Angiogenesis 00: 000000, 2004.

1
AUTHORS PROOF!

2004 Kluwer Academic Publishers. Printed in the Netherlands.

Dierent mechanisms lead to the angiogenic process induced by three

adenocarcinoma cell lines

Lilia E. Davel1, Laura Rimmaudo1, Alejandro Espanol1, Eulalia de la Torre1, Mar a Adela Jasnis1, Mar a
Laura Ribeiro2, Tomomi Gotoh3, Eugenia Sacerdote de Lustig1 & Mar a Elena Sales1
1
Instituto de Oncologa Angel H. Roo, Facultad de Medicina, UBA. Av. San Martn, Buenos Aires, Argentina;
2
Centro de Estudios Farmacologicos y Botanicos (CEFYBO)-CONICET, Serrano, Buenos Aires, Argentina;
3
Department of Molecular Genetics, Kumamoto University, School of Medicine Honjo 2-2-1, Kumamoto, Japan

Received 7 October 2003; accepted in revised form 6 January 2004

Key words: angiogenesis, arginase, COX, mammary tumor cell lines, NOS, VEGF

Abstract

Neoangiogenesis is essential for tumor and metastasis growth, but this complex process does not follow the same
activation pathway, at least in tumor cell lines originated from dierent murine mammary adenocarcinomas.
LMM3 cells were the most potent to stimulate new blood vessel formation. This response was signicantly reduced
by preincubating cells with indomethacin and NS-398, non-selective cyclooxygenase (COX) and COX-2 selective
inhibitors, respectively. COX-1 and COX-2 isoenzymes were both highly expressed in LMM3 cells, and we observed
that indomethacin was more eective than NS-398 to inhibit prostaglandin E2 (PGE2) synthesis. In addition, nitric
oxide synthase (NOS) inhibitors, Nx monomethyl L -arginine and aminoguanidine, also reduced LMM3-induced
angiogenesis and nitric oxide (NO) synthesis as well. NOS2 > NOS3 proteins and arginase II isoform were detected
in LMM3 cells by Western blot. The latter enzyme was also involved in the LMM3 neovascular response, since the
arginase inhibitor, Nx hydroxy L -arginine reduced the angiogenic cascade. On the other hand, parental LM3 cells
were able to stimulate neovascularization via COX-1 and arginase products since only indomethacin and Nx
hydroxy L -arginine, which diminished PGE2 and urea synthesis, respectively, also reduced angiogenesis. In turn,
LM2 cells angiogenic response could be due in fact to PGE2-induced VEGF liberation that stimulated
neoangiogenesis at very low levels of NO.

Abbreviations: COX cyclooxygenase; PGE2 prostaglandin E2; NOS nitric oxide synthase; NO nitric oxide;
VEGF vascular endothelial growth factor

