(Received for publication, December 19, 1996, and in revised form, February 14, 1997)
From the Department of Medicine, Cedars-Sinai Research Institute-UCLA School of Medicine, Los Angeles, California 90048
Using murine AtT20 pituitary cells transfected
with a rat pro-opiomelanocortin (POMC) promoter (706/+64) linked to
the luciferase reporter, we showed leukemia inhibitory factor (LIF) to
strongly potentiate corticotropin-releasing hormone (CRH) induction of POMC gene expression. We therefore tested mechanisms for molecular interactions between LIF and CRH. Although LIF and CRH synergized to
induce an 8-fold induction of POMC transcription, CRH alone (but not
LIF) induced cAMP response element-binding protein phosphorylation (5-fold) or an increase of c-fos mRNA levels
(>100-fold), suggesting that these pathways are not implicated in LIF
transcriptional synergistic effects. Using a DNase I footprint assay,
POMC promoter regions protected by AtT20 cell nuclear extracts were
identified (
121/
109, and
143/
134, and
173/
160). The
protected
173/
160 element fused to a heterologous promoter
conferred LIF-CRH synergy (6.5-fold induction of POMC) and formed a
specific complex with AtT20 cell nuclear extracts. This complex was
supershifted by an anti-phosphoserine antibody, and a serine/threonine
kinase inhibitor also altered both this complex and LIF-CRH
transcriptional synergy on the POMC promoter-luciferase reporter
construct, indicating that these events depend on post-translational
serine phosphorylations. LIF-CRH synergy on POMC transcription is
therefore mediated at least in part by
173/
160 sequences conferring
confluent transcriptional activity of both peptides.
CRH,1 a well characterized 41-residue
neuropeptide, is the principal central regulator of the stress response
(1). It stimulates transcription and biosynthesis of POMC, leading to
increased adrenal glucocorticoid production (2). CRH acts through
specific CRH receptors (G protein-coupled receptors) to activate
adenylate cyclase, thereby increasing intracellular cAMP levels (3). Intriguingly, although cAMP agonists also induce POMC transcription, the POMC gene does not contain a consensus cAMP response element (4).
However, CRH also activates the c-fos proto-oncogene, implying cAMP and protein kinase A or Ca2+ and calmodulin
kinase signaling (5). Furthermore, a c-fos-binding element
in POMC exon 1 (indistinguishable from an AP-1-binding site) has
recently been identified to be responsible for part of the CRH
transactivating effect on POMC transcription (4). Another CRH-inducible
element centered at position 166 on the POMC promoter has also been
described and could be responsible for c-fos-independent
induction of POMC transcription by CRH (6).
The development and function of differentiated pituitary neuroendocrine
cells are regulated both by hypothalamic trophic hormones and by
intrapituitary cytokines and growth factors. Several cytokines stimulate the hypothalamic-pituitaryadrenal axis in
vivo and pituicyte ACTH production in vitro (7-10).
LIF is a pleiotropic cytokine, whose gene expression has recently been
reported in human fetal and adult and murine pituitary cells (11, 12).
The biological functions of LIF are mediated through its binding to a
high-affinity cell-surface LIF-receptor complex, consisting of a
low-affinity LIF-binding subunit (LIF-R) and a gp130 subunit (13).
Ligand activation of LIF-R is followed by gp130 and LIF-R
heterodimerization and subsequent activation of tyrosine kinases (14).
The Janus-activated kinase family of transmembrane receptor-associated
tyrosine kinases is phosphorylated by LIF (15), leading to subsequent
activation and translocation into the nucleus of STAT1
(ignal ransducer and ctivator of
ranscription) and STAT3 (16-19).
LIF also synergizes with CRH to induce secretion of ACTH (20-fold) and
to stimulate POMC gene expression (8-fold) in AtT20 cells (16).
