers
responsiveness to OSM. It was demonstrated that a 4-fold and 5-fold
activation of the MMP-1 promoter (525/+15) by OSM in transfected NHDF
cells and human astrocytes, respectively, could be attained (Fig. 2). Recently, the JAK-STAT signaling pathway has been clearly linked to
OSM-mediated cell activation (32). Our sequence analysis has revealed
the presence of an element homologous to the STAT binding site (SBE)
(33, 43, 44, 45) in the 5
-flanking region of the human MMP-1 gene between
53 and
45. It should be pointed out that the SBE is only partially
responsible for OSM-mediated activation of the MMP-1 promoter, since
mutation of this site did not completely remove OSM inducibility (Fig. 2). Thus, SBE appears to function in conjunction with other elements of
the MMP-1 promoter, which, together with SBE, could be required in
order to achieve a full responsiveness to OSM. It has been demonstrated
previously that the positive and negative effects of tumor promoters,
oncogene products, and growth factors on MMP-1 transcription is at
least in part mediated by the proximal AP-1 site (12, 13, 14, 46). We
demonstrated that the multimerized wild-type AP-1-SBE element
conferred a very strong responsiveness to OSM, up to 7-fold or
16-fold in NHDF cells and human astrocytes, respectively (Fig. 2).
Using the probe containing both the SBE and AP-1 sites of the MMP-1
promoter and different oligonucleotides for competition, we mapped the
OSM-induced pattern of DNA-binding proteins. The most retarded bands
contained SBE element-binding proteins, as these components were easily
out-competed when oligonucleotides with an SBE site or SIE were present
in excess during DNA-protein complex formation (Fig. 3).
Significantly, the OSM-induced protein binding to the SBE element was
activated very early, indeed within minutes. On the other hand, faster
migrating complexes were composed of the AP-1-binding proteins, and
addition of any oligonucleotide with a typical AP-1 site resulted in
their complete disappearance. Increase in binding to the AP-1 site was
also noted, although not as rapid as to the SBE element, and the
maximal level of induction was reached in approximately 1 h. The
binding of transcription factors to SBE site in response to OSM was not
dependent on binding to the AP-1 site (and vice versa), as
was demonstrated in EMSA using as probes AP-1/SBE motifs with either
AP-1 or SBE sites mutated (data not shown). Very similar results were
demonstrated in parallel experiments, where a DNA fragment
corresponding to 97/
75 DNA sequence in the human TIMP-1 promoter
was used as a probe (Fig. 3B), although some differences in
binding to SBE motifs were detected.
Activation of JAK-STAT Signaling Pathway by OSM
In order to
examine intracellular signaling in NHDF cells induced by OSM, we
investigated the activation of an OSM receptor-associated tyrosine
kinase. Tyrosine phosphorylation of Jak1 and Jak2, but not Tyk1, was
observed as a result of cytokine activation (Fig. 4A).
In DNA gel shift assays (Fig. 3), the activation of the SBE-binding proteins by OSM was rapid and had a rather transient characteristic. In fact, the elevated DNA binding activity to the SBE element was observed only 2 min after OSM stimulation, being sustained for the next hour. However, after 1 h of incubation of NHDF cells with OSM, we could not detect any more activated STAT proteins. We also observed both Stat1 and Stat3 are nucleus translocated in response to OSM in human fibroblasts (data not shown), which correlated very well with the EMSA results.
Wagner et al. (43) and Zhong et al. (47) have
determined the pattern and composition of proteins binding to the
sis-inducible element of the fos promoter (SIE)
in DNA mobility gel shift assays using cell extracts from interferon
- or IL-6-induced cells. The composition analysis of these bands
revealed that SIFA consists of a Stat3 homodimer, while SIFC contains a
Stat1 homodimer and SIFB is a Stat1/Stat3 heterodimer (47). In Fig. 4
we show that OSM is able to induce SIFA, SIFB, and SIFC complexes in
gel mobility shift assays where the SIE element (5
-TTTCCCGTAAA-3) (43)
was used as a probe. We then examined OSM-induced DNA-binding proteins, testing as a probe the SBE element of the MMP-1 promoter
5
-ATTTCTGGAAG-3
. It appeared that the Stat1 homodimer had the highest
affinity to this element although we also observed some formation of a Stat3 homodimer and, presumably, a Stat1/3 heterodimer DNA complex. In
addition, we noticed a new OSM-induced DNA binding complex, which was
completely supershifted with antiStat1, while anti-Stat3 antibodies
failed to interact. This complex could contain a multimer of Stat1 with
another member of the STAT family or could represent an example of
cross-interaction between a STAT protein and different transcriptional
factors. In contrast, the SBE element of the TIMP-1 promoter
5
-CATCCAGGAAG-3
preferably bound the Stat3 homodimer and, to a lesser
extent, the Stat1 homodimer, but bound very little if any of the
Stat1/3 heterodimer. The transcriptional discrepancy between MMP-1 and
TIMP-1 responsiveness to OSM is at least to some extent the reflection
of differences in binding competence of these two SBE sites, resulting
from changes in DNA sequences, but the mechanism underlying such an
effect remains obscure.
