Atherosclerosis 144 (1999) 49 – 57

Nitric oxide inhibits proliferation of human endothelial cells via a

mechanism independent of cGMP

R. Heller *, T. Polack, R. Gräbner, U. Till

Center of Vascular Biology and Medicine, Friedrich-Schiller-Uni6ersity of Jena, Nordhäuser Str. 78, 99089 Erfurt, Germany

Received 2 March 1998; accepted 8 October 1998

Abstract

Endothelial cell-derived nitric oxide (NO) has been suggested to inhibit smooth muscle cell proliferation, resulting in the
reduction of intimal hyperplasia during atherogenesis. The present study investigates the role of NO from exogenous and
endogenous sources on the proliferation of human umbilical vein endothelial cells (HUVEC) and human coronary artery
endothelial cells (CAEC). Three different NO-generating compounds [sodium nitroprusside (SNP), S-nitroso-glutathione (GSNO)
and S-nitroso-acetylpenicillamine (SNAP)] were found to inhibit endothelial cell proliferation measured with three independent
methods (cell counting, [3H]thymidine incorporation, DNA histograms) with significant inhibition occurring at concentrations
]100 mM. Growth-inhibiting effects were observed after long-term treatment (18 – 96 h) as well as after short stimulation with NO
donors (10 min with a subsequent NO donor-free culture period of 18 h) and were comparable in culture medium (20% serum,
growth factor supplementation) and serum-deficient medium (1% serum). The NO donor effects were mediated by the release of
NO as they were prevented by NO scavenging. Superoxide dismutase (SOD) was found not to interfere with these effects
suggesting that peroxynitrite formation was unlikely to be involved. 1H-[1,2,4]Oxadiazolo[4,3,-a]quinoxalin-1-one (ODQ), a
specific inhibitor of the soluble guanylate cyclase, was observed not to alter the antiproliferative effects of NO donors although
it completely prevented NO-mediated increase of cyclic guanosine 3%,5%-monophosphate (cGMP), suggesting that the NO-induced
growth inhibition was not mediated by cGMP. Furthermore, inhibition of endogenous NO production by N-nitro-L-arginine
methylester (L-NAME) did not affect endothelial cell growth regardless of using serum plus growth factor supplement, growth
factor supplement alone, or thrombin to stimulate proliferation. We suggest that constitutively synthesized NO may not regulate
endothelial cell proliferation whereas the growth-inhibiting NO effects may occur when an inducible NO synthase associated with
a persistently high NO production is expressed in the atherosclerotic vessel wall. © 1999 Elsevier Science Ireland Ltd. All rights
reserved.

Keywords: Endothelium; Guanylate cyclase; Nitric oxide; Proliferation

1. Introduction key processes in atherogenesis and a decreased NO

formation in dysfunctional endothelial cells is thought
Vascular endothelial cells constitutively synthesize to be an early event in atherogenesis (reviewed in [5]).
NO, a short-lived radical, which exerts vasoprotective Indeed, clinical studies have confirmed an impaired
activities such as smooth muscle relaxation, inhibition endothelium-dependent vasodilatation in coronary ar-
of platelet activation, as well as regulation of endothe- teries of patients with atherosclerosis (reviewed in [6]),
lial cell permeability and adhesivity [1 – 4]. A number of even before morphological changes in coronary an-
in vitro studies suggest that NO interferes with several giograms could be detected [7].
Endothelium-derived NO is suggested to be involved
* Corresponding author. Tel.: +49-361-741-1437; fax: +49-361-
741-1103. in the regulation of smooth muscle cell proliferation
E-mail address: heller@zmkh.ef.uni-jena.de (R. Heller) [8–12]. Both NO-releasing compounds and gene trans-

0021-9150/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved.
PII: S 0 0 2 1 - 9 1 5 0 ( 9 9 ) 0 0 0 4 1 - 6
50 R. Heller et al. / Atherosclerosis 144 (1999) 49–57

fer of endothelial NO synthase (ecNOS) have been (M199), human serum (HS), fetal calf serum (FCS),
shown to inhibit vascular smooth muscle cell growth in collagenase and human serum albumin (HSA) came
vitro [8–10]. Moreover, supplementation of dietary from Boehringer Ingelheim Bioproducts (Heidelberg,
arginine, application of NO donors or transfection of Germany). Endothelial cell basal medium (EBM) and
ecNOS genes into the arterial wall were reported to endothelial cell growth medium (EGM) were purchased
reduce intimal hyperplasia in animal studies in vivo from Clonetics Corporation (San Diego, USA).
[11,12]. The antiproliferative effects of NO have been [methyl-3H]Thymidine ([3H]thymidine), L-[2,3,4,5-
partly attributed to the NO-induced cGMP increase [ H]arginine monohydrochloride ([3H]arginine) and
3

