ESE-1 transactivates the human COX-2 promoter To elucidate whether the COX-2 gene could be an ESE-1 target gene in inflammatory processes, we screened the regulatory region of the COX-2 p
Trang 1of the COX-2 gene by LPS in monocytes
Franck T Grall, Wolf C Prall, Wanjiang Wei, Xuesong Gu, Je-Yoel Cho, Bob K Choy,
Luiz F Zerbini, Mehmet S Inan, Steven R Goldring, Ellen M Gravallese, Mary B Goldring,
Peter Oettgen and Towia A Libermann
New England Baptist Bone and Joint Institute and BIDMC Genomics Center, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, USA
Cyclooxygenase (COX) is an enzyme that converts
arachidonic acid into the prostaglandin H2 This
prod-uct is the critical point of the synthetic pathway of
numerous members of the prostaglandin family COX
exists as two major isoforms derived from two separate
genes: COX-1 and COX-2 COX-1 is constitutively
expressed, whereas COX-2 expression is inducible
Pro-inflammatory substances are some of the major
activators of COX-2 Examples include interleukin (IL)-1 [1], tumor necrosis factor (TNF)-a [2], and bac-terial lipopolysaccharide (LPS) [3] A third isoform COX-3has also been reported [4]
The mechanisms leading to COX-2 expression involve various combinations of different transcription factors, depending on the cell type and stimulus The members of the C⁄ EBP family have been identified as
Keywords
COX-2; ESE-1; Ets; gene expression; LPS
Correspondence
T A Libermann, New England Baptist Bone
& Joint Institute, Department of Medicine,
Beth Israel Deaconess Medical Center,
Harvard Institutes of Medicine, 4 Blackfan
Circle, Boston, MA 02115, USA
Fax: +1 617 975 5299
Tel: +1 617 667 3393
E-mail: tliberma@bidmc.harvard.edu
(Received 21 July 2004, revised 19 January
2005, accepted 2 February 2005)
doi:10.1111/j.1742-4658.2005.04592.x
Cyclooxygenase-2 (COX-2) is a key enzyme in the production of prosta-glandins that are major inflammatory agents COX-2 production is trig-gered by exposure to various cytokines and to bacterial endotoxins We present here a novel role for the Ets transcription factor ESE-1 in regula-ting the COX-2 gene in response to endotoxin and other pro-inflammatory stimuli We report that the induction of COX-2 expression by lipopoly-saccharide (LPS) and pro-inflammatory cytokines correlates with ESE-1 induction in monocyte⁄ macrophages ESE-1, in turn, binds to several E26 transformation specific (Ets) sites on the COX-2 promoter In vitro analysis demonstrates that ESE-1 binds to and activates the COX-2 promoter to levels comparable to LPS-mediated induction Moreover, we provide results showing that the induction of COX-2 by LPS may require ESE-1,
as the mutation of the Ets sites in the COX-2 promoter or overexpression
of a dominant-negative form of ESE-1 inhibits LPS-mediated COX-2 induction The effect of ESE-1 on the COX-2 promoter is further enhanced
by cooperation with other transcription factors such as nuclear factor-jB and nuclear factor of activated T cells Neutralization of COX-2 is the goal
of many anti-inflammatory drugs As an activator of COX-2 induction, ESE-1 may become a target for such therapeutics as well Together with our previous reports of the role of ESE-1 as an inducer of nitric oxide syn-thase in endothelial cells and as a mediator of pro-inflammatory cytokines
in vascular and connective tissue cells, these results establish ESE-1 as an important player in the regulation of inflammation
Abbreviations
Ad, Adenovirus; ChIP, chromatin immunoprecipitation; CMV, cytomegalovirus; COX, cyclooxygenase; CRE, cAMP responsive element; ESE, epithelium specific Ets factor; Ets, E26 transformation specific; GAPDH, glyceraldehydes-3-phosphate dehydrogenase; HRP,
horseradish peroxidase; ICAM, intercellular adhesion molecule; IL, interleukin; iNOS, inducible nitric oxide synthase; LPS, lipopolysaccharide; MMP, matrix metalloproteinase; NFAT, nuclear factor of activated T cells; NF-jB, nuclear factor-jB; TNF, tumor necrosis factor.
