| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Pharmazentrum Frankfurt, Klinikum der Johann Wolfgang Goethe-Universität, Frankfurt, 60590 Frankfurt am Main, Germany
1Correspondence: Pharmazentrum Frankfurt, Klinikum der Johann Wolfgang Goethe-Universität, Theodor Stern Kai 7, 60590 Frankfurt am Main, Germany. E-mail: geisslinger{at}em.uni-frankfurt.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONCYCLOOXYGENASE-INDEPENDENT...CYCLOOXYGENASE-INDEPENDENT...REFERENCES |
|---|
B and AP-1. The respective NSAIDs might interfere
directly with the transcription factors, but their effects are probably
mediated predominantly through alterations of the activity of cellular
kinases such as IKKß, Erk, p38 MAPK, or Cdks. These effects
apparently are not shared by all NSAIDs, since indomethacin failed to
inhibit NF-B and AP-1 activation as well as Erk and Cdk activity. In
contrast, indomethacin was able to activate PPAR
, which was not
affected by sodium salicylate or aspirin. The differences in
cyclooxygenase-independent mechanisms may have consequences for the
specific use of these drugs in individual patients because additional
effects may either enhance the efficacy or reduce the toxicity of the
respective compounds.—Tegeder, I., Pfeilschifter, J., Geisslinger, G.
Cyclooxygenase-independent actions of cyclooxygenase inhibitors
Key Words: aspirin • NSAIDs • apoptosis • metalloproteinase
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONCYCLOOXYGENASE-INDEPENDENT...CYCLOOXYGENASE-INDEPENDENT...REFERENCES |
|---|
2
3)
and cardiovascular
disease (4
5
6
7)
. Recently, celecoxib, a selective COX-2
inhibitor, was approved by the FDA for the prevention of colon cancer
in patients with familial polyposis coli, which is a hereditary
precancerous disease due to a loss of the adenomatous polyposis coli
(APC) tumor suppressor gene. NSAIDs may help to prevent or delay the
onset of Alzheimer’s disease (AD) (8
, 9)
. It is generally
accepted that the anti-inflammatory and analgesic efficacy of NSAIDs
arises mainly from inhibition of the enzymatic activity of
cyclooxygenases (COX), which convert arachidonic acid to prostaglandins
(PGs) (10)
. Two COX isoforms have been described. COX-1 is
constitutively expressed in almost all tissues and has characteristics
of a housekeeping enzyme whose activity depends solely on the
availability of its substrate. COX-1 supplies tissues with
prostaglandins required to maintain physiological organ function, such
as cytoprotection of the gastric mucosa and regulation of renal blood
flow. Conversely, COX-2 behaves as an immediate early gene
(11)
and is subject to rapid regulation at the
transcriptional level. COX-2 is up-regulated in peripheral tissues and
the spinal cord in response to tissue injury and is thought to produce
excessive amounts of prostaglandins that serve to sensitize nociceptors
and stimulate the release of other inflammatory mediators.
Prostaglandins also contribute to tumor growth by inhibiting apoptosis
(12)
and inducing the formation of new blood vessels that
sustain tumor cell viability and growth (13
, 14)
. COX
inhibition may thus explain part of the antitumor activity of certain
NSAIDs. However, some lines of evidence suggest that NSAIDs also
modulate cyclooxygenase-independent signal transduction pathways, which
may be involved in both the anti-inflammatory and antitumor activity of
these drugs (15
16
17)
.
For example, doses of aspirin necessary to treat chronic inflammatory
diseases are much higher than those required to inhibit PG synthesis.
Whereas aspirin inhibits COX activity by acetylating the enzyme
salicylic acid, which lacks the acetyl group and is ineffective as a
COX inhibitor at therapeutic doses, it is nevertheless able to reduce
inflammation (18
19
20)
. We have recently shown that
R-flurbiprofen, which was generally considered the
noncyclooxygenase-inhibiting ‘inactive’ counterpart of
S-flurbiprofen, reduces inflammation (21)
, hyperalgesia
(22)
, and nociception-induced release of
PGE2 from the spinal cord (23)
. In
addition, R-flurbiprofen was reported to reduce tumor formation and
progression in APCMin/+ and TRAMP mice, which are
in vivo models for colon and prostate cancer, respectively (24
, 25)
. It is important to note that mechanisms that have been
suggested to contribute to the antitumor activity of NSAIDs such as
inhibition of cell cycle progression (17
, 26)
, induction
of apoptosis (27
28
29)
, and inhibition of angiogenesis
(30
, 31)
have been observed only at high concentrations of
the respective NSAIDs, which are 100- to 1000-fold higher than those
needed to inhibit prostaglandin synthesis (Table 1
). Furthermore, NSAIDs inhibit tumor formation and growth in
COX-deficient cell lines (16
, 32)
. These data leave little
doubt that cyclooxygenase-independent mechanisms account for part of
the effects of certain NSAIDs, at least at high doses.
|
|
CYCLOOXYGENASE-INDEPENDENT MECHANISMS OF NONSELECTIVE NSAIDs |
|---|
|
TOPABSTRACTINTRODUCTIONCYCLOOXYGENASE-INDEPENDENT...CYCLOOXYGENASE-INDEPENDENT...REFERENCES |
|---|
B)
first reported that sodium
salicylate and acetylsalicylic acid (aspirin) inhibit the activation of
NF-B, much interest has focused on this transcription factor as a
potential target for certain NSAIDs. The vertebrate Rel/NF-B
transcription factor family includes five cellular proteins: c-Rel,
RelA (p65), RelB, p50, and p52. As homodimers and heterodimers,
Rel/NF-B proteins bind to DNA target sites, collectively called B
sites, and directly regulate gene transcription. A single cell can have
an array of diverse dimeric complexes, the most common of which is
called NF-B and consists of a p50/RelA heterodimer. Since NF-B
regulates the expression of proinflammatory enzymes, cytokines,
chemokines, immunoreceptors, and cell adhesion molecules, it has often
been termed a "central mediator of the immune response" (34
, 35)
. Because of this key role, it was suggested that the
inhibition of NF-B observed might contribute considerably to the
anti-inflammatory effects of salicylates (33
, 36)
.
