| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Anesthesiology, University of Illinois at Chicago, Chicago, Illinois 60612, USA
1Correspondence: Anesthesiology M/C519, University of Illinois at Chicago, 1819 W. Polk St., Chicago, IL 60612, USA. E-mail: egalea{at}uic.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSIGNAL TRANSDUCTION PATHWAYS...EFFECTS OF INTRACELLULAR cAMP...MOLECULAR MECHANISMS UNDERLYING...CONCLUDING REMARKSREFERENCES |
|---|
B, and C/EBP. Cyclic AMP must therefore
be considered a modulator rather than a suppressor of NOS2 expression.
This review summarizes evidence derived from in vitro
studies, considers regulation of NOS2 by cAMP in vivo,
and discusses possible therapeutic applications of cAMP
treatment.—Galea, E., Feinstein, D. L. Regulation of the
expression of the inflammatory nitric oxide synthase (NOS2) by cyclic
AMP.
Key Words: transcription factors • gene expression • promoter • gene structure • mRNA expression • cytokines • endotoxin
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSIGNAL TRANSDUCTION PATHWAYS...EFFECTS OF INTRACELLULAR cAMP...MOLECULAR MECHANISMS UNDERLYING...CONCLUDING REMARKSREFERENCES |
|---|
, 2
). On the one hand, it can eliminate infiltrating
microorganisms, reduce thrombosis, and improve blood supply to injured
tissues. In contrast, excess production of NO can cause tissue damage
and contribute to the development of a wide spectrum of diseases
including septic shock, rheumatoid arthritis, cerebral ischemia,
multiple sclerosis, and diabetes. It has been speculated that NO is
such an efficient mechanism of protection that evolution has provided
most cells with the capacity to express NOS2 despite the associated
risk of damage (2)
. Conceivably, regulatory mechanisms
exist that control the timing and intensity of NO production by NOS2 in
order to outweigh protective actions against detrimental ones.
For more than 10 years, 3', 5' cyclic adenosine monophosphate (cyclic
AMP, cAMP) has been considered a ubiquitous regulator of inflammatory
and immunological reactions. Increases of intracellular cAMP block T
lymphocyte activation (3)
, inhibit interleukin 1ß
(IL-1ß) and tumor necrosis factor 
(TNF-) release from
macrophages (4
, 5)
, and, in endothelial cells, reduce
expression of the adhesion molecules that mediate lymphocyte
infiltration into damaged or infected tissues (6)
. In
brain glial cells, the elevation of cAMP by the endogenous
neurotransmitter noradrenaline inhibits LPS-induced IL-1ß expression
(7)
and interferon 
(IFN) -induced MHC class II
expression (8)
. This evidence has led to the notion that
one physiological role of cAMP-dependent signaling is to suppress
immune responses, thus minimizing damage to the host. However,
regulation of NOS2 by cAMP does not conform to this general pattern,
since cAMP can either stimulate or inhibit NOS2 expression, depending
on the cell type. Here we will review the biochemical evidence for the
dual regulation of NOS2 by cAMP, analyze possible molecular mechanisms
involved, and discuss whether such regulation is of biological or
therapeutical relevance.
|
SIGNAL TRANSDUCTION PATHWAYS MEDIATED BY cAMP |
|---|
|
TOPABSTRACTINTRODUCTIONSIGNAL TRANSDUCTION PATHWAYS...EFFECTS OF INTRACELLULAR cAMP...MOLECULAR MECHANISMS UNDERLYING...CONCLUDING REMARKSREFERENCES |
|---|
, 10
). Binding of ligands to
G-protein-coupled receptors induces release of stimulatory Gs proteins
that activate the enzyme adenylate cyclase, which converts bound ATP to
cAMP (11)
. Increase of intracellular cAMP results in
activation of protein kinase A (PKA), a tetramer of two catalytic (C)
and two regulatory (R) subunits. On binding of cAMP, the R and C
subunits rapidly dissociate, allowing C subunits to phosphorylate a
wide spectrum of proteins including membrane receptors, cytoskeleton
proteins, and a wide variety of phosphatases, kinases, and other
enzymes (12)
. Active PKA C subunits can also rapidly enter
the nucleus (13)
, where they phosphorylate and
activate transcription factors (9)
including CREB (cAMP
response element binding protein), CREM (cAMP response element
modulator), ATF-1 (activating transcription factor-1), and AP-2
(activator protein-2). Signaling by cAMP characteristically is
transitory and followed by a refractory period (14)
.
Deactivating mechanisms are multiple and include desensitization of
G-coupled receptors, degradation of cAMP by phosphodiesterases,
inhibition of PKA by endogenous inhibitors, and activation of
phosphatases that remove phosphorylations carried out by active PKA.