Trang 2important regulators of COX-2 expression in
osteo-blasts [5], T lymphocytes [6], amnion epithelial cell
WISH [7], macrophages [8,9], and chondrocytes [10]
The nuclear factors of activated T cells (NFAT) are
essential for COX-2 activation in T lymphocytes [6]
The role of nuclear factor-jB (NF-jB) appears to be
more cell-specific According to Allport et al [7], the
mutation of the major NF-jB site of the COX-2
pro-moter abolishes IL-1-induced COX-2 expression in the
human amnion epithelial cell line WISH Crofford
et al [11] successfully employed p65 antisense
oligo-nucleotides to inhibit IL-1-mediated induction of the
COX-2promoter in synoviocytes Furthermore, a
frag-ment of the COX-2 promoter starting downstream of
the NF-jB sites could not be activated by IL-1 in
chondrocytes [10] However, in macrophages,
Wad-leigh et al [9] mutated the NF-jB site of the murine
COX-2 promoter without loss of LPS mediated
COX-2 activation The cAMP responsive element
(CRE) site overlapping an E-box element is another
important site for transcription factors as the mutation
of the CRE site within the COX-2 promoter or the
expression of a dominant negative mutant of CREB
reduces inducibility of the COX-2 promoter [9,12–14]
The transcription factors binding this site include
CREB and cJun as well as USF-1 for the E-box [12]
Some Ets factors have also been suggested to play a
role in the regulation of COX-2 such as Ets-1 [14,15],
PEA3 [16–18] and Pu.1 [19]
The multiple activation modalities observed across
the different studies as well as the similarity of the
recognition sites of the NFAT and Ets factors led us to
investigate the potential involvement of the epithelial
specific Ets factor 1 (ESE-1⁄ ESX ⁄ ELF3 ⁄ ERT ⁄ JEN) in
this process This factor is a likely regulator of COX-2
expression, as we recently discovered that ESE-1 plays
an important role in the responses of various cell types
to inflammatory mediators [20,21] We reported the
induction of ESE-1 in response to IL-1, TNF-a and
LPS in vascular and connective tissue cells This
induc-tion was mediated through the activainduc-tion of NF-jB
[21] We showed that ESE-1 could activate the
expres-sion of another important player in the inflammatory
response: inducible nitric oxide synthase (iNOS) [20]
Although ESE-1 was initially described as exclusively
expressed in epithelial cells in a variety of tissues
[22–25], we subsequently observed a broader
expres-sion pattern for ESE-1 under inflammatory conditions
[21]
We now report that ESE-1 can bind to the promoter
of COX-2 and that the integrity of the Ets sites is
required for LPS-induced COX-2 expression ESE-1
can activate the COX-2 promoter in the monocyte cell
line RAW 264.7 as well as chondrocytic cells where it acts synergistically with NF-jB and NFAT Moreover,
we show the capacity of a dominant-negative form
of ESE-1 to diminish COX-2 promoter induction in response to LPS or IL-1 exposure
Results
COX-2 induction by pro-inflammatory stimuli correlates with ESE-1 induction
Our previous data had indicated that ESE-1 expression
is rapidly and transiently induced by pro-inflammatory cytokines in a variety of vascular and connective tis-sue cell types [20,21] We also demonstrated that the iNOS gene, a target for pro-inflammatory cytokines, is
a downstream target for ESE-1 To further our under-standing of ESE-1 function during inflammatory processes, we have now explored the involvement of ESE-1 in the regulation of another inflammation-related gene, COX-2, in monocytic cells and chondrocytes We previously demonstrated that LPS stimulation of human monocytic THP1 cells leads to an induction of ESE-1 mRNA expression within 1 h of exposure, reaching a peak at 4 h and leveling off after 24 h [12]
To investigate the level of ESE-1 protein following LPS exposure, we performed western blot analysis in murine monocytic RAW 264.7 cells The intensity of the ECL signal was determined using the alphaease
fc sofware and divided by the protein concentration
of the sample ESE-1 protein was detected 4 h after LPS stimulation and increased levels were observed until 10 h Analysis of COX-2 protein expression in response to LPS in RAW 264.7 cells by western blot revealed that the temporal pattern of COX-2 pro-tein induction upon stimulation by LPS correlated with the expression pattern of ESE-1 (Fig 1) Thus, COX-2 may be a potential target for ESE-1 during inflammation
ESE-1 transactivates the human COX-2 promoter
To elucidate whether the COX-2 gene could be an ESE-1 target gene in inflammatory processes, we screened the regulatory region of the COX-2 promoter for potential ESE-1 binding sites Inspection of the COX-2 promoter sequence (GeneBank accession num-ber AY229989) revealed the presence of five possible Ets binding sites within the first 200 bp upstream of the transcription initiation site, one of which overlaps with a potential NFAT site (Fig 2A)
To determine whether the COX-2 gene may be regulated by ESE-1, we constructed two human
Trang 3COX-2 promoter luciferase reporter plasmids, pXP2⁄
COX-2–170 and pXP2⁄ COX-2–831, starting at )170
and )831, respectively, upstream of the
tran-scription start site, which we transiently transfected
into RAW 264.7 cells Cotransfections of these
pro-moter plasmids together with the ESE-1 expression
vector, pCI⁄ ESE-1, enhanced COX-2 promoter
activ-ity 10- and 30-fold, when the long or the short
pro-moter constructs, respectively, were used (Fig 2B)
Stimulation with LPS resulted in a more than
20-fold induction of the COX-2 promoter (Fig 2B)
This ESE-1-mediated transactivation of the COX-2
promoter was not restricted to RAW 264.7 cells,
since transfection of pCI⁄ ESE-1 also stimulated
tran-scription of the )170 bp COX-2 promoter in the
human chondrocyte cell line T⁄ C28a2 (Fig 2C), a
cell type shown to express ESE-1 in response to
IL-1 [21]
As another Ets factor, PEA3, has previously been
shown to activate the COX-2 promoter, we compared
the relative activities of ESE-1 and PEA3, cloned
downstream of the cytomegalovirus (CMV) promoter,