because permeable cGMP analogues [8,9], the specific cGMP [3H] Biotrak radioimmunoassay system were
inhibition of cGMP hydrolysis [13] or a cGMP eleva- from Amersham International (UK). S-nitroso-glu-
tion by natriuretic peptides [14] could mimic NO ef- tathione (GSNO), ODQ and L-NAME came from
fects. So far, however, a clear definition of the role of Alexis Corporation (Läufelfingen, Switzerland). En-
soluble guanylate cyclase in NO-mediated growth inhi- dothelial cell growth supplement (ECGS), sodium ni-
bition has been hampered by the lack of a potent and troprusside (SNP), S-nitroso-acetylpenicillamine
selective inhibitor of this enzyme. (SNAP), 3-isobutyl-1-methylxanthine (IBMX), propid-
Growth-inhibiting effects of NO have also been re- ium iodide, superoxide dismutase (SOD), ribonuclease
ported for fibroblasts, lymphocytes and mesangial cells A, thrombin and other reagents were purchased from
[15 – 17], but the effects described in endothelial cells are Sigma Chemical Co. (Deisenhofen, Germany). Stock
unclear [18–25]. NO-generating compounds have been solutions of SNAP and ODQ were prepared in
shown to inhibit [19,22,23], to stimulate [18,25] or even methanol.
to have no effects on the proliferation of endothelial
cells [20]. Similarily, the effects of inhibition of endoge- 2.2. Cell cultures
nous NO synthesis are controversial with studies re-
porting no [20,21], increased [24] or reduced actions [25] Human umbilical vein endothelial cells (HUVEC)
on endothelial cell proliferation. Some of the variations were obtained by treating human umbilical cord veins
with collagenase and cultured in 75 cm2 plastic flasks in
reported may reflect the source and tissue origin of the
M199 containing 15% FCS, 5% HS and 15 mg/ml
cells used, in addition to variations in cell culture
ECGS. CAEC were obtained from Clonetics Corpora-
conditions [18–25].
tion and grown in EGM. Confluent cultures were de-
The present study was untertaken to investigate the
tached by trypsin–EDTA (0.05/0.02% v/v) and seeded
influence of exogenously added and endogenously pro-
on six-well plates for purposes of cell counting and
duced NO on proliferation of cultured human endothe-
determination of cGMP, on 24-well plates for evalua-
lial cells using different approaches in the same cellular
tion of [3H]thymidine incorporation, on 60-mm diame-
models. We were especially interested to see if NO
ter dishes for citrulline measurements and on 90-mm
could affect the growth of human coronary artery diameter dishes for preparation of DNA histograms.
endothelial cells (CAEC) which has not been investi- Generally, the seeding density was 10 000 cells/cm2.
gated so far. These cells are involved in atherogenesis Experiments were performed in HUVEC of the first or
and might be subjected to proliferation during repair second passage and in CAEC of the fourth to sixth
processes on the one hand and to an impairment of passage.
endogenous NO synthesis on the other hand. More- To evaluate effects on proliferation and DNA syn-
over, CAEC might also be exposed to higher quantities thesis of HUVEC and CAEC, NO donors were added
of NO if during atherogenesis in the arterial wall an to the respective culture medium only 6 h after seeding,
inducible NO synthase (iNOS) is expressed which, un- when the cells had adhered, and coincubated with cells
like ecNOS, exerts full catalytic activity at low basal for the indicated times. Short stimulations of endothe-
levels of cytosolic calcium [26]. We furthermore ex- lial cells with NO donors were performed in M199 or
plored if cGMP-dependent mechanisms are involved in EBM (for HUVEC or CAEC, respectively) containing
NO effects on proliferation by using 1H- 0.25% HSA (medium–HSA). In some experiments cells
[1,2,4]oxadiazolo[4,3,-a]quinoxalin-1-one (ODQ), a new remained in serum-deficient medium (M199 plus 1%
specific inhibitor of the soluble guanylate cyclase [27]. FCS for HUVEC; EBM plus 1% FCS for CAEC) for 2
days to yield mitogenically quiescent cells. Subse-
quently, NO donors, mitogens or L-NAME were added
2. Materials and methods to serum-deficient medium and coincubated with the
cells as described. If any compounds were dissolved in
2.1. Materials organic solvents, the control cells received the same
volume of solvent. The viability of cells determined by
Plastic-ware for cell culture was from Greiner trypan blue exclusion ranged from 94% to 98% under
Labortechnik (Frickenhausen, Germany). Medium 199 the different conditions described.
 R. Heller et al. / Atherosclerosis 144 (1999) 49–57 51

