 |
INTRODUCTION |
Angiogenesis, the formation of capillaries derived from a
pre-existing vaculature, is central to a wide range of debilitating human pathologies, including solid tumor growth, arthritis, corneal ulceration, and proliferative retinopathies (1, 2). Accumulating evidence implicates members of the matrix metalloproteinase
(MMP)1 family in this
process, in particular, gelatinase A (MMP-2) and gelatinase B (MMP-9)
(3). These enzymes may contribute to growth of new capillaries in
several ways including activation of growth factors that stimulate
endothelial cell migration and tube formation and dissolution of
endothelial basement membranes at the sprouting capillary tips (4-6).
Gelatinase A (MMP-2) is constitutively present in tissues in a latent
72-kDa proenzyme form (7). When required, progelatinase A is
proteolytically activated to a 62-kDa form by cleavage of the N
terminus by membrane-type MMPs (8); in endothelial cells, this occurs
through ligation to the integrin receptor
v
3 (9). In contrast, gelatinase B (92-kDa
MMP-9), like most other MMPs, is expressed only upon demand, and
regulation occurs at the level of gene transcription (10). Mice made
genetically deficient in gelatinase A, although revealing no
developmental abnormalities, display impeded angiogenic response to
tumor stimulus (11). Targeted inactivation of gelatinase B causes a
transient delay in bone development because of defective vascular
invasion (12). These gene knockout studies provide the most definitive evidence to date for the importance of the gelatinolytic MMPs in angiogenesis.
Much attention is currently focused on development of synthetic agents
that block MMP enzymatic activity as a means to pharmacologically manage diseases involving these enzymes (13). However, considering that
most MMPs are expressed only on demand, an alternative approach might
target mechanisms for MMP transcriptional activation (10, 14). A highly
conserved AP-1 transcription factor DNA-binding site is found in many
MMP gene promoters and is thought to be a common thread to their
co-ordinate induction in response to diverse stress stimuli (10). AP-1
is activated in response to hypoxia (15), a physiological stimulator of
angiogenesis (16). Fibroblast growth factor-2 (FGF-2) is another
important angiogenic stimulator (17, 18) that initiates signaling
mechanisms ultimately activating AP-1 (19-21). A response element for
a second transcription factor activated by stress stimuli, NF-
B, has
been found only in the gelatinase B promoter thus far and is one factor
determining the unique expression pattern of this gene (10). NF-B,
an oxidative-response transcription factor, also triggers gene
expression associated with the angiogenic response (22) by engaging
inflammatory cytokine signaling (23). Therefore, targeted inhibition of
AP-1 or NF-B might be a logical step in attempting to modulate the
angiogenic response.
Curcuminoids, natural products of the Indian spice turmeric, are potent
antioxidant and antiinflammatory agents that have been entered into
Phase I clinical trials for chemo-preventation by the National Cancer
Institute (24). At least part of their biological activity can be
attributed to their capacity to inhibit activation of AP-1 and NF-B
transcription factors. Thus, curcuminoids have been shown to inhibit
activation of these transcription factors in response to phorbol
myristate acetate (PMA), tumor necrosis factor-, and hydrogen
peroxide (25, 26). In vascular endothelial cells, curcuminoids reduce
the activation of tissue factor gene expression induced by tumor
necrosis factor-, lipopolysaccharide, PMA, and thrombin resulting
from the coordinate inhibition of AP-1 DNA binding activity and nuclear
translocation of NF-B (27, 28).
In this study, we investigate curcuminoid effects on FGF-2 activation
of the gelatinase B transcriptional promoter and on angiogenesis in a
corneal micropocket assay. We report that curcuminoids inhibit
FGF-2-induced angiogenesis when delivered locally or in the diet in
coordination with their down-regulation of transcription factor AP-1
DNA binding activity and of gelatinase B promoter activity.
| |
EXPERIMENTAL PROCEDURES |
Cell Culture and Treatments--
Stromal cells were isolated
from rabbit cornea and subcultured as described previously (29). For an
experiment, cells were plated, then changed the next day to serum-free
medium, and incubated for 24-36 h to reduce basal activity of the
transcription factors under investigation. Cells were then pretreated
with 20 µM curcuminoids for 30 min, followed by
stimulation with either PMA (1 µM) or FGF-2 (50 ng/ml)
plus heparin (30 ng/ml). For electrophoretic mobility shift assay
(EMSA), treatments were performed for 2-4 h in serum-free medium. For
transfection analysis, treatments were performed for 24 h.
