| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Zentrum der Pharmakologie, Klinikum der Johann Wolfgang Goethe-Universität, D-60590 Frankfurt am Main, Germany
2Correspondence: Institut für Allgemeine Pharmakologie und Toxikologie, Klinikum der JWG-Universität Frankfurt/M., Theodor-Stern-Kai 7, D-60590 Frankfurt/M., Germany. E-mail: Pfeilschifter{at}em.uni-frankfurt.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-, and
interferon-
-induced VEGF mRNA and protein synthesis in
keratinocytes. We could demonstrate that high-level expression of
cytokine-induced VEGF mRNA in keratinocytes is dependent on
endogenously produced NO, as inhibition of the coinduced iNOS by
NG-monomethyl-L-arginine (L-NMMA) markedly decreased
cytokine-triggered VEGF mRNA levels in the cells. We also established
an in vivo model in mice to investigate the role of NO
during wound healing. During excisional wound repair, mice were treated
with L-N6-(1-iminoethyl)lysine (L-NIL), a selective
inhibitor of iNOS enzymatic activity. Compared to control mice,
L-NIL-treated animals were characterized by markedly reduced VEGF mRNA
levels during the inflammatory phase of repair. Immunohistochemistry
demonstrated reduced VEGF protein expression and a completely
disorganized pattern of VEGF-expressing keratinocytes within the
hyperproliferative epithelium at the wound edge in L-NIL-treated mice.
We demonstrate that triggering of VEGF expression is a crucial
molecular mechanism underlying NO function during wound
healing.—Frank, S., Stallmeyer, B., Kämpfer, H., Kolb, N.,
Pfeilschifter, J. Nitric oxide triggers enhanced induction of vascular
endothelial growth factor expression in cultured keratinocytes (HaCaT)
and during cutaneous wound repair.
Key Words: gene expression regulation • skin • inducible nitric oxide synthase
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
.
Inflammation, re-epithelialization, and granulation tissue formation
are driven in part by a complex mixture of growth factors and
cytokines, which are released coordinately into the area of injury.
Besides these protein-type factors and mitogens, evidence is emerging
for an important role of small diffusible molecules in wound repair.
One of them is nitric oxide (NO), a free radical gas that has become
one of the most studied molecules in biomedical sciences during the
past decade.
NO is synthesized by three types of nitric oxide synthases (NOS).
These NOS isoenzymes represent two constitutive isoforms in neuronal
and endothelial tissue and an inducible NOS (iNOS) that releases large
amounts of NO into the surrounding tissue during cytokine-triggered
inflammatory processes (2
3
4
5
6
7)
. All three NOS isoforms are
homodimeric proteins dependent on NADPH, reduced flavins, heme-bound
iron, and an 6(R) 5,6,7,8-tetrahydrobiopterin as essential cofactors
(8
, 9)
. The biological activities of NO in human skin
include regulation of vasodilatation, melanogenesis, and protection
against invading pathogens (10
11
12)
. Moreover, an
important role of NO is emerging for the process of cutaneous wound
healing. A very recent in vivo study demonstrated a severely
delayed cutaneous wound closure in iNOS-deficient mice after skin
injury. This deficiency was completely reversed by a single application
of an adenoviral vector system carrying a cDNA fragment encoding human
iNOS into the wound (13)
. Furthermore, we demonstrated a
large and coordinate induction of iNOS expression during the normal
repair process. Glucocorticoids down-regulated iNOS expression in
wounded skin, suggesting that aberrant expression of the iNOS gene in
glucocorticoid-treated mice might be associated with the wound healing
defects seen in these animals (14)
. However, little is
known about the functional role of NO during the healing process. Among
the different functions mediated by NO, it has become clear that NO
affects expressional regulation of an increasing number of genes
(15)
, including the gene encoding for vascular endothelial
growth factor (VEGF) (16
17
18
19)
.
VEGF represents the most recently discovered endothelial growth factor.
VEGF is a dimeric glycoprotein occurring in four different isoforms in
humans—VEGF121, VEGF165,
VEGF189, and VEGF206—which
arise by alternative splicing of mRNA (20)
. VEGF acts as a
highly specific mitogen for endothelial cells in vitro and
reveals angiogenic properties in vivo (21
22
23
24)
.
Furthermore, it enhances the permeability of local blood vessels. VEGF
is discussed to play a crucial role in tissue repair, as angiogenesis
and increased vascular permeability are important events during wound
healing (25
, 26)
. Expression of VEGF mRNA has been
demonstrated in proliferating keratinocytes at the wound edge during
cutaneous repair, revealing keratinocytes as an important source of
VEGF production within this process (27)
. Recently, we
reported induction of VEGF expression by serum growth factors and
inflammatory cytokines in keratinocytes in vitro and a
defect in VEGF regulation during wound healing in genetically diabetic
db/db mice (28)
, which are characterized by a severe delay
in skin repair (29)
. These findings suggest that
keratinocyte-derived VEGF might stimulate angiogenesis during wound
healing.