2.3. Cell counting quently, the medium was aspirated and 0.5 ml of 96%
ethanol were added to the culture dishes. When the
After experimental incubations with NO donors or ethanol had evaporated, 0.3 ml of buffer (50 mM Tris,
L-NAME, the cells were detached by trypsin – EDTA 4 mM EDTA, pH 7.5) was added and a rapid freeze–
(0.05/0.02% v/v) and resuspended in a small volume of thawing was performed. The cellular extracts were cen-
M199 or EBM. Four independent aliquots of each trifuged for 5 min at 14 000× g. cGMP was measured
suspension were counted using a Neubauer chamber. in the supernatant fraction by radioimmunoassay fol-
lowing the instructions of the manufacturers. Cell
2.4. [ 3H]Thymidine incorporation monolayers incubated in parallel under the same condi-
tions were trypsinized and counted. Intracellular cGMP
[3H]Thymidine (37 kBq/ml) was directly added to the concentration was expressed in pmol/106 cells.
culture medium with the test compounds and incubated
with endothelial cells for 18 h. Alternatively, the en- 2.7. Measurement of citrulline formation
dothelial cells were stimulated with test compounds for
10 min and then grown in freshly added culture Citrulline synthesis represents a sensitive marker of
medium containing 37 kBq/ml [3H]thymidine for 18 h. NO formation and was measured by a modification of
In experiments with mitogenically quiescent cells 37 a previously described technique [29]. Cell monolayers
kBq/ml [3H]thymidine was added to serum-deficient were stimulated with thrombin in Hepes–HSA (145
medium with NO donors or mitogens and incubated mM NaCl, 5 mM KCl, 1 mM MgSO4, 10 mM Hepes,
with cells for 24 h. Experimental incubations were 10 mM glucose, 1.5 mM CaCl2, pH 7.4, containing
stopped by removing the medium, washing the cells 0.25% HSA) in the presence of 10 mM [3H]arginine
with ice-cold 5% trichloroacetic acid (TCA) and incu- (12.2 Bq/pmol) for 15 min. The reaction was stopped
bating them in 5% TCA on ice for 20 min. After two by washing the cells with cold PBS containing 5 mM
additional washings with cold 96% ethanol, precipitates L-arginine and 4 mM EDTA and by adding 96%
were solubilized with buffer (0.1 M NaOH, 2% ethanol. After evaporation of ethanol, the soluble cellu-
Na2CO3, 1% sodium dodecyl sulfate) and their radioac- lar components were dissolved in 20 mM Hepes
tivity was analyzed by liquid scintillation counting. (sodium salt, pH 5.5) and transferred to 2-ml columns
Data are expressed as cpm [3H]thymidine incorporated of Dowex AG50WX-8 (Na + -form). The formed
per well. [3H]citrulline was eluted with water. The radioactivity
corresponding to the [3H]citrulline content of the eluate
2.5. DNA histograms was measured by liquid scintillation counting. The
amount of proteins in cell monolayers incubated in
After experimental incubations (18 h coincubation of parallel under the same conditions was measured ac-
endothelial cells with NO donors in culture medium or cording to Lowry [30] after dissolving the cells in buffer
10 min stimulation with NO donors followed by 18 h (0.1 M NaOH, 2% Na2CO3, 1% sodium dodecyl sul-
culture in the absence of the compounds) the endothe- fate); the citrulline production was expressed in pmol/
lial cells were detached by trypsin – EDTA (0.05/0.02% mg cell protein.
v/v) and suspended in a small volume of cold phos-
phate-buffered saline (PBS). Subsequently, the cells 2.8. Preparation of red blood cells (RBC)
were permeabilized by adding cold ethanol (70%) in
PBS for 30 min, RNA was decomposed with ribonucle- RBC were prepared by centrifugation (1500×g, 10
ase A (110 ng/ml, 30 min) and DNA was stained with min) of fresh human blood treated previously with
propidium iodide (11 ng/ml, 30 min). DNA fluores- heparin (5 U/ml). The plasma and the white cell layer
cence of 10 000 cells per sample was analyzed in a were removed and the RBC washed three times
FACScan flow cytometer (Becton Dickinson, San Jose, (1500× g, 10 min). To obtain packed RBC, the cells
USA) using CellQuest software. The proportion of cells were centrifuged at 5000×g for 10 min.
included in each fraction of the G0 (resting phase)+ G1
(phase ready for DNA synthesis) and S (DNA synthe- 2.9. Statistical analysis
sis) +G2 (phase ready for mitosis)+M (mitosis) was
calculated from FL-2 histograms [28]. Experiments were performed in quadruplicate (cell
counting, [3H]thymidine incorporation, DNA his-
2.6. Determination of cGMP tograms) or duplicate (cGMP, citrulline). Results are
shown as mean9 S.E.M. of three to five independent
Monolayers of endothelial cells were incubated in experiments. To determine the statistical significance of
M199–HSA containing 0.5 mM IBMX for 30 min and the described results, analysis of variance with Bonfer-
then stimulated with NO donors for 10 min. Subse- roni’s correction for multiple comparisons or Student’s
52 R. Heller et al. / Atherosclerosis 144 (1999) 49–57

t-test for unpaired data were used. All statistical com-

parisons were carried out with original data. For rea-
sons of clarity, however, normalized data are shown in
some cases. A value of P B0.05 was accepted as statis-
tically significant.