in a dose–response curve Different amounts of ESE-1 and PEA3 expression vector DNA were cotransfected with the COX-2 promoter luciferase construct, main-taining the total amount of transfected DNA con-stant by adding the parental pCI vector As illustrated in Fig 2D, ESE-1 at all concentrations was more effective than PEA3 in transactivating the COX-2 promoter This result does not appear to be due to a higher production of the ESE-1 protein Western blot analysis of ESE-1 and PEA3 expression after transfection of equal amounts of expression vec-tor into 293ft cells, shows that ESE-1 protein expres-sion is lower than PEA3 (Fig 2E) Indeed we have observed that generally ESE-1 protein expression after transfection is significantly lower than most other Ets factors
ESE-1 binds to the human COX-2 promoter
To investigate whether this induction could be due to
a direct effect of ESE-1 on the COX-2 promoter, we assessed the ability of ESE-1 to bind to the COX-2 promoter by performing an EMSA We used as probes the five putative Ets binding sites of the COX-2 pro-moter present within the )170 COX-2 promoter We tested their ability to form a complex with in vitro translated ESE-1 protein As shown in Fig 3A, sites 1 and 3 formed strong protein⁄ DNA complexes with ESE-1 The specificity of this complex was confirmed
in supershift assays using two different anit-ESE-1 Igs (Fig 3) Site 4, which overlaps with the NFAT binding site, did not appear to bind specifically to ESE-1 in this assay The mobility of the ESE-1 complex was consistent with that reported in Rudders et al and was absent when unprogrammed reticulocyte lysate was used
To further determine whether ESE-1 binds to the COX-2 promoter in vivo, we performed a chromatin immunoprecipitation experiment in chondrocyte cells which express both ESE-1 and COX-2 in response
to IL-1 (Fig 3B) T⁄ C28a2 chondrocyte cells were transfected with either pcDNA3-1⁄ Flag-ESE-1 or pcDNA3-1⁄ Flag After cross-linking the proteins bound to DNA with formaldehyde followed by soni-cation, the cell extracts were immunoprecipitated using the anti-Flag Ig or nonspecific serum As shown
in Fig 4B, a precipitate specifically retaining the COX-2 promoter region spanning all five Ets sites shown in Fig 2 ()186 to +56 of the transcription start site) was only obtained in cells transfected with the plasmid containing ESE-1 This experiment most clearly demonstrates that ESE-1 directly binds to the COX-2 promoter in vivo
A
B
Fig 1 Induction of ESE-1 and COX-2 expression in monocytic
cells RAW cells were grown in the absence or presence of LPS
(100 ngÆmL)1) for 0, 2, 4, 10 or 12 h ESE-1 and COX-2 protein
lev-els were measured by western blotting using ESE-1 and COX-2
specific antibodies (A) ECL signal on a photographic film (B)
Lev-els of ESE-1 and COX-2 shown as integrated intensity divided by
the protein load for each lane.
Trang 4Mutation of multiple ESE-1 binding sites
drastically reduces activation of the COX-2
promoter by ESE-1 and by LPS
To examine whether the Ets sites in the COX-2
promo-ter are responsive to ESE-1 and to depromo-termine whether
LPS induction of the COX-2 promoter is mediated via
ESE-1, we introduced mutations into individual or
multiple Ets sites of the COX-2 promoter Wild type
or mutant constructs of pXP2⁄ COX-2–170 were
cotransfected into RAW 264.7 cells in the absence or
presence of pCI⁄ ESE-1 These experiments indicated
that individual mutations of the Ets binding sites 2, 3 and 4 led to more than 50% reduction of ESE-1-medi-ated COX-2 promoter activation (Fig 4A) Simulta-neous mutation of site 3 along with sites 1, 2 or 4 almost completely eliminated inducibility by ESE-1 suggesting that ESE-1 acts on the COX-2 promoter via multiple Ets sites (Fig 4A), but that sites 2, 3 and 4 are crucial for inducibility, since their mutations gave the strongest reductions of activity compared to other isolated mutations This is in line with the findings of Liu et al [18] who showed that the Ets site number 3 was critical for COX-2 induction by NO No inducibility
A
Fig 2 The COX-2 promoter is a target for ESE-1 (A) Sequence of the COX-2 promoter The five putative Ets binding sites present in the )170 COX-2 construct (starting at the asterisk) are highlighted as well as additional NFAT and NF-jB sites within the COX-2 promoter sequence Two extra Ets sites described in Howes et al [16] are underlined (B) Transcriptional activation of the COX-2 promoter by ESE-1 and LPS RAW cells were cotransfected with the pXP2 luciferase construct containing the COX-2 promoter (pXP2 ⁄ COX-2) starting either at )831 or at )170 and the pCI ⁄ ESE-1 expression vector and incubated in the absence or presence of LPS Luciferase activity in the lysates was determined 16 h later, as described Data shown are means of duplicate measurements from one representative transfection The experiment was repeated three times with different plasmid preparations with comparable results Error bars represent the SD for the two replicates (C) Transcriptional activation of the COX-2 promoter by ESE-1 in chondrocytes T ⁄ C28a2 cells were cotransfected with the pXP2 luciferase construct containing the )170 COX-2 promoter (pXP2 ⁄ COX-2) and the pCI ⁄ ESE-1 expression vector Luciferase activity in the lysates was determined 16 h later, as described Data shown are means of duplicate measurements from one representative transfection Error bars represent SD of the two replicates (D) Transcriptional activation of the COX-2 promoter by ESE-1 and PEA3 RAW cells were cotransfected with the COX-2 promoter luciferase construct (pXP2⁄ COX-2–170) and different amounts of expression vectors for ESE-1 or PEA3 maintaining constant (700 ng) the total amount of DNA with pCI vector Luciferase activity in the lysates was determined 16 h later,
as described Data shown are means of duplicate measurements from one representative transfection The experiment was repeated three times with different plasmid preparations with comparable results Error bars represent SD for the two replicates (E) Protein expression levels of ESE-1 and PEA3 after transfection into 293ft cells Myc-tagged Ets factors were transfected into 293 cells The cells were lysed
16 h later and equal amounts of lysate were loaded on a gel for a western blot analysis using anti-myc Ig.