Both stimulatory and inhibitory components of the cAMP-dependent
cascade exist in several isoforms with different tissue distribution
and regulatory properties (8
, 11
, 15)
, providing a wide
assortment of endogenous means to regulate cAMP signaling. Experimentally, activation of cAMP-dependent signaling has been achieved by 1) activation of G-coupled receptors, 2) direct activation of adenylate cyclase with forskolin, 3) inhibition of cAMP degradation using selective phosphodiesterases inhibitors (i.e., rolipram or IBMX), 4) administration of membrane permeable cAMP analogs resistant to degradation by phosphodiesterases, including dibutyryl cAMP (dbcAMP) and 8-bromo-cAMP, and 5) inhibition of phosphatases with okadaic acid.
|
EFFECTS OF INTRACELLULAR cAMP ON NOS2 EXPRESSION IN VARIOUS CELL TYPES |
|---|
|
TOPABSTRACTINTRODUCTIONSIGNAL TRANSDUCTION PATHWAYS...EFFECTS OF INTRACELLULAR cAMP...MOLECULAR MECHANISMS UNDERLYING...CONCLUDING REMARKSREFERENCES |
|---|
. Characteristic inducers are IL-1ß, TNF-,
and IFN, which often act synergistically. Sustained expression of
NOS2 results in the production of vast amounts of NO (16)
,
which is able to kill bacteria, viruses, and tumor cells, but
eventually is deleterious to the host cells, thus contributing to the
pathogenesis of diseases like septic shock (17)
and
rheumatoid arthritis (18)
. The pathophysiological
relevance of NOS2 expression in macrophages has prompted numerous
studies regarding its regulation by cAMP. As a result, a puzzling
picture has emerged: as many studies report an inhibitory role as
describe a stimulatory effect of cAMP.
Evidence that cAMP can up-regulate NOS2 expression was initially
reported in rat Kupffer cells, the resident macrophage in liver
(19
, 20)
, and then extended to rat peritoneal macrophages
(21
22
23)
, bone marrow-derived macrophages
(24)
, and macrophage cell lines (25
26
27
;
Table 1
). The analogs dbcAMP and 8-bromo-cAMP, as well as forskolin, IBMX,
rolipram, and cholera toxin, were used to elevate intracellular cAMP
and cannabinol was used to decrease it. The overall conclusion drawn
from these studies is that cAMP-dependent pathways can stimulate
macrophage NOS2 expression up to fivefold. Additional evidence for a
potentiating action of cAMP comes from studies of endogenously released
prostaglandins. Thus, antibodies against prostaglandin
E2 (PGE2) (19)
or inhibition of the enzyme that synthesizes
PGE2, cyclooxygenase (24
, 25)
,
decreased the endotoxin or cytokine-induced expression of NOS2. Since
PGE2 increases cAMP levels in macrophages
(64)
, these results suggest that endotoxin or cytokines
induce the release of PGE2, which potentiates
NOS2 expression via elevation of intracellular cAMP.
|
It should be noted that most of the early studies of NOS2 expression
relied on measurements of nitrite (a stable metabolite of NO) levels in
the cell culture media from using Griess reagent (65)
.
Although not highly sensitive, the measurement of nitrites by this
method provides an adequate measurement of NOS2 expression because the
continuous production of NO by NOS2, combined with the stability of
nitrites, results in sufficiently high nitrite concentrations. With the
availability of NOS2 cDNA clones and selective anti-NOS2 antibodies,
NOS2 expression has been more directly assessed by measurements of NOS2
mRNA levels and immunoblot analysis.
In direct contradiction to the above results, several other studies
have described an inhibitory effect of PGE2,
8-bromo-cAMP, IBMX, and forskolin on NOS2 expression. For example, in
the mouse macrophage J774 cell line, induction of NOS2 by LPS was
decreased in a dose-dependent manner by coincubation with nanomolar
amounts of PGE2 or iloprost (a stable
PGI2 analog), which increase cAMP, but not by
incubation with PGF2a, LTB4, or LTC4, prostanoids, which do not elevate
intracellular cAMP (30)
. Bulut et al. (31)
also observed inhibition of NOS2 activity by PGE2
in J774 cells, but required coincubation with IBMX, perhaps necessary
to achieve higher and prolonged elevation of cAMP levels. In these
studies (31)
, incubation with Rp-cAMP, a PKA inhibitor,
reversed the effects of PGE2, indicating a role
for cAMP and PKA. More recently it was shown that
PGE2, PGI2, and forskolin
could also decrease NOS2 expression in J774 cells in the absence of
IBMX (32)
, possibly due to having preincubated the cells
with cAMP-elevating agents before the addition of LPS. Furthermore,
inhibitory effects were not observed if the prostanoids were added
after LPS addition, suggesting that cAMP interfered with early events
in NOS2 expression, such as gene transcription, rather than with later
events like mRNA turnover or protein translation.
Apart from the macrophage J774 cell line, inhibitory effects have also
been described in macrophages from various tissue origins. In
peritoneal macrophages, both LPS-induced NO production and the
cytostatic action against tumor cells were increased by the
cyclooxygenase inhibitor indomethacin and reduced by coincubation with
PGE2 (28)
. It was concluded that
this effect was mediated by cAMP, since treatment of cells with IBMX or
8-bromo-cAMP also reduced NO production. In rat liver macrophages,
PGE2 dose-dependently suppressed LPS-dependent
nitrite accumulation (66)
. Finally, in a more
comprehensive study, forskolin treatment inhibited induction of NOS2
mRNA, protein, and nitrite accumulation in Kupffer cells
(29)
. Inhibition was also observed using dbcAMP, cholera
toxin, and isoproterenol, although to a lesser extent than with
forskolin; in all cases, the phosphodiesterase inhibitor IBMX further
increased the inhibitory effects.