3. Results

3.1. Effect of NO donors on endothelial cell

proliferation

The addition of 1 mM SNP, GSNO or SNAP to

HUVEC cultures 6 h after seeding led to a significant
reduction of cell growth after 96 h (68, 84 and 81%
inhibition, respectively) (Fig. 1A). As shown for SNP,
the inhibitory effect proved to be dose-dependent (Fig.
1A). In the presence of SNP (0.1 or 1 mM) the HUVEC
growth curve shifted downward (Fig. 1B). The cell
count increased with time, as in the control group, but
the degree of increase was significantly lower. In HU-
VEC treated with 1 mM SNP for 48 h and then placed

Fig. 2. Effect of SNP, GSNO and SNAP on [3H]thymidine incorpora-

tion in endothelial cells. SNP, GSNO or SNAP (10 mM to 1 mM) and
37 kBq/ml [3H]thymidine were added to culture medium 6 h after
seeding of HUVEC (A) or CAEC (B). [3H]Thymidine incorporation
was measured 18 h later. Alternatively, cells were stimulated with NO
donors for 10 min and then grown in culture medium containing 37
kBq/ml [3H]thymidine for 18 h. Results are shown as percentage of
untreated control cells (incorporation into controls =100%). Mean 9
S.E.M. of five independent experiments are given. *PB 0.05 versus
untreated control cells.

in fresh medium for another 48 h, a similar growth rate

as in control cells was seen (Fig. 1B). SNP (1 mM) was
also shown to reduce the proliferation of CAEC to a
cell number of 3.149 0.13×105 after 96 h treatment
compared to 4.909 0.18× 105 in untreated control cells
(n=3, PB 0.05).

3.2. Effect of NO donors on [ 3H]thymidine

incorporation in endothelial cells

We next investigated if the growth reduction by NO

donors was preceded by an inhibition of DNA synthe-
sis. Fig. 2 shows that the addition of SNP, GSNO or
Fig. 1. Effect of SNP, GSNO and SNAP on endothelial cell prolifer-
ation. (A) SNP (1 mM to 1 mM), GSNO (1 mM) or SNAP (1 mM) SNAP (10 mM to 1 mM, 18 h) to the endothelial cell
were added to HUVEC culture medium 6 h after seeding. Cells were culture medium induced a concentration-dependent re-
detached and counted after 96 h. (B) SNP (0.1 and 1 mM) was added duction of [3H]thymidine incorporation into the en-
to HUVEC culture medium 6 h after seeding and coincubated with dothelial cells. The inhibitions for SNP, GSNO and
cells for the indicated times. In a series of incubations, medium with
SNAP amounted to 73, 83 and 86% (HUVEC) or 61,
1 mM SNP was replaced by fresh culture medium after 48 h. 48, 72
and 96 h after seeding cells were detached and counted. Data are 96 and 89% (CAEC), respectively. NO donor concen-
expressed as mean9 S.E.M., n= 3. *PB 0.05 versus untreated con- trations 5 1 mM were without any effect. Interestingly,
trol cells. a 10-min stimulation with 1 mM GSNO or 1 mM
 R. Heller et al. / Atherosclerosis 144 (1999) 49–57 53