In this study, we have identified NO as a potent enhancer of
growth factor- and cytokine-induced VEGF expression in keratinocytes.
Since keratinocytes are a major source of VEGF production during the
healing process, we speculated that expression of VEGF might be
impaired during cutaneous wound repair in animals treated with
L-N6-(1-iminoethyl)lysine (L-NIL), a
selective inhibitor of iNOS enzymatic activity (30)
. We
provide evidence for a dramatic decrease in VEGF mRNA and protein
expression during healing in L-NIL-treated mice. Furthermore, histology
revealed a disorganized pattern of VEGF-expressing keratinocytes during
re-epithelialization. The aberrant expression pattern of VEGF during
skin repair in these animals suggests that regular expression of VEGF
during wound healing is dependent on the presence of NO and, hence, on
a functionally active iNOS in the regenerating tissue.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. At this
concentration, L-NIL almost completely blocked the in vivo
enzymatic activity of iNOS in lymph nodes and peripheral skin lesions
of Leishmania-infected mice. This inhibition of iNOS has
been shown to be potent and presumably irreversible in vivo
and is not associated with weight loss or reduced water and food
consumption by the treated animals (31
, 32)
. Control mice
were injected with PBS. L-NIL was from Alexis Corporation
(Grünberg, Germany).
Wounding and preparation of wound tissues
To examine VEGF expression during the wound healing process, six
full-thickness wounds were created on each animal and skin biopsy
specimens from four animals were obtained 1, 3, 5, 7, and 13 days after
injury. To investigate the effect of L-NIL on VEGF expression, wounds
from four L-NIL-treated mice were obtained 1, 3, 5, 7 and 13 days after
injury. Mice were anesthetized with a single i.p. injection of Ketamin
(80 mg/kg body weight)/Xylazin (10 mg/kg body weight). The hair on the
back of these mice was cut and the back was subsequently wiped with
70% ethanol. Six full-thickness wounds (~4–6 mm in diameter, 3–4
mm apart) were made on the backs of these mice by excising the skin and
the underlying panniculus carnosus. The wounds were allowed
to dry to form a scab. An area of 7–8 mm in diameter, which included
the scab and the complete epithelial margins, was excised at each time
point. As a control, a similar amount of skin was taken from the backs
of four nonwounded mice. For every experimental time point, the wounds
from four animals (n=16 wounds) and the nonwounded back skin
from four animals, respectively, were combined, frozen immediately in
liquid nitrogen, and stored at -80°C until used for RNA isolation.
All animal experiments were carried out according to the guidelines and
with permission from the local government of Hessen.
RNA isolation and RNase protection analysis
RNA isolation was performed as described (33)
.
Thirty micrograms of total RNA from wounded or nonwounded skin or 20
µg of RNA from the cell culture experiments, respectively, was used
for RNase protection assays. RNase protection assays were carried out
as described previously (34)
. Briefly, DNA probes were
cloned into the transcription vector pBluescript II KS (+) (Stratagene,
Heidelberg, Germany) and linearized. An antisense transcript was
synthesized in vitro using T3 or T7 RNA polymerase and
[-32P]UTP (800 Ci/mmol). RNA samples were
hybridized at 42°C overnight with 100,000 cpm of the labeled
antisense transcript. Hybrids were digested with RNases A and T1 for
1 h at 30°C. Under these conditions, every single mismatch is
recognized by the RNases. Protected fragments were separated on 5%
acrylamide/8M urea gels and analyzed using a PhosphorImager (Fuji). All
protection assays were carried out with at least three different sets
of RNA from independent wound healing or cell culture experiments. A
316 bp fragment corresponding to nucleotides 139–455 of the murine
VEGF120 cDNA (35)
and a 159 bp
fragment corresponding to nucleotides 339–498 of human
VEGF121 cDNA (36)
were used as
templates. RNases A and T1 were from Boehringer Biochemicals (Mannheim,
Germany).
Cell culture
The human keratinocyte cell line HaCaT (37)
was
cultured in Dulbecco's modified Eagle's medium (DMEM) with 10% fetal
calf serum (FCS). Under these culture conditions, the keratinocytes
remain proliferative and undifferentiated. For the VEGF induction
experiments, cells were grown to confluency without changing the medium
and rendered quiescent by a 24 h incubation in serum-free DMEM.