What is the basis for the conflicting effects of cAMP? Consideration of
the following studies may help reconcile the discrepancies. In the
first one (25)
, low concentrations of
PGE2 increased NOS2 expression in J774 cells
whereas higher doses caused inhibition. Second, Mullet et al.
(26)
reported that in the ANA-1 macrophage cell line,
dbcAMP and the PKA activator Sp-cAMP lost the capacity to stimulate
NOS2 or were clearly inhibitory when used at high (>100 µM)
concentrations (26)
. Together, these results suggest that
the extent of activation of cAMP-dependent pathways dictates whether
NOS2 expression is increased or decreased. Since the production of
IL-1ß and TNF- (which can synergistically stimulate NOS2
expression) is blocked by increased intracellular cAMP (5
, 67)
, it is possible that the inhibitory effects of cAMP on NOS2
expression are the result of reduced release of proinflammatory
cytokines. If so, a cAMP inducer like PGE2 would
be at the core of an autocrine regulatory loop: causing limited
increases in cAMP at low levels, PGE2 would
enhance NOS2 expression directly; at higher levels, associated with
larger cAMP increases, it would decrease NOS2 expression due to
inhibition of cytokine release. This may explain why studies in which
NOS2 expression is activated with LPS, which induces maximal
stimulation and release of high amounts of PGE2, show inhibition of
NOS2 by cAMP (29
, 66)
whereas studies using single
cytokines, leading to less release of PGE2, report a stimulatory effect
of cAMP (19
, 22
, 24
, 39)
. This interpretation implies that
the PGE2/cAMP system may have a dual biological role in macrophages,
which is to enhance initial inflammatory responses while inhibiting
later events, lest excessive NO production cause tissue damage.
The complexity of the findings in vivo mirrors that in
vitro. Stimulation of the sympathetic system, believed to exert
control over leukocyte activation (reviewed in ref 68
),
resulted in decreased NO production in alveolar macrophages
(69)
. In contrast, intrathecal administration of dbcAMP in
rats induced NOS2 mRNA and protein in alveolar macrophages
(33)
. One intriguing difference between the latter study
and in vitro analyses is that in cultured macrophages,
dbcAMP on its own exerted no significant effect, whereas dbcAMP alone
in vivo elicited NOS2 expression comparable to that induced
by LPS. This suggests that LPS and cAMP trigger independent signal
transduction mechanisms leading to NOS2 expression and that
cAMP-dependent signaling may be attenuated in vitro and,
thus, appear mostly as modulatory of the endotoxin- or
cytokine-elicited response.
Smooth muscle and smooth muscle-derived cells
Aorta smooth muscle
IFN, IL-1ß, and TNF- induce the appearance of NOS2 mRNA
and protein in cultured smooth muscle cells from rat aorta. This is
believed to reproduce events associated with sepsis, atherosclerosis,
and arterial injury after balloon angioplasty, where NOS2 may be
induced in arterial walls by cytokines released by locally infiltrated
leukocytes (70)
.
In contrast to results observed in macrophages, all reports to date
indicate that cAMP potentiates NOS2 expression in smooth muscle cells.
Koide et al. (35)
demonstrated not only that cultured rat
aortic smooth muscle cells could respond (albeit to a small extent) to
both forskolin and dbcAMP, but, more important, that these reagents
increased (by up to fourfold) NOS2 expression induced by IFN. These
findings have been reproduced using IL-1ß or TNF- to induce NOS2
expression, and phosphodiesterase inhibitors or stimulators of PKA to
elevate cAMP (40)
. The fact that comparable potentiation
by cAMP was observed with either IFN, IL-1ß, or TNF- indicates
that, in smooth muscle cells, cAMP is most likely up-regulating a step
common to all cytokines.
Agonists of G-coupled receptors, including the ß-adrenoreceptor
agonist isoproterenol (36)
, CGRP (71)
,
prostanoids (35)
, and adenosine (42
, 72)
,
also potentiated the induction of NOS2 by cytokines, although the
effect was moderate compared to the one elicited by forskolin or
lipophilic cAMP analogs. This is not surprising, as the accumulation of
intracellular cAMP after stimulation of G-protein-coupled receptors is
lower than the accumulation achieved with direct stimulation of
adenylate cyclase or the use of nonmetabolizable cAMP analogs
(35)
. Hence, the effect of cAMP on expression of NOS2 in
smooth muscle is, as in macrophages, dependent on the intensity of
stimulus and can be incrementally regulated. However, unlike
macrophages (25
, 26)
, the highest concentrations of cAMP
analogs tested do not result in inhibition of NOS2 expression in smooth
muscle cells.