However, the mechanism for this potent synergy is difficult to explain
as no specific cytokine response elements have thus far been identified
in the POMC gene. In this study, we therefore explored mechanisms of
interaction between the LIF and CRH signaling cascades to determine
points of confluence that might explain the synergy exerted by both
peptides on POMC transcription. We report here that LIF does not
influence the cAMP/c-fos-dependent activation of
POMC transcription by CRH. However, a LIF- and CRH-responsive element
has been identified in the POMC gene (173/
160) and is defined to be
responsible for mediating LIF-CRH synergy on POMC transcription.
AtT20 cells, obtained from the American Type Culture Collection (Rockville, MD), were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 2 mM L-glutamine, 100 units/ml streptomycin, 100 units/ml penicillin, and 0.25 µg/ml amphotericin B (Life Technologies, Inc.).
Western AnalysisAtT20 cells were grown in 10-ml dishes to
80% confluency and serum-deprived 24 h before a 30-min
treatment with LIF (R&D Systems, Minneapolis, MN), CRH (American
Peptide Co., Sunnyvale, CA), LIF and CRH, forskolin, or phorbol
12-myristate 13-acetate (Sigma). Cells were then lysed by scraping into
2 × SDS-polyacrylamide gel electrophoresis loading buffer.
Samples were separated by electrophoresis on 10% SDS-polyacrylamide
gel, and proteins were transferred to polyvinylidene difluoride
membrane (Millipore Corp., Milford, MA). The membrane was blocked in
5% nonfat milk. Anti-phospho-CREB detection was carried out with
polyclonal anti-phosphorylated CREB antibody (Upstate Biotechnology,
Inc., Lake Placid, NY) in blocking buffer for 14 h at 4 °C.
Anti-rabbit Ig conjugated to horseradish peroxidase (Amersham Corp.)
was incubated with the membrane in blocking buffer, and detection was
accomplished using ECL reagent (Amersham Corp.) as suggested by the
manufacturer. To reprobe the membrane, the antibody was stripped using
100 mM -mercaptoethanol, 2% SDS, and 62.5 mM Tris-HCl (pH 6.7) at 50 °C for 30 min.
Poly(A) RNA was prepared from AtT20 cells treated with LIF, CRH, or both using trizol reagent (Life Technologies, Inc.). 10 µg of RNA/lane was size-fractionated under denaturing conditions using formaldehyde-agarose (1.2%) gel, transferred to nitrocellulose, and hybridized as described (20). A cDNA fragment of c-fos was used as a template for a 32P-labeled probe generated using random primers.
Nuclear ExtractsAtT20 cells were grown to 80% confluency and serum-deprived 24 h before a 1-h treatment with or without LIF, CRH, or both. Cells were then harvested in cold phosphate-buffered saline. Nuclear extracts were prepared as described (21). Final concentrations were typically 3 µg/µl as determined by Bio-Rad protein assay.
DNase I FootprintingA double-stranded DNA fragment was
prepared by polymerase chain reaction. Briefly, the antisense
primer was end-labeled using T4 polynucleotide kinase and
[-32P]dATP (6000 mCi/mmol; DuPont NEN), purified
through a Bio-Spin 30 chromatography column (Bio-Rad), and used in the
polymerase chain reaction. The
351/
7 region of the rat POMC
promoter was also amplified and purified on a 1.3% agarose gel. DNA
was recovered using a USBioclean kit.
Nuclear extracts (0-50 µg) were incubated for 20 min at room temperature with 1-2 ng of 32P-labeled DNA (25,000 cpm) in binding buffer (22 mM HEPES (pH 7.9), 6 mM KCl, 10 mM dithiothreitol, 5 mM spermidine, 8% glycerol, and 2% Ficoll) with 2 µg of poly(dI-dC). Samples were briefly treated (1 min at room temperature) with DNase I (Pharmacia Biotech Inc.), titrated to determine a partial digestion pattern. Digestion was terminated by addition of 3 volumes of 768 mM sodium acetate, 128 mM EDTA, 0.56% SDS, and 256 µg/ml yeast RNA and by treatment (5 min at 55 °C) with 2 µg of proteinase K. DNA was also extracted with phenol/CHCl3 and ethanol-precipitated. Reaction products were then analyzed on a standard 8 M urea, 8% polyacrylamide gel adjacent to a sequencing ladder (G + A) generated from the same end-labeled DNA fragment to precisely identify borders of DNase protection.
