(Received for publication, August 19, 1996)
From the Division of Gastroenterology, Department of Internal Medicine, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania 19104-6144
Interleukin-8 (IL-8), a potent neutrophil
chemoattractant, can be expressed at high levels by many different cell
types after immune stimulation. In contrast, expression of IL-8 in
these same cells is virtually absent in the unstimulated state,
demonstrating the tight regulation of the IL-8 gene. Although much is
known about how this gene is transcriptionally activated after immune stimulation, little is known about the regulation of the IL-8 promoter
in the absence of immune activation. In this study we examine how the
IL-8 promoter is transcriptionally regulated in the uninduced state and
how these mechanisms are altered in response to immune stimulation by
IL-1. Electrophoretic mobility shift assay and transfection studies
show that the IL-8 promoter is transcriptionally regulated by both
positive and negative elements. Although the nuclear factor-
B
(NF
B) element regulates only inducible activity of the IL-8 promoter
in response to stimulation with IL-1
, the AP-1 and
CCAAT/Enhancer-binding Protein (C/EBP) elements influence both basal
and inducible activities. In contrast to these three positive
regulatory elements, the binding of the ubiquitously expressed
POU-homeodomain transcription factor, Oct-1, strongly represses
transcriptional activity of the IL-8 promoter by binding independently
to an element overlapping that of C/EBP.
Interleukin-8 (IL-8),1 a member of the
C-X-C family of chemokines, has been shown to have a diverse
spectrum of biological activities including T-cell, neutrophil, and
basophil chemotactic properties (reviewed in Ref. 1). IL-8 can be
produced by a wide variety of cell types and is believed to play a
significant role in the initiation of the acute inflammatory response
(2). Much is known about the transcriptional induction of the IL-8 gene
in response to inflammatory stimulation. IL-8 mRNA levels are
rapidly induced in the absence of new protein synthesis by proinflammatory cytokines such as IL-1 and tumor-necrosis factor- (3). Nuclear run-on assays have shown that this induction is regulated
at the level of gene transcription (3). Elements important for this
response have been identified within the first 135 bp of the 5
-flank.
This promoter region is regulated in a cell line-specific fashion
requiring a NF
B element plus either an AP-1 or a C/EBP element (Fig.
1). For example, the AP-1 element along with the NF
B
site are sufficient for the full expression of this promoter in gastric
cell lines (4), whereas in a fibrosarcoma cell line, only the C/EBP and
NF
B sites are required (5). Recently, several investigators have
shown that NF
B/rel transcription factors can interact with those of
the C/EBP family through a rel domain-bZIP interaction to alter the
transcriptional activation of the IL-8 promoter (6-8).
In contrast to transcriptional induction, there are no previous reports
which describe how the IL-8 gene is regulated in the absence of
inflammatory stimulation, a state which we will refer to as basal
transcriptional activity. In this study we investigate how the IL-8
promoter is transcriptionally regulated in two epithelial cell lines,
Caco-2 and HepG2, in the unstimulated state and in response to a
physiologically relevant mediator of inflammation, IL-1. These two
cell lines constitutively express certain bZIP transcription factors,
which play a role in the transcriptional activation of the IL-8
promoter. The AP-1 and C/EBP elements bind proteins that significantly
influence not only induced transcription activity of the IL-8 promoter
but also activate this promoter in the absence of immune stimulation.
Despite this, these two cell lines have no detectable expression of
IL-8 mRNA by Northern blot analysis in the absence of an
inflammatory stimulus (9). We provide the first evidence that the
ubiquitously expressed transcription factor, Oct-1, binds to a motif
overlapping that of the C/EBP element and acts as a potent
transcriptional repressor, which may help to prevent expression of the
IL-8 promoter in the uninduced state.
Caco-2 and HepG2 cells (from ATCC), used for both nuclear protein isolation and in the transfection experiments described below, were plated at a density of 4 × 104 cells/cm2 in 10-cm dishes containing Dulbecco's modified Eagle's medium with 10% fetal bovine serum and penicillin/streptomycin as described previously (10, 11).