However, this was not confirmed using the murine air pouch as a model
of inflammation. In mice deficient in p105 (the precursor of the p50
component of NF-B), aspirin and sodium salicylate retained their
anti-inflammatory efficacy (37)
. Inhibition of
inflammation was also independent of prostaglandin synthesis but was
completely reversed by injection of an adenosine A2 receptor antagonist
into the air pouch (37)
, suggesting that modulation of
adenosine release or action was somehow involved. Indeed, aspirin and
sodium salicylate caused cells to release adenosine into the
inflammatory exudate, which was associated with a reduction of
leukocyte accumulation (37)
.
In contrast to salicylates, the anti-inflammatory effects of
dexamethasone, which is known to inhibit the activation of NF-B
(38
39
40)
, were completely abolished in p105 knockout mice
(37)
. Thus, effects of salicylates and dexamethasone
apparently were mediated through different mechanisms in this model.
However, it cannot be excluded that NF-B is a major target of NSAIDs
in other models. Recently, we found that NF-B was inhibited by R-
and somewhat less by S-flurbiprofen, which was associated with a
decrease in NF-B-dependent gene transcription (21)
.
Although R-flurbiprofen does not inhibit cyclooxygenase activity and is
not epimerized to S-flurbiprofen in rats, it reduced
zymosan-induced paw inflammation more potently than S-flurbiprofen and
was almost as effective as dexamethasone, suggesting that NF-B may
be a primary target of R-flurbiprofen.
Inhibition of immune cell maturation and cytokine production
The functions of NF-B in inflammation also include the
regulation of immune cell differentiation and survival. The maturation
of dendritic cells in particular depends on the activation of NF-B
(41)
. Immature dendritic cells differentiate on exposure
to inflammatory signals such as cytokines. Matured cells are then able
to initiate immune responses and are thought to contribute to the
pathogenesis of autoimmune diseases by inappropriate presentation of
self-antigens. When immature human myeloid dendritic cells were
differentiated in vitro in the presence of aspirin or sodium
salicylate, they were unable to stimulate T cell proliferation
(42)
. This was associated with a change of cell surface
marker proteins. Levels of CD83 and interleukin 12 (IL-12), both
markers of mature dendritic cells, were markedly reduced, indicating
that salicylates inhibited the differentiation process. The effect was
mediated through inhibition of NF-B and was independent of
prostaglandin synthesis (42)
. The inhibitory effect
occurred at concentrations (IC50: 2.5 mM) similar
to those found in plasma after administration of anti-inflammatory
doses (1–2 mM). Thus, inhibition of immune cell differentiation may
contribute to the anti-inflammatory effects of salicylates. Dendritic
cell maturation was also inhibited by dexamethasone (43)
whereas other NSAIDs, including ketoprofen, indomethacin, and NS398, a
COX-2-selective agent, had no effect (42)
, which suggests
that these NSAIDs did not inhibit NF-B activation up to the
concentrations tested in that study (100, 200, and 900 µM,
respectively).
Inhibition of NF-B in immune cells also results in a decrease in
cytokine production (44
45
46)
. Aspirin and sodium
salicylate were recently found to inhibit the transcription of IL-4 in
CD4+ human T lymphocytes independent of NF-B inhibition and PG
synthesis (47)
. The effect occurred at concentrations of 1
mM and was caused by a reduced binding of a
Ca2+-inducible complex to an IL-4 promotor
element upstream and not overlapping with the NF-B binding site
(47)
. Analysis of the sequence suggested that an Ets
family member and/or a zinc binding protein might be involved in this
complex (47)
. In contrast to salicylates, indomethacin and
flurbiprofen had no effect (47)
. IL-4 is the prototypical
cytokine of T helper 2 cells and plays a major role in the regulation
of hematopoiesis and immune responses (48)
. It is also
involved in the pathogenesis of autoimmune diseases such as juvenile
rheumatoid arthritis and Kawasaki syndrome, where the effectiveness of
aspirin is well documented (49
, 50)
. Thus, inhibition of
IL-4 production may be an additional mechanism by which salicylates
modulate immune function.
Decision between induction of apoptosis or cell survival
NF-B is involved in the control of the transcription of many
genes whose function extends beyond the immediate immune response.
Similarly, several activators are not bacterial or viral pathogens.
Therefore, rather than being a central mediator of immune responses,
NF-B probably represents a ‘regulator of stress responses’. Hence,
NF-B also functions as a regulator of the apoptotic program either
for induction of apoptosis or, more commonly, as its inhibitor. Whether
NF-B promotes or inhibits apoptosis appears to depend on the cell
type and the type of inducer (Fig. 1
). In numerous human cancer cells, NF-B is persistently active, which
results from a constitutive activation of upstream signaling kinases or
mutations that inactivate IB subunits (51)
. The
persistent activation of NF-B renders these cells more resistant to
chemo- or radiotherapy (52
, 53)
or other agents that
induce apoptosis. The salicylate-induced inhibition of NF-B
translocation might reconstitute the sensitivity of these cancer cells
to apoptosis inducing treatment (54)
and thereby
contribute to the antitumor activity of salicylates. NF-B activity
is also required for the oncogenic transformation by ras and raf
(55
, 56)
. Mutations of ras are found in 
50% of colon
cancers and play an important role in the multistage process of
carcinogenesis. Salicylates were found to inhibit the growth of v-ras
transformed fibroblasts. This effect depended on the ability to inhibit
NF-B but was independent of prostaglandin synthesis
(57)
. In rat intestinal epithelial cells, ras-induced
transformation led to an up-regulation of COX-2, which is
constitutively expressed in > 80% of colorectal cancers and
appears to contribute to colon carcinogenesis.