Cells were then incubated for varying periods on fresh DMEM containing
serum, purified growth factors, cytokines, or reagents. Aliquots of
cells and cell culture supernatants were harvested before and at
different time points after treatment with these agents and used for
RNA isolation or enzyme-linked immunosorbent assays (ELISA),
respectively. Each experiment was done at least in triplicate. FCS and
DMEM were purchased from Life Technologies, Inc. (Eggenstein, Germany),
growth factors and cytokines were from Boehringer Mannheim
Biochemicals, cycloheximide, actinomycin D,
N6,2'-O-dibutyryladenosine 3',5' cyclic
monophosphate (cAMP), and 8-bromoguanosine 3',5' cyclic monophosphate
(cGMP) were from Sigma Biochemicals (Deisenhofen, Germany), and
NG-monomethyl-L-arginine (L-NMMA) was from Alexis
Biochemicals.
ELISA
Keratinocyte-conditioned cell culture supernatants (10 ml) from
the individual experimental time points was pooled and cleared by
centrifugation. 200 microliters of the cell culture supernatants were
subsequently analyzed for the presence of immunoreactive VEGF protein
by ELISA using the Quantikine human VEGF kit (R&D systems, Wiesbaden,
Germany) as described by the manufacturer.
GSNO synthesis
S-Nitroso-glutathione was synthesized as described previously
(38)
. Briefly, glutathione was dissolved in 0.625 N HCl at
0°C to a final concentration of 625 mM. NaNO2
was added equimolar and the mixture was stirred at 0°C for 40 min.
After the addition of 2.5 volumes of acetone, stirring continued for
another 20 min, followed by filtration of the precipitate. GSNO was
washed once with 80% acetone, twice with 100% acetone, three times
with diethylether, and dried under vacuum. GSNO was characterized by UV
spectroscopy.
Preparation of wound lysates and nitrite determination
Nitrite, a stable NO oxidation product, was determined in wound
lysates using the Griess reaction. Normal and wounded back skin
(n=8 wounds per experimental time point) was frozen in
liquid nitrogen. Skin samples were homogenized in 2 x lysis
buffer (1xlysis buffer: 1% Triton X-100, 20 mM Tris/HCl, pH 8.0, 137
mM NaCl, 10% glycerol, 5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride,
15 µg/ml leupeptin). The tissue extract was cleared by centrifugation
and the supernatant was diluted 1:1 with water. Lysate (1 ml) from
nonwounded skin and from wounds harvested 3, 5, 7, and 13 days after
injury was cleared by a centrifugation step at 100,000 x
g for 30 min. Cleared wound lysates (200 µl) and cell
culture supernatants (200 µl), respectively, were adjusted to 4°C
and mixed with 20 µl sulfanilamide (dissolved in 1.2 M HCl) and 20
µl N-naphtylethylendiamine dihydrochloride. After 5 min at
room temperature, the absorbance was measured at 540 nm with a
reference wavelength at 690 nm. Phenylmethylsulfonyl fluoride,
leupeptin, sulfanilamide, and N-naphtylethylendiamine
dihydrochloride were from Sigma Biochemicals.
Immunohistochemistry
Mice were wounded as described above. Animals were killed at
days 3 and 5 after injury. Complete wounds were isolated from the
middle of the back, bisected, and frozen in tissue-freezing medium. Six
micrometer frozen sections were fixed with acetone and treated for 10
min at room temperature with 1%
H2O2 in PBS to inactivate
endogenous peroxidases. They were subsequently incubated for 60 min at
room temperature with a polyclonal anti-serum against murine VEGF
(Santa Cruz, Heidelberg, Germany) (1:100 diluted in PBS, 0.1% bovine
serum albumin). The slides were subsequently stained with the
avidin-biotin-peroxidase complex system from Santa Cruz (Heidelberg,
Germany) using 3-amino-9-ethylcarbazole as a chromogenic substrate.
After development, slides were rinsed with water, counterstained with
hematoxylin, and mounted.
Statistical analysis
Data are shown as means ± SD. The data are
presented either as x-fold induction compared with the unstimulated
control (100%) or as mean concentrations as picograms per milliliter.