Introduction present in neural and endothelial cells, respectively, and

Angiogenesis, the process that leads to tumor neovascu-
larization by new blood vessel formation, is essential for
tumor and metastasis growth [1]. It is a multistep and
highly regulated process in which genetic factors, angio-
genic and anti-angiogenic peptides and reactive species
are involved. Several mediators have been identied as
promoters of neovascularization such as VEGF, arginine
catabolites and prostanoids. The enzymatic degradation
of the aminoacid arginine can yield citrulline and NO
via NOS. NO production requires the presence of one
or more of the three NOS isoforms: neuronal (nNOS or
NOS1), endothelial (eNOS or NOS3) and inducible
(iNOS or NOS2) [2]. Generally, NOS1 and NOS3 are
Correspondence to: Prof Mar a Elena Sales, PhD, Guemes 3144 1 B CP
1425, Buenos Aires, Argentina. Tel: +54-11-48251780; Fax: +54-11-
jjjj; E-mail: mesales@2vias.com.ar
Journal : AGEN SPS Article No. : 03031
PIPS No. : 5266904
MS Code : AGEN 03031
PDF-OUTPUT
are constitutively expressed. On the other hand, NOS2
needs to be induced by cytokines or other inammatory
stimuli in essentially every cell type, and can generate
locally high concentrations of NO for prolonged periods
of time [3]. Several amounts of evidence relate iNOS
expression and activity with pathological states such as
cancer. L -Arginine is the common substrate for two
enzymes, NOS and arginase. Arginase is a hydrolase that
metabolizes arginine to urea and ornithine, which is the
precursor of polyamines, essential components of cell
proliferation. Two isoenzymes of arginase have been
described: arginase I, present in liver cell cytosol and
arginase II, ubiquitously localized in mitochondria [4].
There is growing interest in the potential role of arginases
as regulators of tumor growth.
Prostaglandins are a series of compounds that can
enhance tumor development, and their synthesis can be
attributed to both COX-1 and COX-2 enzymes [5].
Dispatch : 3-2-2004 Pages : 7
h LE h TYPESET
h CP
2 L.E. Davel et al.
COX-1 is constitutively expressed in a number of cell
types, whereas COX-2 is inducible by a variety of
soluble mediators, including cytokines and growth
factors [6]. It has been postulated that COX can regulate
tumor-induced angiogenesis by two mechanisms: COX-
2 regulates the production of nearly all angiogenic
factors in cancer cells, whereas COX-1 regulates prolif-
eration of endothelial cells [7]. The candidate to mediate
both these actions is PGE2.
We have previously characterized three dierent
tumor cell lines, LMM3, LM3 and LM2 derived from
mammary adenocarcinomas that spontaneously arose in
BALB/c female mice. These cell lines presented dier-
ences in their local invasive capacity as well as in their
spontaneous metastatic ability [8, 9].
Here we investigate and heft the participation of NOS,
arginases and COX as well as of their metabolic products,
NO, polyamine precursors and PGE2, in the angiogenic
activity of three dierent mammary tumor cells.
Materials and methods
Cell cultures
Tumor cell lines LM2, LM3 (derived from two dierent
spontaneous murine mammary tumors M2 and M3) and
LMM3 (derived from pulmonary metastasis from M3)
were obtained and established in our laboratory [8, 9].
Cells were maintained in MEM (Gibco, BRL jj) with
3 mM L -glutamine, 80 lg/ml gentamycin supplemented
with 5% heat-inactivated fetal calf serum (FCS) at
37 C in a humidied 5% CO2 air atmosphere. Serial
passages were performed by detaching tumor cells with
PBS or trypsinization (0.25% trypsin and 0.02% EDTA
in Ca2+ and Mg2+ free PBS) of conuent monolayers.
The medium was replaced every day. Cell viability was
assessed by trypan blue exclusion test, and the absence
of mycoplasma was conrmed by Hoechst method.
Tumor-induced angiogenesis
Tumor-induced angiogenesis was quantied using an
in vivo bioassay previously described [10]. Briey,
2 105 LM2, LM3 and LMM3 cells were inoculated
i.d. Prior to the inoculation, cells were treated for 1 h
with: (a) COX inhibitors, 10)6 M indomethacin (INDO)
or 10)5 M NS-398; (b) NOS inhibitors, 2 10)3 Nx
monomethyl L -arginine (L -NMMA) or 10)3 M amino-
guanidine (AG); (c) arginase inhibitors, 5 10)2 M
valine (VAL) or 10)4 M Nx-hydroxy-L -arginine
(NOHA). On day 5, animals were sacriced with ether.
The inoculated sites were photographed, and the slides
were projected onto a reticular screen to count the
number of vessels/mm2 skin.
PGE2 radioimmunoassay
The prostanoid levels in the dierent tumor cells were
determined by radioimmunoassay according to the