Gel-shift AssayAll nucleotides were synthesized by the
Cedars-Sinai Molecular Biology Core Facility. The superscript numbers
are in reference to POMC promoter sequence with the transcription start
site as position +1: oligonucleotide
A,5-
175tcgaTCTGCTGTGCGCGCAGCC
158-3
;
oligonucleotide B,
5
-
189ACTTTCCAGGCACATCTGCTGT
168-3
; and
oligonucleotide C,
5
-
167GCGCGCAGCCCCGACCGGGAAG
146-3
.
10 pmol of double-stranded oligonucleotide were end-labeled with
[-32P]dATP using T4 polynucleotide kinase,
purified through a Bio-Spin 30 chromatography column, resolved on a
10% acrylamide gel, and eluted from the gel in 10 mM
Tris-HCl (pH 8) and 1 mM EDTA. Approximately 30 pg of
labeled DNA (20,000-30,000 cpm) were added to the nuclear extracts.
For the binding assays, 7.5 µg of nuclear extracts were preincubated
for 15 min at room temperature in 20 µl of binding buffer with 1 µg
of poly(dI-dC). 32P-Labeled probe was added, and the
binding reaction was left at room temperature for 15 min. In
competition experiments, a 100-fold molar excess of unlabeled
competitor oligonucleotides was added to the preincubation reaction.
Antibody experiments were performed using 1 µl of
anti-phosphotyrosine antibody (Santa Cruz Bioreagents, Santa Cruz, CA)
or 1 µl of anti-phosphoserine antibody (Sigma), which was added to
the preincubation reaction and incubated for 1 h at 4 °C. The
protein-DNA complexes were resolved on a 4% polyacrylamide gel in
0.5 × Tris borate/EDTA. Gels were dried and autoradiographed with
intensifying screens at 70 °C with Amersham Hyperfilm-MP films.
Deletion of the
167/
117 region of the POMC promoter was performed with an ExSite
polymerase chain reaction-based mutagenesis kit (Stratagene Cloning
Systems, La Jolla, CA) according to manufacturer's instructions. The
173/
160 element-luciferase plasmid was made by synthesizing two
complementary DNA sequences (5
-tcgaTCTGCTGTGCGCGCAGC-3
), annealing,
and ligating them into pGL3 (Promega) linearized at the
KpnI/SacI site. All constructs were sequenced for
verification.
AtT20 cells were plated
in 2-ml dishes (100,000 cells/well) and allowed to adhere for 24 h. Cells were transfected as described (16), and 24 h after
transfection, they were treated in triplicate for 6 h. After cell
lysis, the luciferase activity was measured in a Berthold Lumat LB 9501 luminometer (Wallac, Gaithersburg, MD). Cotransfection of the
-galactosidase reporter showed transfection efficiency to vary
<15% within a given experiment.
