(Received for publication, January 14, 1997)
From the Molecular Neurobiology Laboratory, Massachusetts General Hospital, Department of Neurology, Harvard Medical School, Boston, Massachusetts 02114
Ciliary neurotrophic factor (CNTF)-dependent induction of expression of the neuropeptide vasoactive intestinal peptide (VIP) gene is mediated by a 180-base pair cytokine response element (CyRE) in the VIP promoter. To elucidate the molecular mechanisms mediating the transcriptional activation by CNTF, intracellular signaling to the CyRE has been studied in a neuroblastoma cell line. It has been shown previously that CNTF induces Stat proteins to bind to a site within the CyRE. CNTF also induces a second protein to bind to a C/EBP-like site within the CyRE. In this report, we show that this inducible CyRE binding protein is composed of the AP-1 proteins c-Fos, JunB, and JunD. These proteins bind to a non-canonical AP-1 site located near the previously characterized C/EBP site. The serine/threonine kinase inhibitor H7 prevents CNTF-dependent induction of AP-1 binding and CyRE-mediated transcription, suggesting that an H7-sensitive kinase is important to mediating CNTF effects on VIP transcription. The integration at the VIP CyRE of the Jak-Stat and AP-1 signaling pathways with other pre-existing proteins provides a cellular mechanism for cell- and cytokine-specific signaling.
Ciliary neurotrophic factor (CNTF)1 is
a member of the neuropoietic cytokine family, which includes
interleukin-6, leukemia inhibitory factor (LIF), oncostatin M,
interleukin-11, and cardiotrophin-1 (1-4). These structurally related
cytokines utilize a common signal transducing subunit, gp130 (5-9).
Cytokine-receptor interaction induces homo- or heterodimerization of
gp130, leading to tyrosine phosphorylation of a number of intracellular
substrates (10, 11). The transmembrane components of the CNTF receptor,
gp130 and LIFR, have no intrinsic kinase activity (9, 12) but associate with the Jak/Tyk tyrosine kinases (13-15). Activation of
these kinases by ligand-induced receptor dimerization is thought to
initiate signal transduction and induction of gene expression (13, 16,
17). Cytokine stimulation induces members of the STAT transcription
factor family, Stat1 and Stat3, to "dock" onto receptor
phosphotyrosines, enabling their own tyrosine phosphorylation (17-20).
Subsequently, STAT proteins translocate to the nucleus and bind to
conserved genomic regulatory sequences to provide a rapid means of
activating gene transcription (21, 22). These cytokines also activate
components of other intracellular signaling pathways including Ras,
mitogen-activated protein kinase (MAPK), and the Fos-Jun transcription
factors (7, 23-27). How Jak-Stat activation interacts with other
signaling pathways and transcription factors to regulate
cytokine-mediated transcription is poorly understood.
CNTF and LIF induce expression of the neuropeptide vasoactive
intestinal peptide (VIP) in sympathetic neurons in culture (28, 29) and
in a human neuroblastoma cell line, NBFL (30, 31). To gain insight into
the mechanisms underlying cytokine-mediated activation of neuronal gene
expression, we have investigated the genomic regulatory elements
mediating the induction of the VIP gene by CNTF and related cytokines.
We previously identified a large, 180-base pair cytokine responsive
element (CyRE) within the 5-flanking region of the VIP gene, which is
located 1330 bp from the transcription start site (32). The CyRE is
necessary and sufficient to mediate transcriptional activation by CNTF, LIF, IL-6, and oncostatin M (32, 33). CNTF treatment induces binding of
Stat1
and Stat3 proteins to a region of the 5
end of the CyRE (32).
This STAT site is important to transcriptional activation mediated by
the CyRE as a mutation in this site reduces the
cytokine-dependent induction of a CyRE-linked luciferase
reporter by 80% (32). As mutation or deletion of regions distinct from the STAT site also attenuate transcription mediated by the CyRE (33,
34), other regions within the CyRE appear to be important to
CNTF-mediated induction of VIP gene expression. Serine-threonine kinases are important to CNTF-dependent increases in VIP
transcription as induction of VIP mRNA by CNTF in NBFL cells is
prevented by the serine-threonine kinase inhibitor H7 (24). It is not
known whether the H7-sensitive regulation of VIP gene expression by CNTF in NBFL cells is mediated by the CyRE.