SNAP 6 h after seeding was already sufficient to induce induced by 1 mM SNP, 0.1 mM GSNO or 0.1 mM
a significant reduction of [3H]thymidine incorporation SNAP was almost completely inhibited when cells were
into endothelial cells during the subsequent 18 h growth stimulated in the presence of ODQ, a specific inhibitor
period in the absence of compounds (Fig. 2). Experi- of the soluble guanylate cyclase (1 mM, 30-min preincu-
ments were also carried out in serum-deficient medium bation) (Fig. 3B). Similar results were obtained when
to test if the NO donors exhibited any stimulating the same experiment was performed in culture medium.
effects on endothelial cell growth. However, none of the cGMP increased from 0.969 0.34 pmol/106 cells to
agents used (SNP, GSNO or SNAP, 0.01 mM to 1 mM, 9.939 0.98 pmol/106 cells, 9.3090.08 pmol/106 cells
24 h) was able to induce [3H]thymidine incorporation and 6.069 0.34 pmol/106 cells with 1 mM SNP, 0.1
into HUVEC or CAEC, but similar inhibition rates as mM GSNO and 0.1 mM SNAP, respectively, and
seen with serum-rich culture conditions were found cGMP accumulation was completely prevented by 1
(data not shown). mM ODQ (0.87 9 0.39 pmol/106 cells in controls,
To clarify if NO release and/or peroxynitrite forma- 1.0890.51 pmol/106 cells, 0.99 0.42 pmol/106 cells and
tion mediated growth-inhibitory effects of NO donors, 0.849 0.09 pmol/106 cells after stimulation with the
we performed additional experiments in HUVEC. respective NO donors).
Packed RBC (15 ml), known to scavenge NO, were able
to prevent the effect of SNP, GSNO or SNAP (1 mM, 3.5. Effect of guanylate cyclase inhibition on NO
18 h) on [3H]thymidine incorporation, suggesting that donor-inhibited [ 3H]thymidine incorporation
the inhibition was mediated by NO released from the
compounds (the [3H]thymidine incorporation in % of To examine whether DNA synthesis inhibition was
controls was 39.39 0.88, 4.39 0.58 and 14.391.58 mediated by the NO-induced cGMP increase, effects of
after SNP, GSNO and SNAP, respectively; 99.393.18, 1 mM SNP, 0.1 mM GSNO and 0.1 mM SNAP (18 h)
109.79 3.48 and 105.79 1.76 after incubation with the on [3H]thymidine incorporation were measured in the
respective compounds in the presence of RBC; n =3). absence or presence of ODQ (1 mM, 30-min preincuba-
The formation of peroxynitrite from NO and superox- tion). Fig. 4 shows that NO-releasing compounds inhib-
ide was unlikely to contribute to growth inhibition as ited [3H]thymidine incorporation into HUVEC to the
the presence of SOD (100 U/ml) did not affect the NO same extent, regardless whether ODQ was present or
donor induced reduction of DNA synthesis (the not (inhibition after SNP, GSNO and SNAP, respec-
[3H]thymidine incorporation in % of controls was tively: 59, 58 and 81% without ODQ; 61, 56 and 80%
26.9 9 1.84, 17.390.10 and 12.391.10 after 18 h treat- with ODQ). Similar results were obtained with CAEC
ment with 1 mM SNP, GSNO or SNAP, respectively;
28.2 9 1.01, 16.89 0.35 and 10.69 0.45 after incuba-
tion with the respective compounds in the presence of Table 1
Effect of SNP, GSNO and SNAP on the proportion of HUVEC in
SOD; n=3).
the S+G2+M phase of the cell cyclea

3.3. Effect of NO donors on the distribution of Proportion of cells in the S+G2+M phase
endothelial cells within the cell cycle (%)

18 h coincubation 10 min stimulation

The inhibition of DNA synthesis by NO donors was
confirmed by examining the percentage of cells in the Control 53.6 92.10 49.2 9 3.19
DNA-rich cell cycle phases S+G2 +M. Table 1 shows SNP 100 mM 47.9 93.17
that 18 h coincubation of HUVEC with SNP, GSNO SNP 1000 mM 28.0 92.89b 43.39 3.03
or SNAP (0.1–1 mM) in culture medium caused a Control 52.4 92.30 53.4 9 2.19
reduction in the proportion of cells in the S +G2 + M GSNO 100 mM 46.4 94.50
phases. Similar results were obtained when the cells GSNO 1000 mM 17.9 92.01b 41.3 9 2.07b
were stimulated for 10 min with 1 mM GSNO or Control 51.7 92.29 58.5 9 4.70
SNAP followed by an 18-h culture period in the ab- SNAP 100 mM 33.4 94.66b
sence of compounds. SNAP 1000 mM 13.0 9 2.04b 32.5 94.24b

a
NO donors were coincubated with the cells in culture medium
3.4. Effect of NO donors alone or in combination with between 6 and 24 h after seeding. Alternatively, endothelial cells were
ODQ on intracellular cGMP le6els stimulated with NO donors for 10 min and then grown in culture
medium. Cells were harvested 24 h after seeding and stained with
Incubation of HUVEC with SNP, GSNO or SNAP propidium iodide. DNA fluorescence was determined by flow cytome-
try. The proportion of cells in the S+G2+M phase was calculated
(1 mM to 1 mM, 10 min) in M199 – HSA containing 0.5 from FL-2 histograms. Data are expressed as mean 9S.E.M. of five
mM IBMX led to an up to 13-fold increase of intracel- independent experiments.
lular cGMP levels (Fig. 3A). The cGMP accumulation b
PB0.05 versus untreated control cells.
54 R. Heller et al. / Atherosclerosis 144 (1999) 49–57