Trang 5was left when sites 1, 2, 3 and 5 were mutated in
combination
LPS response was also significantly affected when
sites 3, 4 or 5 were mutated individually (Fig 4B)
Combined mutation of the Ets sites 1, 2, 3, and 5,
leaving the NFAT element in site number 4 intact, led
to a drastic inhibition of promoter activation in
response to LPS (Fig 4B) This experiment
demon-strates that LPS activation of the COX-2 promoter is
at least partially mediated via ESE-1 or a related Ets
factor
The mutation of the C⁄ EBPb site that inhibited the
activity of PEA3 [16] led to only a diminution of the
activity of ESE-1
ESE-1 and NFAT act synergistically on the COX-2
promoter
As the NFAT factors have been reported as activators
of COX-2 [6], we evaluated whether ESE-1 and NFAT
could cooperate in the context of the COX-2 promoter
The )170 COX-2 promoter luciferase construct was
cotransfected into RAW cells together with pCI⁄ ESE-1
or a constitutively active form of one member of the NFAT family, NFAT3, cloned into pRK5, or a combi-nation thereof, and either empty pRK5 or pCI, respect-ively (Fig 5A) ESE-1 enhanced COX-2 promoter activity 10-fold compared to only 2.5-fold activation by NFAT3 Combined expression of ESE-1 and NFAT3 synergistically enhanced COX-2 promoter activity more than 20-fold These results indicate that ESE-1 and NFAT3 most likely act via different sites or different DNA sides at the same sites and may actually cooper-ate in transactivation of the COX-2 promoter
ESE-1 cooperates with NF-jB in the
transactivation of the COX-2 promoter
In addition to the Ets binding sites the COX-2 promo-ter contains at least four putative NF-jB binding sites
As NF-jB has been demonstrated previously to play a role in regulating COX-2 promoter activity in at least some cell types [7,11] and NF-jB cooperates with ESE-1 in transactivating the iNOS promoter in
endo-A
B
Fig 3 ESE-1 binds to several Ets sites in the COX-2 promoter (A) Interaction of ESE-1 with Ets binding sites in the COX-2 promoter EMSA using either unprogrammed reticulocyte lysate [23] or in vitro translated ESE-1 (ESE-1) and five labeled oligonucleotide probes containing dif-ferent COX-2 promoter Ets sites The in vitro translated proteins were alternatively preincubated with antibody (Ab1, east-acres Biological; Ab2, QED; C, negative control, normal rabbit serum) The white arrow indicates the specific ESE-1 DNA–protein complex and the black arrow shows the supershift form with the antibody (B) Binding of ESE-1 to endogenous human COX-2 promoter by ChIP The anti-Flag Ig was used to specifically enrich COX-2 promoter DNA sequences in a ChIP assay Chromatin proteins from IL-1 treated T ⁄ C28a2 cells trans-fected with either pcDNA3 ⁄ Flag (lanes 1–3) or pcDNA ⁄ Flag-ESE-1 (lanes 4–6) were crosslinked to DNA with formaldehyde, and purified nucleoprotein complexes were immunoprecipitated using either anti-Flag Ig or nonspecific rabbit serum The precipitated DNA fractions were analyzed by PCR for the presence of the COX-2 promoter region encompassing the Ets sites The input and genomic DNA (gDNA) were used as a positive control, and water was used as a negative control for PCR.
Trang 6thelial cells [20], we evaluated whether ESE-1 and
NF-jB cooperate in the context of the COX-2
promo-ter as well The )831 COX-2 promoter luciferase
construct was cotransfected into RAW cells together with either ESE-1, NF-jB p50 or p65 alone as well as with various combinations thereof (Fig 5B) Whereas
Fig 4 ESE-1 and LPS transactivate the COX-2 promoter through
multiple Ets binding sites (A) Mutation of multiple Ets binding sites
within the COX-2 promoter inhibits induction by ESE-1 RAW cells
were cotransfected with the pCI ⁄ ESE-1 expression vector and the
COX-2 promoter luciferase constructs containing either wild-type
(WT) or multiple mutants of potential binding sites (mut) alone or in
combination Luciferase activity in the lysates was determined 16 h
later, as described elsewhere [23] Data shown are means of
dupli-cate measurements from one representative transfection The
experiment was repeated four times with different plasmid
prepara-tions with comparable results Error bars represent SD of the two
replicates (B) Mutation of the Ets binding sites reduces
LPS-induced transactivation of the COX-2 promoter RAW cells were
transfected with the wild-type or Ets mutant COX-2 promoter
luci-ferase constructs and then stimulated with LPS Luciluci-ferase activity
in the lysates was determined 16 h later Error bars represent the
SD of the two replicates.