We have previously shown that three regions within the VIP CyRE have sequence homology to C/EBP binding sites and bind purified C/EBP proteins (33). While these C/EBP-related sites are important to the CNTF-mediated induction of CyRE transcriptional activity, no evidence was found that C/EBP proteins interact with these sites in NBFL cells. CNTF induces a nuclear protein complex to bind to one of these C/EBP-related sites within the CyRE. CNTF-dependent induction of this DNA-protein complex was protein-synthesis dependent, in contrast to the protein synthesis-independent induction of STAT proteins. In this report, we show that the CNTF-induced nuclear protein complex that binds to the C/EBP-related site in the CyRE is composed of AP-1 proteins and that AP-1 activation represents an H7-sensitive nuclear signaling pathway required for CyRE-mediated transcriptional activation by CNTF.
Cell culture reagents were obtained from Life
Technologies, Inc. (Gaithersburg, MD), fetal bovine serum was from
Sigma, and culture plates were from Becton Dickinson
Labware (Lincoln Park, NJ). Recombinant human CNTF was a gift from
Regeneron Pharmaceuticals (Tarrytown, NY). Oligonucleotides encoding
the consensus sites for the transcription factors AP-1, NF-B, and
CREB were purchased from Promega (Madison, WI). The remaining
oligonucleotides were synthesized in our laboratory on an Applied
Biosystems 380B DNA synthesizer. Antisera to c-Jun, c-Fos, fosB, and an
epitope common to all Jun family members were from Santa Cruz
Biotechnology (Santa Cruz, CA). Antisera to JunB and JunD were a gift
of Dr. Rodrigo Bravo (Bristol Myers Squibb, NJ). H7 was purchased from
Calbiochem (La Jolla, CA).
12-O-Tetradecanoylphorbol-13-acetate was obtained from
Sigma
NBFL cells were maintained and transfected as described previously (31). Cells were transfected by the calcium phosphate co-precipitation method. Each 10-cm plate received 20 µg of luciferase reporter plasmid and 3 µg of RSVCAT plasmid. Cytokine and/or inhibitors were added 24 h after transfection; cells were harvested 6 or 36 h later and assayed for luciferase (35) and CAT (36) activities. Luciferase activity was normalized to CAT activity to control for transfection efficiency.
PlasmidsCy1luc contains the entire 180-bp CyRE fused to
eRSVluc (32). m2CyBluc was constructed by PCR site-directed
mutagenesis as described by Ho et al. (37) with the
oligonucleotides 5
-GGTAACTGGATTAGAAAATACACTTAAGCATAGCAGG-3
and
5
-CCTGCTATGCTTAAGTGTATTTTCTAATCCAGTTACC3-
. These oligonucleotides were paired with oligonucleotide A1 or A4 (32), with Cy1luc as a
template, to create new fragments. The fragments were gel-purified and
used as template in a subsequent PCR, with oligonucleotides A1 and A4
as primers, to create m2CyBluc. Cy1mG3luc was constructed by PCR
amplification of Cy1luc with the primers
5
-CCGGGTACCTAAAAAAGATGTACTGGTATTAAGCCACAGGAACTCTGG3-
and A4. The
resultant 180-bp fragments were digested with KpnI and
PstI, gel purified, and ligated into
KpnI/PstI-digested
eRSVluc (32) to create
plasmids containing the mutant sites. Both plasmids were sequenced to
confirm their fidelity.