In the absence of cytokines, elevation of cAMP alone could also
increase expression of NOS2 mRNA and protein in smooth muscle cells,
although the observations are conflicting, ranging from no or modest
effects (35
, 36
, 40
, 72)
, to a 13-fold increase in nitrite
accumulation when using 8-bromo-cAMP (1 mM) (38)
. IL-1ß
by itself caused a modest but significant increase in intracellular
cAMP (40)
, and the induction of nitrite release due to
IL-1ß was blocked by a selective inhibitor of PKA. This indicates
that cAMP may be involved in IL-1ß-dependent signaling pathways
leading to NOS2 expression in smooth muscle cells.
Whether pharmacological stimulation of cAMP-dependent signaling in smooth muscle cells is therapeutically appropriate is complicated by the dual protective/deleterious effects of NO. For example, the general hypotension occurring in septic shock may be due to vasodilatory actions of NO released by NOS2, and therefore an increase of NOS2 by cAMP would exacerbate the disease. In contrast, in atherosclerosis and balloon-induced injury, increased production of NO by cAMP could prove beneficial by inhibiting platelet and leukocyte adhesion as well as the proliferation of smooth muscle cells.
Renal mesangial cells
The state of contraction of mesangial cells is a finely regulated
parameter that determines the glomerular filtration rate. In normal
conditions, contraction is induced by hormones like angiotensin II and
relaxation is caused by NO released from endothelial cells. In cultured
mesangial cells, incubation with cAMP analogs, forskolin, or cholera
toxin resulted in significant NOS2 expression when administered alone
and dramatically potentiated the induction of NOS2 protein and mRNA by
IL-1ß or TNF- (45)
. As in arterial smooth muscle
cells, the effect of cAMP analogs was dose dependent, and stimulation
of NOS2 expression was maximal at high concentrations of the
nonmetabolizable dbcAMP (45)
. In contrast, exposure to
PGE2, which elicits a lower increase of cAMP than
forskolin or dbAMP, or to concentrations of dbcAMP less than 50 µM
did not induce expression of NOS2 (47)
. This suggests that
the potentiation of NOS2 by cAMP in mesangial cells may be a
pharmacological but not physiological phenomenon. Nonetheless,
pharmacological increase of cAMP may be of therapeutic value in certain
forms of glomerulonephritis where induction of NOS2 in mesangial cells
may lead to excessive release of NO and vasodilation, thereby
exacerbating the disease (73)
.
Cardiac myocytes
It has been proposed that NOS2 expression in cardiac myocytes may
contribute to inflammatory damage during sepsis, after transplantation,
and after ischemia-reperfusion and cause myocardial depression
(74)
. In cultured cells, the NOS2 expression induced by
IL-1ß and TNF- was enhanced severalfold by forskolin, dbcAMP, and
isoproterenol (42
, 48
49
50)
, when assessed at both the
nitrite and mRNA levels. In these reports, little or no NOS2 expression
was detected in the absence of cytokines, although cAMP alone induced
NOS2 expression in neonatal cardiac myocytes when measured by NO
production and electron microscopy (75)
.
Astrocytes and microglia
Astrocytes can express the NOS2 protein in demyelinating diseases
(76
77
78)
, cerebral ischemia (79)
, viral
infection (80)
, and Alzheimer’s disease
(81)
. Likewise, NOS2 has been detected in microglia during
demyelinating diseases (78)
, viral infection
(82)
, and cerebral ischemia (83
, 84)
,
although it is often difficult to ascertain whether the NOS2-expressing
cells are infiltrated macrophages instead of resident microglia.
Induction of NOS2 in glial cells has been widely accomplished in
vitro by exposing cell cultures to agents purported to cause brain
diseases, such as cytokines, endotoxins (85)
, and the
ß-amyloid protein precursor (86)
.
Exposure of rat astrocyte cultures (58)
or rat C6 glioma
cells (62)
to the neurotransmitter noradrenaline
significantly decreased LPS-dependent NO release and NOS2 mRNA
expression. The inhibition was mediated by ß-adrenergic receptors,
since the effects of noradrenaline were reversed by ß- but not
-antagonists and were replicated by the ß-agonist isoproterenol
(58)
. The effect of noradrenaline could be mimicked by
dbcAMP, 8-bromo-cAMP, and forskolin or by direct activation of PKA
(59)
and could be blocked by inhibition of PKA, but not
PKC (62)
. Apart from noradrenaline, other
natural-occurring cAMP-elevating agents such as endothelin-3
(60)
, PGE2 (63)
, and
dopamine (62)
have been shown to inhibit the cytokine- or
LPS-stimulated NOS 2 expression in mixed glial or microglial cultures.
In contrast, in the spontaneously transformed mouse D30 astrocyte cell
line, 8-bromo cAMP increased IL-1ß plus IFN-dependent NOS2
expression, thus suggesting that results obtained using immortalized
cell lines may not reproduce events occurring in primary cultures
(61)
.
The above observations indicate that cAMP-dependent signaling inhibits
NOS2 expression in astrocytes. Furthermore, in contrast to studies done
in most other cell types, the effect of cAMP on astroglial NOS2
expression has been achieved using endogenous ligands to elevate
intracellular cAMP and at concentrations well within the physiological
range. Although this suggests that the in vitro findings
reveal a response of biological relevance, it has not been studied
whether cAMP-elevating neurotransmitters can regulate glial NOS2
expression in vivo. Considering that in some brain areas
astrocytes are the main target of noradrenergic neuronal networks, such
as the locus ceruleus (87)
, it is conceivable
that expression of glial NOS2 could be suppressed by centrally released
noradrenaline.