Nuclear Protein Isolation and Electrophoretic Mobility Shift Assay (EMSA)Caco-2 cells, day 5 post plating, were stimulated
with complete medium containing 5 ng/ml IL-1 for 45 min. Nuclear
proteins were isolated from unstimulated and IL-1
-stimulated Caco-2
cells using a modification (12) of the method of Dignam et
al. (13). The protein concentrations of the extracts were
determined (14), and aliquots were stored at
80 °C. Complimentary
oligonucleotides with overlapping ends were synthesized, annealed, and
labeled with 32P using Klenow enzyme as described
previously (11). Binding reactions, each containing 10 µg of nuclear
extract, were performed as described previously (11) and subsequently
separated on a 4% polyacrylamide gel. Competition and antibody
supershift experiments were performed in a fashion described previously
(15). Supershift assays used 1 µl of each antibody for the three
C/EBP isoforms
,
, and
(Santa Cruz Biotechnology, Santa Cruz,
CA), Oct-1 (antiserum was a gift of Dr. Winship Herr, Cold Spring
Harbor Laboratory), and c-fos (Upstate Biotechnology, Lake
Placid, NY).
The sense strand sequences of the various oligonucleotides used for
EMSA (described under "Results") are listed in Fig. 1. In
particular, the C/EBP.cons oligonucleotide contains a previously described consensus sequence for C/EBP (16). The Oct.cons
oligonucleotide contains a octamer consensus sequence which was
commercially available (Santa Cruz Biotechnology) and was labeled using
T4 kinase. Two additional oligonucleotides, not listed in Fig. 1,
include AP-1 (the AP-1 element of the IL-8 promoter, designated in
boldface type) and mAP-1 (a 2-bp mutation, underlined, of the IL-8 AP-1 element): AP-1, 5-gtgagatctGAAGTGTGATGACTCAGG-3
and mAP-1, 5
-gtgagatctGAAGTGTGAT
CTCAGG-3
.
The IL-8
promoter region (bp 135 to +46) was amplified by polymerase chain
reaction using human liver genomic DNA as a template with the primers
IL-8(
135) and IL-8(+46) (Scheme 1
![]() |
![]() |
![]() |
). This fragment was
ligated into the BglII and HindIII sites of the
luciferase reporter plasmid pGL2-Basic (Promega, Madison, WI) yielding
the reporter construct (wt)LUC. The same 3-bp mutation of the NFB
binding element in the IL-8 promoter used in EMSA described above was
incorporated into the IL-8 promoter using the method of overlap
extension polymerase chain reaction described previously (17). The EMSA
oligonucleotides mNF
B as well as IL-8(
135) and IL-8(+46) were used
in these reactions. Ligation of the resulting product into pGL2-Basic
yielded the reporter construct (mNF
B)LUC. Site-directed mutations of
the AP-1 (bp
126 to
120) and C/EBP (bp
94 to
80) binding
elements in the IL-8 promoter were amplified using IL-8(+46) with the
primers
135IL8(mAP-1) and
135IL8(mC/EBP), respectively. These same
mutations, which have been characterized previously in other cell lines
(4, 6), were used in the EMSA studies described umder "Results."
Ligation of these amplified products into pGL-2Basic yielded the
reporter constructs (mAP-1)LUC and (mC/EBP)LUC. The binding elements
are in boldface type with the point mutation underlined:
135IL8(mAP-1),
5
-GTGAGATCTGAAGTGTGAT
CTCAGG-3
;
135IL8(mC/EBP),
5
-GTGAGATCTGAAGTGTGATGACTCAGGTTTGCCCTGAGGGGATGGGCCATCAG
T
C
A
TCG-3
.
The 3-bp mutation of the C/EBP element used in the IL-8(Oct.cons)
EMSA oligonucleotide was incorporated into the IL-8 promoter in a
similar fashion to yield the reporter construct, (Oct.cons)LUC.
Similarly, the 5-bp mutation of the C/EBP element used in the
IL-8(C/EBP.cons) EMSA oligonucleotide was incorporated into the IL-8
promoter to yield (C/EBP.cons)LUC. The fidelity of each polymerase
chain reaction product was verified by dideoxy chain-termination
sequencing using Sequenase 2.0 (U. S. Biochemical Corp.). Each
reporter plasmid was amplified in DH5
Escherichia coli
and purified by alkaline lysis followed by two successive bandings of
plasmid DNA in cesium chloride gradients (18).