|
Conversely, NF-B activation does not inhibit but promotes apoptotic
cell death in a subset of neuronal cells including hippocampal neurons
(58)
and cerebellar granule cells (59)
. In
these cells, salicylate and aspirin were shown to reduce
glutamate-induced neurotoxicity, which was closely related to their
ability to inhibit NF-B activation (59)
. However,
several lines of evidence suggest that NF-B can also promote
survival in neurons (60
61
62)
. For example, the ß-amyloid
peptide associated with AD appears to cause its neurodegenerative
effects by down-regulation of NF-B activity (63)
probably through up-regulation of IB
mRNA and protein expression
(60)
. In contrast, exposure of astroglial cultures to
ß-amyloid peptide resulted in an activation of NF-B and
NF-B-dependent gene transcription, including IL-1ß and IL-6
(60
, 64)
. Thus, alterations of NF-B-directed gene
expression may contribute to both the neurodegenerative and
inflammatory response that occur in AD. The contradictory functions of
NF-B in cortical neurons and astroglial cells may partly explain the
somewhat disappointing clinical results with NSAIDs in the treatment of
AD (65
66
67
68)
. Unfortunately, some of these studies were
performed with diclofenac (67)
and indomethacin
(68)
, which probably do not significantly inhibit NF-B
activation (see above).
The mechanisms by which NF-B promotes cell survival are due in part
to the up-regulation of antiapoptotic genes such as members of the bcl
and IAP families (69
70
71)
. As a result, salicylate-induced
cell death was associated with a down-regulation of MCL-1, a Bcl-2
family member (72)
. Less is known about target genes that
may be involved in cells where NF-B induces apoptosis. Since NF-B
activates the genes encoding Fas and Fas ligand (73
74
75
76
77)
,
NF-B activation might sensitize cells to agents that act through or
cooperate with the Fas triggered death pathway.
Mechanisms involved in NF-B inhibition
Inhibitor -B kinase (IKK) complex
NF-B transcription factors are regulated primarily by
interaction with inhibitor IB proteins. Thus, in most cells NF-B
exists in the cytoplasm in an inactive complex bound to IB. Most
agents that activate NF-B do so through a common pathway, which is
based on the phosphorylation-induced, proteasome-mediated proteolysis
of IB. The key regulatory step in this pathway involves activation
of an IKK complex. Liberated NF-B then translocates from the
cytoplasm to the nucleus, where it binds to the B-sites in the
promoter region of target genes and regulates their transcription.
IKK inhibition through salicylates
When cells were stimulated with lipopolysaccharides (LPS) in the presence of sodium salicylate or
aspirin, the LPS-induced proteolysis of IB was abolished, which
suggests that the observed NF-B inhibition was mediated through
inhibition of the phosphorylation and/or the subsequent proteasome
cleavage of IB (33)
. The effects of sodium salicylate
and aspirin were observed at concentrations of 2–20 mM and 2–5 mM,
respectively (33)
. In serum, salicylate concentrations of
1–2 mM are required for anti-inflammatory activity whereas
concentrations of > 6 mM are toxic (47
, 78)
. At
concentrations in the range of 2–20 mM, salicylate also inhibited the
activity of several other cellular kinases. Thus, the specificity of
the observed NF-B inhibition has been questioned (79)
.
Nevertheless, subsequent studies have confirmed that both aspirin and
sodium salicylate inhibit LPS- or cytokine-induced nuclear
translocation of NF-B by preventing IB phosphorylation and
degradation (80)
. Furthermore, it was shown that aspirin
could directly bind to and inhibit the kinase activity of IKKß by
reducing its ability to bind ATP (81)
. The competitive
inhibition of the ATP binding to IKKß provides a potential
explanation for the ability of sodium salicylate to inhibit the
activity of multiple kinases. However, this ability makes it more
difficult to decipher which pathways are mediating the observed
effects. The simultaneous inhibition of multiple signaling pathways may
be important for the ability of salicylates to inhibit tumorigenic cell
growth.
IKK activity with other NSAIDs
Other NSAIDs including ibuprofen (1–3 mM; ref 82
, 83
), sulindac, sulindac
sulfide (1 and 0.2 mM, respectively; ref 84
), and R- and
S-flurbiprofen (0.01–1 mM and 1 mM, respectively; ref 21
)
were also able to inhibit NF-B activation. Indomethacin (0.025–0.25
mM; see ref 42
, 84
), ketoprofen, and ketorolac (up to 1
mM; ref (42)
, however, were inactive (Table 1)
. Among the
NSAIDs found to inhibit NF-B, R-flurbiprofen and sodium salicylate
are the only substances that do not significantly inhibit
cyclooxygenase activity and thus do not cause gastrointestinal
toxicity. In contrast to salicylate (81)
, ibuprofen
(83)
and sulindac (84)
, which were shown to
inhibit IKKß, R-flurbiprofen affected neither the phosphorylation,
degradation, nor expression of IB, suggesting that the effects of
R-flurbiprofen were independent of IKK activity and IB gene
transcription (21)
. Thus, flurbiprofen apparently acts
through a different mechanism than sodium salicylate. It is possible
that R-flurbiprofen directly interacts with NF-B, as has been
described for glucocorticoids (38)
, or targets proteins
that facilitate the nuclear translocation of NF-B such as heat shock
protein 70 (hsp70) (85
, 86)
, which has been shown to play
an important role in the immunosuppressive activity of
15-deoxyspergualin (87
, 88)
.