Data were analyzed by unpaired Student's t test on raw data
using Sigma Plot (Jandel Scientific, Erkrath, Germany).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 39)
. To identify potential functions of NO action
during wound healing, we studied the regulation of VEGF gene expression
in the presence of S-nitroso-glutathione (GSNO) in the human
keratinocyte cell line HaCaT (37)
, since keratinocytes
represent a major source of VEGF production during the process of
cutaneous repair (27
, 28)
. As shown in Fig. 1
A, low basal levels of VEGF mRNA expression were detected in
quiescent keratinocytes. Upon addition of 500 µM GSNO, an induction
of VEGF mRNA expression was observed. Within 2 h after GSNO
stimulation, VEGF mRNA levels were sixfold higher compared to the basal
level. The effect of GSNO on VEGF mRNA expression was transient, with
maximal VEGF mRNA levels occurring between 2 h and 8 h after
stimulation (see also Fig. 2
A, Fig. 3
A, Fig. 4
A, B). Induction of VEGF expression could also be
observed using the NO donor spermine-NONOate (data not shown). The
GSNO-mediated increase in VEGF mRNA levels could be blocked completely
in the presence of the transcription inhibitor actinomycin D (2
µg/ml), which was added to the cells 45 min prior to GSNO treatment
(Fig. 1A
). This finding indicates that GSNO acts by
transcriptional activation of the VEGF gene in keratinocytes. The rapid
induction of VEGF mRNA expression by GSNO suggested that VEGF mRNA
induction could occur in the absence of de novo protein
synthesis. To address this question, the effect of the protein
synthesis inhibitor cycloheximide (25 µg/ml) on GSNO-stimulated cells
was analyzed. Indeed, cycloheximide did not suppress GSNO-induced VEGF
mRNA expression, but resulted in a tremendous superinduction (120-fold)
of VEGF mRNA expression in response to the NO-donating agent (Fig. 1A, B
). Treatment of cells with cycloheximide in the absence
of GSNO did not alter VEGF expression (data not shown). Therefore,
GSNO-triggered VEGF gene activation is independent from de
novo protein synthesis; therefore, NO-induced VEGF expression
resembles characteristics of primary response gene activation.
|
|
|
|
Finally, we investigated whether GSNO-triggered VEGF mRNA induction
might use cAMP- or cGMP-mediated signaling pathways in keratinocytes,
since involvement of both second messenger molecules has been described
for VEGF gene activation in different cell lines (16
, 40)
.
Surprisingly, neither cAMP nor cGMP treatment altered VEGF mRNA
expression levels at all, indicating that NO-triggered VEGF gene
activation is most likely independent of protein kinase A and
G-mediated signaling pathways in keratinocytes (Fig. 1A
).
Serum- and serum growth factor-mediated induction of VEGF
expression in the presence of GSNO
Since serum-derived growth factors are present in the early wound
tissue after local hemorrhage and iNOS expression could be detected as
early as 12 h after injury (14)
, we first tested the
potency of GSNO to modulate serum-mediated VEGF expression in
keratinocytes. As shown in Fig. 2A
, low levels of VEGF mRNA
were detected in quiescent keratinocytes. As previously published,
serum, epidermal growth factor (EGF), and transforming growth
factor-ß1 (TGF-ß1) were strong inducers of VEGF expression, with
TGF-ß1 being less potent than serum and EGF. Platelet-derived growth
factor BB (PDGF), the major mitogen in platelets, had no effect (data
not shown). In the presence of 500 µM GSNO, serum- and EGF-mediated
VEGF expression was not affected (Fig. 2A
), suggesting that
the GSNO stimulus is not able to further increase the large induction
in VEGF mRNA induction seen in keratinocytes as a response to serum-
(50-fold) or EGF- (70-fold) treatment. However, Fig. 2A
demonstrates that TGF-ß1-stimulated VEGF mRNA expression (threefold)
(panel marked by an arrow) is clearly enhanced in the
presence of 500 µM GSNO 2 to 8 h after induction (five- to
sevenfold). Moreover, GSNO is able to sustain TGF-ß1-mediated VEGF
mRNA expression when compared to control cells treated with TGF-ß1
only (Fig. 2B
). This is consistent with data obtained at the
protein level, since concentrations of VEGF-specific proteins are
highest in cell culture supernatants of TGF-ß1/GSNO-treated cells, as
assessed by ELISA (Fig. 2C
).
Keratinocyte growth factor-stimulated VEGF mRNA and protein
expression is enhanced by GSNO
KGF has been demonstrated to be a potent mitogen for keratinocytes
(41)
. Furthermore, KGF is highly expressed in the dermis
during wound healing, stimulating wound reepithelialization in a
paracrine manner (34
, 42)
. We addressed the question
whether KGF-induced VEGF expression in keratinocytes might be enhanced
by exogenously applied NO. As shown in Fig. 3A, B
, GSNO and
KGF were potent inducers of VEGF mRNA in keratinocytes. Compared to
GSNO and KGF treatment alone, combined
GSNO/KGF stimulation of keratinocytes revealed clearly increased VEGF
mRNA levels, resulting in high levels of VEGF protein as assessed by
ELISA (Fig. 3C
).