Using the rat POMC
promoter (706/+64) fused to the luciferase reporter gene, we
previously showed that LIF induced CRH action on rat POMC transcription
(16). To determine the specificity of LIF action, the transcriptional
effects of other cytokines, notably TNF
and IL-1
, which are also
induced by acute inflammatory insults (22, 23), were examined. AtT20
cells transiently transfected with the POMC promoter-luciferase
construct were treated for 6 h with LIF (1 nM), CRH
(10 nM), TNF
(1 ng/ml), IL-1
(100 pg/ml), or
combinations of CRH and one of the other cytokines. Neither TNF
nor
IL-1
alone exerted transcriptional effects, nor did they alter CRH
induction of POMC (Fig. 1). Thus, there does not appear
to be redundancy in the response of POMC to these inflammatory cytokines, i.e. the LIF synergistic effect with CRH on POMC
transcription appears selective. Presumably, the specificity of the LIF
response is determined by LIF receptor expression (11). Urocortin, a novel neuropeptide related to both urotensin and CRH, activates CRH
receptors and, in vivo, is a more powerful stimulant of the hypothalamic-pituitary-adrenal axis than CRH (24). We therefore examined urocortin transcriptional activity and demonstrated that it
could also activate POMC transcription as well as potentiate LIF
effects on POMC (Fig. 1). These results provide evidence for the role
of CRH receptor subtype I (25) in modulating LIF action. Therefore, to
determine points of confluence between the two signaling cascades, we
explored the hypothesis by which LIF could enhance CRH signaling
pathways leading to POMC transcription and ACTH secretion.
LIF and CRH Effects on the cAMP Pathway
CRH induction of POMC transcription was reported to be mediated through a cAMP pathway (3, 26-29), although LIF itself did not alter AtT20 cAMP levels (16). cAMP signaling leads to phosphorylation of CREB and to a transient increase of c-fos mRNA levels (30-32). We therefore tested whether LIF could activate these two events.
CREB PhosphorylationAs CRH stimulates accumulation of cAMP
and protein kinase A activation in target cells (16, 26), we tested
whether LIF altered CREB phosphorylation. Using an antibody that
exclusively recognizes the phosphorylated form of CREB (33), CRH (10 nM), but not LIF (1 nM), induced a striking
increase (>5-fold) in the proportion of intracellular phosphorylated
CREB within 30 min (Fig. 2): a doublet that migrated
with the 43-kDa marker, the expected size for CREB, was detected by an
anti-phospho-CREB antibody. These two inducible bands may be CREB
and CREB
, which differ in sequence by 14 amino acids (33). These
changes were not accompanied by increases in the amount of total
intracellular CREB, as measured by an antibody recognizing all forms of
CREB. Interestingly, although LIF and CRH powerfully synergize to
stimulate POMC expression, LIF does not further enhance CREB
phosphorylation over that observed with CRH treatment. We concluded
that CREB phosphorylation is therefore not a molecular event
responsible for LIF synergistic effects on CRH-induced POMC
transcription.
c-fos Expression
c-fos is a member of a family of
related acute-phase transcription factors that form heterodimers to
bind DNA and activate transcription (34). c-fos mRNA
accumulation occurs during treatment of AtT20 cells with CRH, and
overexpression of c-fos results in enhanced POMC
transcription (5). In addition, LIF may, in some systems, enhance
c-fos transcription (35). Therefore, changes in
c-fos mRNA were analyzed after cells were treated for 15 min with LIF and CRH (1 and 10 nM, respectively). Striking
inductions (>100-fold) of c-fos mRNA after CRH
treatment were observed, in keeping with previous reports (5). However,
in these cells, LIF alone failed to induce c-fos expression
and furthermore failed to enhance the CRH effect on c-fos
mRNA levels (Fig. 3). Therefore, although
c-fos may play a role in mediating CRH action, it does not
appear to be important for LIF action or, more significantly, for
mediating the strong synergy between LIF and CRH.
In addition, c-fos has been shown to stimulate POMC transcription through an AP-1 site located in exon 1 of the POMC promoter. Using this site as a probe in the gel-shift assay or a construct containing the POMC promoter mutated at this site and fused to the luciferase reporter, we showed this site not to be critical in the context of LIF and CRH synergy on POMC transcription (data not shown).
LIF Action on CRH-induced POMC Transcription, Implying a c-fos-independent PathwayCRH may also induce POMC transcription
and ACTH secretion by a c-fos-independent pathway (4). A
region in the POMC promoter (171/
160), different from the exon 1 AP-1 site, was shown to confer strong CRH stimulation on POMC
transcription (6). We therefore explored whether this element could be
the point of convergence of the CRH and LIF pathways.
