(Received for publication, November 21, 1996, and in revised form, March 7, 1997)
From the Vascular Medicine and Atherosclerosis Unit, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts 02115
In the lesions of atherosclerosis, vascular
smooth muscle cells (SMC) display many functions characteristic of
cytokine activation that likely contribute importantly to ongoing
inflammation during human atherogenesis. The transcription factor
nuclear factor -B (NF
B) often mediates the effects of cytokines
on target cells, but the identity of Rel family members important in
human SMC activation remains uncertain. In vitro, human SMC
express multiple Rel family members. Of these, dimers of p65 and p50,
but not a putative SMC-Rel, comprise basal and inducible NF
B binding
activities. SMC express two inhibitor proteins I
B
and I
B
.
Interleukin-1
stimulation caused transient loss of I
B
and a
sustained decrease of I
B
that correlated with increased and
persistent levels of p65/p50 protein and binding activity in the
nucleus. SMC cultured under serum-free conditions displayed little
NF
B activity, but addition of serum or platelet-derived growth
factor did activate NF
B. In situ analyses showed no
evidence for basal NF
B activity in SMC in vivo as
nonatherosclerotic arteries did not contain nuclear p65 or p50 protein.
However, the nuclei of intimal SMC within human atheroma did contain
both Rel proteins. We conclude that (i) dimers of p65 and p50, but not
SMC-Rel, comprise NF
B complexes in human SMC; (ii) stimulatory
components in serum activate NF
B and likely account for previously
reported "constitutive" NF
B activity in cultured SMC; and (iii)
exposure to inflammatory cytokines may produce prolonged NF
B
activation in SMC because of sustained decreases in the inhibitory
subunit I
B-
.
Vascular smooth muscle cells (SMC)1 at
sites of atherosclerotic lesions express features of an inflammatory
process, such as increased expression of genes encoding growth factors,
inducible surface proteins, and molecules involved in extracellular
matrix remodeling (1-3). The nuclear factor -B (NF
B) family of
transcription factors has emerged as a regulator of many of these
molecules by vascular cells. Inflammatory cytokines, oxidized lipids,
and oxidative stress, factors or events present in human atheroma, can
activate NF
B in vitro and also elicit specific functions in SMC (1, 4-6). Several genes up-regulated in SMC during atherogenesis, including vascular cell adhesion molecule-1 (VCAM-1), interleukin-1 (IL-1), tumor necrosis factor-
(TNF-
), and
c-myc also contain functional
B elements in their
promoter/enhancer regions (5, 7). Moreover, SMC, macrophages, and
endothelial cells within human atheroma exhibit nuclear localization of
the NF
B subunit p65 (Rel A) in situ (8). Thus, in
vitro and in situ studies underscore the potential
importance of the NF
B pathway in dysfunction of vascular cells
during atherogenesis.
NFB exists in the cytosol of many cell types as an inactive complex
of Rel-related factors, bound to a member of inhibitor proteins termed
I
B (reviewed in Refs. 5 and 7). Rel family members include p65 (Rel
A), Rel (c-Rel), Rel B, and the Drosophila homolog
dorsal, each of which contain transactivation domains necessary for gene induction. Other members, p50 (NF
B1) and p49 (NF
B2), are synthesized, respectively, as p105 and p100 precursors, and transactivate only weakly, but can form functional dimers with
members of the first group. NF
B is sequestered in the cytosol in an
inactive heteromeric complex by associating with one of several
inhibitors denoted I
B, most commonly I
B
or I
B
, or with
Rel precursor proteins. Activation of NF
B follows phosphorylation of
I
B or p105 on serine residues, possibly by a ubiquitin-regulated Ser/Thr kinase (9). Phosphorylated I
B or p105 is enzymatically degraded or processed to p50, respectively, by the multicatalytic proteasome complex (10), and liberated NF
B dimers then translocate to the nucleus and promote transactivation of target genes. One of
these target genes encodes the inhibitor I
B
that binds to and
thus limits further NF
B activity and gene expression (11, 12).