Fig. 3. Effect of SNP, GSNO and SNAP alone or in combination with ODQ on intracellular cGMP levels in HUVEC. (A) Cells were stimulated
for 10 min with SNP, GSNO or SNAP with the indicated concentrations in M199/HSA containing 0.5 mM IBMX. cGMP was measured in
cellular extracts by radioimmunoassay. Data are expressed as mean 9S.D., n = 3. *P B 0.05 versus unstimulated control cells. (B) Selected NO
donor concentrations (1 mM SNP, 0.1 mM GSNO, 0.1 mM SNAP) were incubated for 10 min with HUVEC in M199/HSA containing 0.5 mM
IBMX in the absence or presence of ODQ (1 mM, 30 min preincubation). Data are expressed as mean 9 S.D., n = 3, untreated and ODQ-treated
cells were compared. *PB0.05.

coincubated with 1 mM SNP for 18 h (inhibition in %: To investigate if stimulation of endothelial NO pro-

46.59 1.48 without ODQ; 41.991.20 with ODQ). duction by a calcium-increasing agonist would influence
endothelial proliferation, we next performed experi-
3.6. Effect of endogenous NO on endothelial cell ments with thrombin in the presence or absence of
proliferation and [ 3H]thymidine incorporation L-NAME. Thrombin induced the [3H]thymidine incor-
poration (24 h, serum-deficient medium) into quiescent
To examine if NO basally synthesized by endothelial HUVEC in a concentration-dependent way (in % of
cells would influence endothelial cell proliferation, we controls: 114.89 2.60, 124.99 4.03*, 143.49 2.39* af-
performed experiments with L-NAME, an inhibitor of ter 0.1, 0.5 and 1.0 U/ml thrombin, respectively; n=5,
endothelial cell NO synthase activity. It had been *PB 0.05). The presence of L-NAME (1 mM, 30-min
shown in preliminary experiments that the addition of preincubation before the addition of 1 U/ml thrombin),
L-NAME (1 mM, 24 h) to the culture medium com- however, did not alter the mitogenic effect of thrombin
pletely blocked calcium-dependent citrulline synthesis although the thrombin-stimulated citrulline synthesis
in HUVEC subsequently stimulated with 1 mM iono- was completely blocked (Fig. 6).
mycin for 15 min (32.7 93.81 pmol citrulline/mg Additionally, we checked if the prolongation of the
protein in untreated cells, not detectable after L- NO half-life by the addition of SOD (100 U/ml, 18 h)
NAME). HUVEC grown in culture medium in the
presence of 1 mM L-NAME, however, showed the same
growth behaviour as untreated control cells (Fig. 5).
Accordingly, L-NAME (1 mM, 18-h coincubation with
cells in culture medium) did not affect the
[3H]thymidine incorporation in HUVEC (Fig. 5) or
CAEC (incorporation in cpm× 103: 14.1 91.43 in con-
trol cells, 13.691.41 in L-NAME-treated cells, n= 3).
Similar results were obtained in HUVEC and CAEC
grown in serum-deficient medium demonstrating that
L-NAME (1 mM, 24 h) did not affect either the basal
or the ECGS-stimulated DNA synthesis.
3 3
[ H]Thymidine incorporation (in cpm× 10 ) in un-
treated versus L-NAME-treated HUVEC was 5.99
0.26 versus 6.39 0.49 under basal conditions and Fig. 4. Effect of ODQ on NO donor-inhibited [3H]thymidine incorpo-
10.9 9 1.43 versus 11.892.68 after ECGS stimulation ration in HUVEC. HUVEC were preincubated 6 h after seeding with
(30 mg/ml) (n=4). The respective data for CAEC (in 1 mM ODQ for 30 min in culture medium. Subsequently, 1 mM SNP,
0.1 mM GSNO or 0.1 mM SNAP and [3H]thymidine (37 kBq/ml)
cpm× 103) were 5.6 9 0.96 versus 5.3 9 0.49 under were added. [3H]Thymidine incorporation was measured 18 h later.
basal conditions and 10.3 91.33 versus 9.8 9 1.62 after Data are expressed as mean 9 S.E.M., n = 3. *P B0.05 versus unstim-
ECGS stimulation (n=4). ulated control cells.
 R. Heller et al. / Atherosclerosis 144 (1999) 49–57 55