Using a DNase I footprint assay, we examined POMC promoter regions
whose DNase sensitivity was protected by bound nuclear proteins present
in AtT20 whole cell extracts previously treated for 1 h with CRH
(10 nM), LIF (1 nM), and LIF + CRH (1 and 10 nM, respectively) (Fig. 4). We focused on
351/
7 footprinting sequences as this region contains the important
regulatory element for the synergy of LIF-CRH on POMC transcription
(16).
50 µg of AtT20 cell nuclear extracts protected three elements
(109/
121,
134/
143, and
173/
160) in this region, both in the
basal state and after stimulation of the cells by LIF, CRH, or both
agents together. These regions have previously been designated as
potential overlapping AP-2- and NF-
B-binding sites (
134/
143 element) (6) and as a mouse metallothionein metal regulatory element
(
173/
160 element) (6). However, we also found that no additional
regions were further protected after treatment of the cells with both
LIF and CRH compared with LIF or CRH alone.
To further analyze whether these regions could be important for the
synergy of LIF-CRH on POMC transcription, we performed functional
assays by transfecting AtT20 cells treated with LIF (1 nM)
and CRH (10 nM) for 6 h with the rat POMC promoter
lacking the 167/
117 region and fused to a luciferase reporter gene
(Fig. 5). This deletion reduced basal promoter activity
and also severely reduced the LIF-CRH synergy (61.5%) on reporter
activity. We therefore fused the
173/
160 element to the same
luciferase reporter gene with a heterologous SV40 promoter (Fig. 5). We
focused on this element first because of its identification during
footprint analysis and also because of previous work demonstrating that
this region was bound by a CRH-inducible factor and was the major
CRH-responsive element in the POMC promoter (6). As expected, CRH
exerted its positive regulation of transcription, but surprisingly, LIF and CRH together strongly enhanced this induction (6.5-fold), thus
conferring LIF responsiveness to the heterologous construct.
To confirm these results, we analyzed binding protein states in
the 173/
160 region by using AtT20 whole cell extracts (after cell
treatment with LIF, CRH, or both) and a radioactive probe comprising
this element using a gel-shift assay. Nuclear extracts from control,
CRH-treated (10 nM, 60 min), LIF-treated (1 nM, 60 min), or (LIF + CRH)-treated (1 and 10 nM, respectively;
60 min) AtT20 cells were prepared as described under "Materials and Methods." These extracts were also shown to bind efficiently to oligonucleotide A, containing the region (
173/
160) important for
the LIF + CRH induction of POMC transcription (Fig.
6A). In fact, three complexes were shifted,
and therefore, to determine the exact binding localization of each on
the A probe, two oligonucleotides were designed: B and C. After
addition of B and C sequences, the entire A nucleotide sequence was
covered without overlap, while A-C were the same size. From these
experiments, we concluded that only the slowest migrating band
(designated Y) was specific for the
167/
158 region as it appeared
only with the A and C probes (Fig. 6, A and C,
respectively). The "intermediate" complex (designated I) seemed to
bind the
175/
168 region (overlapping region between the A and B
probes) (Fig. 6, A and B), but was not further
considered because of its diffuse nature. The faster migrating band
(designated L) was not specific for any region as it appeared with all
three probes (A-C) (Fig. 6, A-C).
To analyze the binding specificity of the Y complex on the A probe,
competition experiments were performed (Fig. 7). The Y complex was shown to be efficiently competed by unlabeled
oligonucleotide A (100-fold molar excess) (Fig. 7, lanes 5 and 7), whereas an AP-1 oligonucleotide was inefficient in
competing (lanes 9-11). The binding specificity of the Y
complex on the 167/
158 motif was further confirmed by a less
efficient competition with a unlabeled oligonucleotide similar to
oligonucleotide A except that it contained a mutation at GCGC
(
167/
164) (Fig. 7, lanes 12-14). This
167/
164 region must therefore be at least a component of the binding site of
the Y complex. In addition, competition experiments with unlabeled oligonucleotides B and C further strengthened the binding specificity of the Y complex on the
167/
158 region as oligonucleotide B did not
compete this binding, whereas oligonucleotide C competed effectively
(Fig. 6A, lanes 6 and 7).