Inhibition of activator protein 1 (AP-1)
Inhibition of NF-B through R-flurbiprofen resulted in a reduced
expression of COX-2 and tumor necrosis factor (TNF-), whereas
inducible nitric oxide synthase (iNOS) was unaffected
(21)
. Aspirin and sodium salicylate also reduced the
expression of adhesion molecules in endothelial cells (80
, 89)
and chemokines such as monocyte chemoattractant protein 1
and interferon (IFN-) -inducible protein 10 in monocytes and
macrophages (90
91
92)
.
The transcriptional response to NF-B activation depends in part on
the coactivation of other transcription factors, since more than one
transcription factor generally is required to induce effective gene
transcription. Thus, although NF-B participates in the
transcriptional regulation of > 150 target genes, salicylates and
other NSAIDs will probably affect only some of them. The selection of
target genes might depend on the cell type, type of stimulus, and
additional effects on other transcription factors, such as AP-1. AP-1
is a protein complex consisting of products of the jun and fos oncogene
families, which are activated in response to a number of stimulants
including UV irradiation, growth factors, TNF-, and IL-1. Some of
the genes known to be regulated by AP-1 are involved in the immune and
inflammatory responses or tumor formation and progression. Thus, AP-1
and NF-B target genes partially overlap and most target genes are
activated by both AP-1 and NF-B. However, for some genes the effects
may be contrary. For example, iNOS transcription is activated by
NF-B (93)
but negatively regulated by AP-1
(94)
. These opposite effects may explain why
R-flurbiprofen, which inhibited both NF-B and AP-1, had no effect on
iNOS expression (21)
.
Aspirin and sodium salicylate have been shown to inhibit epidermal
growth factor and UV-induced AP-1 activation in mouse epidermal cells
(95)
. Furthermore, R-flurbiprofen was shown to inhibit
LPS-induced AP-1 activation in mouse macrophages (21)
. In
Epstein-Barr virus-infected tumor cells, expression of the EBV-latent
membrane protein, the primary oncoprotein of EBV, was associated with
increased activity of both AP-1 and NF-B, which resulted in
up-regulation of metalloproteinase 9 (MMP-9) and enhanced tumor
invasiveness (96)
. Aspirin and sodium salicylate were able
to reduce MMP-9 levels and tumor invasion through concomitant
inhibition of NF-B and AP-1 (97)
. Metalloproteinases
also play an important role in the degradation of cartilage and bone in
chronic arthritis (98
, 99)
. High NF-B and AP-1 DNA
binding activity was found in the synovium of patients with rheumatoid
arthritis and osteoarthritis (100)
, which resulted in an
up-regulation of metalloproteinases (100)
. A major
metabolite of aceclofenac was recently shown to inhibit the production
of pro-MMP-1, pro-stromelysin, and pro-MMP-2 in rheumatoid synovial
cells (101)
. These effects may contribute considerably to
the therapeutic efficacy of the respective NSAIDs.
Alterations of the MAP kinase cascade
It is not known whether salicylates or R-flurbiprofen directly
inhibit AP-1 or influence other upstream targets. AP-1 is a downstream
target of MAP kinase family members including extracellular signal
regulated kinases (Erk-1 and -2; p42/p44 MAPK), Jun kinases (JNKs), and
p38 MAPK, which are believed to be activated in response to stress
signals. Aspirin and sodium salicylate were shown to inhibit UV-induced
AP-1 activity in mouse epidermal cells through blocking the activity of
Erk-1 and -2 (102)
. Other studies have confirmed that
activation of Erk-1 and -2 can be blocked by sodium salicylate and
aspirin under certain circumstances (103
, 104)
. Effective
concentrations were in the range of 1–5 mM. It was recently shown that
inhibition of angiogenesis by COX-2-selective and unselective
NSAIDs was mediated through direct effects on endothelial cells
involving inhibition of Erk2 activity and interference with its nuclear
translocation (31)
. In contrast to Erk-1 and -2, p38 MAPK
was reported to be activated by sodium salicylate in human fibroblasts
(105)
and was associated with induction of apoptosis.
Since activation of p38 leads to a down-regulation of NF-B activity
in these cells (105)
, the apoptosis may have been mediated
in part through inhibition of NF-B. The same authors found that
salicylate inhibited the TNF--induced activation of c-Jun
NH2-terminal kinase (JNK) in human fibroblasts
(106)
. On the other hand, in HT-29 colon cancer and COS-1
cells, salicylate treatment resulted in activation of JNK
(107)
. The contradictory results suggest that the effects
of salicylate on MAPK family members may depend on the cellular context
and may result from unspecific effects on kinase activity.
Inhibition of ribosomal S6 kinase
Another kinase shown to be inhibited by sodium salicylate is the
ribosomal S6 kinase 2 (p90RSK2) (108)
; aspirin (1–3 mM)
and ibuprofen (1–2 mM) had similar effects. This kinase plays a
critical role as effector of the Ras-mitogen-activated protein kinase
pathway and as regulator of immediate early gene transcription. Since
p90RSK2 is phosphorylated and activated by Erk-1 and Erk-2, these MAPK
family members may have been the primary targets of salicylates.
Nevertheless, sodium salicylate inhibited p90RSK2 activity and thereby
suppressed the phosphorylation and activation of the p90RSK2
substrates, cAMP response element binding protein (CREB) and IB.