Caco-2 or HepG2 cells were plated at a density of 4 × 104 cells/cm2 in 10-cm dishes containing
complete medium. After 24 h, 7 µg of reporter plasmid were
co-transfected with 1 µg of SV40 -galactosidase (transfection
control) using the calcium phosphate precipitation method of Chen and
Okyama (19). To determine the effect of IL-1
on reporter gene
expression, 5 µg/ml of this cytokine was added to the culture medium
48 h after transfection. After 6 h, the cytoplasmic extract
was then assayed for luciferase and
-galactosidase activity as
described previously (10). A minimum of three replicates for each
experiment was performed with results expressed as mean light units per
unit
-galactosidase activity.
In order to determine the
functional effect of previously characterized DNA binding elements in
the IL-8 promoter in the colon cancer cell line, Caco-2,
transcriptional activation of the wild-type IL-8 promoter
((wt)LUC) or the promoter with point mutations in the NFB
((mNF
B)LUC), C/EBP ((mC/EBP)LUC), or AP-1 ((mAP-1)LUC) binding
elements were determined by transient transfection into Caco-2 cells as
described under "Experimental Procedures" (Fig. 2).
After stimulation with IL-1
, the activity of this promoter is
increased by 114-fold. Although a 3-bp mutation in the NF
B element
of the IL-8 promoter does not affect the basal activity of this
promoter, it does eliminate induction of promoter activity by IL-1
.
A 4-bp mutation of the C/EBP binding element increases basal activity
by 60% and decreases, but does not eliminate, the induction of the
IL-8 promoter in response to IL-1
. In contrast, mutation of the AP-1
binding element decreases the basal activity of the IL-8 promoter by
3-4-fold, but does not affect the inducible response (98-fold) of the
promoter to IL-1
. These results demonstrate that, although protein
binding to the NF
B element of the IL-8 promoter is critical for
transcriptional induction in response to IL-1
, the AP-1 and C/EBP
binding elements are required for the full expression of the IL-8
promoter.
The Proto-oncogene Product fos Is Part of the Protein Complex That Binds Constitutively to the AP-1 Element of the IL-8 Promoter
EMSA using a double-stranded oligonucleotide spanning
the AP-1 element of the IL-8 promoter shows that nuclear proteins from Caco-2 cells bind constitutively to this motif (Fig. 3).
Effective competition using the unlabeled wild-type AP-1
oligonucleotide demonstrates the specific nature of this DNA-protein
interaction. In contrast, ineffective competition using the mutant
oligonucleotide, mAP-1, demonstrates that a 2-bp mutation of the AP-1
element significantly reduces protein binding to this element.
Supershift of this band with the addition of fos antiserum
shows that the proto-oncogene product, fos, is a component
of the protein complex that binds to the AP-1 element of the IL-8
promoter. The specific nature of this supershift is demonstrated by
effective competition with the addition of excess unlabeled wild-type
AP-1 oligonucleotide.
Transcription Factors of Both the Octamer and C/EBP Families Bind Independently to the C/EBP Element of the IL-8 Promoter
The C/EBP
binding motif of the IL-8 promoter, located between bp 82 and
94,
also contains an overlapping octamer motif located on the complementary
strand in the opposite orientation (Fig. 1). Pictured in Fig.