In addition, Figs. 6A and 7 show that greater amounts of the
shifted Y complex were detected in assays that contained nuclear extracts from AtT20 cells treated with CRH for 60 min as compared with
untreated cell extracts. No quantitative enhancement of the shifted Y
complex was observed with extracts from AtT20 cells treated with LIF + CRH for 60 min as compared with those treated with CRH alone for 60 min. We also concluded that LIF, at the concentrations used and for the
duration of the treatments, did not further enhance CRH-induced complex
shifts or the amount of Y complex. LIF and CRH may also act through the
same binding element (173/
160) by activating similar transcription
factor(s) by modifying their conformation or phosphorylation state.
Indeed, no significant correlation was observed between the
quantitative binding state of these proteins on the
173/
160 element
after cell treatment with CRH, LIF, or both or in the corresponding
stimulation of POMC transcription in experiments where the
173/
160
element was fused to the heterologous SV40-luciferase reporter
gene.
The time-dependent activation of the Y complex by both
cytokines suggested a post-translational regulation. To explore this hypothesis, we analyzed POMC transcription in the luciferase assay as
well as Y complex formation in the gel-shift assay with or without
pretreating AtT20 cells for 30 min with 10 µg/ml cycloheximide. This
inhibitor did not alter these LIF- and CRH-induced events (data not
shown), demonstrating that ongoing protein synthesis is not required.
To investigate whether phosphorylated events are critical, supershift
experiments were performed using anti-phosphoprotein antibodies and
showed the Y complex to contain protein(s) phosphorylated on serine
residues, but not on tyrosine. Indeed, the Y complex was supershifted
by an anti-phosphoserine antibody (Fig. 8A,
lane 12) and not by an anti-phosphotyrosine antibody
(lane 13) or by preimmune serum (lane 11).
Several complexes were in fact supershifted by the anti-phosphoserine
antibody; however, only the two upper bands (designated and
)
are specific since they did not appear in a control reaction containing
the A probe with the anti-phosphoserine antibody but without AtT20 cell
nuclear extracts (data not shown). The appearance of at least two
supershifted complexes suggests that different serine-phosphorylated
proteins are components of the Y complex with DNA. Furthermore, the
presence of these complexes even in the untreated reaction reveals that
the signaling pathway leading to these serine phosphorylations is
constitutively activated. To provide further functional confirmation
for these observations, AtT20 cells were pretreated for 1 h with a
0.5 µM concentration of a serine/threonine inhibitor,
staurosporine, before adding LIF or CRH, and Y complex formation was
tested or luciferase activity was measured (Fig. 8, B and
C, respectively). LIF-, CRH-, or (LIF + CRH)-induced POMC
transcription was abolished by staurosporine, and Y complex formation
was severely reduced after staurosporine treatment, whereas this agent
only modestly altered these parameters in control cells. Furthermore,
treatment of AtT20 cells with 100 µg/ml genistein, a specific
tyrosine kinase inhibitor, did not alter LIF and CRH induction (data
not shown). These results indicate that LIF- and CRH-triggered
signaling pathways involve the specific action of serine/threonine
protein kinases.
Phosphorylation events on serine/threonine
residues are required for LIF- and CRH-enhanced POMC transcription.