method previously described [11]. Cells (106/ml MEM)
with or without COX inhibitors, INDO (10)6 M) or NS-
398 (10)5 M), were incubated for 90 min at 37 C in a
Dubno bath with carbogen (5% O2 in 95% CO2). After
incubation, cells were centrifuged for 10 min at 900 rpm,
and supernatants were frozen at )80 C until the RIA
assay was carried out. Samples or standards (100 ll)
were incubated for 30 min with 500 ll rabbit anti-PGE2
antiserum (SIGMA, jj) at 4 C and 5 picograms (pg)
[3H]-PGE2 (specic activity: 154 Ci/mmol) (New Eng-
land Nuclear, jj) were added to each tube. All dilutions
were done in 0.01 M PBS, pH 7.4, containing 0.1%
bovine serum albumin and 0.1% sodium azide. After
incubation, a dextran-coated charcoal suspension was
added to separate the bound from the free fraction.
Supernatants were removed from each tube and a
scintillation solution (Optiphase Hisafe 3 Wallac, jj)
was added to determine the amount of radioactivity
present. Results are expressed in picograms per million
cells (pg/106 cells).
Identication of COX isoforms by Western blot
Cells (5 106) were seeded in 100-mm petri culture
dishes in MEM with 5% FCS. When cells reached 80
90% of conuence, they were deprived of FCS for 24 h.
Cells were rinsed twice with ice-cold PBS and then
scraped into 1 ml lysis buer: 20 mM TrisHCl (pH 7.4),
1 mM EDTA, 10 lg/ml leupeptin, 2 lg/ml aprotinin,
10 lg/ml DTT, 100 lg/ml trypsin inhibitor and 1 mg/ml
benzamidine. Lysates were centrifuged at 5000 rpm for
5 min, and protein content was determined by Lowrys
method [12]. Samples (30 lg) were subjected to 7.5%
sodiumdodecyl sulfatepolyacrylamide minigel electro-
phoresis (SDSPAGE). Standards of known molecular
weight (BIO-RAD) were also seeded. After electropho-
resis, proteins were transferred to a nitrocellulose mem-
brane (BIO-RAD) and washed with distilled water.
Nitrocellulose strips were blocked in 20 mM TrisHCl
buer, 500 mM NaCl, 0.05% Tween 20 (TBST) with 5%
skimmed milk for 1 h at room temperature. Membranes
were incubated with the rst rabbit polyclonal anti-
mouse COX-1 and COX-2 antibodies (Cayman, Chem-
ical Co, jj) for 90 min in TBS at room temperature.
Then, an alkaline phosphatase-conjugated secondary
antibody (anti-rabbit IgG) was added overnight. Bands
were visualized with a mixture of nitro blue tetrazolium
chloride (NBT)/5-bromo-4-chloro-3-indolyl phosphate-
p-toluidine salt (BCIP). Quantication was performed
using a computerized densitometer connected to an
image analyzer (BIO-RAD GS 700).
NO production
NO production was measured as nitrite by the Griess
reaction as was previously described [13]. Tumor cells
(105) were seeded in triplicate in 24-well plates with
500 ll of MEM plus 5% FCS. Supernatants were
replaced by fresh MEM without FCS, and nitrite
Angiogenic pathways in adenocarcinoma cell lines
production was evaluated after 24 h. To inhibit NO
synthesis, L -NMMA (10)4 M) or AG (2 mM) was
added for the rst 30 min. Of the culture supernatants
100 ll were combined with equal volume of Griess
reagent (1% sulphanylamine in 30% acetic acid with
0.1% N-(1-naphtyl) ethylenediamine dihydrochloride in
60% acetic acid). Absorbance was measured at 550 nm
with an ELISA reader. Nitrite concentration was
determined with respect to a standard curve of NaNO2
diluted in MEM. Results were expressed as nanomoles
of nitrite per million cells (nmol/106cells).
Identication of NOS isoforms by Western blot
Tumor cells (107) were plated in 100 mm Petri culture
dishes in MEM with 5% FCS. When cells reached 80
90% of conuence, they were deprived of FCS for 24 h.
After washing the cells with PBS at 4 C, 1 ml of lysis
buer (TrisHCl 50 mM, pH 8.0 containing 10 mM
EDTA, 10 mM EGTA, 100 mM NaCl, 1 mM PMSF,
10 lg/ml aprotinin, 10 lg/ml leupeptin) was added in an
ice-cold bath, and cells were detached with a scraper.
After 1 h, lysates were centrifuged at 10,000 g for
15 min at 4 C. Supernatants were used to detect NOS
expression. Samples were subjected to 7.5% SDS-PAGE,
seeding 40 lg protein in each lane. Proteins were
transferred to nitrocellulose membranes, and after wash-
ing were blocked in 20 mM TrisHCl buer, 500 mM
NaCl, 0.05% Tween 20 (TBST) with 5% skimmed milk
for 1 h at room temperature and subsequently incubated
overnight with rabbit polyclonal anti-mouse NOS1 and
NOS3 antibodies and goat anti-mouse NOS2 antibody
(Santa Cruz Biotechnology Inc., jj) all diluted 1:100 in
TBST. After several rinses with TBST, strips were
incubated at 37 C for 1 h with alkaline phosphatase-
conjugated secondary antibody (anti-rabbit IgG or anti-
goat IgG, respectively), diluted 1:4000 in TBST. Bands