A
B
C
Fig 5 ESE-1 cooperates with NFAT and NF-jB in transactivating the COX-2 promoter (A) RAW cells were cotransfected with the pCI ⁄ ESE-1 expression vector or the pRK5 ⁄ NFAT3 expression vec-tors or a combination thereof and the )170 COX-2 promoter luci-ferase construct Luciluci-ferase activity in the lysates was determined
16 h later, as described elsewhere [23] Data shown are means of duplicate measurements from one representative transfection The experiment was repeated twice with different plasmid preparations with comparable results Error bars represent the standard devi-ation of the two replicates (B and C) RAW cells were cotransfected with the pCI ⁄ ESE-1 expression vector, the NF-jB p50 and p65 expression vectors or a combination thereof and the )831 or )170 COX-2 promoter wild-type (B) or )170 mut ets 1 +2 + 3 +4 + 5 (C) luciferase constructs Luciferase activity in the lysates was deter-mined 16 h later, as described elsewhere [23] Data shown are means of duplicate measurements from one representative trans-fection The experiment was repeated twice with different plasmid preparations with comparable results Error bars represent the standard deviation of the two replicates.
Trang 7p50 alone did not significantly enhance COX-2
promo-ter activity, p65 increased promopromo-ter activity by three
to fourfold The p50⁄ p65 combination enhanced
trans-activation of the COX-2 promoter to 10-fold, similar
to the effect of ESE-1 alone ESE-1 cooperated with
both p50 and p65, which enhanced COX-2 promoter
activity by 30-fold in cotransfection with ESE-1 This
cooperativity was most striking when ESE-1 was
cotransfected with the p50⁄ p65 heterodimer, which
increased transactivation of the COX-2 promoter by
70-fold Mutation of the Ets sites within the COX-2
promoter did not significantly affect NF-jB mediated
transactivation, but strongly reduced cooperativity
between ESE-1 and NF-jB (Fig 5C)
Dominant-negative ESE-1 mutants inhibit LPS
and IL-1 mediated induction of COX-2 gene
expression
To evaluate whether ESE-1 is indeed involved in
regu-lation of inducible COX-2 gene expression we used
dominant-negative mutants of ESE-1 as tools to block
endogenous COX-2 gene expression We constructed
two dominant-negative forms of ESE-1 One of these
two constructs, Dominant Negative 1 (DN1),
encom-passes the carboxy-terminal Ets DNA binding domain
of ESE-1 and competes with intact endogenous ESE-1
for binding to target gene promoters The second
dominant-negative mutant, Dominant Negative 2
(DN2), encompasses the amino-terminal
transactiva-tion domain and Pointed domain fused to a nuclear
localization signal and presumably acts as a dominant
negative ESE-1 due to its ability to interact with
coac-tivators and other cofactors needed for transactivation
of ESE-1, thereby depriving intact ESE-1 of its factors
needed for transactivation DN1 and DN2 were cloned
into adenovirus vectors and the expression plasmid
pCI
We tested the effect of dominant-negative ESE-1 on
inducible COX-2 gene expression in the human
chond-rocyte cell line T⁄ C28a2 Very little COX-2 mRNA
expression was detected in unstimulated cells, but a
strong induction was observed upon exposure to IL-1
(Fig 6A) Infection with AdE1-DN1 or AdE1-DN2
inhibited IL-1-induced expression of COX-2 mRNA
by 50–70% compared to Ad-bGal infection (Fig 6A)
As RAW monocytic cells are difficult to infect with
adenoviruses and also do not transfect with high
effi-ciency, they are not suitable for assessing the effects of
the dominant negative ESE-1 on endogenous COX-2
mRNA levels Therefore, we evaluated the effects of
the dominant-negative ESE-1 mutants in RAW cells
transfected with the pXP2⁄ COX-2 promoter luciferase
construct ESE-1 DN1 expression completely blocked LPS-mediated induction of the COX-2 promoter, again confirming the involvement of ESE-1 in LPS-mediated activation of the COX-2 promoter (Fig 6B)
These data most vividly indicate that ESE-1 may play a critical role in induction of the COX-2 gene by pro-inflammatory stimuli
Discussion
The regulation of COX-2 expression during inflamma-tion has been the focus of numerous studies The het-erogeneity of parameters such as the cell type and the stimulus used has made it difficult to describe precisely
Fig 6 Dominant-negative mutants of ESE-1 reduce expression of the endogenous COX-2 gene in response to LPS and inhibit the transactivation its promoter (A) T ⁄ C28a2 chondrocyte cells were infected with the adenoviruses Ad ⁄ ESE-1 DN1, DN2 or b-galac-tosidase (beta-gal) and 16 h later treated with IL-1 (500 pgÆmL)1) The RNA was harvested 28 h later and used for real-time PCR using COX-2 specific primers Data shown are means of duplicate measurements from one representative transfection representing the ratio of the measurements of COX-2 to GAPDH mRNA Error bars represent the SD of the two replicates (B) RAW cells were cotransfected with the pCI expression vector containing the dom-inant-negative mutant of ESE-1 (ESE-1 DN1) and the COX-2 pro-moter Luciferase ( )170) construct Cells were grown in the absence or presence of LPS (500 ngÆmL)1) for 16 h Luciferase activity in the lysates was determined 16 h after addition of LPS Data shown are means of duplicate measurements from one rep-resentative transfection Error bars represent the SD of the two replicates.