I
B
, another I
B member, is not resynthesized following its
degradation, and may mediate prolonged NF
B activity in lymphocytes exposed to lipopolysaccharide or IL-1 (13). The diversity of dimeric
complexes formed by Rel factors requires definition of the NF
B
system in the context of each cell type examined.
In endothelial cells, inducible expression of leukocyte adhesion
molecules requires participation of a well characterized NFB system
(14, 15). Bovine SMC were recently found to exhibit basal, constitutive
NF
B activity in vitro (16). A novel, putative Rel protein
termed "SMC-Rel" is thought to comprise constitutive NF
B
activity in SMC and may permit cell division in serum-containing medium
(16, 17). However, the repertoire of Rel proteins or their inhibitors
expressed in SMC, the identity of NF
B members involved in DNA
binding, or whether constitutive NF
B activity exists in human SMC
in vivo has not been fully delineated. In view of the
potential importance and unresolved issues regarding the role of NF
B
in SMC, we have addressed the identity of Rel family members and the
issue of their constitutive expression in vitro and in
vivo in human SMC cultures and in normal and diseased arterial
specimens. As the inhibitory limb of control of NF
B activity appears
critical, we also tested the hypothesis that I
B
and I
B
may
play distinct roles in the control of cytokine activation of this cell
type central to the pathogenesis of atherosclerosis.
Rabbit affinity-purified polyclonal antisera to
NFB proteins p65, p50 (NF
B1), p49/52 (NF
B2), cRel, Rel B,
I
B
/MAD-3, and I
B
were obtained from Santa Cruz
Biotechnology, Inc. (Santa Cruz, CA). Monoclonal antibody HHF-35 (mouse
IgG1) recognizing muscle-specific actin was purchased from
Enzo Diagnostics (Sysosset, NY). Polyclonal rabbit neutralizing
antibody to human PDGF-BB was purchased from Genzyme (Cambridge, MA). A
nonimmune rabbit IgG fraction was from Dako (Carpenteria, CA).
Smooth muscle cells obtained from explanted sections of human saphenous veins were grown in Dulbecco's modified Eagle's Medium (Life Technologies, Inc., Grand Island, NY) supplemented with 20 mM Hepes, 10% fetal calf serum (Hyclone), 5 mM L-glutamine, plus 50 units/ml penicillin and 50 µg/ml streptomycin in a humidified atmosphere of 5% CO2, 95% air. Cells were passaged by brief trypsinization and were used through passage 5. The cells were characterized by phase contrast microscopy and were routinely screened for mycoplasma contamination by polymerase chain reaction using a commercially available kit (ICN Biomedicals Inc., Aurora, OH). Human aortic SMC and saphenous vein endothelial cells were isolated enzymatically and cultured as described previously (18).
Western Immunoblot AnalysisWhole cell lysates were
prepared as described elsewhere (19) in 2 × SDS lysis buffer
(1 × = 125 mM Tris·HCl, pH 6.8, 10% glycerol, 2%
sodium dodecyl sulfate, 5% 2-mercaptoethanol). Equivalent amounts of
whole cell lysates (30-40 µg of protein) or nuclear or cytosolic
fractions (prepared as described below) were resolved on 10 or 12%
SDS-PAGE gels, followed by electrophoretic transfer to polyvinylidene
difluoride membranes (Millipore, Bedford, MA). Membranes were incubated
in PBS-T (phosphate-buffered saline, 0.1% Tween-20), containing 5%
non-fat dry milk for 1 h at 37 °C, and then incubated for
1 h with primary antibodies used at 0.4 µg/ml, with the
exception of IB antisera, which was used at 0.5 µg/ml. Membranes
were washed with PBS-T and incubated with horseradish peroxidase-conjugated donkey anti-rabbit IgG as a secondary antibody (Jackson Laboratories, West Grove, PA), diluted 1:15,000 in PBS-T, 5%
dry milk. Immunocomplexes were visualized using Renaissance chemiluminescence reagents (DuPont NEN) and exposure to x-ray film.