Because the contribution of NOS2 to brain diseases is controversial,
the therapeutic potential of the decrease of NOS2 by cAMP in primary
glial cultures is uncertain. This is best illustrated in experimental
allergic encephalomyelitis, an animal model for multiple sclerosis,
where inhibition of NOS2 expression by different approaches has
resulted in both improvement (77)
and worsening
(88)
of the disease. The conflict may arise from the fact
that NOS2 is expressed in different glial populations as well as in
infiltrated leukocytes during the course of the disease. Hence,
depending on its source and timing, NO release may be
protective or destructive. Studies in which NOS2 expression is
manipulated in a time- and cell-specific manner are necessary to
establish the specific role of NO in brain pathologies
Adipose tissue, pancreatic cells, hepatocytes, skeletal muscle, and
endothelial cells
Brown adipose tissue (BAT) stands out as an example of a
modulatory role of cAMP-dependent signaling on NOS2 expression in
vivo under nonpathologic conditions. BAT is involved in
thermogenesis in mammals and thus is found in immediate contact with
large blood vessels, so that the heat it releases is rapidly
transmitted to the blood. The thermogenic activity of BAT is regulated
by the sympathetic release of noradrenaline, which in turn is modulated
by changes in temperature. Activation of the sympathetic system in rats
by exposure to low temperatures resulted in the appearance of NOS2
protein and activity in BAT (56)
. These findings were
reproduced in cultured adipocytes, mimicking the sympathetic activity
with noradrenaline at low concentrations, ß-adrenergic agonists, and
forskolin, all of which induced NOS2 expression in the absence of
cytokines or endotoxin (56)
. The cAMP-dependent release of
NO from adipocytes in vivo could account for the known
vasodilation that accompanies noradrenaline-induced thermogenesis and
thus facilitate delivery of heat.
Expression of NOS2 in pancreatic cells has been implicated in islet
cell death during diabetes, since inhibitors of NOS protect islet cells
from IL-1ß-induced damage and loss of insulin release
(52)
. In these cells (51)
as well as in the
rat insulinoma ß-cell line RINm5F (52)
, exposure to cAMP
analogs decreased NOS2 expression, suggesting a protective role for
cAMP-dependent regulation of NOS2.
NOS2 expression may also contribute to the damage that occurs in liver
hepatocytes during sepsis via inhibition of glycogenolysis and
gluconeogenesis. Since glucagon is released during sepsis, the effects
of this hormone on NOS2 induction have been characterized. Using
primary rat hepatocyte cultures, Harbrecht et al. (54)
and
Smith et al. (53)
demonstrated that the endotoxin or
cytokine-mediated induction of NOS2 is decreased by coincubation 1 µM
glucagon (which activates cAMP signaling), that the inhibition is
replicated by dbcAMP or forskolin, and that the effect of glucagon is
reversed by incubation with an adenylate cyclase inhibitor. These
findings are consistent with a protective role of the glucagon released
during sepsis by counteracting the actions of NOS2.
NOS2 appears in skeletal muscle during myositis (89)
. In
addition, brain endothelial cells express NOS2 protein during embryonic
and postnatal development (90)
, after cerebral ischemia
(83)
, and in Alzheimer’s disease (91)
.
Evidence for a role of cAMP in regulating NOS2 expression in skeletal
muscle or brain endothelium is limited at present to studies in cell
lines that show a stimulatory effect of forskolin or dbcAMP (55
, 57
, 92)
.
Studies in human cells
Few studies exist concerning the effect of cAMP on human NOS2
expression. In isolated human myocytes, ß-endorphins, acting via
cAMP-dependent pathways, increase expression of NOS2 (34)
.
In contrast, in the human DLD epithelial cell line, neither forskolin
nor 8-dibutyryl-cyclic AMP induced NOS2 mRNA by themselves, nor do they
alter NOS2 mRNA induction in response to cytokines (93)
.
|
MOLECULAR MECHANISMS UNDERLYING cAMP EFFECTS |
|---|
|
TOPABSTRACTINTRODUCTIONSIGNAL TRANSDUCTION PATHWAYS...EFFECTS OF INTRACELLULAR cAMP...MOLECULAR MECHANISMS UNDERLYING...CONCLUDING REMARKSREFERENCES |
|---|
, and
IFN, it is likely that a step common to these various inducers is
targeted. In some cases (e.g., in alveolar macrophages in
vivo and mesangial cells), cAMP alone is sufficient to stimulate
NOS2 expression.