This resulted in an inhibition of CREB and NF-B-dependent gene
transcription. Effective concentrations of salicylate were in the range
of 5–20 mM.
Activation of signal transducer and activator of transcription
(STAT1)
In contrast to the inhibitory activity of salicylates on NF-B
and AP-1, salicylates were recently found to enhance the
IFN--induced activation of STAT1 (signal transducer and activator of
transcription) in mouse macrophages (RAW 264.7). Salicylates alone did
not enhance the activity of this transcription factor, but prolonged
its tyrosine phosphorylation and enhanced its nuclear translocation and
DNA binding activity after IFN- treatment of the cells
(109)
. STAT1 is phosphorylated by Janus kinases after
ligand binding to cytokine receptors. It has multiple transcriptional
functions. Upon activation, it drives the expression of many genes such
as p21waf1 (110)
, intercellular
adhesion molecule-1 (111)
, or RANTES (112)
but also suppresses the transcription of others such as collagenase 3
(113)
. These opposing characteristics apply to its role in
facilitating the cross-talk between signal transduction pathways, as it
participates in both synergistic activation and inhibition of gene
expression. Using targeted gene disruptions in mice, it was shown that
STAT1 mediates growth inhibitory signals and contributes to the host’s
rejection of tumors (114)
.
Regulation of the expression of a transforming growth factor ß
(TGF-ß) family member
Another protein that might be involved in the antitumoral effects
of NSAIDs was recently identified in HCT-116 colon cancer cells
(115)
. The gene, designated ‘NSAID-activated gene’
(NAG-1) has an identical sequence with a recently described novel
member of the TGF-ß superfamily. In these studies, the protein was
designated ‘macrophage inhibitory cytokine’ (116)
and
‘placental transforming growth factor ß’ (PTGF-ß)
(117)
. NAG-1 possesses proapoptotic and antitumorigenic
activity. Incubation of HCT-116 cells with different NSAIDs, including
sulindac sulfide, indomethacin, diclofenac, aspirin, and others,
increased the expression of NAG-1 and initiated apoptosis, sulindac
sulfide being the most potent NSAID tested (115)
. In
antisense NAG-1 cells, NSAID-stimulated apoptosis was attenuated,
suggesting that the proapoptotic effects of NSAIDs were mediated at
least in part through NAG-1 up-regulation (115)
. Further
studies demonstrated that NAG-1 (called PTGF-ß in these studies) was
regulated by the p53 tumor suppressor gene (118
, 119)
,
suggesting that NAG-1 induction by NSAIDs might depend on wild-type
p53. Since NAG-1 expression also resulted in growth arrest and
apoptosis in MCF-7 breast cancer cells (118)
, the
NAG-1-mediated effects of NSAIDs are probably not restricted to cells
of colonic origin.
Inhibition of cell cycle progression
Inhibition of p70S6 kinase
Salicylate was recently shown to inhibit the activation of p70S6
kinase at concentrations of 2–10 mM, which was associated with a
down-regulation of c-myc, cyclin D1, cyclin A, and
proliferating cell nuclear antigen (120)
. These target
genes are known to play an important role in cell proliferation, and
their down-regulation might contribute to salicylate-induced growth
arrest. p70S6 kinase is a mitogen-activated kinase that is important
for protein synthesis and G1 cell cycle
progression, and has been identified as a target of the
immunosuppressant drug rapamycin (121)
. Inhibition of
p70S6 kinase through sodium salicylate was associated with
dephosphorylation on Thr-389, the major rapamycin-sensitive site
(120)
. A rapamycin-resistant mutant of p70S6 kinase was
also resistant to salicylate-induced dephosphorylation
(120)
. Rapamycin inhibits cell cycle transition from
G1 to S, mediated partly through a reduction of
cyclin D1 mRNA and protein stability (121)
. Likewise,
salicylate inhibited the progression from G1 to S
and reduced cyclin D1 levels (120)
. The latter effect was
also observed in human pancreatic cancer cells (122)
.
Expression and activity of cyclins and cyclin-dependent kinases
Progression through the various phases of the cell cycle is
regulated mainly by cyclins and cyclin-dependent kinases (Cdks)
(123)
. The Cdks associate with cyclins before they are
activated in discrete phases of the cell cycle. Cdks that participate
in regulating cell cycle progression include 1) Cdk4
(124)
and Cdk6 (125)
, which control the
progression through G1 together with D-type
cyclins; 2) Cdk2, which regulates the
G1/S transition in association with cyclin E
(126
127
128
129)
; and 3) Cdc-2, which in concert with
type A and B cyclins controls the G2/M phase
transition (130
131
132
133)
. The function of cyclins is
controlled primarily by changes in cyclin levels whereas Cdks are
regulated through phosphorylation. The activity of the Cdk/cyclin
complex is further negatively regulated by a number of Cdk inhibitors,
including p21 (waf1) and p27 (kip1). Cdks hyperphosphorylate the
retinoblastoma protein (pRb). pRb is present in quiescent cells in a
hypophosphorylated state and sequesters E2F transcription factors
(134)
. Phosphorylation of pRb at multiple sites releases
E2Fs. These in turn activate the transcription of genes required for
the progression of the cell cycle such as cdc2 and cyclin A
(135)
. Recently, sodium salicylate (5–10 mM) was shown to
inhibit the proliferation of vascular smooth muscle cells (SMCs)
through up-regulation of the Cdk inhibitors
p27kip1 and p21waf1. This
was associated with a decrease in Cdk2 and to a lesser extent in Cdk6
activity, thus preventing hyperphosphorylation of pRb and cell cycle
progression. Cyclin protein levels were unchanged (136)
.