Saphenous vein SMC were
grown to confluence (~5 × 105 cells) in 10-cm Petri
dishes in serum-containing medium. In some experiments, cells were
preincubated 48 h in serum-free IT medium (Dulbecco' modified
Eagle's/Ham's F-12, 1:1, insulin 5 µg/ml, transferrin 5 µg/ml) to
reduce exposure to serum components before addition of stimuli (20).
Cells were stimulated with human recombinant cytokines IL-1,
TNF-
, or PDGF-BB (Endogen, Cambridge, MA) for the indicated times,
and were collected into chilled Microfuge tubes. Nuclear and cytosolic
extracts were prepared according to Dignam et al. (21), with
the additional step of washing nuclear pellets in low salt buffer prior
to high salt extraction of nuclear proteins to remove any residual
cytosolic contamination. Aliquots were assayed for protein
concentration, dithiothreitol was added to a final 1 mM
concentration, and extracts were stored at
80 °C until analysis.
Oligonucleotides corresponding to the
B site of the murine
immunoglobulin
-light chain enhancer
(AGTTGAGGGGACTTTCCCAGGC), mutant NF
B
(AGTTGAGGcGACTTTCCCAGGC; substitution in lowercase), and
GAS/ISRE (AAGTACTTTCAGTTTCATATTACTCTA) (Santa Cruz
Biotechnology) were radiolabeled with [
-32P]ATP and T4
polynucleotide kinase (New England Biolabs, Beverly, MA) and purified
by gel filtration through a 1.0-ml column of Bio-Gel P-6 polyacrylamide
beads (Bio-Rad). Nuclear extracts (5-10 µg) were incubated in a
total volume of 20 µl containing 32P-oligonucleotide
(20,000 cpm), 2 µg of poly(dI·dC) (Boehringer Mannheim), 10 µg of
bovine serum albumin, 10 mM Tris·HCl, pH 7.5, 50 mM NaCl, 1 mM dithiothreitol, 1 mM
EDTA, and 5% glycerol. Reactions were incubated at room temperature
for 20 min, then at 4 °C for 10 min. In some experiments, increasing
amounts of cold competitor oligonucleotides were added 5 min prior to
addition of labeled probe. For supershift analysis, 2 µg of the
indicated antibodies were incubated with nuclear extracts for 15 min
prior to addition of labeled probe. Binding complexes were resolved on
5% nondenaturing polyacrylamide gels via electrophoresis in 0.5 × Tris/borate/EDTA buffer. Gels were dried and exposed to film for
16-24 h.
As a measure of NFB activity,
luciferase reporter constructs containing the thymidine kinase promoter
(pTK) alone or downstream of three tandem NF
B binding sites
(p
B-TK; AGCTTGGGACTTTCCATGGGACTTTCCTAGGGATTCCCC) were used. SMC were
seeded at 2 × 104 cells/cm2 in 6-well
plates and allowed to attach overnight. Cells were incubated 5 h
in serum-free Opti-MEM medium (Life Technologies, Inc.) containing 1 µg of reporter plasmid and 5 µl of LipofectAMINE (Life
Technologies, Inc.). Co-transfection with pRSV-
-Gal (0.2 µg) was
used in all experiments to correct for variations in transfection efficiency.