EMSA--
Nuclear lysates were prepared as described previously
(30). Protein concentrations of the nuclear extracts were determined using the Bio-Rad reagent, and aliquots of equal protein were frozen at
70 °C. Double-stranded oligonucleotides containing the
consensus DNA-binding sites for transcription factors AP-1 (5'-CGCTTGATGAGTCAGCCGGAA-3') and NF-B (5'-
AGTTGAGGGGACTTTCCCAGG-3') served as probes for EMSA (Promega Co.,
Madison, WI). An oligonucleotide containing a mutated and nonfunctional
AP-1-binding site (AP-1*; 5'-CGCTTGATGAGTTGGCCGGAA-3') was
employed in competion controls. The oligonucleotides were labeled with
[
-32P]ATP, diluted, and used as probes in EMSA,
according to standard methods (30). Antibodies to NF-B family
members p50, p65, and c-Rel were purchased from Santa Cruz
Biotechnology (Santa Cruz, CA). For supershift EMSA, 1 µg of one of
these antibodies was added to the nuclear lysate containing the
radiolabeled probe and incubated for 40 min before electrophoresis.
Antibodies to the unrelated transcription factor AP-2 served as a
negative control.
Transfection Experiments--
Rabbit corneal fibroblasts plated
in 6-well dishes were transfected with 1 µg of a specific gelatinase
B promoter-CAT plasmid, the contruction of which was previously
described (31). To normalize transfection efficiencies, cultures were
co-transfected with 0.5 µg of a cytomegalovirus promoter/-gal
reporter construct (Stratagene, La Jolla, CA). An exception was the
experiments involving curcuminoid treatment; these were not normalized
to -gal, because the curcuminoids inhibited cytomegalovirus promoter
activity (data not shown). One day after transfection, cell cultures
were placed in serum-free medium, and treatments were added as
appropriate to the experiment. Cultures were then harvested, and CAT
and -gal activities were assayed from triplicate samples (31).
Rabbit Corneal Micropocket Angiogenesis Assay--
Pellets of
curcuminoids (2 mg, 1.5 mm in diameter) were made using a pellet press
(Parr Instrument Co., Moline, IL). Slow release (sucralfate)-FGF-2 (80 ng) pellets were prepared as described (32) and implanted into the
right eye of rabbits. New Zealand rabbits were purchased from Charles
River labs (Cambridge, MA). All ocular surgical procedures were
performed with the aid of a dissecting microscope after total
anesthesia with xylazine and ketamine and topical application of 0.5%
proparacaine (Ophthatetic, Alcon, TX). For the co-implantation of a
curcuminoid pellet with the FGF-2 pellet, the surgical procedure was
modified; a second micropocket was created 3 mm from the one containing
FGF-2 pellet, and a single curcuminoid pellet was placed in the second
micropocket 5 mm from the limbus. Animals were monitored daily for for
signs of eye irritation, inflammation, or infection. On day 12 after surgery, rabbits eyes were photographed and then sacrificed by an
intracardiac injection of phenobarbitol. Corneas were immediately removed and frozen at 70 °C for later use in zymographic analyses.
Tissue Extracts and Gelatin Zymography--
A circular disc of
tissue 5 mm in diameter from pellet-implanted rabbit corneas was
isolated with the aid of a trephine; any slow release pellets were then
removed. The corneal discs were minced into small fragments and
pulverized with a homogenization plunger in 1.5-ml microfuge tubes
containing extraction buffer (2% SDS, 50 mM Tris-Cl, pH
7.5, and 1 mM CaCl2). Protein concentrations were determined from trichloroacetic acid-precipitated material. Equal
amounts of protein from each rabbit cornea was mixed with Laemeli gel
loading buffer (without added reducing agents) and subjected to
electrophoresis on an 7.5% polyacrylamide gel containing 2.5%
gelatin. Gels were processed according to standard zymographic procedures (29). In brief, SDS was extracted from the gels by washes in
2.5% Triton X-100 for 1 h, and subsequently the gel was incubated
in 50 mM Tris-Cl, pH 7.5, 1 mM
CaCl2 for 16 h at 37 °C to permit in-gel refolding
and autoactivation of the metalloproteinases. Coomassie Blue staining
revealed the presence of the gelatinolytic enzymes as clear bands
against the blue background. Gelatinase B proenzyme (MMP-9), gelatinase
A proenzyme (MMP-2), and the proteolytically cleaved, activated form of
gelatinase A (62 kDa) were identified by their co-migration with rabbit
gelatinase standards prepared from conditioned media of phorbol
ester-stimulated rabbit corneal fibroblasts as described (33).
Preparation and Provision of Curcuminoid Diet--
Curcuminoids
purchased from Sigma, sold under the name curcumin, contain
approximately 77% curcumin, 17% demethoxycurcumin, and 3%
bisdemethoxycurcumin (34). The curcuminoid-supplemented diet was
prepared by mixing 1 g (experiment 1) or 2 g (experiment 2)
of curcuminoids/kg of standard rodent chow. In brief, the rodent chow
was pulverized into a powder and thoroughly blended with the
curcuminoid mixture. The chow mix was wetted with sterilized distilled
water, recompressed into nugget size pieces, and immediately frozen at
20 °C. Adequate amounts of the curcuminoid-supplemented chow were
brought to room temperature and provided fresh each day to the mice.