We have already demonstrated that both AP-1 and
SBE sites in MMP-1 and TIMP-1 promoters are receiving an OSM-induced
signal (Figs. 2 and 3). In order to demonstrate that OMRE is a mediator of transcriptional activation in response to OSM, we constructed a
luciferase reporter gene with five repeats of OMRE (from MMP-1 and
TIMP-1 promoters) in head-to-tail orientation inserted pGL2-Promoter: pMP1-AP1-SBE(5x)-luc and pT1-AP1-SBE(5x)-luc. Human astrocytes were
transfected with these constructs and challenged with either OSM or PMA
or both. PMA was chosen, since protein kinase C was shown to be an
inducer of the AP-1 site in the MMP-1 promoter (12). We found that
pMP1-AP1-SBE(5x)-luc and pT1-AP1-SBE(5x)-luc were maximally induced
after 24 h of treatment with OSM, resulting in 16-fold and 8-fold
stimulation of transcription, respectively (Fig. 5).
Remarkably, PMA, a protein kinase C activator, failed to cooperate with
OSM in the transcriptional activation of the OMRE-driven luciferase
expression vector. We do not understand the basis of this phenomenon,
but it is possible that an elevated PMA-induced signaling pathway is
competing out a common coactivator of transcription, as was
demonstrated in case of the AP-1 and a nuclear receptor (48). Next, we
addressed the question as to whether cis-acting regulatory sequences
within OMRE of MMP-1 and TIMP-1 promoters cooperate with each other to
achieve maximal induction of transcription. Mutant constructs were
prepared, which had five repeats of OMRE in a head-to-tail orientation
inserted into the pGL2-Promoter plasmid but were defective either in
the AP-1 site or in the SBE element in a manner so that they failed to
bind transcription factors (Table I, Figs. 2
(A and B) and 5). We showed that both cis-acting
motifs in OMRE are fundamental to achieving an OSM-induced maximal rate
of transcription of OMRE-driven genes (Fig. 5). When the SBE element
was mutated, the responsiveness to OSM was reduced drastically,
although not completely. The basal level of the gene expression and
PMA-inducibility remained unchanged, indicating that the SBE site is
not involved in PMA-mediated activation. On the other hand, the AP-1
binding site is essential for both OSM- and PMA-activation and seems to
be indispensable for the basal level of gene expression in this model.
These data leave no doubt that both the AP-1 and SBE sites are
necessary to confer OSM responsiveness in OMRE-driven genes, and their
synergistical cooperation is crucial to accomplishing maximal rates of
transcription.
OSM-induced Activation of OMRE Is Raf-dependent
Since the activity of both AP-1 and
STAT can be modulated by the MAP kinase pathway, we investigated
whether there was a requirement for one of the members of the growth
promoting signal cascade, Raf-1 (49), during OMRE activation by OSM.
Raf-1 is believed to be a central cytoplasmic signal transducer in the
signaling pathway induced by many extracellular activators, but its
direct effect on OSM-induced transcriptional activation has not been previously studied. In order to check this possibility, human astrocytes were co-transfected with pMP1-OMRE(5x), with a dominant negative mutant of Raf-1, pRSV-RafC4, or pRSV-RafC4PM17 serving as a
control (34), and cells were challenged with OSM. It was demonstrated
that the Raf-1 dominant negative expression vector suppressed OSM
induction of the OMRE-driven luciferase expression vector, whereas
pRSV-RafC4PM17 was not effective (Fig. 7), which is indicative of a
Raf-1 involvement in OSM-induced signal transduction.