-Gal activity was invariant with experimental
treatments. After overnight recovery in Dulbecco's modified Eagle's
medium, 10% serum, cells were incubated 48 h in serum-free IT
medium or maintained in serum-containing medium. Cells were then
stimulated with 10% serum, PDGF-BB, IL-1
, or vehicle (IT medium)
and incubated for an additional 24 h. Experiments were terminated
by two washes with ice-cold PBS and addition of 200 µl/well lysis
buffer (100 mM KPO4, pH 7.8, 0.2% Triton
X-100, 1 mM dithiothreitol). Luciferase and galactosidase
activities were measured in 20-µl aliquots over a period of 5 s
using the Tropix detection system (Bedford, MA). Relative luciferase
activity was calculated by dividing luciferase by
-Gal activity. For
neutralization experiments, preparations of 10% FCS, 30 ng/ml PDGF-BB,
or IT medium were incubated with or without 200 µg of PDGF-BB
neutralizing antibody for 2 h at 37 °C prior to addition to
cell cultures.
Sections of human aortae and
atherosclerotic carotid artery were obtained from transplantation
donors and at endarterectomy, respectively. Paraffin-embedded sections
were deparaffinized with xylene and rehydrated with graded steps of
ethanol, then incubated with 2.5% normal goat serum in PBS. p65c and
p50 polyclonal antisera or nonimmune rabbit IgG (1.5 µg/ml) diluted
in 2.5% goat serum in PBS were added to sections for 2 h at room
temperature. After three washes with PBS, sections were layered with
biotinylated goat anti-rabbit IgG (Vector Laboratories) for 45 min.
Complexes were detected with streptavidin-conjugated Texas Red or
fluorescein isothiocyanate (Amersham Corp.). Some sections were
counterstained for cell nuclei with 0.5 mg/ml bis-benzimide (H-33258)
(Calbiochem) in PBS. Antibody specificity for immunostaining was
assessed by immunocytofluorescent staining of TNF--stimulated human
saphenous vein endothelial cells with p65c and p50 antisera, with or
without prior absorption with a 100-fold excess corresponding cognate peptides. Fluorescence was observed using an Olympus BX60F microscope (Olympus Optical Co., Ltd, Japan).
Because of the unsettled nature of the identification of
Rel subunits utilized by SMC we determined the spectrum of Rel family members expressed by saphenous vein SMC by immunoblot analysis of whole
cell lysates. SMC contained prominent levels of p65 (Rel-A) compared
with c-Rel or Rel B, detected as weaker bands of similar size on
SDS-PAGE gels (Fig. 1). c-Rel and Rel B in SMC
comigrated with c-Rel and Rel B in Jurkat cell lysates and were
abolished by preincubation of antisera with the corresponding cognate
peptides (data not shown). Antisera to p50 (NFB1) recognized a
~50-kDa protein, whereas antisera to p49 (NF
B2) recognized a
single 100-kDa protein, suggesting that cultured SMC expressed p49
mostly as the unprocessed precursor p100. Human aortic SMC exhibited a
similar profile of Rel protein expression, indicating similar
expression of Rel proteins by SMC cultured from differing human
vascular beds.
Human SMC Cultured in Serum-containing Medium Exhibit Two Complexes of Constitutive and Inducible NF
Previous studies reported that cultured bovine SMC
exhibit constitutive NFB binding activity composed of p50 complexed
with a putative Rel protein termed SMC-Rel (16). To identify the Rel
proteins in human SMC that participate in binding DNA under basal and
cytokine-stimulated conditions, nuclear extracts were prepared from
saphenous vein or aortic SMC following 2 h of incubation with or
without IL-1
, a potent stimulus of NF
B activation. As the NF
B
proteins involved in DNA binding have been identified in human
endothelial cells (14), for comparison nuclear extracts were prepared
from saphenous vein endothelial cells treated with or without TNF-
.