Any remaining chow that was not consumed was discarded and replaced
with a fresh batch the next day.
Mouse Corneal Micropocket Angiogenesis Assays and
Treatment--
An 80-ng slow release (sucralfate)-FGF-2 pellet was
surgically implanted in one cornea in each of ten CD-1 female mice
(Charles River Labs, MA) according to the well established procedure of J. Folkman and co-workers (32). One group of five mice harboring the
FGF-2 pellets was allowed to feed ad libitum on 1 g
curcuminoids/kg chow for a period of 10 days starting on the day of the
surgery. The control group of five mice harboring the FGF-2 pellets was allowed to feed on regular rodent chow ad libitum. On days 6 and 10, mice were anesthetized, and their eyes were photographed under a dissecting stereo microscope. In a second experiment, six CD-1 female
mice were allowed to feed ad libitum on 2 g
curcuminoids/kg chow for 7 days prior to surgical implantation of the
80-ng FGF-2 pellets and continued on the curcuminoid chow for an
additional 6 days post-surgery. The control group of six mice was
allowed to feed ad libitum on regular chow over the entire
period, prior to, and post-surgical implantation of FGF-2. The eyes of
the control and curcuminoid-treated mice were photographed on day 6 post-surgery.
In a separate angiogenesis experiment, transgenic mice (line 3445; CD-1
background) that contain the DNA sequences between 522 and +19 of the
rabbit gelatinase B gene promoter fused transcriptionally to the coding
sequences of the bacterial LacZ gene were used. The
generation of these mice and the embryonic developmental and wound-induced stimulation of reporter gene expression has been reported
(30). Line 3445 transgenic mice were implanted with 80-ng FGF-2 corneal
pellets. Developing new blood vessels were followed by examination
between days 3 and 7, when some mice were sacrificed during this time
period and enucleated eyes were stained for -gal expression. In some
cases, neovascularizing eyes were also photographed before sacrifice.
Mouse eyes were enucleated, fixed in cold 4% paraformaldehyde for 30 min, and washed three times for 10 min each at 4 °C in phosphate
buffered saline. The eyes were stained in 2% X-gal staining solution
for 12 h at 30 °C and photographed. X-gal-stained eyes were
also embedded in paraffin and sectioned at 6 µm, and the sections
were counter-stained with eosin.
Quantitation of Neovascularization in Mouse Corneas--
We
attempted to quantitate corneal blood vessel density by the
computer-assisted method of J. Folkman and co-workers (35); however,
because CD-1 mice lack a pigmented iris, we could not obtain high
contrast photographic images for direct quantitation of blood vessels
from photographic prints. Therefore, 35-mm color slide images of
neovascularized eyes of each mouse were obtained, and the total number
of blood vessels in each cornea was quantified by analyzing traced
images of blood vessels. In brief, using a slide projector the slide of
each neovascularized cornea was projected at a fixed distance to
produce an enlarged picture that fit within a piece of 8.5 × 11-inch white paper. Using a black ink fine-tip ball point pen, each
corneal blood vessel from the enlarged image was traced by hand onto
the sheet of paper. The slides were coded to allow a masked analysis by
the investigator, and the tracings were repeated by a second
investigator blinded to the study. The traced image of blood vessels of
each mouse cornea was scanned using a laser scanner (ScanJet 4C) into a
Macintosh computer using the Deskscan and Photoshop software programs
in grayscale format at 8-bits/pixel resolution. The imported images
from Photoshop were directly transferred without any alterations into
the NIH Image analysis software, and the sum of all vessels in a cornea were quantified as pixels. The data in the control and
curcuminoid-treated groups were decoded, and results of the analysis
performed by the two researchers were averaged. Student's t
test was applied to the data to obtain significance values.