Synthetic
complementary oligonucleotides with GGG or GATC overhangs were annealed
and labeled with [-32P]dCTP using Superscript reverse
transcriptase (Life Technologies, Inc.) or Klenow fragment. AP-1
oligonucleotide had no overhang and was therefore labeled with
[
32P]ATP using T4 polynucleotide kinase. Nuclear
extracts were isolated and binding reactions were performed as
described previously (32). Nuclear extracts (approximately 15 µg of
protein) were incubated with 0.5 ng of labeled probe (approximately
200,000 cpm) for 20 min at room temperature before electrophoretic
separation on a 5% non-denaturing polyacrylamide gel (37.5:1) in
0.5 × Tris-borate-EDTA at 200 V. Antibodies, when used, were
added 10 min prior to the addition of the probe. The following pairs of
complementary oligonucleotides were used in DNA mobility shift assays
(mutated residues are underlined): NF-
B,
5
-AGTTGAGGGGACTTTCCCAGGC3-
and 3
-TCAACTCCCCTGAAAGGGTCCG5-
; CyB,
5
-GGGAAAATATGATTAAGCATAG3-
and 3
-TTTTATACTAATTCGTATCGGG5-
; m2CyB,
5
-GGAAAATATTAAGCATAGG3-
and
3
-TTTTATAATTCGTATCCG5-
; m3CyB,
5
-GGAAAATATGATTAAGCGG3-
and
3
-TTTTATACTAATTCGCCG5-
; m4CyB,
5
-GGATATGATTAAGCATAGG3-
and
3
-TATACTAATTCGTATCCG-5
; AP-1,
5
-CGCTTGATGAGTCAGCCGGAA3-
and 3
-GCGAACTACTCAGTCGGCCTT5-
; and cAMP
response element (CRE), 5
-AGAGATTGCCTGACGTCAGAGAGCTAG3-
and
3
-TCTCTAACGGACTGCAGTCTCTCGTCGATC5-
.
Total cytoplasmic RNA was isolated from NBFL cells and transfered to nylon membranes as described previously (33). Northern blots were hybridized with either a 1.2-kilobase NaeI fragment of human c-fos or a 0.39-kilobase BamHI fragment of jun-B and rehybridized with a probe for the unregulated internal reference gene cyclophilin to correct for loading differences.
We have previously shown that the CyB site, one of three C/EBP-like sites within the VIP CyRE, is functionally important to CNTF-dependent activation of transcription mediated by the CyRE (33). In EMSAs, we previously showed that the CyB site binds two protein complexes from NBFL cells. The more slowly migrating DNA-protein complex (referred to previously as complex III) is present in nuclear extract from untreated NBFL cells. The faster migrating complex (complex IV) is only present in nuclear extracts prepared from NBFL cells that have been treated with CNTF, LIF, or OM. Neither of these nuclear protein complexes is recognized by antibodies against known C/EBP proteins. Complex IV is induced to bind to the CyB probe within 1 h of CNTF treatment. This induction is inhibited by the protein synthesis inhibitor cycloheximide. The rapid induction and protein synthesis dependence of the induction suggested that the protein complex may be composed of immediate early gene products such as AP-1 proteins.
To test the hypothesis that AP-1 proteins are present in CyB complex
III, the CyB DNA-protein complexes were competed with excess unlabeled
consensus AP-1 binding sites. Complex IV was competed by excess
unlabeled AP-1, CRE, and CyB oligonucleotides (Fig. 1).
In contrast, the uninduced complex III is competed only by excess cold
CyB oligonucleotide (Fig. 1). These results suggest that complex III
and complex IV are composed of distinct nuclear proteins and that
proteins that bind to AP-1 and CRE oligonucleotides also bind to the
CyB site.
Inspection of the sequence of the CyB oligonucleotide revealed an
imperfect homology with a consensus AP-1 site (see Fig. 4A).