Both venous and aortic SMC expressed two bands of specific DNA-binding
complexes under unstimulated conditions, referred to as complexes I and
II (Fig. 2a). Stimulation with IL-1
for
2 h markedly increased binding of complex I and slightly increased
binding of complex II. Inclusion of polymyxin B (50 µg/ml) in the
experiments did not inhibit binding activity, arguing against
activation of NF
B by lipopolysaccharide contaminants (data not
shown). In comparison to SMC, unstimulated endothelial cells showed
little to no complex I and low levels of complex II, and binding
activity was readily induced by stimulation with TNF-
. Of note,
complexes I and II in SMC and EC comigrated in nondenaturing gels,
suggesting a similar subunit composition. Unlabeled wild type NF
B
probe inhibited binding activity to a much greater extent than mutant
NF
B or GAS/ISRE probe, indicating the specificity of the DNA-binding
complexes for NF
B motifs (Fig. 2b). A third, faster
migrating complex (NS) likely represents a low affinity
protein interaction with the NF
B probe used in these experiments as
indicated by 1) lack of regulation by cytokine stimulation, 2) less
competition or a smearing of the complex by cold competitor
oligonucleotide, and 3) no interaction with recombinant I
B
added
to the binding reaction (data not shown).
The subunit composition of basal and cytokine-induced DNA-binding
complexes in human SMC was ascertained by supershifting with a panel of
antisera to Rel family members (Figs. 3, a and b). Complex I in both unstimulated and
IL-1-stimulated cells interacted with two different antisera to the
carboxyl or amino terminus of p65 (p65c and p65a, respectively), and
with antisera to p50. Complex II interacted only with antisera to p50.
Moreover, addition of both p65 and p50 antisera abolished all binding
activity in nuclear extracts from unstimulated SMC. Antisera to p49,
c-Rel, Rel B, or nonimmune rabbit IgG did not affect either complexes I
or II. Thus, in cultured human SMC, complex I likely contains heterodimers of p65 and p50, and complex II likely contains homodimers of p50. Human endothelial cells stimulated with TNF-
have
identical subunit composition (14) (data not shown).
Loss of I
NFB signaling in cell types such as lymphocytes and
macrophages is self-limited due to postinduction synthesis of the
NF
B inhibitor I
B
(7); however, such an autoregulatory system has not been identified in SMC. To characterize further NF
B
signaling in human SMC, the fate of inhibitor proteins I
B
and
I
B
was followed over time in nuclear and cytoplasmic fractions
from SMC stimulated with IL-1
or TNF-
. I
B
occurs as a
~37-kDa protein in the cytosol of unstimulated SMC (Fig.
4). I
B
rapidly (
0.5 h), yet transiently
disappeared following stimulation with IL-1
or TNF-
, and returned
to detectable levels at 2 h. The nuclear fraction contained no
I
B
up to 24 h after cytokine stimulation, indicating that
the disappearance of I
B
from the cytosol resulted from
proteolysis rather than translocation to the nucleus. Inhibition of
IL-1
- or TNF-
-induced depletion of I
B
by MG132, a potent inhibitor of proteasome activity, confirmed this interpretation (data
not shown) (22). I
B
occurs as a ~46-kDa protein in unstimulated SMC. Two hours following treatment with IL-1
, levels of I
B
markedly decreased, but remained detectable. In contrast to I
B
, levels of I
B
remained low throughout the period of treatment with
IL-1
. By comparison, TNF-
affected levels of I
B
little for
up to 24 h (Fig. 4). Loss of cytosolic I
B
correlated
temporally with induction of NF
B binding activity in nuclear
extracts from SMC cultured in parallel. Interestingly, the duration of
increased NF
B activity correlated inversely with levels of I
B
protein. Loss of cytoplasmic I
B
in response to IL-1
associated
with NF
B-DNA binding activity that persisted for 24 h. TNF-
stimulation, which affected I
B
levels little, induced transient
NF
B-DNA binding activity that peaked by 1 h and declined to
near basal levels by 6 h (Fig. 4).