| |
RESULTS |
Gelatinase B Is a Downstream Target in the FGF-2 Angiogenic Gene
Expression Program--
We have recently reported our development of a
transgenic mouse line (3445) harboring a lacZ reporter gene
driven by 522 to +19 of the rabbit gelatinase B transcriptional
promoter (30). We showed that this promoter fragment is sufficient to
drive embryonic and injury-induced reporter gene expression in a manner
that mimics the pattern of the endogenous gelatinase B gene. In this
study, we used line 3445 mice to investigate activation of the
gelatinase B promoter to an FGF-2 angiogenic stimulus using the corneal
micropocket assay. Implantation of slow release pellets containing 80 ng of FGF-2 into the mouse cornea stimulated growth of corneal
capillaries from the limbal vasculature located at the transitional
zone between the cornea and sclera. These new blood vessels were
abundant by 7 days post-surgery when observed under a dissecting
microscope (Fig. 1A,
small arrows). The neovascularized eyes were analyzed by
whole mount histochemical staining for reporter gene expression. -Gal activity was observed around the FGF-2 pellets and in a punctate pattern paralleling the capillary network growing toward the
pellets (Fig. 1B, small arrows). Finally,
staining was seen within the area of the pre-existing limbal vessels
(Fig. 1B, large arrow); however, dissection of
the tissue revealed that most of this staining occurred in underlying
structures in the angle of the anterior chamber (such as the ciliary
body), which also stained in eyes without implanted pellets (data not
shown). In thin tissue sections of the X-gal-stained neovascularized
portions of the cornea, -gal activity was revealed to occur in cells
composing the vessel lumen and cells that had accumulated around the
new capillaries (Fig. 1C). A time course study revealed that
gelatinase B promoter activity was transient within the vascularizing
tissues; migratory endothelial cells stained by day 2 and outlined
newly formed blood vessels till day 7, becoming sporadic after this time point (data not shown). The fellow eye of the same mouse that
lacked FGF-2 pellet or eyes of transgenic mouse line 3445 implanted
with pellets lacking growth factor did not develop corneal blood
vessels and failed to show corneal -gal staining at 7 days post-surgery (data not shown). Together, these results indicate that
activity of the gelatinase B transcriptional promoter is spatially and
temporally regulated during the process of angiogenesis that occurs in
response to FGF-2 stimulus.
View larger version (55K):
[in this window]
[in a new window]
|
Fig. 1.
Activation of
-galactosidase expression in FGF-2-induced corneal
blood vessels in gelatinase B promoter/lacZ transgenic
mice. The corneas of transgenic mice (line 3445) containing the
522 to +19 rabbit gelatinase B promoter/lacZ fusion gene
(30) were implanted with a single slow release 80-ng FGF-2 pellet.
Angiogenesis that proceeded was photographed on the 7th day
(A), and expression of the reporter gene was analyzed by
staining this eye in X-gal solution (B). The new corneal
blood vessels are indicated by the small arrows, the
pre-existing limbal blood vessels are shown by the larger
arrows, and growth factor pellet is shown by the
asterisk (A and B). Note that not all
blood vessels in the cornea (A) stain for -gal expression
(B). Paraffin-embedded thin-tissue sections of the
X-gal-stained eye were also obtained and counter-stained with eosin.
-Gal expression was localized to endothelial cells within vessel
lumen (small arrow, C) as well as those migrating
outside. Some endothelial cells in the capillary lumen also failed to
stain (asterisk, C).
|
|
Role of AP-1 DNA-binding Elements in Response of the Gelatinase B
Promoter to FGF-2 and Inhibition by Curcuminoids--
The effect of
curcuminoids on the FGF-2-mediated activation of two transcription
factors known to contribute to gelatinase B promoter activity, AP-1 or
NF-B, was investigated by EMSA. As a positive control, we confirmed
the known activity of curcuminoids against transcription factors
activated in response to PMA (25, 26). Cells were serum-starved for
24-36 h to reduce transcription factor activities to basal levels
prior to initiating an experiment.
When radiolabeled consensus oligonucleotide probes for AP-1 or NF-B
were incubated with nuclear lysates prepared from untreated serum-free
cells, a single major DNA-protein complex formed in each case
(arrowheads), as revealed by the presence of a band with
retarded electrophoretic mobility in comparison to the free probe (Fig.
2, A and B,
SF lanes). The specificity of the complex that formed on the
radiolabeled AP-1 oligo was revealed by its abrogation when competed
with a 50-fold excess of cold AP-1 oligo, but not by a 50-fold excess
of the mutant AP-1 oligo (AP-1*) containing two nucleotide changes in
the consensus DNA sequence (Fig. 2A, 50 × oligo
lanes). The specificity of the major complex that formed on the
radiolabeled NF-B oligo was similarly revealed by competition with a
50-fold excess of cold probe (Fig. 2A, 1:50
lane), but not by a nonspecific binding control, cold AP-1 probe
(Fig. 2A, NSB lane). Furthermore, antibodies to
the NF-B subunits, p50 and p65, each supershifted a subcomponent of
the major complex, but antibodies either to NF-B subunit c-Rel or to
the unrelated transcription factor AP-2 had no effect (Fig.
2A). The nonspecific nature of the faster migrating minor
complex (Fig. 2A, asterisk) in NF-B EMSA is
indicated by its unresponsiveness to competition or antibody
supershift.
View larger version (51K):
[in this window]
[in a new window]
|
Fig. 2.