Within the CyB site, the AP-1 sequence is located 5 to the previously
identified homology to a C/EBP site (33). To ascertain whether
complexes III and IV had different DNA recognition requirements, a
series of mutations were made in the CyB region. EMSAs performed with
these CyB mutations and nuclear extracts prepared from CNTF-treated NBFL cells showed that the two CyB binding complexes were
distinguishable (Fig. 2). Mutation of three base pairs
in the AP-1 homology region abolished binding of the inducible complex
IV (m2CyB). In contrast, mutation of three base pairs immediately 3
to
this site (m3CyB) reduced the binding of complex III but did not affect
complex IV (Fig. 2B). Mutation of three base pairs 5
to the
AP-1 site (m4CyB) did not affect the binding of complexes III or IV.
Thus, complexes III and IV bind to different regions of the CyB probe, and the CNTF-inducible complex IV binds to the DNA sequence most closely resembling the AP-1 site.
To determine whether complex IV is composed of AP-1 proteins, antisera
to members of the Fos and Jun families of proteins were incubated,
prior to running an EMSA, with nuclear extract prepared from
CNTF-treated NBFL cells. Binding to the m3CyB probe was examined to
allow easier visualization of the inducible complex IV. Complex IV, but
not complex III, binding was altered by several Fos and Jun antisera
(Fig. 3). Antisera that recognize an epitope common to
Jun proteins inhibited the binding of complex IV (Fig. 4). When antisera specific to JunB and JunD were added to the AP-1
binding reaction, the more slowly migrating complex ("supershift") was formed (Fig. 4). The addition of c-Fos antisera to the m3CyB EMSA
inhibited complex IV binding. Antisera to c-Jun and fos-B did not
affect complex IV binding to the m3CyB probe. Similar results were
obtained when the native CyB site was used as a probe (data not shown).
Thus, CNTF induces an AP-1 protein complex to bind to the CyB site
within the VIP CyRE. This AP-1 complex contains c-Fos, JunB, and
JunD.
To characterize the differences between m3CyB and a consensus AP-1 site, the binding of CNTF-induced complex IV to both probes was examined. Complex IV binding to m3CyB was competed with equal efficiency by itself, AP-1, or CRE-unlabeled oligonucleotide. In contrast, CNTF-induced AP-1 binding to the consensus AP-1 site was most strongly competed by unlabeled AP-1 and most weakly by m3CyB (Fig. 4B). These data suggest that AP-1 proteins have a lower affinity for the m3CyB site than for a typical AP-1 site. The mobilities of the CNTF-induced AP-1 complex to the AP-1 consensus site, the CyB, and the m3CyB probe were identical (data not shown). Complex III did not bind to the consensus AP-1 site, nor did excess AP-1 and CRE sites interfere with complex III binding to the CyB probe (Figs. 2 and 4B), thus demonstrating that complexes III and IV have different DNA binding specificities.
We next sought to determine whether the AP-1 site within the CyRE was
important to the CNTF-dependent activation of transcription mediated by the 180-base pair CyRE linked to the luciferase reporter. A
luciferase reporter plasmid was constructed that had three base pairs
of the CyB AP-1 site mutated to produce the m2CyB site (Fig. 2) within
the context of the CyRE (m2CyBluc). The m2CyB mutation did not bind the
CNTF-induced AP-1 proteins (Fig. 2) or compete for binding of these
proteins to the CyB probe in competition assays (data not shown).
However, the m2CyB mutation did not alter the binding of the
non-inducible complex III proteins (Fig. 2). Therefore, the m2CyB
mutation was used to assess the importance of the AP-1 complex to
CNTF-mediated transcriptional activation through the CyRE. Mutation of
the CyB AP-1 sequence within the CyRE reduced CNTF-mediated induction
of luciferase by 50% compared with the native CyRE contained in the
Cy1luc reporter (Fig. 5). These data demonstrate that
the AP-1 site within the CyB region of the CyRE is required for
CNTF-dependent transcriptional activation mediated by the
CyRE and further suggest that AP-1 protein binding to this site is
important to this activation. However, a luciferase reporter containing
three multimerized copies of the CyB site directing transcription from
a basal promoter was not induced by CNTF (data not shown). Therefore,
the CyB AP-1 site appears to act in concert with other sites within the
CyRE to mediate the full CNTF-dependent transcriptional
activation mediated by the CyRE.