Consistent with the loss of IB inhibitor proteins and increased
NF
B binding activity, nuclear fractions contained increased amounts
of p65 and p50 protein following IL-1
stimulation, determined by
immunoblotting (Fig. 5). Increased levels of nuclear
p65/p50 persisted for 24 h compared to unstimulated SMC,
consistent with sustained NF
B binding activity observed during this
time (Fig. 4). Nuclear extracts from unstimulated SMC contained low
amounts of both p65 and p50, consistent with basal levels of DNA
binding activity (Figs. 2 and 3). Levels of cytosolic p65 and p50 did not change appreciably during the experiments, suggesting that a small
fraction of the total pool of these NF
B dimers translocated to the
nucleus in response to cytokine stimulation. In striking contrast to
the changes in p65 and p50, c-Rel remained cytosolic, and no nuclear
accumulation of c-Rel was observed in response to IL-1
, indicating
selectivity in mobilization of Rel dimers in human SMC.
Serum Constituents Regulate NF
The above experiments examined NFB family members in
smooth muscle cells cultured in serum-containing medium. Since serum may contain or induce production of factors that can activate NF
B,
experiments were performed in SMC cultured in a defined serum-free
medium (IT medium). Removal of serum for 48 h strikingly decreased
both basal NF
B-DNA binding as well as luciferase activity from a
heterologous promoter construct containing three tandem repeat
NF
B-binding elements (Fig. 6, a and
b). Reintroduction of serum (10%) restored
NF
B binding and luciferase activity to "basal" levels, which
persisted for at least 24 h. Platelet-derived growth factor
(PDGF), a serum-associated mitogen also expressed in human atheroma,
alters many functions of SMC relevant in atherogenesis. PDGF-BB
increased NF
B-DNA binding and luciferase activity in IT-cultured
SMC, but did not further increase basal levels of NF
B-DNA binding in
SMC maintained in serum (Figs. 6, a and b). A
neutralizing PDGF antibody abolished the stimulatory effect of PDGF-BB
on
B-dependent luciferase activity but failed to block the stimulatory effect of 10% serum (Fig. 6c), indicating
that factors other than PDGF contribute to constitutive NF
B activity present in SMC cultured in serum. Human SMC responded similarly to
IL-1
either in serum-free or serum-containing medium. Thus, "constitutive" NF
B activity in cultured human SMC likely results from the presence of serum constituents or, perhaps, the proliferative status of the cell population.
Expression of p65 and p50 in Normal and Atherosclerotic Human Vessels
In the normal vessel wall, SMC do not encounter mitogens
or certain other constituents of serum, an unphysiologic fluid. We therefore tested whether SMC in the normal artery wall exhibit activation of NFB. To this end, we examined sections of aortae obtained from transplantation donors for the localization of p65 and
p50 by indirect immunofluorescence. Staining with rabbit polyclonal antisera to p65 or p50 (red fluorescence) and counterstaining with
bis-benzimide to visualize nuclei (blue fluorescence) yielded diffuse
red cytoplasmic staining around blue nuclei of medial SMC (Fig.
7a). Staining with nonimmune rabbit IgG
yielded no signal. A filter that allowed single transmission of red
fluorescence revealed dark, nonfluorescent nuclei surrounded by red
fluorescence, indicating no nuclear staining with either p65 or p50
antisera. In contrast, sections of human atheroma obtained at carotid
endarterectomy showed cells with clear nuclear red fluorescence when
stained with anti-p50 and anti-p65 antibodies. Co-staining with
-actin antibodies (HHF-35) identified these cells as smooth muscle
(green-yellow color). Specificity of p65c and p50 antisera was assessed
in TNF-
-stimulated human saphenous vein endothelial cells, as the
identity of Rel subunits mobilized by TNF-
in this cell type is
established (14). Immunofluorescent staining with p65c and p50 antisera
yielded diffuse cytoplasmic but little nuclear fluorescence.
Stimulation with TNF-
for 1 h yielded intense nuclear
fluorescence for p65 and increased nuclear fluorescence for p50,
indicative of nuclear translocation of these Rel subunits.
Preabsorption of antisera with a 100-fold excess of corresponding
cognate peptide abolished fluorescence of TNF-
-stimulated
endothelial cells (Fig. 7b). Thus, in medial SMC of the
normal vessel wall, NF
B proteins p65 and p50 are restricted to the
cytosol, in contrast to SMC cultured in serum or in atherosclerotic
lesions.