Differential inhibition of transcription
factors by curcuminoids in corneal fibroblasts stimulated by FGF-2 and
PMA. Serum-starved corneal fibroblasts left untreated in
serum-free medium (SF) or treated with 20 µM
curcuminoids (Cur) were treated 30 min later with 1 µM PMA and incubated for 4 h (A) or 50 ng/ml FGF-2 and incubated for 2 h (B). Nuclear extracts
prepared from these cells were incubated with radiolabeled AP-1 or
NF-B consensus oligonucleotides and analyzed by EMSA. For the AP-1
analysis, competition was performed with a 50-fold excess of cold AP-1
probe or a mutant AP-1 oligonucleotide (AP-1*). For the NF-kB analysis,
competition was performed with a 50-fold excess of cold nonspecific
probe AP-1 (NSB) or with a 10-fold (1:10) or 50-fold (1:50)
excess of cold NF-kB probe. Supershift analysis was performed with
antibodies to the NF-kB proteins p50, p65, and c-Rel. An antibody to
transcription factor AP-2 served as a negative control. The position of
supershifted complexes is indicated by the long arrows. The
position of the specific oligonucleotide binding complexes identified
by these controls are indicated by the arrowheads. The
nonspecific protein-DNA complex that binds the NF-B probe is denoted
by an asterisk.
|
|
PMA treatment of cells stimulated AP-1 and NF-B DNA binding
activities as previously reported (25, 26) (Fig. 2A). In contrast, FGF-2 treatment stimulated AP-1 DNA binding activity, but
NF-B DNA binding activity was not altered (Fig. 2B).
Curcuminoid treatment of otherwise untreated cells for 30 min did not
alter the basal DNA binding activity to the two probes under
examination (Fig. 2, A and B, compare
SF and Cur lanes). However, curcuminoid treatment
inhibited the PMA stimulation of DNA binding activities for AP-1 and
NF-B, as previously reported (Fig. 2A). The new finding
of this study was that curcuminoid treatment had a similar effect on
transcription factor activation by FGF-2. Thus, activation of AP-1 DNA
binding activity stimulated by FGF-2 was inhibited, but basal DNA
binding activity was unaffected (Fig. 2B). These findings
indicate that curcuminoids distinguish basal and PMA- or
FGF-2-stimulated signaling pathways. Furthermore, although FGF-2
activates different signaling pathways than PMA in corneal stromal
cells, curcuminoids are as effective in blocking the FGF-2 pathway as
they are in blocking the PMA-stimulated signaling pathway.
To determine the role of NF-B and AP-1 response elements for
gelatinase B promoter activity, we performed transfection experiments with a series of gelatinase B promoter-reporter gene fusion constructs (Fig. 3A). The first
construct, 519 CAT (encompasing promoter sequences between 519 and
+19) lacks a functional NF-B site since the 5' trucation removes
four of the ten consensus nucleotides required for NF-B-mediated
transcription (31). However, this construct retains the distal and
proximal AP-1 site (30, 31). Treatment of Pr 21-transfected corneal
fibroblasts with 50 ng/ml FGF-2 induced reporter gene expression by
over 2-fold. In contrast, a cut-back construct lacking the distal AP-1
site (423 CAT) failed to respond to FGF-2 stimulation, although
substantial basal levels of promoter activity were observed.
Mutagenesis of the proximal AP-1 DNA-binding site in 519 CAT
(519(AP1-Prox) CAT) (31) also abrogated FGF-2-stimulated reporter
gene expression and inhibited basal promoter activity as well. This
indicates that the proximal AP-1 site is necessary for gelatinase B
promoter activity and is required along with the distal AP-1 element to
confer transcriptional response to FGF-2.
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 3.
AP-1 mediates FGF-2-stimulated gelatinase B
promoter activity and inhibitory response of curcuminoids on gelatinase
B promoter activation. Rabbit corneal fibroblasts were transfected
with Gel B/CAT promoter constructs and assayed for CAT activity after
treatments with FGF-2 and/or curcuminoids. A diagram of the full-length
519 CAT construct is shown, with the two AP-1 response elements
indicated. A, transfected cells were incubated in triplicate
with or without 50 ng FGF-2 for 24 h. The construct 519 CAT
contains two AP-1 response elements. The distal AP-1 element is deleted
in 423 CAT. The proximal AP-1 element mutated in the -519 (AP1-Prox)
CAT construct (31). B, cells were transfected with 519 CAT
and then treated for 24 h with 50-nM curcuminoids, 50 µg of FGF-2, or a combination of the two agents.
|
|
We next addressed the question of whether gelatinase B promoter
activation induced by FGF-2 could be inhibited by curcuminoids (Fig.
3B). Addition of as little as 50 nM curcuminoids
to 519 CAT-transfected cultures resulted in complete abrogation of
FGF-2-stimulated promoter activity, bringing reporter gene expression
back down to the basal level. In contrast, treatment of 519
CAT-transfected cultures with 50 nM curcuminoids did not
alter basal levels of promoter activity. Thus, curcuminoids distinguish
between AP-1-dependent promoter activation by FGF-2 and
basal activity of the promoter in agreement by our EMSA results.