Cytokine-induced VEGF expression is increased by GSNO
Since the infiltration of polymorphonuclear leukocytes followed by
macrophages and T cells is another crucial event in wound repair, we
tested the ability of cytokines produced by theses cells to induce VEGF
expression in the presence or absence of the NO donor GSNO. As shown in
Fig. 4
, interleukin 1ß (IL-1ß) and interferon (IFN-)
-mediated VEGF mRNA induction are synergistically increased in the
presence of 500 µM GSNO (indicated by arrows). For
IL-1ß-induced VEGF mRNA levels, the synergistic effect of
GSNO/IL-1ß treatment was highest after 5 h of stimulation
(16-fold) compared to the stimulatory potency of GSNO (2.5-fold) and
IL-1ß (7-fold) alone (Fig. 4A
). Simultaneous stimulation
of cells with GSNO (2-fold induction) and IFN- (14-fold) revealed an
even stronger synergism in inducing VEGF mRNA levels (120-fold) after
24 h. However, we observed a different situation at the protein
level for NO-enhanced IL-1ß- or IFN--mediated VEGF expression. The
observed synergism of GSNO/IFN- at the mRNA level is clearly
followed by a synergistic VEGF protein secretion into the cell culture
supernatants (Fig. 4B
, right panel). This is in contrast to
the situation found in GSNO/IL-1ß-treated cells, where the
synergistic stimulatory effect at the mRNA level translates only in an
additive production of VEGF protein (Fig. 4A
, far right
panel). For tumor necrosis factor (TNF-) -stimulated
keratinocytes, we observed an 23-fold induction of VEGF mRNA levels,
which could be further increased in the presence of GSNO (32-fold)
(Fig. 4B
, left panel).
Stimulatory effect of GSNO on IFN--induced VEGF mRNA and protein
expression is dose dependent
We treated IFN--stimulated keratinocytes with increasing
amounts of the NO-donating agent GSNO (100, 250, 500 µM). As shown in
Fig. 5
, GSNO enhanced IFN--induced VEGF expression at the mRNA and protein
level in a dose-dependent manner. The maximum response was obtained
using 500 µM GSNO, but an enhancement of IFN--mediated VEGF
induction could be detected even at the lowest concentration of GSNO
(100 µM) (Fig. 5A, B
).
|
Cytokine-induced VEGF expression is increased by endogenously
produced NO
Since keratinocytes of the wound margins and the
hyperproliferative epithelium have been shown to strongly express iNOS
during the repair process (14)
, we investigated whether
endogenously produced NO had the potency to trigger an enhanced VEGF
expression in these cells. We stimulated keratinocytes using a cytomix
(2 nM IL-1ß, 2 nM TNF-, 100 U/ml IFN-) to coinduce iNOS
(39)
and VEGF expression in keratinocytes, since purified
growth factors and cytokines alone did not induce iNOS and subsequent
nitrite accumulation in the cell culture supernatants
(39)
. As shown in Fig. 6A
, the cytomix rapidly induced VEGF expression (four- to
sixfold) after 3 h (Fig. 6B
). This rapid and first
cytomix-induced increase in VEGF mRNA levels was independent of the
availability of NO, as expression of iNOS could not be detected at all
at the early 3 h time point of cytokine stimulation (Fig. 6A
, right panels). Remarkably, VEGF mRNA levels showed a
second strong increase after 8–16 h of stimulation (20- to 55-fold)
(Fig. 6B
), when cytokine-induced iNOS mRNA could be detected
in keratinocytes (Fig. 6A
, right panels). Addition of
L-NMMA, a potent inhibitor of all NOS isoenzymes, completely blocked
the enhanced induction of cytokine-induced VEGF mRNA associated with
the expression of iNOS. Inhibition of NOS enzymatic activities could be
demonstrated by measuring nitrite levels in the cell culture
supernatants of the same experiment (Fig. 6C
). These data
clearly indicate that the enhanced induction of VEGF expression
observed in keratinocytes after cytokine stimulation was dependent on
the presence of endogenously produced NO.
|
Defects in VEGF expression during wound healing in L-NIL-treated
mice
Finally, we speculated that the observed enhancement of VEGF
expression in growth factor- and cytokine-stimulated keratinocytes
might be an important and common mechanism occurring in the in
vivo situation during cutaneous wound healing. If our hypothesis
were true, we should expect a decreased or delayed VEGF expression
pattern in L-NIL-treated animals compared to the control situation,
since L-NIL is a potent and selective inhibitor of iNOS
(30)
. For this purpose, we isolated RNA from excisional
full-thickness wounds at different intervals after wounding from
L-NIL-treated mice and PBS-injected control mice and performed RNase
protection assays. Sixteen wounds from the backs of four mice were
excised for each time point and used for RNA isolation. Nonwounded back
skin from the same area was used as a control. To ensure the inhibitory
activity of the selective iNOS inhibitor L-NIL, we determined
concentrations of nitrite, a stable oxidation product of NO, in wound
lysates from PBS- or L-NIL-injected mice, respectively. As shown in
Fig. 7
A, we observed a strong reduction in wound nitrite
concentrations at day 3 and day 5 after injury, clearly indicating an
inhibitory effect of L-NIL on iNOS enzymatic activity at times when
iNOS expression levels are highest during repair (14)
.