4 is an EMSA of Caco-2 nuclear extracts using a
double-stranded oligonucleotide as a probe, named C/EBP, that spans
this region. This sequence binds nuclear transcription factors of both
the octamer (slower mobility complex) and C/EBP families in both
unstimulated Caco-2 cells and those stimulated with IL-1
, lanes 2 and 3, respectively. Near complete
competition of both the bands by addition of 100-fold excess unlabeled
C/EBP oligonucleotide in lane 4 demonstrates the
specific nature of the binding. Competition experiments using
double-stranded oligonucleotides encoding consensus sequences for
C/EBP (C/EBP.cons) and octamer (Oct.cons) in lanes 5 and 6, respectively, show that these two
transcription factors can bind to the C/EBP probe independently. As
expected, competition with both of these consensus oligonucleotides in
lane 7 eliminates all specific binding. Identical results
were observed in EMSAs performed using nuclear extracts isolated from
both HepG2 cells and Hela cells (data not shown). Supershift
experiments using antiserum specific for the nuclear transcription
factor Oct-1 in lane 8 (20) demonstrate that the slower
mobility complex is due to binding of Oct-1. Competition with the
octamer consensus oligonucleotide in lane 9 shows the
specific nature of the supershifted band. In a similar fashion,
supershift experiments using antisera specific for the C/EBP isoforms
,
, and
(in lanes 10, 11, and
12, respectively) show that all three isoforms are present in Caco-2 cells and can bind to C/EBP probe.
Adjacent C/EBP and NF
Physical interaction between NFB and C/EBP
transcription factors through the rel and bZIP domains of these
proteins lead to synergistic activation of the IL-8 promoter (6-8). In
order to demonstrate this interaction in Caco-2 cells and to determine if Oct-1 binding occurs within the context of this interaction, EMSAs
of nuclear extracts isolated from both control (C) and
IL-1
-stimulated (S) Caco-2 cells were performed using a
double-stranded oligonucleotide probe, named IL-8wt, which spans both
the C/EBP and NF
B elements of the IL-8 promoter (Fig.
5, lanes 3 through 13). EMSA using the NF
B probe in lanes 1 and 2 demonstrate the
relative mobility of proteins which bind to the NF
B element and
confirm their induction of binding by IL-1
. The specific binding
patterns in both control and IL-1
-stimulated extracts using the
IL-8wt probe, in lanes 3 and 4, respectively, is
demonstrated by effective competition using the IL-8wt oligonucleotide
in lanes 5 and 6. Oct-1 and C/EBP bind
independently to the IL-8wt probe in unstimulated Caco-2 cells
(lane 3). This is confirmed by effective competition with the C/EBP.cons and Oct.cons oligonucleotides in lanes 9 and
11, respectively. Upon stimulation with IL-1
, NF
B and
C/EBP form a heteromeric complex of slower mobility (lane
4). This is confirmed by the loss of this band upon competition
with either the NF
B oligonucleotide (lane 8) or the
C/EBP.cons oligonucleotide (lane 10). Oct-1 binding appears
to be diminished upon cell stimulation in lane 4 as
confirmed by the similarity in binding patterns when the Oct.cons
oligonucleotide is used as a competitor in stimulated cell extracts in
lane 12. Lane 13 represents the IL-8wt probe alone. In
summary, these results demonstrate that both C/EBP and Oct-1 bind to
the promoter in control cells. Upon stimulation with IL-1
, Oct-1
binding may be reduced and the induction of NF
B binding leads to the
formation of a heteromeric complex of C/EBP and NF
B.
Separation of Oct-1 and C/EBP Binding by Site-directed Mutagenesis of the IL-8 C/EBP Element
The EMSA using the IL-8wt
oligonucleotide in Fig. 5 shows the complex nature of the DNA-protein
interactions at this motif involving proteins of the NFB/rel, C/EBP,
and Oct-1 families. The consensus sequence for either C/EBP or octamer
(Fig. 1) were used to create site-directed mutations of this
oligonucleotide to separate C/EBP from Oct-1 binding. The effect of
these mutations on DNA-protein interactions using Caco-2 nuclear
proteins were determined by EMSA. Fig. 6A
shows that the oligonucleotide IL-8(C/EBP.cons), which contains a 5-bp
mutation, enhances C/EBP binding but eliminates Oct-1 binding.