A, gel-shift analyses with nuclear extracts (7.5 µg) from
untreated (control (C)); lanes 1, 2,
and 9), LIF-treated (1 nM, 60 min; lanes
3 and 4), CRH-treated (10 nM, 60 min;
lanes 5 and 6), and (LIF + CRH)-treated (1 and 10 nM, respectively; 60 min; lanes 7, 8,
and 10-13) AtT20 cells. 32P-Labeled
oligonucleotide A was used as the probe. An anti-phosphoserine antibody
(Ab -PSerine) was added to the binding reaction in
lanes 2, 4, 6, 8, and
12 and shifted the Y complex, while an anti-phosphotyrosine antibody (Ab
-PTyrosine; lane 13) or nonimmune
serum (lane 11) added under the same conditions did not.
B, gel-shift analyses with nuclear extracts (7.5 µg) from
untreated (lanes 1 and 2), LIF-treated (1 nM, 60 min; lanes 3 and 4),
CRH-treated (10 nM, 60 min; lanes 5 and
6), and (LIF + CRH)-treated (1 and 10 nM, respectively; 60 min; lanes 7 and 8) AtT20 cells.
Before cytokine treatment, cells were preincubated for 60 min with 0.5 µM staurosporine (Sigma) (lanes 1,
3, 5, and 7). 32P-Labeled
oligonucleotide A was used as the probe. C, rat POMC reporter-luciferase-transfected AtT20 cells untreated or treated for
6 h with 1 nM LIF, 10 nM CRH, or a
combination of both with (shaded bars) or without
(closed bars) a 1-h pretreatment with 0.5 µM
staurosporine. Luciferase activity was measured.
Our earlier work had shown the pleiotropic cytokine LIF to demonstrate a very robust interaction with CRH in potentiating its effects on POMC transcription (16). However, these two peptides exerted opposing effects on a number of indices of cell proliferation, including S phase accumulation (36). These either synergistic effects on POMC transcription or antagonistic effects on the cell cycle indicated complex interactions between both peptides.
It is also relevant that the CRH type II receptor ligand urocortin (24)
has similar synergistic effects on the rat POMC promoter, presumably
acting through the type I receptor (25). However, the synergistic
effect appears to be highly specific to LIF as two unrelated
inflammatory cytokines, IL-1 and TNF
, had no effect on POMC alone
and did not alter the stimulation caused by CRH.
These studies were performed to dissect interactions between the LIF and CRH signaling cascades and to identify potential intracellular loci where the observed alterations in the target gene (i.e. POMC) activity could occur. Cross-talk between signaling cascades is being increasingly recognized as important for fine-tuning cellular responses to extracellular signals (37). The regulation of POMC expression by both CRH and LIF appears to be a novel model for studying cross-talk between two different signaling pathways. Previously, inhibition of cytokine signaling by the cAMP pathway has been described (38), making the potentiated responses we observed of even greater interest. The results shown here indicate a complex pattern of signaling interaction between LIF and CRH in AtT20 cells.
CRH is the major positive regulator of the POMC transcriptional response (2). Although its primary effects are an increase in cAMP levels and activation of protein kinase A (16, 26), leading to phosphorylation of CREB, the induction of phospho-CREB by CRH has not previously been reported. However, our results show that CRH signaling through phospho-CREB is not altered by LIF alone or in combination with CRH, and consequently, this signal is apparently not a mechanism for both peptides to exert their synergy on POMC transcription.
A putative downstream target of both LIF and CRH is the proto-oncogene c-fos. c-fos mediates a component of the CRH activation of POMC transcription through a c-fos-responsive sequence in exon 1 of the POMC promoter and was therefore a rational end point to examine. We were surprised that LIF had no effect on c-fos mRNA in pituicytes, whereas CRH clearly was a powerful stimulant at the mRNA level. LIF did not enhance c-fos mRNA and, in fact, inhibited the marked increase observed with CRH. Thus, LIF and CRH exhibit antagonistic effects on c-fos mRNA, similar to the discordant effect we have previously observed for cell cycle progression in AtT20 cells (36).