Curcuminoids Reduce the Angioproliferative Response in Vivo and
Inhibit the Appearance of Specific Gelatinase Forms in the
Cornea--
To begin to examine the effects of curcuminoids on
angiogenesis stimulated in response to FGF-2, we again utilized the
corneal micropocket assay. This time, however, we used rabbits because of the larger corneal size and thus availability of more tissue for
biochemical analysis. Three groups of four rabbits each were analyzed.
In group one, each rabbit was implanted with a single 80-ng FGF-2
pellet. In group two, each rabbit was implanted with a 2-mg curcuminoid
pellet placed in a micropocket adjacent to the FGF-2 pellet. In group
three, each rabbit was implanted with a single 2-mg curcuminoid pellet.
The rabbit corneas are considerably larger than mouse corneas;
therefore, the 80-ng FGF-2 pellets were not as effective in stimulating
angiogenesis as in the experiment described above. Nevertheless,
rabbits implanted with FGF-2 pellets had developed limbal vessel
dilatation in response to the angiogenic stimulus of the growth factor
when examined 7 days post-surgery. By the end of 12 days, the group of
rabbits treated with only FGF-2 continued to show limbal vessel
dilatation, and corneal angiogenesis was most noticeable at the limbus
(Fig. 4B), where as rabbits
harboring FGF-2 and curcuminoid pellets showed no limbal vessel
dilatation and angiogenesis (Fig. 4C). In fact, these
corneas looked the same as they did prior to implantation of FGF-2
pellets. Because vessel growth was not optimally stimulated by FGF-2 in this model, we did not attempt to quantify angiogenesis. However, we
observed that all the rabbit corneas implanted with FGF-2 and curcuminoid pellets demonstrated attenuated angiogenic response that
was consistent within this group. These data also suggested that
curcuminoids produced a vasosuppressive effect on FGF-2-induced vessel
dilatation. The rabbits treated with curcuminoid pellets alone (Fig.
4A) showed no signs of eye irritation, inflammation, or
corneal opacity, suggesting that curcuminoids as implantable drugs were
relatively safe to the anterior ocular surface.
View larger version (72K):
[in this window]
[in a new window]
|
Fig. 4.
Localized delivery of curcuminoids inhibits
the angioproliferative response to FGF-2 stimulation in rabbit
corneas. Photograph of rabbit eyes showing the cornea-scleral
region of neovascularized corneas 12 days post-surgery. The slow
release 80-ng FGF-2 pellet in B and C are
indicated by the asterisks, and the 2-mg curcuminoid pellet
in A and C appear as yellow circular
discs.
|
|
We next investigated the effects of curcuminoid pellet implantation on
gelatinase B expression in the rabbit corneas. Analysis of extracts by
zymography from central regions of the four corneas implanted with only
curcuminoid pellets revealed a gelatinolytic protein that migrated at
65 kDa (Fig. 5). We identify this protein as pro-gelatinase A based on its co-migration with rabbit
pro-gelatinase A produced by cultured corneal stromal cells. The
presence of progelatinase A is consistent with our previous studies
that have identified this MMP as a normal component of the corneal
stroma (29, 36, 37). A minor gelatinolytic protein was also present at
62 kDa; this size is appropriate for the proteolytically activated form
of gelatinase A. Extracts from FGF-2-pellet implanted corneas also
contained the 65-kDa and minor 62-kDa gelatinase A forms. However, a
new gelatinolytic activity migrating at 92 kDa was also apparent. This
new activity was clearly present in all four of the FGF-2-treated
corneas. Based on co-migration with the rabbit progelatinase B
standard, we identify this enzyme as progelatinase B. Addition of
curcuminoid pellets did not alter wound healing of rabbit corneas
harboring FGF-2 pellets to any extent apparent by gross observation.
However, FGF-2-stimulated appearance of gelatinase B was completely
inhibited in all four curcuminoid-implanted corneas. These results
indicate that curcuminoids inhibit FGF-2-stimulated expression of
gelatinase B, in parallel with inhibition of FGF-2 stimulation of the
angiogenic response.
View larger version (73K):
[in this window]
[in a new window]
|
Fig. 5.
Localized delivery of curcuminoids reduces
FGF-2-induced expression of gelatinase B in neovascularized corneal
tissues. A circular region of neovascularized tissue 5 mm in
diameter encompassing the FGF-2 pellets was isolated from the corneas
of each of the four rabbits in the three different treatment groups.
Equal amounts of SDS-extracted proteins from each tissue specimen were
subjected to gelatin zymography. The presence of specific gelatinases
was revealed by the clear bands in the stained gelatin background.