Competition with the C/EBP.cons oligonucleotide eliminates all specific
binding in extracts from unstimulated cells (lane 7) and
permits binding of NF
B/rel factors in the stimulated state
(lane 8). In contrast, the DNA-protein interactions are not
altered by competition with the Oct.cons oligonucleotide in extracts
isolated form either control or IL-1
stimulated extracts (compare
lanes 1 and 2 to lanes 9 and
10). This oligonucleotide still permits the formation of the
slower mobility heteromeric C/EBP-NF
B complex upon IL-1
stimulation (lane 2). This complex is lost upon competition
with oligonucleotides specific for either NF
B (lane 6) or
C/EBP (lane 8) but not Oct-1 (lane 10). Although this mutation clearly increases binding of nuclear proteins specific for a C/EBP element, it is not known if the proportion of the various
C/EBP isoforms which bind has been altered compared to those which bind
to the wild-type C/EBP element (Fig. 4).
The oligonucleotide IL-8(Oct.cons) contains a 3-bp mutation which
enhances Oct-1 binding and eliminates C/EBP binding (Fig. 6B). Competition with an oligonucleotide specific for
octamer eliminates the predominant band (lanes 11 and
12) in contrast to competition with C/EBP.cons which does
not alter the DNA-protein interactions (compare lanes 3 and
4 to lanes 9 and 10). Although proteins can still bind to the NFB element in this oligonucleotide (lane 12), there is no heteromeric complex formation between
these proteins and Oct-1. Pictured in Fig. 6C is an EMSA
using the oligonucleotide IL-8(mC/EBP). This oligonucleotide, which
contains a 4-bp mutation of the C/EBP element (Fig. 1) that increases
basal activity of the IL-8 promoter in both Caco-2 and HepG2 cells
(Figs. 2 and 7C), eliminates both C/EBP and
Oct-1 binding while still permitting NF
B to bind. The identical
binding pattern was also observed using nuclear extracts isolated from
HepG2 cells (data not shown).
Oct-1 Is a Potent Repressor of Both Basal and Activated Transcription of the IL-8 Promoter in Caco-2 and HepG2 Cells
In
order to determine the effect of Oct-1 and C/EBP binding on both basal
and stimulated activity of the IL-8 promoter, two site-directed mutants
of the wild-type IL-8 promoter, (C/EBP.cons)LUC and (Oct.cons)LUC, were
transfected into both Caco-2 and HepG2 cells. Enhanced binding of C/EBP
in the absence of Oct-1 (the C/EBP.cons mutant) increases the
unstimulated transcriptional activity in both Caco-2 and HepG2 cells by
severalfold (Fig. 7, A and C, respectively)
compared to the wild-type IL-8 promoter. In contrast, enhanced binding
of Oct-1 in the absence of C/EBP (the Oct.cons mutant) virtually
abolishes the basal activity of this promoter in both these cell lines.
Fig. 7C also shows that elimination of both C/EBP and Oct-1
binding with the (mC/EBP)LUC mutant actually enhances the basal
activity of the IL-8 promoter in HepG2 cells compared to the wild-type;
a finding also observed in Caco-2 cells (Fig. 2). These results
demonstrate that C/EBP is an activator of basal transcription in
contrast to Oct-1 which acts as a repressor. Fig. 7, B and
D, shows that, upon stimulation of Caco-2 and HepG2 cells,
respectively, with IL-1, transcriptional activation of the IL-8
promoter is preserved with enhanced C/EBP binding(C/EBP.cons) but is
virtually eliminated upon binding of Oct-1 (Oct.cons).
Transcriptional regulation of the IL-8 promoter involves the
interaction of transcription factors which are activated by immune stimulation, such as those of the NFB/rel family, and constitutively expressed proteins of the bZIP family, such as C/EBP and AP-1 (3). The
transcriptional regulation of this promoter in the basal state and
after induction by the proinflammatory cytokine, IL-1
, was studied
in two epithelial cell lines. As anticipated, inducible activation of
NF
B transcription factors by IL-1
is critical for the
transcriptional induction of the IL-8 promoter but plays no role in
uninduced transcriptional activity (Fig. 2). Full transcriptional
activation of this promoter, however, requires the binding of nuclear
transcription factors to both the C/EBP and AP-1 elements (Fig. 2).