Trang 8the mechanisms by which the promoter of COX-2 is
activated Several transcription factor families have
been shown to be involved in this process such as
C⁄ EBP [5,9,10,26], NF-jB [7,10,11], NFAT [6,10] and
Ets [16,18,27]
We now report the involvement of another Ets
tran-scription factor ESE-1 in the regulation of COX-2
expression in monocytes⁄ macrophages and
chondro-cytes ESE-1 (also named ELF3, Jen, ERT, ESX) is an
Ets family transcription factor, recently discovered by
us and others [22,23,28,29], whose expression under
normal physiological conditions is restricted to
epi-thelial cells However, we uncovered an unexpected
function for ESE-1 in the vascular system and in
con-nective tissue cells where its expression is induced
fol-lowing exposure to pro-inflammatory stimuli such as
IL-1, TNF-a, and LPS [20,21]
We show here that LPS-mediated induction of COX-2
gene expression is, at least partially, dependant upon
ESE-1 upregulation ESE-1 binds to the promoter of
COX-2 on several sites and activates its expression
The integrity of these sites is required for full COX-2
promoter activation by LPS, since mutation of two or
more sites markedly attenuates the response This is
true even when the NFAT site is left intact (identical to
Ets site number 4) ESE-1 mediated transactivation of
the COX-2 promoter is synergistic with NFAT and
NF-jB, since ESE-1 can enhances the activity of COX-2
promoter due to NFAT or NF-jB transactivation more
than additive This cooperativity can be due to the
pre-viously demonstrated [20] direct binding of ESE-1 to
the NF-jB family members p50 and p65 NF-jB itself
also upregulates endogenous ESE-1 expression in
response to pro-inflammatory stimuli via a high affinity
NF-jB binding site within the ESE-1 promoter [20,21]
Thus, ESE-1 may be involved in a positive feedback
loop during the inflammatory response to enhance the
transactivation effect of NF-jB on some of its target
genes by cooperating with ESE-1 The mutation of the
Ets site 4 decreases the ESE-1 activation of COX-2,
even though no binding could be detected in vitro by
EMSA This site may be functionally mixed or it is
possible that these two factors interact with each other
and bind as a complex on this site That would explain
the further increase of COX-2 activation when ESE-1 is
transfected together with NFAT
The involvement of an Ets factor in the regulation
of COX-2 was reported by Howe et al [16], who
dem-onstrated that overexpression of another member of
the Ets family, PEA3, could activate the COX-2
pro-moter When we compared the activities of PEA3 to
ESE-1, we found that ESE-1 is a more effective
trans-activator of the COX-2 promoter than PEA3, further
supporting the notion that ESE-1 may be relevant for COX-2 regulation However, in contrast to ESE-1, PEA3 does not appear to be regulated by pro-inflam-matory stimuli Surprisingly, the effect of PEA3 seems
to be mediated via the C⁄ EBPb site Our data show that mutation of the C⁄ EBPb site also affects ESE-1 mediated transactivation, but only partially This sug-gests that ESE-1 may also cooperate with C⁄ EBPb or another factor binding to this site in this process but that does not account for the entire activation activity
as the mutation of Ets sites leads to a more drastic abolition of this induction
It has also been shown that the pattern of expression
of PEA3 in breast cancer samples correlates with the patterns of expression of HER-2⁄ neu and COX-2 [17] suggesting that the levels of COX-2 may results from
an HER-2⁄ neu stimulation of PEA3 Interestingly, a similar correlation between the expression of HER-2⁄ neu and ESE-1 has also been observed [28,30,31] and HER-2⁄ neu could itself activate the expression of ESE-1 [31]
The role of Ets factors in inflammation and in the regulation of cytokine-responsive genes has not been studied in detail However, several genes, including urokinase-type plasminogen activator, matrix metallo-proteinase (MMP)-1, MMP-3, TNF-a, scavenger receptor, intercellular adhesion molecule (ICAM)-1, ICAM-2, and IL-12 have been shown to depend on Ets factors for their inducibility by cytokines such as IL-1 or TNF-a [32–36] Many additional cytokine-responsive genes contain putative Ets binding sites within their regulatory regions, including COX-2, iNOS, and MMP-13
We demonstrate here that endogenous COX-2 gene expression can be inhibited by using dominant negative forms of ESE-1 The inhibitory effect of these domin-ant negative ESE-1 mutdomin-ants on COX-2 expression, confirm the importance of ESE-1 or a related factor that would bind to these Ets sites These results give some insight in the potential use of new therapeutic approaches manipulating the activity of ESE-1 or other Ets factors as a tool to reduce inflammation Indeed, several pro-inflammatory agents such as IL-1 [21], and TNF-a [21] can also induce ESE-1, and ESE-1 target genes with regard to inflammation so far identified by us include iNOS, COX-2 and potentially MMP-1 and MMP-13 (X Gu, F Grall, M Joseph,