Regulation of NOS2 mRNA steady-state levels and stability
Unlike the two other NOS isoforms, the NOS2 enzyme is
constitutively active due to an essentially irreversible binding of
calmodulin to the NOS2 protein (94)
, which renders NOS2
activity insensitive to changes in calcium concentrations within the
physiological range of 100 nM to 1 µM. [The calcium independence of
NOS2 is not absolute; high concentrations (5 mM) of the calcium
chelator EGTA cause up to 30% inhibition of NOS2 isolated from
macrophages (95)
.] NOS2 activity is therefore dependent
primarily on NOS2 protein concentration, as determined by NOS2 mRNA
contents, which in turn is dependent on NOS2 gene transcriptional rate
and mRNA stability. There is general consensus that the cAMP-induced
changes in NOS2 protein are accompanied by equivalent changes in the
steady-state levels of NOS2 mRNA, both in the stimulatory and
inhibitory modes. Thus, NOS2 mRNA levels are down-regulated by cAMP in
astrocytes (59
, 62)
, Kupffer cells (29)
,
hepatocytes (54)
, and islet cells (51)
and
up-regulated in macrophages (24
, 26
, 33
, 39)
, smooth
muscle cells (35
36
37
38
, 42)
, mesangial cells (45
, 46)
, myocytes (41
, 42)
, an astrocyte cell line
(61)
, adipocytes (56)
, and skeletal muscle
cells (55)
.
Most studies report no effect of cAMP on NOS2 mRNA stability, and
nuclear runoff experiments as well as promoter studies have confirmed
that cAMP regulates NOS2 gene transcription both when acting in a
stimulatory (45
, 96)
and inhibitory (29
, 62)
mode. However, it has been reported that elevations of cAMP can
stabilize NOS2 mRNA (41
, 45)
, even in cells where cAMP
reduces transcription of the NOS2 gene, and its net effect is
inhibitory (62)
. Together, this evidence points to
transcription as the primary target of cAMP-activated signaling, but
opens the possibility of a role for regulation of message stability in
NOS2 expression.
Structure of the NOS 2 promoter
Changes in gene transcription are ultimately the result of changes
in promoter activation. The NOS2 promoter has been cloned from mouse
(97
, 98)
, human (99
100
101)
, and rat
(102
, 103
; V. Gavrilyuk and D. L. Feinstein,
unpublished results). However, analysis of the effects of cAMP has so
far been carried out only with rat and mouse promoters. Since
differences in promoter sequence among species suggest different
transcriptional regulation, findings obtained with rodent promoters
should be extrapolated to other species only with caution.
The NOS2 promoter contains consensus binding sites for numerous
transcription factors that could mediate the effect of cAMP (Fig. 1
). For the purposes of this review, we will focus our attention on three
transcription factors that have been well characterized: NF-B, CREB,
and C/EBP. NF-B is the best-characterized activator of gene
expression during inflammatory and immunological events
(104)
. It consists of dimers of members of the Rel protein
family, and in resting conditions is maintained in the cytoplasm due to
tight association with inhibitory IB proteins (105)
. On
exposure to LPS or cytokines, IB is phosphorylated and degraded,
thus allowing NF-B to migrate to the nucleus and trigger
transcription.
|
There are two binding sites for NF-B within the rodent NOS2
promoters: a downstream, or proximal, site located 80 to 90 bases
upstream from the transcriptional start site, and a second, distal site
located roughly 980 bases upstream. DNA footprinting analysis of the
mouse macrophage (106)
and rat cardiac myocyte
(50)
NOS2 promoter region has provided direct proof of
occupation of NF-B binding sites after LPS stimulation. The relative
importance of the two NF-B binding sites for NOS2 induction is
controversial. In the initial characterization done in macrophage RAW
264.7 cells (97)
, removal of the proximal B site
abrogated all response to LPS induction, whereas an upstream construct,
containing the distal but not the proximal B site did not respond to
LPS plus IFN. This suggested that in macrophages, the distal B
site is not critical for induction by LPS. In contrast, using the same
mouse promoter in rat A7r5 vascular smooth muscle cells revealed that
deletion of the distal B site by mutagenesis reduced induction by
cytokines only ~35%; moreover, a 112 bp promoter region containing
only the upstream B site could confer NOS2 induction by cytokines
(107)
. Finally, it has been reported (108)
that both proximal and distal B sites were necessary for full
induction by a combination of LPS plus IFN. Together, these results
support the idea that NF-B-dependent pathways involved in NOS2
expression may vary depending on cell type and the particular
combinations of cytokines and LPS used.
A central role in inflammation for transcription factors other than
NF-B is now recognized. This is the case of C/EBP, which consists of
a family of leucine zipper transcription factors comprising six members
(,ß,,
,
,
) ubiquitously expressed in mammalian cells
that play pivotal roles in cellular differentiation, energy metabolism,
and inflammatory and immunological functions (for a review, see ref
109
). The C/EBP isoforms interact with each other to
produce multiple combinations of dimers, which are capable of
regulating transcription by binding to cognate sites in gene promoters.
The rat NOS2 promoter contains four putative C/EBP binding sites (Fig. 1)
.