This study investigated the role of the NFB signaling system in
the regulation of functions of human SMC of importance in human
atherosclerosis. Recent studies suggest that SMC in culture express
basal, constitutive NF
B activity (16, 23). Prior to this, only
lymphocyte cell lines and neurons were known to exhibit constitutive
NF
B activity (24-26). Curiously, the Rel proteins that comprise
this basal activity differ in each cell type. In lymphocytes and
neurons, the basal complexes contain Rel-B heterodimers and p65/p50
dimers, respectively. In SMC, the basal complexes are thought to
contain p50 and a putative Rel protein termed SMC-Rel (16). The results
reported here suggest rather that the basal NF
B complexes in
cultured human SMC contain p65/p50 heterodimers and p50/p50 homodimers,
based on immunoblot and gel shift analyses. This discrepancy could
reflect species differences between the bovine SMC used in the former
study and the human SMC used herein. Indeed, bovine aortic SMC exhibit
considerable levels of basal NF
B activity compared with human SMC,
indicating that at least regulation of the NF
B system differs
between these two cell types.2
Alternatively, SMC-Rel-containing complexes may not
bind to the
immunoglobulin enhancer motif used in the current
experiments. This is unlikely, however, since basal NF
B activity in
bovine SMC activated transcription from a reporter construct containing multimerized elements of the
light chain enhancer (16). Moreover, experiments in human SMC using the tandem
B motif from the VCAM-1 promoter also identified basal complexes as containing p65 and p50
(23). Thus, the data suggest that, unlike bovine SMC, basal NF
B
complexes in human SMC cultured in serum contain "classical" NF
B, i.e. p65 and p50.
Since the aforementioned studies examined SMC in culture, it is
possible that culture conditions provoked a low level of basal NFB
activity, an activity lacking or not present in SMC of the vascular
wall. Previous reports indicate that basal NF
B activity in cultured
bovine SMC does not depend on serum growth factors (16). In human SMC,
removal of serum considerably decreased "constitutive" NF
B-DNA
binding as well as transcriptional activity from a reporter plasmid
containing tandem
B-binding elements. NF
B activity was restored
within hours upon reintroduction of either serum or PDGF, a growth
factor considered important in atherogenesis. It is likely that serum
constituents other that PDGF-BB chain contribute to the stimulatory
effect of serum since a neutralizing PDGF-BB antibody failed to reduce
serum-induced increases of NF
B activity, despite abolishing the
stimulatory effect of PDGF. This finding suggests that constitutive
NF
B activity in vitro results from exposure to serum, an
unphysiologic medium, rather than being an intrinsic feature of SMC.
Moreover, SMC exposed to growth factors such as PDGF may result in low
level, persistent activation of NF
B that may contribute to sustained
activation of this cell type at sites of vascular lesions (33). It
should be noted that some residual NF
B activity remained even in SMC cultured without serum. This residual activity likely resulted from
autocrine production of growth factors or IL-6, the latter of which is
released from cultured SMC and can increase NF
B activity (17, 27).
Nevertheless, these experiments cannot exclude this residual activity
as basal NF
B activity in SMC in serum-free culture. A more
definitive answer was provided by in situ analyses of human
aortae from transplantation donors for the presence of Rel proteins p65
or p50. Nuclei within medial SMC showed no immunoreactivity with Rel
antibodies, arguing against the presence of basal NF
B activity in
SMC of the normal vessel wall. Together, the data support the view that
NF
B is restricted to the cytosol in quiescent SMC, as occurs in most
cell types examined, and that nuclear translocation of Rel proteins and
increased NF
B activity occurs in response to cytokine as well as
mitogenic stimulation.