Identification of enzymes was made by comparison of migration position
to a rabbit corneal cell gelatinase standard (Std):
gelatinase B at 92 kDa (GelB), gelatinase A proenzyme at 65 kDa (pGelA), and a proteolytically activated form of
gelatinase A at 62 kDa (GelA). The arrow
indicates the glycosylated form of gelatinase B, and the
arrowhead indicates the unglycosylated form.
|
|
Dietary Curcuminoids Inhibit Corneal Angiogenesis in Mice--
We
returned to the mouse model for a more quantitative analysis of the
effects of curcuminoids on corneal angiogenesis. In this study, we also
chose to examine the efficacy of an oral route for curcuminoid
delivery, through dietary supplementation. After implantation of a
single 80-ng slow release FGF-2 pellet in one eye of each mouse, one
group of mice was placed on the curcuminoid-supplemented (1 g
curcuminoids/kg chow) diet, whereas the control group was fed regular
rodent chow. Mice were allowed to feed and drink water ad
libitum. The implanted corneas of mice were photographed on days 6 and 10 after FGF-2 pellet implantation to visualize the development of
new blood vessels. Angiogenesis was quantified in each cornea by
measurement of total sum of blood vessel number and length. Mice fed a
diet containing curcuminoids were found to have fewer and less tortous
blood vessels compared with the control group on day 6 post-surgery
(Fig. 6, A and B).
In addition, the limbal and corneal blood vessels of the
curcuminoid-fed group were not as dilatated as the control group. An
even greater difference between the experimental and control groups in
these parameters was observed in mice fed curcuminoids for an
additional 4 days (Fig. 6, C and D). Quantitation
of the angiostatic effect of curcuminoids revealed a greater than 60%
inhibition of new blood vessel growth (Table
I), which was significant
(p < 0.007).
View larger version (141K):
[in this window]
[in a new window]
|
Fig. 6.
Inhibition of corneal angiogenesis by dietary
curcuminoids. Corneas of CD-1 mice (n = 10) were
implanted with a single slow release 80-ng FGF-2 pellet, and one group
of five mice was placed on normal chow, whereas the second group of
five mice was fed rodent chow containing 1 g curcuminoids/kg chow.
Representative photographic images of mouse corneas taken at day 6 (A and B) and day 10 (C and
D) reveal the presence of corneal blood vessels.
A and C are representatives from the control
group, whereas B and D are from the
curcuminoid-fed group. The asterisk indicates the position
of the FGF-2 pellet.
|
|
We also evaluated the possibility that prefeeding mice a diet
containing curcuminoids prior to FGF-2 stimulation might have therapeutic effects. In this experiment, the test group of mice was fed
a diet of curcuminoids (2 g/kg chow) for 1 week prior to surgical
implantation of FGF-2 pellets and continued on this chow for 6 days
post-surgery; the control group was maintained on normal chow
throughout the entire period. By as early as day 6 post-surgery, new
blood vessel growth was inhibited by greater than 50% (Table I) in the
curcuminoid-treated mice compared with the control group, which was
also significant (p < 0.0002). Together, these
findings reveal that curcuminoids harbor potent angiostatic activity
also when provided in the diet.
| |
DISCUSSION |
We have employed a gelatinase B promoter/lacZ
transgenic mouse (line 3445) and a corneal micropocket angiogenesis
model to demonstrate that gelatinase B is a downstream target in the
FGF-2-regulated angiogenic pathway. Using cultured corneal cells, we
show that FGF-2-stimulates DNA binding activity of transcription factor AP-1 but not NF-B and that stimulation of AP-1 is inhibited by curcuminoids. We further show that induction of gelatinase B promoter activity in response to FGF-2 is dependent on AP-1, but not NF-B, response elements and that promoter activity is also inhibited by
curcuminoids. When curcuminoids were delivered locally to the cornea
via an implantable pellet, the angiogenic response to FGF-2 was
inhibited, including induction of endogenous gelatinase B expression.
This finding is in keeping with previous studies demonstrating targeting of gelatinase B expression by curcuminoids (45, 46). The
angiostatic activity of curcuminoids on FGF-2-stimulated angiogenesis was further demonstrated by dietary provision to mice. To our knowledge, this is the first study showing the efficacy of curcuminoids as implantable drugs for inhibition of angiogenesis locally or as
orally active drugs to inhibit angiogenesis systemically.
Accumulating evidence implicates both of the gelatinolytic matrix
metalloproteinases, gelatinase A and gelatinase B, in the process of
angiogenesis. Although these two enzymes have very similar
specificities for extracellular components of the basement membrane,
their activities are regulated quite differently. Thus, gelatinase A is
constitutively present in tissues, primarily in the proenzyme form, and
is activated upon demand (36). In contrast, gelatinase B proenzyme
expression is transcriptionally regulated. Here, we utilized line 3445 mice to study activation of the gelatinase B promoter during
angiogenesis in situ. We previously documented the validity
of this mouse model to study the regulation of gelatinase B promoter
activity in tissue remodeling events of development and wound healing
(30). In this study, we demonstrate that the gelatinase B promoter is
spatially and temporally regulated in response to FGF-2 stimulation in
a manner that supports the requirement for focal expression of
gelatinase B in angiogenesis. Thus, cells of invading corneal
capillaries in line 3445 transgenic mice displayed discrete regions of
promoter activity, in a manner similar to the localized activity
observed in basal cells of the actively migrating epithelial sheet in
wound healing skin (30).