Therefore, all three previously described cis-acting elements in the
IL-8 promoter, AP-1, C/EBP, and NF
B, act as positive elements in the
activation of the IL-8 promoter in Caco-2 cells. This is in contrast to
previously published results obtained from other cell lines where the
NF
B element along with either the C/EBP or AP-1 element are
sufficient for full activation of this promoter (4, 5). It is likely
that different sets of transcription factors may be involved in the transcriptional regulation of the IL-8 promoter in a cell type-specific manner.
In contrast to the inducible binding of NFB, EMSAs of Caco-2 nuclear
proteins show that transcription factors such as fos bind
constitutively to the AP-1 element of the IL-8 promoter (Fig. 3). It is
not surprising, therefore, that the transfection of AP-1 mutants of the
IL-8 promoter into Caco-2 cells show that this site has no effect on
the inducibility of this promoter but is rather involved in the basal
level of transcription (Fig. 2). It is interesting to note, however,
that the induced transcriptional activity of this promoter is reduced
in proportion to the reduction in basal activity when the AP-1 element
is mutated. This suggests that the potential for transcriptional
induction by IL-1
is influenced by the basal activity of this
promoter.
In similar fashion, proteins that bind to the C/EBP element of the IL-8
promoter play a role in both the basal and induced transcriptional
activity of this promoter. Several investigators have shown that
physical interactions of C/EBP transcription factors with those of the
NFB/rel family lead to synergistic transactivation of the IL-8
promoter through the adjacent C/EBP and NF
B DNA binding elements
(6-8). EMSAs confirm that stimulation of Caco-2 cells lead to the
formation of a heteromeric complex of NF
B/rel- and C/EBP-binding
proteins (Fig. 5). EMSAs using the C/EBP element show that three
isoforms of C/EBP,
,
, and
, are able to bind to the IL-8
promoter in Caco-2 cells (Fig. 4). Although we did not characterize the
specific C/EBP isoforms, which interact with the NF
B/rel proteins in
Caco-2 cells, Stein et al. (21) have shown that all three
isoforms can physically interact with p65 to enhance gene expression of
a reporter gene regulated by the SRE of the human c-fos
promoter.
In contrast to the role of C/EBP in the induction of IL-8 promoter activity, the role that this element plays in the regulation of this promoter in the absence of immune stimulation has not been previously studied. EMSAs of the C/EBP element clearly demonstrate that not only C/EBP but also Oct-1 can bind to this motif. The binding of Oct-1 to the IL-8 promoter has not been previously reported. Oct-1 belongs to the POU-homeodomain family of transcription factors which bind to the 8-bp octamer sequence ATGCAAAT (22). The IL-8 C/EBP element diverges form this consensus sequence by 1 bp (Fig. 1). It is well known, however, that Oct-1 can bind promiscuously to a divergent array of DNA binding elements (23). The overlapping nature of the C/EBP and Oct-1 binding sequences in the IL-8 promoter is consistent with the observation that these two proteins bind independently to this element (Fig. 4).
Oct-1 is ubiquitously expressed and plays a critical role in basic cellular functions. It has been shown to both activate and repress transcription. For example, Oct-1 can bind to the regions of the granulocyte/macrophage-colony-stimulating factor, IL-3, and IL-5 promoters where it contributes to basal T cell gene transcription (24). Emerging evidence suggests that transcriptional activation via Oct-1 is dependent upon interactions with additional proteins such as Pit-1 (25), OAP (26), or VP-16 (27). In contrast, Oct-1 has also been shown to act as a transcriptional repressor for a number of regulatory regions such as the LCR enhancer of HPV 16 (28), the SV40 enhancer (29), and the HPV 18 enhancer (30). Each one of these regulatory regions contains an Oct-1 element, which overlaps that of a specific transcriptional activator (PEF-1, Sph-I, and KRF-1, respectively). Transcriptional activation by each of these regulatory regions is repressed by displacement of the transcriptional activator by the binding of Oct-1. Therefore, the overall level of transcriptional activity is determined by balance between the binding of Oct-1 and the particular transcriptional activator (28). In order to ascertain the functional significance of Oct-1 binding in the context of the IL-8 promoter, we compared the transcriptional activity of three site-directed mutants of this promoter: 1) IL-8(C/EBP.cons), a 5-bp mutation which enhances C/EBP binding but eliminates the binding of Oct-1; 2) IL-8(Oct.cons), a 3-bp mutation which enhances Oct-1 binding but eliminates the binding of C/EBP; and 3) IL-8(mC/EBP), a 4-bp mutation which eliminates binding of both Oct-1 and C/EBP (Fig. 6, A-C).