The recent demonstration of the activation of the STAT signaling
pathway by the oncogene for Src raises the possibility that STAT
proteins can contribute to oncogenesis by Src (50). Moreover, v-src and other oncogenes were found to activate the MMP-1
promoter (16). Src is a non-receptor tyrosine kinase, which was
demonstrated to activate Raf-1 (51) and mediate regulation of a variety
of transcription factors. Overexpression of the v-src
oncogene in human astrocytes induced a very high transcriptional
activity of the OMRE-driven luciferase expression vector and completely diminished further OSM inducibility. Cotransfection of human astrocytes with the v-src expression vector, the Raf dominant negative
mutant, and the OMRE-driven luciferase expression vector revealed that Raf is only in part involved in v-src OMRE element
activation (Fig. 6). At this point the biological
relevance of this observation is unclear.
Astrocytes constitute approximately 50% of the total cell number
within the central nervous system (52). Aside from their pivotal
function of elaborating extracellular matrix components in the brain,
these cells play essential roles in the regulation of the extracellular
environment by responding and releasing growth factors and cytokines,
which are involved in maintaining homeostasis in the brain
extracellular fluid (1, 53). In our experiments it was found that human
primary astrocytes can express several matrix metalloproteinases:
MMP-1, MMP-3, MMP-2, and MMP-9, as well as their endogenous inhibitors
TIMP-1, TIMP-2, and TIMP-3, although each was apparently under cell
type-specific differential regulation (Fig. 1). In unstimulated cells,
the only detected mRNA of all of the examined metalloproteinases
was for MMP-2 (Fig. 1). Another gelatinase, MMP-9, could, however, be
induced after stimulation with IL-1 or PMA, as was recently
demonstrated in rat astrocytes (54). IL-1
and OSM synergize to
stimulate MMP-1, particularly in human primary astrocytes, where they
show a nearly 90-fold induction over that found with either cytokine
alone (Fig. 1). Conversely, OSM, which is not an efficient inducer of
TIMP expression in human primary astrocytes, has a strong positive effect on TIMP-1 in human fibroblasts. IL-1, synergistically with OSM,
does induce MMP-3 expression in human astrocytes, whereas we have not
observed this effect in NHDF cells. Since TIMP-1, -2, and -3 are not
strongly up-regulated by OSM in human primary astrocytes, a net
positive proteolytic activity will be created upon exposure of these
cells to IL-1
and OSM. However, in NHDF cells, IL-1
and OSM,
together, not only induce MMP-1 but also TIMP-1. Thus MMP regulation
may be more tightly controlled in fibroblasts than in astrocytes,
particularly when OSM and IL-1
work in concert with each other.
There is strong evidence linking neurodegenerative diseases with an imbalanced expression of proteinases and proteinase inhibitors (3, 4, 55), and an elevated metalloendopeptidase activity has been detected in Alzheimer-affected hippocampal tissue in comparison to age-matched controls (56, 57, 58). Since MMPs have a broad specificity, it is not without reason to assume that up-regulation of their synthesis relative to that of controlling inhibitors by elevated level of inflammatory cytokines including OSM could allow then to play a major role in neurodegenerative diseases. This is consistent with the recent finding that overexpression of OSM in the central nervous system of transgenic mice was detrimental and lethal (59).
The data presented indicate that both of the cis-acting sequences, AP-1 and SBE, present in an OMRE not only are the targets for OSM-induced nuclear effectors but also cooperate together in transcriptional activation. We have also found this characteristic composite AP-1/SBE motif in promoters of two distally related human genes encoding human MMP-1 and its endogenous inhibitor TIMP-1. The requirement of AP-1 and SBE sites to confer OSM responsiveness onto the rat TIMP-1 promoter was also reported (60), but neither synergistic cooperation between AP-1 and SBE in response to the cytokine nor AP-1 activation in rat hepatocytes was noted previously. Composition analysis of SBE complexes revealed involvement of Stat1 and Stat3 in OSM-elevated transcriptional activation of MMP-1 and TIMP-1 genes (Fig. 4). Because of different DNA sequences, the SBE of the MMP-1 promoter had the highest affinity to the Stat1 homodimer, while the Stat3 homodimer preferably bound to the TIMP-1 promoter, although both SBE elements could bind to a lesser extent other homo- and heterodimers of STAT family members activated by OSM (Fig. 4). Whether this could explain how the JAK-STAT signal transduction pathway can accomplish specificity remains to be established. Moreover, OSM induces a transcription factor that binds to the AP-1 site (Fig. 2). The activity of AP-1 is the function of a large group of bZip transcriptional factors including those of the Fos and Jun families, which function as Jun-Jun or Jun-Fos dimers, forming a bimolecular DNA binding domains. Cross-family dimerization with other members of the bZip family alters DNA binding specificity (61, 62, 63). It will be interesting to further investigate the composition of OSM-induced AP-1 binding complexes and their cooperation with SBE element-binding proteins in OMRE-driven transcription machinery.