Response elements for transcription factor AP-1 are essential for basal
activity of the gelatinase B promoter and are required for promoter
stimulation by all agents and conditions tested thus far (10, 31, 39,
40). Studies on FGF-2 signaling have identified AP-1 as a final
effector of gene expression activated by this pathway (38). The EMSA
performed in this study demonstrating activation of AP-1 DNA binding
activity by FGF-2 confirmed this report. In contrast, transcription
factor NF-B was not activated by FGF-2 in our cultured stromal cell
model. Our transfection experiments corroborated the EMSA findings,
revealing that the FGF-2-stimulated transcriptional response of the
gelatinase B gene promoter requires AP-1 but not NF-B response
elements. These findings agree with a recent report that activation of
NF-B occurs only on co-stimulation of FGF-2 with inflammatory
cytokines such as interleukin-1 and tumor necrosis factor- (41).
Furthermore, lack of requirement for NF-B is consistent with
FGF-2-stimulated angiogenesis in vitro; Stoltz et
al. (42) have shown that NF-B is not activated in endothelial
cells when stimulated by FGF-2 to form capillary tubes in Matrigel.
Together these findings suggest that AP-1 would offer the best target
to interfere with FGF-2 signaling to abrogate angiogenesis and
expression of gelatinase B.
Curcuminoids exhibit both antioxidant and anti-inflammatory activities.
These activities have been ascribed to their ability to scavenge active
oxygen and nitrogen species and interfere with lipid peroxidation
(reviewed in Ref. 24). Owing to their potent inhibitory activity on
arachidonic acid release and metabolism both cycloxygenase and
lipoxygenase pathways are inhibited by these natural products (43).
Anticancer activity of curcuminoids has been associated with inhibition
of c-myc, c-jun, and c-fos oncogene
expression, and signal transduction studies have revealed that this
inhibition occurs via inhibition of c-Jun N-terminal kinase activation
(44).
Although curcuminoids are known to inhibit AP-1 and NF-B activation
in response to a variety of stimuli (25, 26), we are not aware of
previous studies examining their inhibitory effects on FGF-2-stimulated
activation. Interestingly, the inhibitory activity of curcuminoids on
AP-1 was found to be stimulus-dependent; constitutive DNA
binding activities were unaffected as demonstrated in EMSA. In
addition, we demonstrate that curcuminoids reduce FGF-2-induced but not
basal transcriptional activity of the gelatinase B promoter. These
results suggest that curcuminoids act to block some step in the FGF-2
signaling pathway upstream of DNA-binding that becomes engaged upon
growth factor stimulation. It is now well known that curcuminoids
inhibit production of reactive oxygen species (ROS) that act as
intermediates in many signal transduction pathways (24). ROS are a
likely target in our case because it has been reported that FGF-2
induces AP-1 activation via ROS produced through NADPH oxidase
(38).
Turmeric and its ethanolic extracts (curcuminoids) have been documented
as antiinflammatory and antioxidant agents in treatment of tumors,
arthritis, and wound healing disorders, both through oral intake and
topical application (50, 51). In fact, dietary curcuminoids have been
shown to improve the metabolic status in diabetic condition in rats
(52) and demonstrated anticarcinogenic effects in several preclinical
tests (24, 53, 54). In this study, we show that dietary supplementation
with curcuminoids has significant angiostatic effects in mice. However,
to demonstrate efficacy the timing and duration of treatment was an
important factor. We postulate that curcuminoids acting as antioxidants may exhibit anti-angiogenic activities by stimulating redox-regulatory defense systems reviewed in Ref. 55. Given the finding that FGF-2
induces ROS production (38), we speculate that predosing mice with
dietary curcuminoids would impinge on this redox pathway and interfere
with the ability of FGF-2 to stimulate AP-1. The identification of the
target of curcuminoids should throw more light on the molecular
mechanisms of its inhibitory activity on FGF-2 signaling.
It is of interest that quite a few biologically active natural products
derived from ethnic foods are being explored for the treatment of
cancer, arthritis, and wound healing disorders. A recent study has
revealed that vascular endothelial growth factor-induced corneal
angiogenesis is inhibited by providing green tea extract (containing
epigallocatechin-3-gallate) in drinking water to mice (56). The
curcuminoids, which are the major active ingredients of the spice
turmeric, are considered relatively safe for human consumption by the
joint Food and Agriculture Organization of the United Nations/World
Health Organization committee pending preclinical toxicity studies
(24). This provides an unique opportunity to test the inhibitory
activities of curcuminoids in other angiogenic models as well. We
propose the local delivery and/or systemic administration of
curcuinoids to control fibrovascular proliferative diseases where long
term therapy may be necessitated.
§
TOP
ABSTRACT
INTRODUCTION
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