tion of HUVEC and CAEC. This was demonstrated

during cell growth in normal culture medium (20%
serum, ECGS supplementation) with three independent
methods (cell counting, [3H]thymidine incorporation,
DNA histograms) after long-term incubations as well
as after short stimulations with NO donors. The an-
tiproliferative effect of SNP, GSNO and SNAP was due
to NO release as it was completely prevented by NO
scavenging, whereas peroxynitrite formation was most
probably not involved as SOD did not alter the NO
donor effects. Accordingly, cytotoxic effects were un-
likely to be responsible for growth inhibition. Cell
Fig. 5. Effect of L-NAME on proliferation and [3H]thymidine incor- viability was not altered during the different incuba-
poration in HUVEC. Left panel: 1 mM L-NAME was added to
tions described and, moreover, replacement of SNP-
HUVEC culture medium 6 h after seeding and coincubated with cells
for the indicated times. Cells were detached 48, 72 and 96 h after containing medium by fresh culture medium induced
seeding, counted and compared with untreated control cells which cell growth with a similar rate as the growth rate seen
were incubated in parallel. Right panel: 1 mM L-NAME and 37 in control cells. Our results are in agreement with
kBq/ml [3H]thymidine were added to the culture medium 6 h after studies on long-term effects of NO donors on the
seeding of HUVEC. In parallel, incubations without L-NAME were proliferation of bovine aortic endothelial cells, bovine
performed (control). [3H]Thymidine incorporation was measured 18 h
later. Data are expressed as mean 9 S.E.M., n= 3. retinal endothelial cells, HUVEC and canine endothe-
lial cells [19,22,23]. In contrast to our results, Ziche et
would affect [3H]thymidine incorporation, but no dif- al. [18,25] have observed that SNP (1–100 mM, 24 h)
ference between control and SOD-treated cells was stimulated [3H]thymidine incorporation in endothelial
cells from bovine coronary postcapillary venules with a
found (incorporation in cpm×103: 14.6 90.9 versus
maximum increase of 55% at 100 mM. As these studies
14.7 90.9, n=3).
were performed in serum-deficient medium after syn-
chronization of the cell cycle, we repeated our experi-
ments in HUVEC and CAEC under similar conditions.
4. Discussion
We could, however, not demonstrate any DNA synthe-
sis promoting effects of NO donors. On the contrary,
The present study shows that three different NO-gen-
we found comparable NO-induced growth inhibition
erating compounds dose-dependently inhibit prolifera-
rates in serum-deficient and serum-rich culture condi-
tions (data not shown), suggesting that growth-stimu-
lating effects of NO might occur only in cells derived
from the microvasculature.
The implication of an exposure of endothelial cells to
high growth inhibitory concentrations of NO may be
important in vivo. Recent evidence in animal studies
has indicated that hypercholesterolemia and arterial
injury trigger iNOS expression and that endotoxin-in-
duced iNOS expression is augmented in atherosclerotic
arteries [31–33]. While this might compensate for the
loss of endothelial NO [32], it may also contribute to
vascular injury by direct or peroxynitrite-mediated cy-
Fig. 6. Effect of thrombin in the absence or presence of L-NAME on totoxic effects [33]. From our results we can speculate
citrulline formation and [3H]thymidine incorporation. Left panel: that iNOS expression with high NO output might in-
HUVEC were incubated for 30 min in the absence or presence of 1 hibit CAEC proliferation and reendothelialization after
mM L-NAME and stimulated with 1 U/ml thrombin in Hepes/HSA arterial injury. Nitrovasodilator therapy, on the other
buffer containing 1.5 mM calcium and 10 mM [3H]arginine (12.2
Bq/pmol). After 15 min the formed [3H]citrulline was separated by
hand, is not likely to produce growth-inhibitory NO
cation exchange chromatography and its radioactivity was measured concentrations [34–36].
by liquid scintillation counting (n.d. = not detectable). Right panel: The results presented in our study also show that
Quiescent HUVEC in serum-deficient medium were preincubated inhibition of endothelial NO synthesis by L-NAME
with 1 mM L-NAME for 30 min. Subsequently, 1 U/ml thrombin
does not affect proliferation of HUVEC or CAEC. This
and 37 kBq/ml [3H]thymidine were added. In parallel, cells were
incubated with 1 U/ml thrombin and 37 kBq/ml [3H]thymidine in the was demonstrated during cell growth in normal culture
absence of L-NAME . [3H]Thymidine incorporation was measured 18 medium as well as in serum-deficient medium supple-
h later. Data are expressed as mean 9 S.E.M., n= 3. mented with ECGS. Moreover, the mitogenic effect of
56 R. Heller et al. / Atherosclerosis 144 (1999) 49–57