were visualized with a mixture of NBT/BCIP and
quantied by a densitometric analysis.
Arginase activity assay
Arginase activity was determined in cell lysates accord-
ing to a previously described method [14]. Briey, 105
cells treated with or without NOHA (10)4 M) or VAL
(10)4 M), were lysed with 0.5 ml 0.1% Triton X-100,
25 mM TrisHCl containing 5 mM MnCl2, pH 7.4. The
enzyme was activated at 56 C, and 25 ll of the activated
lysates were incubated with 25 ll of 0.5 M arginine, pH
9.7, at 37 C for 60 min. The reaction was stopped in
acid medium. Urea concentration was measured at
540 nm with a microplate reader. Results are expressed
as lmol of urea per hour per 106 cells (lmol/h/106 cells).
Identication of arginase isoforms by Western blot
With ice-cold PBS 5 106 cells were rinsed twice and
then scraped into 300 ll lysis buer: 50 mM TrisHCl
(pH 7.5), 0.1 mM EDTA, 0.1 mM EGTA, 1 lg/ml
3
leupeptin, 1 lg/ml aprotinin and 0.1 mM PMSF. Lysis
was completed by sonication. Samples (25 lg) were
subjected to 10% SDSPAGE as was stated in a
previous method [15]. Nitrocellulose membranes were
incubated overnight with a monoclonal anti-mouse
arginase I antibody (BD Transduction Laboratories,
jj) or with a rabbit anti-arginase II antibody (kindly
gifted by Dr Masataka Mori) [16]. The secondary
antibody anti-mouse or anti-rabbit IgG conjugated with
alkaline phosphatase was added for 1 h at 37 C.
Proteins were visualized using a mixture of NBT/BCIP
and quantied by a densitometric analysis.
Detection of VEGF by Western blot
VEGF production was determined in supernatants and
lysates from 2 107 cells. Supernatants were obtained
after 24 h cell incubation at 37 C in a humidied 5%
CO2 air atmosphere. For lysates, cells were rinsed twice
with ice-cold PBS and then 800 ll lysis buer, 100 mM
NaCl, 10 mM EGTA, 10 mM EDTA, 50 mM TrisHCl
pH 8, 1% Triton X-100, 1 mM PMSF, 10 lg/ml
leupeptin, 10 lg/ml aprotinin, were added. After 1 h in
an ice-cold bath, lysates were centrifuged at 10,000 rpm
for 10 min at 4 C. Samples (25 lg) were subjected to
10% SDSPAGE. Proteins were transferred to nitro-
cellulose membranes and incubated overnight with a
goat polyclonal anti-VEGF antibody (Santa Cruz Bio-
technology). The secondary antibody anti-goat IgG
conjugated with alkaline phosphatase was added for 1 h
at 37 C. Proteins were visualized using a mixture of
NBT/BCIP and quantied by a densitometric analysis.
Drugs
Fresh dilutions of inhibitors were prepared immediately
before use. Unless stated otherwise, all reagents and
drugs were purchased from SigmaAldrich (jj).
Statistical analysis
All experiments were performed at least three times with
comparable results. In all the assays, the signicance was
determined by Students t-test using the STAT PRI-
MER program. Dierences between means were con-
sidered signicant if P < 0.05. Results were expressed
as means SEM.
Results
COX-derived PGE2 is involved in angiogenic response
LMM3, LM3 and LM2 cells exhibit genotypic dier-
ences that distinctively control tumor progression. Here
we show that although all tumor cell lines induce a
positive vascular response (N vessels/mm2 skin),
LMM3 cells were signicantly more angiogenic
(3.96 0.40) than LM3 (2.73 0.48) and LM2
4 L.E. Davel et al.
(2.82 0.30) when injected in equal number (Figures VEGF production in tumor cell lines
13).
LMM3 cells express the highest level of COX-1 and It is well known that VEGF is able to promote vascular
COX-2 isoforms, which directly correlates with the endothelial cell proliferation and to induce a potent
amount of PGE2 and with the strongest neovascular angiogenic response in dierent pathological conditions
response (Figure 1). The participation of PGE2 in the such as cancer. We demonstrate that LMM3, LM3 and
angiogenic activity of LMM3 cells was corroborated LM2 cells, all triggering strong angiogenic responses, are
using COX inhibitors: INDO as well as NS-398, signif-
icantly inhibited both PGE2 liberation and neovascular-
ization. In LM3 cells, both COX isoforms are also
expressed but signicantly lesser than in LMM3. Both
COX inhibitors reduced PGE2 synthesis but only INDO-
inhibited angiogenesis (Figure 1B). These results demon-
strated that mainly COX-1- derived PGE2 could be
involved in LM3-induced angiogenesis. In LM2 cells,
INDO potently decreased neovascularization as it did in
LMM3 and LM3 cells concordantly with its ability to
reduce PGE2 liberation in a signicative manner. The
eect of NS-398 on neovascularization was less evident
although it signicantly reduced PGE2 production (Fig-
ure 1C).