p21waf1 is a major target gene of the tumor
suppressor p53, which is activated and up-regulated upon cellular
stress, particularly DNA damage and hypoxia. In parallel to
p21waf1, salicylate treatment of SMCs caused an
up-regulation of p53 (136)
, suggesting that the observed
effects might represent a stress response induced by high salicylate
concentrations. Salicylate treatment also resulted in a modest
inhibition of NF-B activity (136)
. Although it has been
shown that NF-B activity is essential for SMC proliferation
(137)
, its modest inhibition could not explain the
profound inhibition of SMC proliferation in that study
(136)
. Neither Cdk nor NF-B inhibition was observed
with indomethacin (136)
. Similar to salicylates, however,
sulindac (1.2 mM) and sulindac sulfide (0.2 mM) reduced the
proliferation rate of HT-29 colon carcinoma cells and caused them to
accumulate in the G0/G1
phase (17)
. This was associated with reduced expression
and reduced catalytic activity of cyclin-dependent kinases (Cdc2, Cdk2,
and Cdk4) (17)
.
Inhibition of the heat shock response
Apart from the effects on kinase activity, sodium salicylate has
the property of partially inducing the human heat shock response
(138
, 139)
. This effect is probably mediated through
activation of the heat shock transcription factor 1 (HSF1) from a
latent cytoplasmic form to a nuclear, DNA binding state (139
, 140)
. HSF1 can function as both an activator of heat shock genes
and a repressor of cytokine genes. Treatment of LPS-stimulated
monocytes with sodium salicylate resulted in an inhibition of cytokine
gene expression similar to that observed after heat shock or
overexpression of HSF1 (140)
. Similar effects were
obtained with aspirin, sulindac, ibuprofen, and piroxicam, which also
induced HSP70 mRNA expression (140)
. Thus, exposure to
certain NSAIDs may lead to a switch in gene expression, with
suppression of cytokine and induction of stress genes
(141)
. HSF1 appears to play a regulatory role in these
effects. Since p90RSK2 may function as a repressor of HSF1
(142)
, the observed inhibition of RSK2 through sodium
salicylate (see above) might be involved in the salicylate-induced
activation of HSF1.
Modulation of the activity of nuclear receptor family members
Activation of peroxisome proliferator-activated receptor
(PPAR)
The multitude of effects observed with sodium salicylate and
aspirin argue in favor of the hypothesis that actions of salicylates at
high concentrations are mediated through effects on several cellular
kinases (79)
. Since the interest has focused on
salicylates and, more recently, COX-2-selective agents, it is not clear
whether the effects observed with salicylates are shared with the
majority of other NSAIDs. The results obtained with indomethacin,
however, which inhibits neither NF-B (84
, 143
, 144)
,
AP-1 (95
, 145)
, p42/p44 MAPK (103)
, nor cell
cycle regulatory proteins (136)
, suggest that there are
substance specific differences.
This is further supported by the finding that indomethacin, but not
sodium salicylate binds to and activates PPAR, a member of a
superfamily of nuclear hormone receptors that function as
ligand-dependent transcription factors (146)
. Three
distinct PPAR isoforms—PPAR, 
, and —have been isolated and
characterized (147)
. PPARs bind to sequence-specific DNA
response elements as a heterodimer with the retinoic acid receptor
(148)
. PPAR is highly expressed in adipose tissue and
plays an important role in regulating genes involved in lipid
utilization and storage, adipocyte differentiation, and insulin action.
The antidiabetic thiazolidinediones such as troglitazone are
high-affinity ligands of PPAR (Kd 40 nM; ref
149
). They enhance insulin sensitivity and promote
preadipocyte differentiation in adipose tissue. Although the affinity
of indomethacin to PPAR is relatively low (Kd 100
µM; ref 150
) vs. that of troglitazone, indomethacin was
able to induce the differentiation of mesenchymal stem cells to
adipocytes in vitro (150)
. Because of this feature,
indomethacin has been used widely to study the process of adipocyte
differentiation (151
152
153
154)
. Some other NSAIDs, including
ibuprofen, fenoprofen, and flufenamic acid, also bind and activate
PPAR as assessed in reporter gene assays, but are less potent than
indomethacin (150)
.
Some arachidonic acid metabolites, especially prostaglandin
D2 metabolites such as
15-deoxy-PGJ2, are natural high-affinity PPAR
ligands (Kd 2 µM) (155
156
157)
that
promote adipocyte differentiation in vitro (158)
.
Prostaglandin D2 metabolites have not yet been
identified in adipose tissue, but are major products of arachidonic
acid metabolism in macrophages (159)
and mast cells
(160)
raising the possibility that they might serve as
endogenous PPAR ligands in these cells. PPAR is expressed in
monocytes (161)
, bone marrow precursors
(162)
, splenocytes, and helper T cells (163)
and was shown to be up-regulated in activated macrophages (161
, 164)
. Activation of PPAR in response to
15-deoxy-PGJ2 and synthetic PPAR ligands
resulted in inhibition of cytokine production in macrophages
(161
, 165
, 166)
and T cells (163)
and reduced
T cell proliferation (163)
. This effect was mediated in
part by antagonizing the activities of the transcription factors AP-1,
STAT, and NF-B (161
, 167)
. Thus,
15-deoxy-PGJ2 can be expected to provide
anti-inflammatory effects. This is supported by a recent study showing
that 15-deoxy-PGJ2 induces apoptosis in
synoviocytes and suppresses adjuvant-induced arthritis in rats
(168)
. Furthermore, inhibition of
15-deoxy-PGJ2 synthesis through COX inhibitors
was associated with an exacerbation of inflammation in the late stage
of carrageenan-induced pleurisy in rats (169)
, i.e., at a
stage dominated by mononuclear cells. Since this effect occurred with
both indomethacin and NS398 (COX-2 selective), the unfavorable effect
is apparently shared by selective and unselective NSAIDs. Since
indomethacin is only a weak PPAR agonist, reduction of the
endogenous PPAR ligand 15-deoxy-PGJ2 is
probably not offset by its activity.