Expression of iNOS and release of NO from macrophages clearly
diminished from day 5 after injury (14
, 43)
. Accordingly,
we could not observe increased nitrite levels 7 days after injury (Fig. 7A
). We were not able to determine nitrite levels in wound
lysates from day 1 wounds, as the samples were highly contaminated with
blood that interfered with the Griess color reaction. The highly
colored lysates were due to early hemorrhage and formation of a blood
clot at the area of injury. This clot and the scab were greatly reduced
at day 3 postwounding, and lysates from 3 day wounds subsequently were
nearly colorless. Thus, we could only determine nitrite levels from day
3 on. As shown in Fig. 7B
, two different protected fragments
were obtained with RNA from normal and wounded skin, corresponding to
different forms of VEGF mRNA. The longer protected fragment is
generated by RNA encoding the murine VEGF120,
whereas the shorter protected fragment is generated by mRNAs encoding
murine VEGF164 and VEGF188.
In PBS-injected control mice, a strong induction of VEGF mRNA
expression was observed within 24 h after injury; highest VEGF
mRNA levels were found between day 1 and 7 after injury. These data are
consistent with previously published data (28)
. However,
VEGF mRNA expression is severely altered in wounds isolated from
L-NIL-treated mice. As shown in Fig. 7B
, the initial
increase and sustained high-level expression of VEGF mRNA found in
PBS-injected control mice were dramatically reduced (50%) in
L-NIL-injected animals. This reduction in VEGF mRNA occurred during the
inflammatory phase and granulation tissue formation occurred during
repair. These findings demonstrate that the wound healing defect seen
in iNOS-deficient mice (13)
might be associated with
reduced VEGF expression during the time when wound angiogenesis
normally occurs.
|
L-NIL-treated mice show a disorganized pattern of VEGF-expressing
keratinocytes during re-epithelialization
Finally, we were interested in the morphology of wounds isolated
from L-NIL-treated animals. For this reason, we isolated total wound
tissue at day 3 and day 5 after injury from PBS- and L-NIL-injected
mice. To confirm the results obtained at the mRNA level, we stained the
wound tissue against VEGF protein using immunohistochemistry. As shown
in Fig. 8
, we could observe a clearly distinct pattern of VEGF expression in
wounds of PBS-treated control mice compared with wounds isolated from
L-NIL-treated animals. For PBS-injected control mice, we observed a
large number of keratinocytes within the hyperproliferative epithelium
that were characterized by an intense staining, indicating a large and
strong expression of VEGF protein. These cells were located at the
margins of the hyperproliferative epithelium at the wound edge,
directly adjacent to the developing granulation tissue. Remarkably, the
immunopositive keratinocytes formed an well-organized cluster of
VEGF-expressing cells (Fig. 8A-C
). This is in contrast to
the situation found in wounds of L-NIL-treated animals. In these mice,
the hyperproliferative epithelium was characterized by a dramatic
reduction of VEGF protein expression. This effect was due to a clearly
reduced number of keratinoctyes expressing VEGF protein. We observed a
completely different pattern of VEGF-positive keratinocytes within the
hyperproliferative epithelium. VEGF-expressing keratinocytes were
`spotted' into this proliferating mass of cells rather than forming a
coherent VEGF-expressing group of cells at the wound edge (Fig. 8D-F
). Therefore, the few VEGF-expressing
keratinocytes found in L-NIL-treated mice were not located close to the
granulation tissue where angiogenesis normally occurs.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. However, it became evident
in recent years that nitric oxide, a free radical gas, might represent
an important signaling molecule in the skin. Under normal physiological
conditions, two NO-generating enzyme systems could be detected in human
skin: the endothelial-type NOS (eNOS) and iNOS. Immunohistochemistry
revealed that both NOS isoforms are expressed in epidermal
keratinocytes (45)
. NO is discussed to be involved in
maintaining certain aspects of skin homeostasis: regulation of
vasodilatation, melanogenesis, or protection against invading pathogens
(10
11
12)
. Besides theses physiological functions, however,
it also turned out that NO is likely to have deleterious effects in
pathophysiological states of the skin: iNOS, which is characterized by
its ability to produce large amounts of NO, is overexpressed in skin
dermatoses and psoriatic skin lesions (46
47
48
49)
. Another
important role of NO seems to emerge for the process of cutaneous wound
healing. A recent study demonstrated a functional role of NO during
wound repair. In this in vivo study, iNOS-deficient knockout
mice are characterized by a severe delay (31%) in wound closure.
Furthermore, this delay could be compensated for by a single low-titer
application of an adenoviral vector system containing the cDNA encoding
human iNOS (13)
. Accordingly, we observed a large and
coordinated induction of iNOS and its tetrahydrobiopterin-synthesizing
cofactor GTP-cyclohydrolase I during excisional wound repair.