NOS is not always involved in tumor angiogenesis

It has been previously described that NO production by

NOS is involved in tumor angiogenesis. Here we demon-
strated that in LMM3 cells, this process could be mainly
linked to NOS2 expression and activity, as angiogenesis
was blunted by LNMMA and AG, a NOS non-selective
and NOS2-selective inhibitors, respectively (Figure 2A,
left panel). LM3 cells only expressed signicative amounts
of NOS1 isoform, as was demonstrated by immunoblot-
ting assays, and produced higher amounts of NO than
LMM3 (Figures 2A and B, right panel). NOS1 activity
was blunted by the preincubation with LNMMA, but
angiogenic response was NO independent since it was not
modied by either LNMMA or by AG. In LM2 cells, low
expression of NOS3 protein was detected, accompanied
with very low levels of NO (Figure 2C, right panel).
Neither NO production nor angiogenesis was modied by
L -NMMA or AG.

Arginase-derived products are involved in tumor-induced

angiogenesis

Arginine metabolism via NOS is involved in angiogenesis

induced only by LMM3 cells. Since arginine could be also
metabolized via arginase producing urea and ornithine,
we tested the role of this enzyme in neovascularization.
Figure 3A shows that arginase inhibitors, VAL and
NOHA, signicantly reduced urea production as well as
neovessel formation by LMM3 cells. Immunoblotting
assays demonstrated that arginase II is the most abundant
isoform expressed in the three cell lines, being more active
in LMM3 cells measured as higher levels of urea (Fig-
ure 3, right panels). In addition, angiogenic response
induced by LM3 and LM2 cells was less sensitive to the
inhibitory action of VAL and NOHA.
Figure 1. (A) LMM3 cells; (B) LM3 cells; and (C) LM2 cells. Left
panel: in vivo angiogenic response measured as number of vessels per
square millimeter of skin (N vessels/mm2 skin) in the absence or
presence of COX inhibitors (INDO and NS-398). Values are
mean SEM of six experiments. **P < 0.0001 and *P < 0.002.
Right panel: detection of COX-1 and COX-2 expression by Western
blot (one representative experiment of three is shown). PGE2 produc-
tion expressed in picograms per million cells (pg/106 cells) in the
absence (basal) or presence of COX inhibitors was measured by
radioimmunoassay. Values are mean SEM of three experiments.
Angiogenic pathways in adenocarcinoma cell lines
Figure 2. (A) LMM3 cells; (B) LM3 cells; and (C) LM2 cells. Left
panel: in vivo angiogenic response measured as number of vessels per
square millimeter of skin (N vessels/mm2 skin) in the absence or
presence of NOS inhibitors (L -NMMA and AG). Values are
mean SEM of six experiments. *P < 0.0001. Right panel: detection
of NOS1, NOS2 and NOS3 by Western blot (one representative
experiment of three is shown). NO production expressed as nanomoles
million cells (nmol/106 cells) in the absence (basal) or presence of NOS
inhibitors was measured using Griess reagent. Values are mean -
SEM of (n) experiments.
able to synthesize and liberate dierent amounts of VEGF
as is demonstrated by Western blot assays (Figure 4).
Discussion
Angiogenesis has become an important area of scientic
research due to its involvement in various physiological
and pathological processes. Neovascularization is usually
a prerequisite for tumor progression, and the develop-
ment and spread of new capillaries is directed and
regulated by a complex network of mechanisms which
control tumor angiogenesis in a positive or negative
manner. Tumor heterogeneity leads to heterogeneity in
the tumor vasculature and just as there are multiple
phenotypes for any given tumor type, so can there be
multiple phenotypes of the tumor angiogenic process [17].
Our results evidence that murine tumor cells origi-
nating from dierent spontaneous mammary adenocar-
cinomas arising in BALB/c mice, exhibit dierent
5
Figure 3. (A) LMM3 cells; (B) LM3 cells; and (C) LM2 cells. Left
panel: in vivo angiogenic response measured as number of vessels per
square millimeter of skin (N vessels/mm2 skin) in the absence or
presence of arginase inhibitors (NOHA and VAL). Values are
mean SEM of six experiments. **P < 0.0001; *P < 0.002. Right