An extensively documented mechanism by which activation of PKA
modulates transcription is phosphorylation of the cAMP binding protein
CREB. A member of the class of DNA binding proteins with a leucine
zipper structure, unphosphorylated CREB is constitutively present in
the nucleus. Phosphorylation by PKA catalytic subunit increases its
trans-activating capacity in two ways. First, it
increases the binding affinity of CREB homodimers for the
cAMP-responsive element (CRE) sites present in the promoter regions of
target genes. Second, it triggers the association of CREB to a CREB
binding protein (CBP), a transcriptional coactivator that acts by
recruiting and/or stabilizing factors at the transcriptional start site
(110)
. Although the rat NOS2 promoter contains two CRE
sites, the corresponding sites in the mouse promoter do not conform to
the consensus CRE motifs, suggesting that binding (or at least
canonical binding) of CREB to these sites does not mediate either the
inhibitory or potentiating effects of cAMP.
In addition to being a coactivator of CREB, CBP interacts with several
other transcription factors including C/EBP, NF-B, c-jun, c-Myb,
GATA-1, c-fos, and steroid receptors (111)
. The discovery
of CBP and its interactions with a wide variety of transcription
factors has reinforced the concept that the regulation of gene
transcription is dependent on physical interactions between multiple
factors within the nucleus.
Increase in NOS2 transcription by cAMP
Since numerous precedents for activation of NF-B by cAMP exist
(112)
, it has been explored whether NF-B activation is
increased in cells in which cAMP potentiates NOS2 expression. However,
electrophoretic mobility shift assays (EMSAs) have revealed that
neither translocation of NF-B to the nucleus nor binding to B DNA
are altered (40
, 113
, 114)
. In mesangial cells, where cAMP
alone triggers NOS2 expression to an extent comparable to that elicited
by IL-1ß, mutation of the proximal NF-B binding site in the NOS2
promoter had no effect on the induction by cAMP but completely
abolished the induction caused by IL-1ß (96)
. Taken
together, these findings suggest that cAMP-dependent potentiation of
NOS2 transcription is independent of changes in NF-B activation.
Nonetheless, it should be stressed that 1) experiments of
mobility shift assays have not addressed the potentially different
contributions of the distal vs. the proximal NF-B binding sites, and
2) cAMP-dependent phosphorylation of NF-B can increase
its transcriptional activity without altering EMSA results
(114)
.
The absence of any overt effect on NF-B activation has prompted
investigations of whether activation of other transcription factors is
altered by cAMP. To narrow down the cis-elements involved in
the induction of NOS2 expression by dbcAMP, Eberhardt et al.
(96)
tested in mesangial cells the activity of serial
deletion constructs of the NOS2 promoter. Identical levels of induction
were achieved using a promoter extending to nucleotide -1700 as
compared to one extending only to -526, indicating that the majority
of important cis-elements for induction by cAMP (as well as
by IL-1ß) are located within 526 bp of the transcriptional start
site. Focusing on the three C/EBP sites present in this region, they
observed that a shorter promoter, lacking the most distal site (CEBP1)
had slightly diminished (75%) induction compared to control, whereas
further removal of the central C/EBP site (C/EBP2) abolished cAMP
induction altogether. Similarly, point mutation of the central C/EBP2
diminished induction to ~50% of control values. These data point to
a role for C/EBP1 and 2 sites in the induction by cAMP, with C/EBP2
playing a predominant role. Using the central C/EBP2 sequence as a
probe, EMSA revealed that cAMP induces binding to this region of a
mixture of heterodimers composed of C/EBP-ß, C/EBP-, and CREB, a
finding also reported previously (50)
. No mobility shifts
were observed using the most proximal C/EBP site (C/EBP3), suggesting
that this site does mediate the cAMP effects.
It thus appears that in mesangial cells, stimulation of the cAMP-PKA
signaling pathways activate C/EBP members as well as CREB, both of
which could enhance NOS2 transcription by binding to C/EBP sites, thus
potentiating the activity of the CCAAT box. The homology between CRE
and C/EBP sites may allow for the binding of CREB to the latter sites;
not surprisingly, C/EBP proteins have also been shown to recognize the
CRE motif (115)
. It remains to be determined, however,
whether activation of the CCAAT box also underlies the actions of
increased cAMP in cells like macrophages or aortic smooth muscle cells,
in which cAMP potentiates the effect of cytokines but has no or low
stimulatory capacity on its own. An equally intriguing question is
whether the CREB-mediated activation of the CCAAT box is absent in
cells where cAMP is inhibitory.
Decrease in NOS2 transcription by cAMP
Analogous to the experiments described above, precedents exist for
inhibition of NF-B activation by elevated cAMP (116)
.
Therefore, the effects of cAMP on NF-B activation have been examined
in cells where cAMP decreases NOS2 induction. In Kupffer cells
(29)
, forskolin abolished two hallmarks of the activation
of NF-B by LPS, namely, the degradation of IB and the
translocation of the NF-B subunit p65 to the nucleus. In addition,
forskolin up-regulated the expression of IB mRNA, which possibly
led to increased cytoplasmic IB and counteracted the LPS-induced
degradation. These findings suggest that the inhibitory effect of cAMP
in macrophages is due to increased stabilization of NF-B in the
cytoplasm by IB. Conceivably, PKA activation may inhibit the
IB kinase activities involved in phosphorylation of IB prior
to its degradation, although the characterization of such reactions has
not been addressed.