A clue to the mechanism by which OSM might activate transcription of OMRE-driven genes is beginning to emerge (Fig. 7). Clearly, OSM elicits cellular responses that cannot be limited to any single signaling cascade (JAK-STAT). We postulate that cells must exert active MAP kinase and JAK-STAT pathways in order to maintain full OMRE inducibility by OSM. Activation and cooperation between these two signaling cascades in response to OSM is required to achieve maximal transcriptional activity of the OMRE-driven luciferase vector. Raf-1 is an activator of multiple nuclear effectors, including Fos, Jun, and activating transcription factors (49, 64). Moreover, phosphorylation on serine 727 in Stat1 and Stat3 is required for their maximal transcriptional activity (65). ERK2, which is a serine/threonine kinase downstream from Raf in the MAP kinase pathway, was reported to directly convert Stat1 to its fully active form by phosphorylating serine 727 (66). Recently, it was shown that Jak2 was required for growth hormone-stimulated activation of ERK2/MAPK (67). This could explain how OSM induced signaling bifurcates to the ERK2/MAPK and STAT signaling pathways and converges with Stat1 and Stat3 phosphorylation by ERK/MAPK, and by activation of OMRE. This is consistent with our observation that suppression of endogenous Raf-1 impaired OSM inducibility of the OMRE-driven luciferase vector (Fig. 6).
The synergistic enhancement of transcription by AP-1 and STAT proteins could be accomplished by utilizing a coactivator of transcription referred to as cointegrator (48) such as a nuclear 265 kDa CREB-binding protein (68), the protein linked to transactivation by CREB, Jun/Fos, and nuclear receptors (48, 69, 70). Moreover, the binding of CREB-binding protein to CREB or AP-1 proteins is serine phosphorylation-dependent (69, 70). The strong activation of OMRE by v-src and suppression of OMRE inducibility by dominant negative Raf-1 mutant (Fig. 6) support this hypothesis, since both Src and Raf-1 kinases trigger serine kinase activity (49, 51, 64).
Alternatively, OSM has been shown to stimulate a rapid but transient elevation of primary response genes including c-Jun (71), and products of this gene can enhance OMRE, already activated by STAT transcriptional factors, as a positive feedback regulatory loop. The mechanism by which OSM activates transcription of immediate-early genes is not always clear, but most likely this induction is mediated by the JAK-STAT pathway, and the newly biosynthesized transcriptional factors could further enhance transcription of later genes such as MMPs and TIMPs.
OSM is expressed by activated monocytes and T-lymphocytes (19, 20), and it may be involved in modulation of stromal cell function at inflammatory sites since fibroblasts, astrocytes, and also endothelium (26) respond very strongly to this cytokine. The ability to bind gp130 directly as well as two other receptor subunits, leukemia inhibitory factor receptor and oncostatin M receptor (72, 73), provides OSM with a variety of activities that any other member of this cytokines family. The recent finding that a tissue-specific targeted bovine OSM in transgenic mice had a profound and often lethal effects (59) raises the possibility that OSM plays a role in the regulation of development and homeostasis. The physiological function of OSM is still unknown, and the targeted gene disruption could be the experiment of choice in order to establish the primary function of this cytokine.
We thank the following people and their co-workers for their generous gifts provided for these studies: Dr. Ulf R. Rapp for Raf-1 dominant negative mutant expression vector; Dr. Tony Hunter (Salk Institute) for v-src expression vector; Dr. B. L. Marner for human MMP-1, MMP-2, MMP-9, and TIMP-2 cDNA probes; and Dr. S. Apte for human TIMP-3 cDNA probe.