In cell types such as macrophages and lymphocytes, activation of the
NFB system is self-limited due to increased synthesis of the NF
B
inhibitor protein I
B
. Such an autoregulatory mechanism ensures
transient activation of NF
B and circumvents a potential NF
B-mediated positive feedback loop of target gene expression in
cells exposed to inflammatory stimuli. In SMC stimulated with IL-1
,
increased levels of p65/p50 in the nucleus persisted for at least
24 h, despite the rapid reappearance of cytosolic I
B
to
prestimulation levels. SMC thus appear to have a limited ability to
curb NF
B activity induced by IL-1
. Another member of the I
B
family may thus regulate NF
B in SMC. Indeed, the current results
show that human SMC express the recently cloned NF
B inhibitor I
B
(13) and that treatment with IL-1
induces a sustained reduction in levels of I
B
protein. Unlike I
B
, cytosolic
pools of I
B
are not restored following activation and thus may
result in persistent activation of NF
B. Indeed, sustained decreases in I
B
appear to cause persistent NF
B activity in T cells
stimulated with IL-1 or lipopolysaccharide and in human endothelial
cells treated with TNF-
(13, 28). Consistent with this view, TNF-
fails to modulate I
B
levels substantially in human SMC during 24 h of incubation, whereas regulation of I
B
levels
resembles that observed in response to IL-1
. In this case, increased
NF
B activity is transient; indeed NF
B activity now declines as
cytosolic levels of I
B
increase. Thus, regulation of I
B
and
possibly the nature of the NF
B response to cytokine stimulation
(i.e. transient versus sustained activation)
likely differs between human smooth muscle cells and endothelial cells.
Potential functional consequences of divergent NF
B regulation in SMC
and endothelial cells are of particular interest based on the close
proximity of these two cell types within the vessel wall.
An alternative explanation for persistent NFB activity is that
p65/p50 may have a relatively long half-life in the nuclear compartment. That I
B
did not appear in the nucleus throughout the
period of treatment precludes disruption of Rel/DNA interactions (29,
30). Modification of Rel subunits, such as phosphorylation following
activation, may be important in this regard (31). Alternatively,
turnover of I
B
increases following cytokine stimulation (32),
which may allow a continuous low level of nuclear translocation of
p65/p50, accounts for prolonged NF
B activity in this cell type.
Although it remains uncertain whether NF
B activity in SMC exposed to
IL-1 results from continued translocation of cytosolic pools of NF
B
or prolonged residence of nuclear NF
B proteins, such sustained
NF
B activity may promote long term expression of gene products and
maintain SMC in an activated or "primed" condition. In this regard,
intimal SMC in atheroma show chronic expression of VCAM-1 (3, 33), an
NF
B dependent process, and the current results show that NF
B is
present in the nucleus of these cells.
Regulation of NFB activity has emerged as a potentially important
pathway in mediating specific functions of vascular endothelium pertinent to atherogenesis, such as expression of some adhesion molecules and PDGF (34). In vascular smooth muscle, activation of
NF
B regulates expression of the adhesion molecule VCAM-1 and occurs
during cell growth induced by serum or thrombin (17, 23, 35). The
presence of the nuclear Rel protein p50 in intimal SMC within human
atheroma shown here and of p65 shown here and elsewhere (8) support a
role for this transcription factor in expression of SMC products of
potential importance to progression of vascular lesions, including
cytokines, growth factors, and proteins involved in coagulation. This
study clarified the nature of the Rel proteins expressed in human SMC
and characterized activation of this pathway in response to pertinent
proinflammatory stimuli, IL-1
and TNF-
. An understanding of the
NF
B system in SMC should allow a more rational basis for elucidating
potential roles of NF
B in mediating specific functions of SMC in
vascular diseases.
We thank Marysia Muszynski and Elissa Simon-Morrissey (Brigham and Women's Hospital) for their skillful assistance, Dr. Edward Murray (Roche, Welwyn Garden City, Herts, UK) for the generous gift of luciferase reporter constructs, and Dr. Rosalind Fabumni for helpful discussions.