panel: detection of arginase I and II by Western blot (one represen-
tative experiment of three is shown). Urea production expressed as
micromoles per hour and per million cells (lmol/h/106 cells) in the
absence (basal) or presence of arginase inhibitors was measured by a
calorimetric method. Values are mean SEM of (n) experiments.
mechanisms involved in angiogenesis. These dierences
could be supported by their dierent genotypes yielding
distinct phenotypes. Many authors have documented
that COX-2 is over-expressed in various types of
malignant tumors as gastric, skin and breast cancer
[1820]. They also provided evidence that COX-2-
derived prostaglandins contribute to tumor growth by
inducing newly formed blood vessels that sustain tumor
cell viability and growth. In this way, we tested the
expression and function of COX isoforms, and we
observed that both proteins were dierentially expressed
in three adenocarcinoma cell lines. LMM3 cells express
6 L.E. Davel et al.
Figure 4. Detection of VEGF in supernatants (S) and cell lysates (L)
from LM2, LM3 and LMM3 cells by Western blot with anti-VEGF
antibody and densitometric analysis of the bands. One representative
experiment of three performed.
higher amounts of COX-1 and COX-2 than LM3 and
LM2 cells. The dierence between LM3 and LM2 cells is
that both express similar amounts of COX-1 protein
whereas COX-2 is 10-fold higher in LM2 than in LM3.
We also demonstrated that, in spite of the fact that
INDOdiminished PGE2 synthesis more eectively than
NS-398 in LMM3 cells, both drugs were equally potent
to reduce neoangiogenesis induced by these cells. In
LM2 cells, angiogenic response was more sensitive to
INDO inhibitory action than to NS-398, pointing
mainly to COX-1 participation in this response. It is
important to note that although this isoform is less
expressed, it may contribute to LM2-induced neovas-
cularization. In turn, LM3 cells showed a band corre-
sponding to COX-1 isoform, and neovascularization
was only reduced preventing COX-1-derived PGE2
formation with INDO. These results could be indicating
that it is necessary to reconsider the role of COX-1 in
this step of tumorigenesis. Sales et al. [21] indicated that
overexpression of COX-1, in HeLa cells, was associated
with enhanced expression of angiogenic factors. This
upregulation was abolished by the dual COX enzyme
inhibitor indomethacin indicating that COX-1 up reg-
ulation modulates the expression of factors that may act
in an autocrine/paracrine manner to enhance and
sustain tumorigenesis in neoplastic cells.
Within the paradigm of multi-step tumor progression,
almost all stages seem to be inuenced by NO levels as
well as by distinct NOS isoforms expression. The
concentration of NO may be a very important factor
in determining its function as well as the genetic makeup
of tumor cells [22, 23]. Our results indicate that LMM3
and LM3 cells produce similar amounts of NO, but
immunoblotting assays show dierences in the expres-
sion of NOS isoforms. NOS2 was expressed in three-fold
higher amounts than NOS3 in LMM3 cells, whereas
NOS1 was not detected. On the contrary, LM3 cells
only expressed the neuronal isoform of NOS. Although
almost undetectable levels of NO were produced in LM2
cultures, endothelial NOS protein was detected by
immunoblotting. Dierences in NO participation in
tumor-induced vascularization were also detected, since
this response was blunted by AG and LNMMA only in
LMM3 cells pointing that iNOS-derived NO seems to be
necessary to stimulate tumor vessel growth.
In the last few years, arginine metabolism has been
reviewed, since arginase catabolism has acquired great
importance in cell biology. NOS and arginases can be
co-expressed in some cell types, and it has been
documented that arginine metabolism to urea and
ornithine via arginase is needed for polyamine synthe-
sis and, as a consequence, for cell proliferation [24, 25].
We have previously demonstrated that peritoneal mac-
rophages in LMM3 tumor-bearing mice may positively
modulate tumor growth by providing polyamine pre-
cursors to malignant and/or endothelial cells to stimu-