Decreased NF-B activation associated with inhibition of NOS2
expression has also been described in primary astrocyte cultures
(59)
. In a second study also using primary astrocytes,
however, no changes in either NF-B nuclear uptake, DNA binding, or
subunit composition were observed up to 7 h after incubation with
noradrenaline to elevate cAMP levels (62)
. A conciliatory
explanation is that the different levels of cAMP achieved by
noradrenaline vs. forskolin or the different NF-B-activating agents
used may account for the conflicting results. Furthermore, the decrease
in NF-B nuclear migration reported was slight and difficult to
reconcile with the up to 80% decrease of NOS2 transcription induced by
cAMP (59)
. This suggests that unlike the studies of
Kupffer cells described above, mechanisms other than regulation of
nuclear uptake of NF-B account for the inhibitory effect of cAMP in
astrocytes. The absence of cAMP effect on NF-B activation has been
also reported in endothelial cells (117)
, again suggesting
that effects on other transcription factors may mediate cAMP inhibitory
actions. However, as is the case for stimulatory actions of cAMP, the
possibility of a PKA-dependent inhibitory phosphorylation of NF-B
cannot be ruled out.
Recently, a hypothesis has been offered to explain how cAMP inhibits
transcription of adhesion molecules in endothelial cells that could be
applicable to the inhibitory actions of cAMP on NOS2 expression,
involving regulation of the transcriptional coactivator CBP.
Overexpression of CBP enhances p65-mediated transcription via binding
of CBP to the carboxyl-terminal region of p65 (118
, 119)
,
thus indicating that CBP is a transcriptional coactivator of p65. Since
CBP can also bind to phosphorylated CREB, CREB and p65 may compete for
the binding of CBP. Thus, increased amounts of activated CREB may
recruit the available CBP, thereby diminishing the CBP available for
activation of p65. In support of this idea, it has been shown in
endothelial cells that the inhibitory effects of cAMP were reversed by
overexpression of CBP (119)
. Similar studies have not yet
been undertaken in the context of NOS2 expression.
To better characterize the areas of the NOS2 promoter sensitive to the
inhibitory actions of cAMP, a functional analysis of the promoter has
been carried out using C6 glioma cells as a model for astrocytes
(62)
. In C6 cells stably transfected with the mouse NOS2
promoter attached to the bacterial CAT gene, noradrenaline reduced the
LPS plus cytokine-induced reporter gene expression. In contrast, this
suppression was lost when a truncated promoter was used (which extended
only to the proximal NF-B binding site). These results confirm that
cAMP-dependent inhibition occurs at the level of NOS2 transcription and
point to an area of the NOS2 promoter located upstream of the proximal
NF-B site as the target of cAMP.
A surprising outcome of the above described studies was that incubation with noradrenaline potentiated nearly threefold the cytokine-dependent induction of NOS2 expression in cells transfected with the truncated promoter. Moreover, noradrenaline alone dramatically induced the activity of this minimal promoter, thus conferring unto these cells a phenotype more similar to that of mesangial cells, which respond to cAMP alone. These results demonstrate that in astrocytes, cAMP-dependent signaling can be both stimulatory and inhibitory of NOS2 expression and that under normal conditions (i.e., in the context of the endogenous gene or of the whole promoter), the inhibitory action prevails. These results are consistent with a model in which a cAMP-activated suppressive factor binds to an upstream location in the NOS2 promoter, a cAMP-activated potentiating factor binds to a downstream area, and, in the presence of both factors, inhibition is observed. Further analysis of these cell lines, which can be manipulated to permit either suppression or potentiation of NOS2 expression, should help clarify the mechanisms and factors mediating the dual actions of cAMP.
|
CONCLUDING REMARKS |
|---|
|
TOPABSTRACTINTRODUCTIONSIGNAL TRANSDUCTION PATHWAYS...EFFECTS OF INTRACELLULAR cAMP...MOLECULAR MECHANISMS UNDERLYING...CONCLUDING REMARKSREFERENCES |
|---|
B, and C/EBP. However, it may be premature to
pronounce cAMP a widespread regulator of NOS2 expression, as most
studies have been performed using cells or cell lines derived from the
rat. It is also important to remember that the majority of existing
evidence derives from cultured cells and that in many experiments,
elevation of intracellular cAMP levels was achieved using stable,
membrane-permeable cAMP analogs such as dbcAMP. The concentrations
tested of these analogs far exceed any increase in intracellular cAMP
due to physiological messengers such as neurotransmitters, the effects
of which have been examined in only a few instances. Therefore, it is
not yet clear that the actions of cAMP in vitro extend to
in vivo. It remains to be explored at large whether central
or peripheral neuronal networks, or cAMP-elevating agents present in
blood, are able to regulate NOS2 expression. Whether or not its effects
are physiological, increases of cAMP can be induced with therapeutic
goals in mind; however, this approach should be carefully examined in
each case, since NO can be cytotoxic or protective, depending on the
context of its release.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSIGNAL TRANSDUCTION PATHWAYS...EFFECTS OF INTRACELLULAR cAMP...MOLECULAR MECHANISMS UNDERLYING...CONCLUDING REMARKSREFERENCES |
|---|
and bacterial lipopolysaccharide. J. Immunol. 152,4110-4118[Abstract]