Besides its role in adipogenesis and inflammation, PPAR is highly
expressed in the normal large intestine and in breast, colon, and
prostate cancer (170
171
172)
. PPAR agonists such as
troglitazone and 15-deoxy-PGJ2 were able to
induce differentiation and apoptosis in tumor cells
(170
171
172
173)
, suggesting that PPAR suppresses tumor cell
proliferation. This is further supported by the finding that human
colon cancer is associated with loss-of-function mutations of PPAR
(174)
. Indomethacin was shown to reduce the clonogenic
activity of prostate cancer cells (172)
and increase the
antiproliferative effect of 5-fluorouracil in colon cancer cells
(175)
.
Inhibition of peroxisome proliferator-activated receptor
(PPAR)
Unlike the PPAR and PPAR receptors, little is known about
the physiological functions of PPAR. Recently, PPAR was shown to
play an important role in the female reproductive process
(176)
. PPAR was activated by prostacyclin
(PGI2), which is the major PG subtype at the
implantation site where levels of PPAR were significantly increased
(176)
. In COX-2 knockout mice, which exhibit multiple
reproductive failures (including implantation defects), the stable
PGI2 analog carbaprostacyclin
(cPGI2) and PPAR agonists were able to reverse
the implantation defect (176)
. Cicaprost, however, which
can activate IP2 receptors but does not activate
PPAR, did not reverse the deficiency, suggesting that the actions of
PGI2 were mediated through activation of PPAR
(176)
. Since PGI2 synthesis is
inhibited by selective and nonselective NSAIDs, PPAR activity can be
expected to be reduced.
Besides its role in reproduction, PPAR has recently been identified
as one of the downstream targets of ß-catenin. The ß-catenin
protein plays a critical role in embryonic development and oncogenic
processes through its effects on E-cadherin-mediated cell adhesion and
as an effector of Wnt-dependent signal transduction. In the absence of
Wnts, ß-catenin is phosphorylated by glycogen synthase kinase 3
(GSK-3ß), which triggers ubiquitination of ß-catenin and
degradation in proteasomes. Conversely, Wnt signaling leads to an
inhibition of GSK-3ß and stabilization of cytosolic ß-catenin
(177)
, which is then transferred into the nucleus where it
binds to members of the Tcf (T cell factor)/Lef (lymphoid enhancing
factor) transcription factor family. The ß-catenin/Tcf complex
up-regulates the expression of ß-catenin/Tcf-dependent genes
including c-myc (178)
, cyclin D1
(179)
, PPAR (180)
, and gastrin
(181)
. The phosphorylation and degradation of ß-catenin
occurs only when ß-catenin is complexed with conductin/axin and the
APC tumor suppressor gene, which acts as a cofactor (182)
.
Several studies have suggested that a major function of APC is to
down-regulate cytosolic levels of ß-catenin (183
184
185)
,
thus preventing the formation of ß-catenin/Tcf complexes and
transcription of ß-catenin-dependent oncogenes. An inactivating
mutation of the APC gene is an early event in the development of colon
and other cancers. Moreover, inherited mutations of APC cause familial
adenomatous polyposis, which is characterized by the development of
hundreds of colorectal adenomas that result in colon cancer in all
affected individuals. The pivotal role played by APC/ß–catenin
interaction is underscored by the observation that in tumors where the
APC gene is not lost or mutated, mutations are frequently found in
ß-catenin (186
, 187)
. Accordingly, restoration of APC
function in colorectal cancer cells with defective APC results in
growth suppression and apoptosis (188)
.
Recently, sulindac and etodolac were shown to increase the expression
of APC mRNA in the colon of rats treated with the colon-specific
carcinogen azoxymethane (189)
. The increase in APC mRNA
expression was associated with a reduction of preneoplastic lesions
(aberrant crypt foci) (189)
. Furthermore, aspirin
(190)
and indomethacin (191)
were shown to
decrease intracellular ß-catenin levels, associated in the case of
aspirin with a reduction of the rate of tumor formation in APC Min/+
mice (190)
. PPAR was recently identified as one of the
target genes of the APC/ß-catenin-Tcf pathway (180)
and
its expression was shown to be up-regulated in colorectal carcinomas
(192)
. As mentioned, PGI2 can act as
an activator of PPAR (176
, 192)
. Thus, NSAID-mediated
inhibition of eicosanoid metabolism is likely to result in a repression
of PPAR activity. This theory was supported by the finding that
sulindac sulfide inhibited carbaprostacyclin
(cPGI2) -stimulated DNA binding activity of the
PPAR/RXR heterodimer, which was associated with induction of
apoptosis in colorectal cancer cells (180)
. A similar
effect was observed with the sulindac sulfide-related compound sulindac
sulfone, which is devoid of COX inhibitory activity (193)
,
suggesting that inhibition of PPAR was in part mediated by a direct,
prostaglandin-independent effect (180)
. Directly or
indirectly, the suppression of PPAR activity associated with the
promotion of apoptosis can compensate for APC or ß-catenin mutations,
which result in an increased expression of PPAR and constitute an
early step in colon carcinogenesis. Thus, inhibition of PPAR
activity may contribute to the chemopreventive activity of NSAIDs.