It has been previously shown that CNTF-mediated induction of VIP
mRNA in NBFL cells is sensitive to the serine/threonine protein kinase inhibitor H7 (24). However, binding of STAT proteins to the STAT
site within the CyRE was not inhibited by H7 treatment (32), indicating
that the H7 sensitivity of VIP mRNA induction was not due to
alteration in STAT protein binding to the CyRE (32, 38, 39). To
determine whether the AP-1-like CyB site was a location at which the H7
sensitivity of CNTF-dependent VIP transcriptional
activation was mediated, we determined whether AP-1 induction and
binding to the CyRE was inhibited by H7 treatment. We first examined
the kinetics and sensitivity of CNTF-mediated induction of mRNA
encoding the immediate early genes c-fos and jun-B, which comprise the
AP-1 proteins binding to the CyB site. H7 treatment abolished
CNTF-mediated c-fos mRNA induction and significantly attenuated
jun-B mRNA induction, also delaying its induction (Fig.
6A). H7 pretreatment of NBFL cells also
attenuated CNTF-dependent AP-1 binding (Fig. 6). H7
treatment alone did not induce binding to the AP-1 site. Similar
results were obtained using the CyB site as a probe (data not shown).
Thus, H7 inhibits the CNTF-mediated induction of c-fos and jun-B
mRNA and formation of the AP-1 complex, suggesting that the CyB
AP-1 site is a possible location through which H7 inhibits the
CNTF-mediated induction of VIP mRNA.
Multiple sites within the 180-bp CyRE or sites outside the CyRE in the
VIP promoter are potential sites of H7 inhibition. To determine whether
H7 inhibited the CNTF-mediated induction of transcription through the
VIP CyRE, NBFL cells transfected with the CyRE-luciferase reporter
Cy1luc were treated with H7, and the effect on CNTF induction was
assessed. H7 completely inhibited CNTF induction of luciferase activity
mediated by Cy1luc (Fig. 7) without significantly
reducing luciferase activity driven by RSVluc (data not shown). To
ascertain whether H7-mediated repression of CNTF-dependent
transcriptional activation was mediated by the CyB AP-1 site, NBFL
cells were transfected with the m2CyBluc reporter, and the effect of H7
on CNTF-induced transcription was assessed. If H7 acted through the CyB
AP-1 site, then mutation of this site should abrogate the H7
inhibition. Similar to its effect on the native CyRE, however, H7
completely inhibited CNTF induction of transcription mediated by a CyRE
containing a mutation at the CyB AP-1 site (m2CyBluc; Fig. 7). This
indicates that the ability of H7 to inhibit CNTF-dependent
transcriptional activation depends on other sites within the CyRE.
As serine-threonine phosphorylation has been shown to affect the
ability of STAT proteins to activate transcription (40), H7 could
influence transcriptional activation by CNTF by altering serine-threonine phosphorylation of STAT proteins. We have shown that
STAT proteins contribute to the CNTF-dependent activation of transcription by interacting with a site within the 5 region of the
CyRE (32). Therefore, we investigated the importance of the CyRE STAT
binding site to H7-mediated repression of CNTF-dependent transcription activation. H7 inhibited CNTF-dependent
transcription mediated by a CyRE reporter (Cy1mG3luc) containing a
mutation in the STAT binding site in NBFL cells (Fig. 7). The similar
inhibitory effects of H7 on the native CyRE and the CyRE STAT site
mutant indicates that H7 repression of CNTF-dependent
transcription depends on other sites within the CyRE.
These experiments demonstrate that CNTF induces an AP-1 complex to bind to a region within the VIP CyRE, the CyB site (33). AP-1 activation is necessary for the full CNTF-dependent activation of transcription mediated by the CyRE. These data present the first functional evidence for the role of AP-1 proteins in CNTF-dependent regulation of transcription. Although CNTF stimulation of AP-1 protein binding and activation contributes to CNTF-dependent transcriptional activation (Figs. 5 and 7), interaction of AP-1 proteins with multimerized CyB or canonical AP-1 sites is not able to mediate transcriptional activation in response to CNTF.2 These data suggest that AP-1 proteins binding to the CyB region participate in CNTF-mediated regulation of VIP gene expression by acting in a combinatorial fashion with other transcription factors acting at other sites within the CyRE.