Our previous studies had indicated that the cytokine-responsive region
on the POMC gene was situated in the 5-promoter sequence between
positions
323 and
166 (16). To further dissect this region, a
footprint analysis was performed and showed three elements whose DNase
sensitivity was protected by AtT20 cell nuclear extracts. Also, further
deletions between positions
167 and
117 led to severe abrogation of
LIF-CRH synergy on POMC transcription. In addition, fusion of the
173/
160 element to a heterologous SV40-luciferase reporter gene led
to enhancement of POMC transcription by both peptides to the same
extent as the wild-type promoter. This implied that both LIF and CRH
act mostly through closely apposed or similar DNA elements present
within the
173/
160 region of the POMC promoter.
To confirm these results, dynamic binding assays, using a radiolabeled
probe covering the 173/
160 sequences and AtT20 cell nuclear
extracts, were performed and showed no significant differences between
binding states induced by CRH alone or in combination with LIF. We
conclude that both peptides may stimulate two signaling pathways,
converging distally and leading to enhancement of similar transcription
factor(s) binding to the
173/
160 element.
Involvement of a similar promoter element mediating two different
signaling pathways to enhance transcription of a common target gene has
been reported (39, 40). For example, the acute-phase response element
fused to a heterologous promoter conferred responsiveness to both
interferon- and IL-6 in HepG2 cells (39). Furthermore, an 18-base
pair core region of the rat
2-macroglobulin gene was sufficient to confer both induction by IL-6 and synergy of IL-6 plus
glucocorticoids to a minimal promoter (40). However, in these latter
cases, two separate transduction pathways leading to the activation of
distinct transcription factors converged on a common element on the
promoter to give a synergistic effect. We questioned therefore whether
LIF and CRH pathways converge upstream of the transcription factor
level, leading to enhancement of the same binding protein, or
downstream, directly on the
173/
160 element by activating different
transcription factors. The notion of different transcription factors
whose homo- or heterodimerization or coordinated effects are
responsible for transcriptional synergistic induction between cytokines
is now well recognized (41, 42). However, LIF-, CRH-, or (LIF + CRH)-induced activation of POMC transcription is insensitive to protein
synthesis inhibition by cycloheximide, occurring therefore at the
post-translational level, and could be prevented by the protein kinase
inhibitor staurosporine, known to inhibit most serine/threonine protein
kinases. Furthermore, a supershift binding assay with an
anti-phosphoserine antibody showed the Y complex to contain several
serine-phosphorylated proteins. However, since these proteins are
phosphorylated and bind DNA constitutively in the basal unstimulated
state, we propose that LIF and CRH synergy occurs through an
enhancement of the phosphorylation level of these proteins and/or
through other conformational modifications. We therefore conclude that
LIF and CRH may activate either a common binding protein or distinct
transcription factors, whose activation through serine/threonine
phosphorylation events may influence their common transcriptional
activity. Requirement of serine phosphorylation events for formation of
the activated homodimer STAT3-STAT3, which binds DNA, has similarly
been reported (43). However, STAT3 and STAT1
are not implicated in
LIF- and CRH-triggered synergy on POMC transcription since neither
STAT3 nor STAT1
antibodies were able to supershift the Y complex in binding assays (data not shown).
In summary, we have demonstrated that LIF and CRH effectively synergize
to induce POMC transcription in AtT20 cells and that LIF and CRH
effects are not dissociated and are centered on the 173/
160
element. Interestingly, this element was previously shown to bind a
novel CRH-inducible transcription factor, a metallothionein protein,
whose binding may transactivate POMC gene expression (6). This protein
could also be a molecular target for the convergence of LIF and CRH
pathways. Furthermore, LIF and CRH synergy appears to occur
post-translationally and to involve serine/threonine-mediated phosphorylation events. We have thus identified and characterized a
novel interaction between a G protein-coupled receptor and a cytokine
receptor that results in synergy at the level of POMC gene
transcription.
We thank Dr. Lin Pei (Cedars-Sinai Medical Center, Los Angeles, CA) for technical advice and critical reading of the manuscript.