Induction of the orphan nuclear receptor NFGI-B (Nur77)
Indomethacin was recently found to induce another member of the
steroid-thyroid hormone receptor family, nerve growth factor-inducible
B (NGFI-B), also termed Nur77 or TR3 (194)
. NGFI-B is one
of the immediate early genes originally identified by virtue of its
rapid activation by nerve growth factor (NGF) in PC12 pheochromocytoma
cells (195)
and by serum in fibroblasts
(196)
. NGFI-B is rapidly synthesized in response to a
variety of growth factors, phorbol esters, and treatments resulting in
calcium influx. The protein is modulated by phosphorylation; the extent
of phosphorylation depends on the stimulus. NGFI-B binds as monomer to
a relatively simple response element, which implies it has the
potential to activate a wide array of target genes (197)
.
In contrast to other nuclear receptor family members, which need their
respective ligands to gain transcriptional competence, NGFI-B is
constitutively active when expressed (197)
. The action of
NGFI-B has been shown to be necessary for activation-induced cell death
in T cells and thymocytes (198
199
200)
. In response to
apoptotic stimuli, NGFI-B translocates from the nucleus to mitochondria
to induce cytochrome c release and apoptosis
(201)
. The induction of NGFI-B by indomethacin was
associated with induction of apoptosis in HCT-15 colon cancer cells
(194)
. Since these cells lack COX-2 expression, NFGI-B
induction appears to be independent of COX-2.
|
CYCLOOXYGENASE-INDEPENDENT ACTIONS OF COX-2-SELECTIVE DRUGS |
|---|
|
TOPABSTRACTINTRODUCTIONCYCLOOXYGENASE-INDEPENDENT...CYCLOOXYGENASE-INDEPENDENT...REFERENCES |
|---|
B through celecoxibB and
NF-B-dependent gene transcription (202)
, suggesting
that its cyclooxygenase-independent actions may also differ from those
of nonselective agents. NF-B activation caused by high
concentrations of celecoxib resulted in a complete loss of its
anti-inflammatory efficacy (202)
. It is not yet known
whether this NF-B-activating effect is shared by all COX-2-selective
agents or whether it is a special feature of celecoxib.
Celecoxib was repeatedly shown to reduce colon carcinogenesis in
animals and humans (203
204
205)
. Thus, it apparently
possesses antitumor activity similar to that of salicylates. These
effects might be due in part to the activation of NF-B since it has
been reported that activation of NF-B in colon cancer cells causes
an up-regulation of FasL and facilitation of Fas-mediated killing
(74)
and a sensitization to apoptosis-inducing treatments
such as chemo- or radiotherapy (206)
. This is supported by
a recent study demonstrating that aspirin activates NF-B in colon
cancer cell lines by increasing the phosphorylation and subsequent
degradation of IB, which was associated with induction of apoptosis
(207)
. Aspirin had no effect in cells of noncolonic origin
(207)
, suggesting that the activation of NF-B by
aspirin is cell type specific and may explain why aspirin mainly
prevents colon cancer. The apparent contradiction that aspirin may both
inhibit and activate NF-B could be explained by the cell type and
stimulus, since NF-B inhibition was observed mainly in immune cells
after cytokine stimulation whereas NF-B activation occurred in colon
cancer cells only when aspirin was administered alone
(207)
. In addition, the outcome is probably influenced by
concomitant effects on cellular kinase activity or other transcription
factors. Thus, the ability of celecoxib to activate NF-B may limit
its usefulness in inflammatory diseases, but may be an interesting
feature concerning colon cancer or Alzheimer disease, where an
activation of NF-B has been shown to reduce ß-amyloid-induced
neuronal apoptosis (63)
.
Inhibition of Akt phosphorylation
Celecoxib was recently shown to induce apoptosis in
COX-2-expressing, androgen-responsive (LNCaP) and nonresponsive (PC-3)
prostate cancer cells (208)
. In contrast to an earlier
report that attributed the apoptotic activity of the
COX-2-selective inhibitor NS398 in LCNaP to a down-regulation of Bcl-2
(209)
, celecoxib did not alter Bcl-2 levels and its
effects were not reversed by overexpression of Bcl-2
(208)
. In that study, celecoxib inhibited the
phosphorylation of protein kinase B (PKB/Akt), thereby blocking its
antiapoptotic activity. The effects of celecoxib on Akt were
independent of phosphatidylinositol-3-kinase activity and not reversed
by okadaic acid, which is an inhibitor of protein phosphatases 1 and
2A, suggesting a direct effect on Akt kinase.
We do not know whether celecoxib also affects other pathways that have
been shown to be modulated by nonselective NSAIDs, especially
salicylates. NAG-1 expression (see above) was not affected by most
COX-2-selective inhibitors with the exception of SC-58125, which
suggests that there might be additional differences between unselective
and COX-2-selective NSAIDs apart from those concerning NF-B.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONCYCLOOXYGENASE-INDEPENDENT...CYCLOOXYGENASE-INDEPENDENT...REFERENCES |
|---|
); inhibition (
); inhibited
by NSAIDs (shaded oval); activated by NSAIDs (oval). Abbreviations:
Akt/PKB, protein kinase B; APC adenomatous polyposis coli tumor
suppressor gene; AP-1 activator protein 1; Cdk, cyclin-dependent
kinase; CREB, cAMP response element binding protein; Erk, extracellular
signal-regulated kinase; GSK3ß, glycogen synthase kinase 3 beta; HSF,
heat shock factor; Hsp, heat shock protein; IKK, I-kappa kinase; JNK,
Jun NH2-terminal kinase; MAPK, mitogen-activated
kinase; MAPKAP, MAPK-activated protein kinase; MEK/MKK,
mitogen-activated protein kinase kinase; MLK, mixed lineage kinase;
MSK, mitogen- and stress-activated kinase; NIK, nuclear factor
kappaB-inducing kinase; NF-