Furthermore, a severe down-regulation of these NO-generating enzyme
systems was found during wound healing in glucocorticoid-treated mice,
which are characterized by wound healing disorders (14)
.
These data implicate a potential role of iNOS for a normal healing
process. This is supported by clinical studies demonstrating that oral
supplementation of L-arginine has a beneficial effect on immune
function and wound healing in patients (50
, 51)
.
Furthermore, arginine deficiency has been shown to be associated with
impaired wound healing in rats (52)
. All three NOS
isoforms catalyze a reaction leading to NO and citrulline, using
L-arginine as a substrate. And indeed, both products of NOS activity
were detected in wound fluid during the early, inflammatory phase of
repair (53)
. In spite of these data, which clearly
demonstrate a role of NO during wound repair, the molecular mechanisms
underlying NO action during repair remain largely unknown. During
repair, iNOS-derived NO might activate or modulate gene transcription
in target cells lying within the wounded tissue. Recent studies
demonstrated the potency of NO to regulate VEGF gene expression in
human glioblastoma and hepatocarcinoma cells and rat renal mesangial
cells (16
, 19)
. By contrast, endothelial cell-derived, NO
down-regulated VEGF expression in underlying smooth muscle cells in
regenerating blood vessels (17)
, thus demonstrating a
tissue- and cell type-specific effect of NO on VEGF gene expression.
We decided to establish a model system in animals that would enable us
to elucidate molecular mechanisms underlying the crucial role of NO for
the process of cutaneous wound healing. In contrast to the
iNOS-deficient mouse model (13)
, we chose to inhibit iNOS
enzymatic activity during wound healing by i.p. application of the
selective iNOS inhibitor L-NIL. This inhibitor is known to be a potent
and selective inhibitor of mouse iNOS (30)
. L-NIL has been
described to be the more potent inhibitor in vivo than
NG-monomethylarginine, an inhibitor of enzymatic
action of all NOS isoenzymes. Furthermore, L-NIL almost completely
blocked iNOS enzymatic activity in lymph nodes and peripheral skin
lesions of Leishmania-infected mice. This inhibition by
L-NIL was presumably irreversible for the in vivo situation
(31
, 32)
. According to these data, we have chosen L-NIL
(2.5 mg/injection, every 12 h) to suppress iNOS enzymatic action
and subsequent NO release during the inflammatory phase of repair
(14)
. In line with published data, we detected strongly
reduced nitrite concentrations in wound lysates 3 and 5 days after
wounding in L-NIL-treated animals compared to nitrite concentrations in
wound lysates isolated from PBS-injected control mice, clearly
indicating a functional active inhibitor at the wound site.
We recently reported an induction of VEGF expression in keratinocytes
by growth factors and inflammatory cytokines. Furthermore, we
demonstrated a severe defect of VEGF regulation during wound healing in
genetically diabetic db/db mice, which are characterized by an impaired
wound healing (28)
. During cutaneous repair, VEGF is
expressed in keratinocytes at the wound margins; hence, keratinocytes
represent the major source of VEGF production within the wounded
tissue. Invading mononuclear cells expressed VEGF only occasionally
(27)
. Furthermore, expression of the flt-1
receptor for VEGF is up-regulated in the sprouting blood vessels at the
wound edge and in endothelial cells of the granulation tissue
(54)
. These findings suggest that keratinocyte-derived
VEGF is most likely to stimulate angiogenesis during wound healing in a
paracrine manner. This hypothesis is further supported by findings
demonstrating that keratinocyte-derived VEGF is indeed a potent mitogen
for dermal microvascular endothelial cells (55)
. Moreover,
VEGF, but not bFGF, mediates angiogenic activity during the
proliferative phase of wound healing in humans (26)
.
Since VEGF is most likely to be an important protein-type factor
for the healing process, and since VEGF expression is strongly
regulated by growth factors and cytokines in keratinocytes, we
speculated that expression of this gene might be influenced by NO in
keratinocytes in vitro and during wound repair in
vivo. This was indeed the case, as NO induced large expression of
VEGF mRNA and protein in cultured keratinocytes. This is consistent
with the observation that NO increased VEGF expression in human tumor
cell lines (16)
. Cycloheximide treatment of cells led to
an tremendous superinduction of VEGF mRNA by NO in keratinocytes,
indicating that de novo synthesis of protein-type factors is
not necessary for NO-triggered induction of the VEGF gene. This might
enable keratinocytes to respond rapidly to the triggering NO stimulus.