late angiogenic cascade [26]. Moreover, a screening
performed in dierent human adenocarcinoma cell lines
showed a positive correlation between arginase expres-
sion and cell proliferation [27]. Our present results show
that arginase II is expressed in all tumor cell lines while
arginase I protein is only present in LM3 cells, although
at low levels. In spite of the presence of the mitochon-
drial arginase isoform, which seems to play the major
role in tumor angiogenesis, vessel formation was pre-
vented with VAL and NOHA in all cell types. We
cannot discard the fact that arginase I products exert a
positive action on neovascularization since LM3 cells
co-expressed both isoforms of this enzyme. More
experiments using selective inhibitors of arginase iso-
forms are needed.
The existence of angiogenic factors was initially
postulated on the basis of the strong neovascular
response induced by transplanted tumors. Many mole-
cules have been implicated as positive regulators of
angiogenesis, including broblast growth factors (acidic
and basic), transforming growth factor, interleukin-8
and granulocyte colony-stimulating factor. For over a
decade, the role of VEGF in the regulation of angio-
genesis was the object of intense investigation [28].
Recent evidence indicates that new vessel growth and
maturation are highly complex and coordinated pro-
cesses, requiring the sequential activation by numerous
ligands. In this way, COX-2 expression and its product
PGE2, have been pointed out as promoters of angio-
genesis by modulating the synthesis of dierent factors
including VEGF [29]. It has been described that COX-1
is also over-expressed in murine lung adenocarcinomas
that also liberate high levels of VEGF [30]. This
mechanism could be occurring in LM3 cells in which
COX-1 protein is upregulated in comparison with COX-
2.
We have previously described an NOS-COX crosstalk
in LMM3 cells that regulated migration and angiogenic
response [31]. We postulated that iNOS-derived NO
down-regulated COX activity. The reduction in PGE2
formation could be responsible of lower VEGF levels in
LMM3 cells in comparison to LM2. We must consider
that LM2 cells produce the highest amounts of VEGF
without exhibiting the most potent angiogenic response.
Kroll and Waltenberger [32] have reported that eNOS
can mediate VEGF-A induced angiogenesis. Since
VEGF-A is the isoform classically linked to angiogen-
esis, further experiments to discriminate the relative
Angiogenic pathways in adenocarcinoma cell lines
amount of this member of the VEGF family in our
preparations must be performed.
Conclusion
We can conclude that three adenocarcinoma cell lines
are able to stimulate blood vessel formation by similar
but not equal mechanisms that distinctly involve COX,
NOS and arginase products.
Acknowledgement
A grant from the University of Buenos Aires (UBACYT
M052) supported this work.
References
1. Folkman J. Tumor angiogenesis. Adv Cancer Res 1985; 43: 175
203.
2. Moncada S, Higgs A, Furchgott R. XIV International Union of
Pharmacology Nomenclature in Nitric Oxide Research. Pharmacol
Rev 1997; 40: 13742.
3. Colasanti M, Suzuki H. The dual personality of NO. Trends
Pharmacol Sci 2000; 21: 24952.
4. Wu G, Morris SM. Arginine metabolism: Nitric oxide and beyond.
Biochem J 1998; 337: 117.
5. Prescott SM, Fitzpatrick FA. COX-2 and carcinogenesis. Biochem
Biophys Acta 2000; 1470: M69M78.
6. Salvemini D. Cyclooxygenase: An important transduction system
for the multifaceted roles of nitric oxide. In Rubanyi GM (ed):
Pathophysiology and Clinical Application of Nitric Oxide, jj
California: Harwood Academic Publishers 1999; Vol. 5, 15570.
7. Tsuji M, Kawano S, Tsuji S et al. Cyclooxygenase regulates
angiogenesis induced by colon cancer cells. Cell 1998; 93: 70516.
8. Galli S, Colombo L, Vanzuli S et al. Characterization of a
broblastoid mammary carcinoma cell line (LM2) originated from
a mouse adenocarcinoma. Int J Oncol 2000; 17: 125965.
9. Urtreger A, Laeda VE, Puricelli L et al. Modulation of bronectin
expression and proteolytic activity associated with the invasive and
metastatic phenotype in two murine mammary tumor cell lines. Int