The sequence within the CyB region to which AP-1 binds is a non-canonical AP-1 site. The two nucleotides that distinguish the CyB AP-1 and canonical AP-1 sites have been shown by mutational analysis to be critical for AP-1 binding (42) although AP-1 proteins can bind to this sequence in other genes (43). The sequence difference between the CyB AP-1 site and the canonical AP-1 site may be reflected in the differences in affinity of AP-1 proteins for these sites (Fig. 4).
There are two general cellular mechanisms by which CNTF may activate AP-1. Expression of the Fos-Jun components of the AP-1 complex may be induced, or phosphorylation of Fos and Jun proteins by CNTF-activated kinases may lead to a more transcriptionally active AP-1 complex. In NBFL cells, we have shown that CNTF induces c-fos and JunB mRNA (Fig. 6A and Ref. 30) and that CNTF induction of AP-1 binding is protein synthesis-dependent (33). The pathways that lead to induction of IEG transcription are complex and involve multiple pre-existing transcription factors such as STATs, serum response factor, and CREB (44-48). STAT proteins themselves appear to interact with the c-fos promoter to contribute to c-fos induction by CNTF (19). Similar pathways may activate transcription of the Jun-B gene (49) in response to CNTF or related cytokines.
The intracellular signaling pathways that lead to regulation of AP-1 activity have been partially characterized (50). The MAP kinases, JNK, ERK, and FRK increase c-Fos and c-Jun synthesis and phosphorylation, thereby increasing AP-1 binding and activation (51-53). These kinases are, therefore, candidate molecules for mediating the CNTF-induced activation of AP-1 in NBFL cells. While it has been shown that neuropoietic cytokines activate ERK 1 and 2 in several different cell types (10, 25, 26), it is not known whether other MAPKs are activated by these cytokines and act to regulate gene expression. Consistent with the involvement of MAPKs in CNTF gene regulation, we have previously shown that CNTF activates Ras, an upstream regulator of MAPKs, in NBFL cells (24). CNTF-mediated induction of AP-1 binding is also partially sensitive to H7, consistent with the role of MAPK or other serine-threonine kinases in AP-1 activation in response to CNTF.
Inhibition of serine-threonine kinases by H7 may not only inhibit AP-1
activation but may regulate the Jak-Stat pathway directly. Serine-threonine phosphorylation of STATs can increase the ability of
these proteins to activate transcription and bind DNA (54). STAT
proteins contain a MAPK phosphorylation site and demonstrate ligand-dependent association with MAPK that is required for
interferon -dependent activation of STATs (55).
Therefore, H7 may attenuate CNTF-dependent transcription by
inhibiting MAPK phosphorylation of Stat proteins and limiting their
ability to activate transcription. In NBFL cells, H7 does not affect
CNTF-induced Stat binding detected by EMSA (32), but serine-threonine
phosphorylation of Stat3 may enhance its ability to activate
transcription without a change in the amount of STAT binding (54).
The sensitivity of CNTF-induction of VIP mRNA to the serine/threonine kinase inhibitor H7 (24) supports a role of kinases, such as MAPKs or protein kinase C, in CNTF-mediated induction of VIP gene expression. We show here that this H7 sensitivity of CNTF-dependent induction of VIP mRNA in NBFL cells is mediated, at least in part, through the CyRE (Fig. 7). H7 does not appear to inhibit transcription one of the CyRE sites (STAT and CyB AP-1 sites) that we have shown to bind inducible proteins (Stat1, Stat3, and AP-1) in response to CNTF. Instead, H7 may be acting at a point further upstream in CNTF-mediated nuclear signaling, inhibiting a kinase or kinases which have effects on multiple regions of the CyRE. If H7 sensitive kinases are acting through both the AP-1 and Stat sites, then the transcriptional activity remaining when one site is mutated would be abolished by H7 effects at the other. If these effects are mediated by two separate H7-sensitive kinases, then use of more specific kinase inhibitors than H7 may assist in clarifying this issue. Alternatively, H7-sensitive kinases may affect proteins binding to regions of the CyRE other than the AP-1 and STAT sites, affecting activity of a transcription factor we have not yet identified.