Furthermore, the observed superinduction indicates a tight
post-transcriptional regulation of NO-induced VEGF mRNA involving
short-lived, negatively regulating protein-type factors within
keratinocytes. NO represents a potent inducer of VEGF expression,
suggesting transcriptional activating mechanisms that are independent
of the pathways used by growth factors and cytokines. Moreover, NO has
the potency to interfere with cytokine-activated signal transduction
pathways, since the presence of either exogenous NO or endogenously
produced NO further enhanced cytokine-triggered VEGF expression. Until
now, however, cellular targets of NO action were poorly defined and
have not been clearly elucidated (15)
.
For the in vivo situation, serum growth factors released on
hemorrhage, inflammatory cytokines produced by mononuclear cells, and
NO are present at the same time and the same location at the wound
site. For this reason, we investigated the potency of NO to influence
growth factor and cytokine-triggered VEGF induction in keratinocytes.
We observed a larger production of VEGF in TGF-ß1-treated cells in
the presence of NO. Since large amounts of TGF-ß1 are released from
platelets on injury, the NO-enhanced, TGF-ß1-mediated VEGF induction
might significantly contribute to the early induction of VEGF
expression after wounding. Polymorphonuclear leukocytes are the
predominant cell type during the early inflammatory phase of repair.
They are followed by infiltrating macrophages and T cells. The
mononuclear cells have been shown to produce a series of inflammatory
cytokines including tumor necrosis factor-,
interleukin-1ß, or interferon-. Not unexpectedly, we observed
increased VEGF expression in TNF--, IL-1ß-, or IFN--stimulated
keratinocytes in the presence of NO compared to the effects triggered
by the cytokines alone. Notably, IFN- action on VEGF expression is
synergistically enhanced by NO. Since macrophages and T cells remain
present during granulation tissue formation and even beyond the
inflammatory phase of repair (1
, 56)
, the NO-enhanced
effects of these cytokines might also contribute to the long-lasting
high expression levels of VEGF during skin repair. A recent study
demonstrated that NO-regulated VEGF expression is not restricted to
keratinocytes, but might represent a common mechanism for regulation of
angiogenesis by VEGF. In analogy to our data, this study
(57)
demonstrated that VEGF-mediated angiogenic properties
of the RAW264.7 macrophage cell line and mouse peritoneal macrophages
were dependent on a functionally active iNOS pathway. Mouse peritoneal
macrophages expressed an antiangiogenic factor that blocks the
angiogenic activity of VEGF when the LPS/IFN--induced iNOS pathway
was inhibited in these cells using specific iNOS inhibitors.
Summarizing these observations, we speculated that VEGF expression
might be reduced during wound repair in L-NIL-treated mice. Consistent
with this hypothesis, we detected a significantly reduced VEGF
expression after injury in L-NIL-injected mice. In PBS-injected control
mice, a significant induction of VEGF expression was observed within
24 h after injury. Expression levels were high during the period
of granulation tissue formation and returned to basal levels only after
completion of skin repair. In contrast, VEGF mRNA levels were clearly
reduced (50%) in the presence of the selective iNOS inhibitor L-NIL,
and reduced levels of VEGF transcripts were present during the period
when wound angiogenesis normally occurs. Moreover, even this phase of
repair (day 1 to day 5) is characterized by the presence of iNOS, which
is expressed mainly by invading neutrophils and macrophages (14
, 43)
, but also strongly by keratinocytes of the
hyperproliferative epithelium (14)
. Accordingly, for the
in vivo situation, it becomes clear that either immune
cell-derived NO or NO produced endogenously in cytokine-induced
keratinocytes might contribute to enhanced VEGF expression in wound
keratinocytes.
Immunohistochemistry not only confirmed data obtained at the mRNA level, but revealed new insights into the role of NO in regulating cellular expression patterns within the regenerating tissue. However, we not only observed a reduced number of VEGF-expressing keratinocytes in the hyperproliferative epithelium but, more important, these few VEGF-expressing cells are located within the cellular mass of proliferating keratinocytes. This situation is in contrast to VEGF-producing keratinocytes lying at the margins of the hyperproliferative epithelium at the wound edge in wounds of control animals. These cells are in direct contact with the neighboring granulation tissue, thus enabling the secretion of the angiogenic VEGF into the site where angiogenesis occurs. For L-NIL-treated mice, one might speculate that expression of VEGF within the hyperproliferative epithelium, and not at its margins, further contributes to a reduced availability of VEGF for endothelial cell proliferation toward the developing granulation tissue.
These findings demonstrate that reduced levels of VEGF mRNA are directly associated with the inhibition of iNOS activity during repair, suggesting that up-regulation of VEGF expression is dependent on the presence of a functionally active iNOS and, hence, the availability of NO during normal wound healing. Our data implicate an important role for NO in regulating growth factor- and cytokine-triggered processes during repair, and therefore identify NO as a potential therapeutic target molecule to improve wound healing disorders.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication March 24, 1999. Revised for publication July 9, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