J Oncol 1997; 11: 48996.
10. Monte M, Davel L, Sacerdote de Lustig E. Hydrogen peroxide is
involved in lymphocyte activation mechanisms to induce angio-
genesis. Eur J Cancer 1997; 33: 67682.
11. Granstrom E, Kindhal H. Radioimmunoassay of prostaglandins
and thomboxanes research. In Frolich JC (ed): Advances in
Prostaglandin and Thromboxanes Research. New York: Raven
Press 1978; 119210.
12. Lowry O, Rosebrough N, Randall R et al. Protein measurement
with Folin phenol reagents. J Biol Chem 1971; 193: 2658.
13. Green LC, Wagner DA, Glogowski J et al. Analysis of nitrate
nitrite and (15N) in biological uids. Anal Biochem 1982; 126: 131
8.
14. Modolell M, Corraliza IM, Link F et al. Reciprocal regulation of
the nitric oxide synthase/arginase balance in mouse bone marrow-
7
derived macrophage by TH1 and TH2 cytokines. Eur J Immunol
1995; 25: 11014.
15. Gotoh T, Sonoki A, Nagasaki K et al. Molecular cloning of cDNA
for nonhepatic mitochondrial arginase (arginase II) and compar-
ison of its induction with nitric oxide synthase in a murine
macrophage-like cell line. FEBS Lett 1996; 395: 11922.
16. Gotoh T, Mori M. Arginase II downregulates nitric oxide (NO)
production prevents NO-mediated apoptosis of murine macro-
phage-derived RAW 264.7 cells. J Cell Biol 1999; 144: 42734.
17. Bergers G, Benjamin LE. Tumorigenesis and the angiogenic
switch. Nature Cancer Rev 2003; 3: 40110.
18. Uefuji K, Ichikura T, Mochizuki H. Cyclooxygenase-2 expression
is related to prostaglandin biosynthesis and angiogenesis in human
gastric cancer. Clin Cancer Res 2000; 6: 1358.
19. Higashi Y, Kanekura T, Kanzaki T. Enhanced expression of
cyclooxygenase (COX)-2 in human skin epidermal cancer cell:
Evidence for growth suppression by inhibiting COX-2 expression.
Int J Cancer 2000; 86: 66771.
20. Subbaramaiah K, Telang N, Ramonetti J et al. Transcription of
cyclooxygenase-2 is enhanced in transformed mammary epithelial
cells. Cancer Res 1996; 56: 44249.
21. Sales KJ, Katz AA, Howard B et al. Cyclooxygenase-1 is up-
regulated in cervical carcinomas: Autocrine/paracrine regulation
of cyclooxygenase-2, prostaglandin E receptors and angiogenic
factors by cyclooxygenase-1. Cancer Res 2002; 62: 42432.
22. Lala PK, Orucevic A . Role of nitric oxide in tumor progression:
Lessons from experimental tumor progression. Cancer Met Rev
1998; 17: 91106.
23. Davel L, Eijan AM, Sales ME et al. Reactivity of tumor lymph
nodes and the nitric oxide pathway. Int J Oncol 2002; 20: 59
67.
24. Bruch-Gerharz D, Schnorr O, Suschek C et al. Arginase 1
overexpression in psoriasis: Limitation of inducible nitric oxide
synthase activity as a molecular mechanism for keratinocyte
hyperproliferation. Am J Pathol 2003; 162: 20321.
25. Li H, Meininger CJ, Kelly KA et al. Activities of arginase I and II
are limiting for endothelial cell proliferation. Am J Physiol Reg
Integr Comp Physiol 2002; 282: R64R9.
26. Davel LE, Jasnis MA, de la Torre E et al. Arginine metabolic
pathways involved in the modulation of tumor induced angiogen-
esis by macrophages. FEBS Lett 2002; 532: 21620.
27. Singh R, Pervin S, Karimi A et al. Arginase activity in human
breast cancer cell lines: N(omega)-hydroxy-L -arginine selectively
inhibits cell proliferation and induces apoptosis in MDA-MB-468
cells. Cancer Res 2000; 60: 330512.
28. Ferrara N. VEGF and the quest for tumor angiogenesis factor.
Nature Cancer Rev 2002; 2: 795803.
29. Joo YE, Rew JS, Seo YH et al. Cyclooxygenase-2 overexpression
correlates with vascular endothelial growth factor expression and
tumor angiogenesis in gastric cancer. J Clin Gastroenterol 2003;
37: 2833.
30. Kisley LR, Barrett BS, Bauer AK et al. Genetic ablation of
inducible nitric oxide synthase decreases mouse lung tumorigen-
esis. Cancer Res 2002; 62: 68506.
31. Davel L, DAgostino A, Espanol A et al. Nitric oxide synthase-
cyclooxygenase interactions are involved in tumor cell angiogenesis
and migration. J Biol Reg Homeost Agents 2002; 16: 181
9.
32. Kroll J, Waltenberger J. VEGF-A induces expression of eNOS and
iNOS in endothelial cells via VEGF receptor 2 (KDR). Biochem
Biophys Res Commun 1998; 252: 7436.