L Zerbini & T Libermann, unpublished data) The kin-etic of activation of ESE-1 seems to implicate ESE-1
at later stages of inflammation These results suggest that ESE-1 regulates a subset of the genes whose inhi-bition could be of significant interest in the manage-ment of the inflammatory reaction
Trang 9In conclusion, this report demonstrates further that
ESE-1 is a relevant player in the inflammatory process
ESE-1 is upregulated in several cell types in response to
pro-inflammatory stimuli through the NF-jB pathway
It can activate some important genes such as iNOS and
COX-2 Our studies also suggest that, by modulating
the activity of ESE-1, we could decrease the
inflamma-tory reaction in response to LPS exposure
Experimental procedures
Cell culture and patient samples
THP-1 (human monocytic) and RAW 264.7 (murine
mono-cytic) cells (ATCC, Manassas, VA, USA) were cultured in
DMEM with 10% serum (Hyclone, Logan, UT, USA) with
undetectable levels of endotoxin Immortalized human
chondrocytes T⁄ C28a2 were grown and treated with
cyto-kines as described [37,38] Lipopolysaccharide was obtained
from Sigma (St Louis, MO, USA) (catalogue number L8274)
RT⁄ PCR analysis
Total RNA was harvested using QIAshreder (Qiagen,
Valencia, CA, USA) and RNeasyMini Kits (Qiagen) The
cDNAs were generated from 1 lg total RNA using
Ready-To-GoTM You-prime First-Strand Beads (Amersham
Phar-macia Biotech Inc., Piscataway, NJ, USA)
SYBR Green I-based real-time PCR was carried out on
the Opticon Monitor (MJ Research, Inc., Waltham, MA,
USA) All PCR mixtures contained: PCR buffer (final
con-centration 10 mm Tris⁄ HCl pH 9.0, 50 mm KCl, 2 mm
MgCl2, 0.1% TritonX-100), 250 lm deoxy-NTP (Roche
Pleasanton, CA, USA), 0.5 lm each PCR primer, 0.5·
SYBR Green I, 5% dimethylsulfoxide, 1 U Taq DNA
poly-merase (Promega, Madison, WI, USA) with 2 lL cDNA
in a final volume of 25 lL The samples were loaded into
wells of Low Profile 96-well microtiter plates After an
ini-tial denaturation step at 95C for 2 min, conditions for
cycling were 38 cycles of denaturation (95C, 30 s),
annealing (54C, 30 s), and extension (72 C, 1 min)
Then, the fluorescence signal was measured immediately
following incubation at 78C for 5 s that follows each
extension step, thereby eliminating possible primer dimer
detection At the end of the PCR cycles, a melting curve
was generated to identify the specificity of the PCR
prod-uct For each run, serial dilutions of human GAPDH
plas-mids were used as standards for quantitative measurement
of the amount of amplified cDNA For normalization of
each sample, hGAPDH primers were used to measure the
amount of hGAPDH cDNA All samples were run as
duplicates and the data were presented as ratios of Cox-2⁄
hGAPDH The primers used for real time PCR are as
follows For hGAPDH forward, 5¢-CAAAGTTGTCATG
GATGACC-3¢; reverse, 5¢-CCATGGAGAAGGCTGGG G-3¢, which will amplify 195 bp of human GAPDH For Cox2 forward, 5¢-TTCAAATGAGATTGTGGGAAAA TTGCT-3¢; reverse, 3¢-ATATCATCTCTGCCTGAGTAT CTT-3¢, which will amplify 304 bp of human Cox2
Expression vector and luciferase reporter gene constructs
Full-length and dominant negative mutant ESE-1 cDNAs were inserted into the EcoRI site of the pCI (Promega) euk-aryotic expression vector downstream of the T7 and CMV promoter as described [39] The dominant negative form of ESE-1, DN1, encodes the ESE-1 peptidic sequence deleted
of the amino acid residues 76–198, and DN2 encodes the amino acid residues 1–231 fused in frame to a nuclear local-ization signal motif repeated three times and a sequence coding for EYFP in the pEYFP-NLS vector from Clon-tech Full length ESE-1 was fused in-frame with the Flag peptide at the amino terminus in the pcDNA3-1 vector The human COX-2 promoter sequences spanning)831 and )170 to +103, kindly provided by L J Crofford, Division
of Rheumatology, University of Michigan [11], were cloned into the pXP2 luciferase vector in the HindIII and XhoI sites (pXP2⁄ COX-2) An expression vector for the mouse PEA3 gene downstream of the CMV promoter (pCANMycPEA3) was a gift from L Howe, Strang Cancer Research Laboratory, Rockefellar University
EMSA
In vitro transcription⁄ translation was performed in TNT rabbit reticulocyte lysate (Promega) using the pCI⁄ ESE-1 vector as described [23] EMSAs were performed using 2 lL
of in vitro translation product and [32P]-labeled double-stranded oligonucleotide probes [40] Supershift assays were performed by preincubating the in vitro translated protein
20 min at room temperature with 2 lL antibody
Oligonucleotides used as probes and for competition studies were: (a) COX-2 promoter Ets site #1, 5¢-GCA CGTCCAGGAACTCCTCAGC-3¢; (b) COX-2 promoter Ets site #2, 5¢-GAGAGAACCTTCCTTTTTATAA-3¢; (c) COX-2promoter Ets site #3, 5¢-CGAAAAGGCGGAAAG AAACAGT-3¢; (d) COX-2 promoter Ets site #4, 5¢-GAGA GGAGGGAAAAATTTGTGG-3¢; 3¢-CTCTCCTCCCTTT TTAAACACC-5¢; (5) COX-2 promoter Ets site #5, 5¢-TCTCATTTCCGTGGGTAAAAA-3¢