In a fashion analogous to the regulatory elements of the HPV and SV40
enhancers, EMSA and site-directed mutagenesis of the C/EBP element of
the IL-8 promoter show that Oct-1 acts as a potent transcriptional
repressor of both the basal and induced IL-8 promoter activity in
Caco-2 and HepG2 cells (Fig. 7). It is likely that the loss of
transcriptional activity observed with the IL-8(Oct.cons) mutation is
due to a combination of the loss of C/EBP binding and the gain of Oct-1
binding. However, the fact that a mutation which eliminates binding of
both C/EBP and Oct-1 to this element actually increases basal activity
while retaining a significant degree of transcriptional inducibility
(Figs. 2 and 7) supports the role of Oct-1 as a transcriptional
repressor. Furthermore, the IL-8(C/EBP.cons) mutation demonstrates that
C/EBP binding can act to increase transcriptional activity of this
promoter in unstimulated cells (Fig. 7, A and C).
Indeed, although C/EBP binding to the IL-8 promoter in the unstimulated
state is relatively weak (Fig. 4), other investigators have shown that
this promoter can be transactivated by severalfold upon co-transfection
of an expression vector for NF-IL6 (C/EBP ) (7, 8). Therefore, in
light of the EMSAs, which show that C/EBP binding is present in the
uninduced state (Figs. 4 and 5), transcriptional repression by Oct-1
would be necessary to maintain the low level of basal activity of the
wild-type IL-8 promoter in the absence of immune stimulation.
The results of this study would support the following model regarding
regulation of the transcriptional activity of the IL-8 promoter in the
uninduced state (Fig. 8). Basal activity is determined by constitutive binding of transcription factors to the AP-1 element and the proportional binding of C/EBP, a transcriptional activator, and
Oct-1, a transcriptional repressor, to overlapping motifs located at
the C/EBP element of this promoter. The site-directed mutants used in
this study provide evidence that protein binding to the C/EBP element,
which can be divided into three different states, has a significant
influence on the basal activity of this promoter. Exclusive binding of
Oct-1 with the (Oct.cons)LUC reporter leads to low basal activity while
exclusive binding of C/EBP with the (C/EBP.cons)LUC reporter results in
high basal activity. Intermediate basal activity is observed with the
(mC/EBP)LUC reporter where protein binding to this element is absent.
The observed basal activity of the wild-type IL-8 promoter is,
therefore, a composite of these three promoter states.
Although it is not clear by EMSA whether or not Oct-1 binding is lost
when Caco-2 cells are stimulated by IL-1 (Fig. 5), the loss of any
detectable transcriptional activation when Oct-1 binding is present
(Fig. 7, B and D) in contrast to the 100-fold transcriptional induction noted with the wild-type promoter (Fig. 2)
strongly suggests that Oct-1 binding is replaced by that of NF
B and
C/EBP upon IL-1
stimulation (Fig. 8). The lack of transcriptional activity with the (Oct.cons)LUC reporter after stimulation with IL-1
is likely an artificial state due to the mutation of the C/EBP element
which disrupts the cooperative binding of C/EBP and NF
B upon cell
stimulation. The physiologic role of Oct-1 binding may, therefore, not
be to inhibit transcriptional induction to an inflammatory stimulus,
but rather to inhibit basal activity perhaps in cell types which
constitutively express high levels bZIP proteins which are capable of
binding to the C/EBP and/or AP-1 elements of the IL-8 promoter. IL-8 is
an initiator of the acute inflammatory response and can be induced in
many different cell types (31). Oct-1, a transcription factor that is
ubiquitously expressed (22), would therefore be well suited as a
transcriptional repressor which inhibits expression of this potent
chemoattractive substance in the absence of immune stimulation.
We thank Drs. Peter G. Traber and Theodore M. Danoff for their critical review of the manuscript.