The mechanism by which Jak-Stat and AP-1 intracellular signals
integrate to regulate CNTF transcription through the CyRE is not known.
AP-1 and STAT transcription factors, which appear to be required for
full CNTF-dependent transcription by the CyRE, may directly
interact to activate transcription (56). A direct interaction between
c-Jun and a STAT protein (Stat3) binding to adjacent AP-1 and STAT
sites in the IL6-responsive element of the
2 macroglobulin gene
promoter has been postulated to result in synergistic transcriptional
activation (56). Direct protein-protein interaction between STATs and
AP-1 proteins could stabilize protein-DNA binding complex formation and
contribute to CNTF-mediated transcriptional activation in NBFL cells.
However, there is approximately 100 bp separating the Stat and AP-1
sites with many other proteins binding to intermediate sites (33, 41).
Therefore, the integration of these two signaling pathways may be
through complex formation with other basal or cell-specific
factors.
The composition of the AP-1 complex with different members of the Fos/Jun family may have important biological consequences (57, 58). Jun family members have similar DNA binding and dimerization domains but differ in their activation domains (59, 60). JunB and JunD are weaker transcriptional activators than c-Jun and may antagonize its function (59, 61, 62). The JunB gene is often activated during differentiation, in contrast to the c-Jun gene that is induced during proliferation and cell death (57, 58, 63). Fos and JunB genes are often co-induced in the nervous system in response to a variety of extracellular signals (64, 65). It has been previously shown that neuropoietic cytokines induce the c-Fos and JunB genes in several types of transformed and primary cells (7, 23, 27, 30, 66). In NBFL cells, cytokine treatment induced an AP-1 complex that contained c-Fos, JunB, and JunD. CNTF does not produce a proliferative response in NBFL cells but induces several neuropeptide genes that mimic the cytokine-mediated differentiation of sympathetic neurons (31, 67). Thus, induction of AP-1 proteins (JunB, in particular) in NBFL cells by CNTF is another example of the participation of this transcriptional activator in differentiation-related process.
A model summarizing CNTF-induced signaling to the VIP CyRE is shown in
Fig. 8. AP-1 and Stat proteins are induced by CNTF treatment to bind to sites in the VIP CyRE. Stat proteins also bind to
the promoters of immediate early genes and may thereby participate in
the transcriptional induction of AP-1 proteins. AP-1 and Stat-mediated
transcriptional activation is sensitive to the protein kinase inhibitor
H7, indicating possible sites where H7-sensitive kinases may act during
CNTF-mediated induction of CyRE transcription. The AP-1 complex
contributes to CNTF-mediated transcriptional activation of the VIP
CyRE, acting in a combinatorial manner with other proteins binding to
the CyRE. Many diverse extracellular stimuli activate AP-1 proteins
(68). Integration of the AP-1 and Jak-Stat signaling pathways at the
CyRE with other DNA binding proteins provides a mechanism whereby the
VIP CyRE mediates transcriptional regulation by neuropoietic cytokines
in a cytokine- and cell-specific manner (34). Integration of multiple
signaling pathways with pre-existing proteins at the CyRE may provide
biological specificity of cytokine action leading to selective effects
of neuropoietic cytokines on differentiation, proliferation, and
survival in susceptible cells in the nervous system.
We thank Drs. Prithi Rajan, Michael Schwarzchild, and Susan Lewis for many helpful discussions, Dr. Peter Sasieni for statistical assistance, and Regeneron Pharmaceuticals for the generous gift of CNTF.