Site-directed mutagenesis Mutations in the different COX-2 promoter ETS sites were generated by site-directed mutagenesis with the QuikChange Site-directed Mutagenesis kit (Stratagene, Cedar Creek, TX, USA) and confirmed by sequencing
Trang 10The following primers were used (the mutated bases are
underscored): (a) COX-2 promoter Ets site #1, 5¢-GCT
GAGGAGTAGCTGGACGTGCTCCTGAC-3¢; (b) COX-2
promoter Ets site #2, 5¢-CAGTCTTATAAAAACCAA
GGTTCTCTCGGTTAGCGACC-3¢; (c) COX-2 promoter
Ets site #3, 5¢-GACGAAATGACTGTTTCTTTGAGCC
TTTTCGTACCCC-3¢; (d) COX-2 promoter Ets site #4,
5¢-AGGGGAGAGGAGGGTTAAATTTGTGGGGGGTA
CGAAAAGGCGG-3¢; (e) COX-2 promoter Ets site #5:
5¢-GGGTTTTTTACCCACGCTAATGAGAAAATCGGAA
ACC-3¢
DNA transfection assays
Cotransfections were carried out in 6-well plates containing
3–8· 105 cells per well using 600 ng of reporter gene
con-struct DNA and 200 ng expression vector DNA using
LipofectAMINE PLUS (Gibco-BRL) for 16 h as described
[23] Transfections were performed independently in
dupli-cate, repeated three to four times with different plasmid
preparations and gave similar results Cotransfection of a
second plasmid for determination of transfection efficiency
was omitted, because potential artifacts with this technique
have been reported [41] and many commonly used viral
promoters contain binding sites for Ets factors
Adenovirus infection
Adenoviruses encoding the dominant negative forms of
ESE-1 were constructed using the Adeno-X expression
system from Clontech TC⁄ 28a2 cells were infected with
adenovirus for 1 h in serum-free medium using a multiplicity
of infection of 300 After infection the cells were washed with
medium and incubated for 16 h in DMEM containing 10%
fetal calf serum in the absence or presence of IL-1
(500 pgÆmL)1) (R & D Systems, Minneapolis, MN, USA)
Western blot analysis
RAW 264.7 cells were plated at 4· 105cells per well 16 h
before being exposed to LPS (100 ngÆmL)1) in fresh
med-ium for different periods of time The cells were rinsed with
NaCl⁄ Pi, harvested in 200 lL RIPA lysis buffer containing
protease inhibitors (Roche) and frozen and thawed once
before being sonicated Forty microliters of lysate were
loa-ded on a 10% polyacrylamide gel containing SDS Proteins
were transferred to a poly(vinylidene difluoride) (PVDF)
membrane and blocked with 5% milk in NaCl⁄ Pi⁄ Tween
(0.2%) A polyclonal antibody directed against the
amino-terminal half of ESE-1 (East Acres Biologicals,
South-bridge, MA, USA) or an anti-COX-2 Ig (Santa-Cruz, Santa
Cruz, CA, USA) were used to detect the presence of these
proteins A secondary antibody labeled with horseradish
peroxidase (HRP) was used for detection by ECL The
signal intensity was determined with the alphaease soft-ware (AlphaInnotech, San Leandro, CA, USA) and then divided for normalization by the protein concentration for each lane
293ft cells (3· 105; Invitrogen, Carlsbad, CA, USA) were transfected with 3 lg ESE-1 or PEA3 expression plasmid using lipofectamine and lysed 16 h later Cell lysate (225 lg) was loaded onto a SDS⁄ PAGE gel Anti-myc ⁄ HRP conju-gated Ig (Santa-Cruz) was used at 1 : 200 dilution for 4 h to detect myc-tagged ESE-1 and PEA3
Chromatin immunoprecipitation (ChIP) ChIP was conducted as previously reported [21] Briefly, TC28⁄ a2 chondrocytes cells (2 · 107) were plated on 150-mm dishes and transfected with either pcDNA3Flag⁄ ESE-1
or pcDNA3Flag and after 24 h stimulated with IL-1 (500 pgÆmL)1) for an additional 3 h A 10-min formaldehyde cross-linking step was stopped by adding glycine (0.125 m,
5 min at room temperature) After two washes, the cells were resuspended in 0.3 mL lysis buffer, sonicated and then centri-fuged at 4C Supernatants were collected and 100 lL of chromatin preparation were aliquoted as the input fraction The remainder of the supernatants was diluted 1 : 10 in dilu-tion buffer for immunoclearing with sheared salmon sperm DNA, normal rabbit serum and protein A–Sepharose for 2 h
at 4C Immunoprecipitation was performed overnight at
4C with 80 lL M2 Agarose (anti-Flag Ig at 50% slurry in TE) (Sigma) or with 5 lL rabbit IgG and 80 lL protein A– Sepharose as a negative control Precipitates were washed se-quentially for 10 min each in 1 mL of TSE buffers [21] Preci-pitates were then extracted three times with 1% SDS, 0.1 m NaHCO3 Eluates were pooled and heated at 65C overnight
to reverse the formaldehyde cross-linking, without proteinase digestion DNA fragments were purified with QIAquick PCR purification Kit (Qiagen) PCR was performed using
5 lL of a 60 lL DNA extraction in TE buffer with Hi-Fi Taq polymerase (Invitrogen) for 28 cycles (95C for 30 s,
52C for 30 s and 68 C for 1 min) The primers used were 5¢-CTGGGTTTCCGATTTTCTCA-3¢ and 5¢-CTGCTG AGGAGTTCCTGGAC-3¢ which amplify 200 bp of the human COX-2 promoter
Acknowledgements
This study was supported by National Institutes of Health Grant RO1⁄ AI49527 and a Brain Tumor Soci-ety Research grant to T A L and by National Insti-tutes of Health Grant KO8⁄ CA 71429 to P O K
References
1 Crawford HC & Matrisian LM (1996) Mechanisms controlling the transcription of matrix metalloproteinase