thrombin was not affected by L-NAME treatment al- not to be involved in NO-induced inhibition of HU-
though L-NAME completely blocked thrombin-induced VEC and CAEC proliferation. Our results are in agree-
citrulline production, a sensitive marker of NO forma- ment with a study by Garg et al. who reported that
tion. Thus, our results suggest that a reduction of BALB/C 3T3 fibroblasts could be growth-inhibited by
endothelial NO formation as occurring in atheroscle- NO-generating compounds even though the cells lack
rotic coronary arteries, does probably not affect en- soluble guanylate cyclase activity [16]. Similarily, a
dothelial growth and regeneration of injured spermine–NO complex has been reported to inhibit
endothelial cell layers. Our data are in agreement with [3H]thymidine incorporation into bovine aortic en-
studies in bovine aortic and retinal endothelial cells as dothelial cells without increasing intracellular cGMP
well as in rat microvascular endothelial cells [20 – 22]. In [22]. Clearly, the precise mechanism by which NO
contrast, endogenously synthesized NO has been sug- donors inhibit DNA synthesis and endothelial cell pro-
gested to exhibit growth-inhibiting [24] or growth-pro- liferation remains to be established. An effect on inhibi-
moting effects [25] in bovine aortic endothelial cells or tion of ribonucleotide reductase, the rate-limiting
bovine coronary venular endothelial cells, respectively. enzyme in DNA synthesis, has been suggested and may
The variations might not only be due to differences in show promise [41,42]. NO from activated macrophages
the source of endothelial cells but also depend on or from NO-generating compounds has been reported
specific growth conditions. Vascular endothelial growth to inhibit ribonucleotide reductase by scavenging the
factor, for example, has been suggested to exert mito- tyrosine radical in the active site of the enzyme [42].
genic effects by stimulating the production of NO while Further, bypassing the ribonucleotide reductase step by
basic fibroblast growth factor might act independently providing deoxyribonucleotides has been shown to re-
of NO formation [37]. duce the NO-mediated growth inhibition of tumor cells
The most important target enzyme of endothelial- [41] and of bovine aortic endothelial cells [22].
derived NO is the soluble guanylate cyclase. By binding In summary, we have demonstrated, that NO donors
to the iron in the heme at the active site of the enzyme, in relatively high concentrations inhibited the prolifera-
NO induces its activation and the generation of cGMP tion of HUVEC and CAEC. The application of a
which mediates major NO effects such as smooth mus- specific guanylate cyclase inhibitor showed that growth
cle relaxation and inhibition of platelet activation [1,2]. impairment was independent of NO-mediated intracel-
Yang et al. [19] proposed that NO-induced growth lular cGMP accumulation. We speculate that an im-
inhibition of bovine aortic endothelial cells was also pairment of CAEC proliferation and thus of
mediated by an increase in cGMP. Similar suggestions reendothelialization after injury of the arterial intima
were made in studies with smooth muscle and mesan- could occur when the iNOS isoform with high NO
gial cells [8,9,13,15]. Indeed, the application of perme- output is induced in the atherosclerotic arterial wall.
able cGMP analogues [8,9,15], a specific inhibition of
cGMP hydrolysis [13] or a stimulation of the particu-
late guanylate cyclase with natriuretic peptides [14] References
have been shown to decrease cell proliferation. The
precise mechanism of cGMP in regulating cell prolifera- [1] Furchgott RF, Zawadski JV. The obligatory role of endothelial
cells in the relaxation of arterial smooth muscle by acetylcholine.
tion has not yet been clarified. Recent studies, however,
Nature 1980;288:373 – 6.
have demonstrated that cGMP-elevating agents af- [2] Alheid U, Frölich JC, Förstermann U. Endothelium-derived
fected the mitogen-activated protein kinase pathway by relaxing factor from cultured human endothelial cells inhibits
preventing Ras-dependent activation of Raf-1 and that aggregation of human platelets. Thromb Res 1987;47:561–71.
this effect was mediated by the cGMP-dependent [3] Westendorp RGJ, Draijer R, Meinders AE, van Hinsberg
VWM. Cyclic-GMP-mediated decrease in permeability of human
protein kinase [13]. To determine if cGMP or NO or
umbilical and pulmonary artery endothelial cell monolayers. J
both mediated NO-induced inhibition of cell prolifera- Vasc Res 1994;31:42 – 51.
tion, cGMP production must be inhibited selectively. [4] Heller R, Bussolino F, Ghigo D, Pescarmona GP, Calvino R,
This has been previously attempted with methylene Gasco A, Till U, Bosia A. Activation of endothelial guanylate
blue [19], but this compound has additional effects such cyclase inhibits cellular reactivity. Agents Actions 1994;45:177–
81.
as the inhibition of prostacyclin synthesis [38], the
[5] Bult H. Nitric oxide and atherosclerosis: possible implications
generation of superoxide anions [39] or a direct inhibi- for therapy. Mol Med Today 1996;2:510 – 8.
tion of NOS [40]. Recently, a specific guanylate cyclase [6] Anderson TJ, Gerhard MD, Meredith IT, Charbonneau F,
inhibitor, ODQ, has become available [27] and this has Delagrange D, Creager MA, Selwyn AP, Ganz P. Systemic
been applied in our studies. We demonstrate that a nature of endothelial dysfunction in atherosclerosis. Am J Car-
diol 1995;75:71B – 4B.
specific inhibition of the soluble guanylate cyclase by
[7] Zeiher AM, Drexler H, Wollschläger H, Just HJ. Modulation of
ODQ, which was sufficient to prevent NO-mediated coronary vasomotor tone in humans: progressive endothelial
cGMP increase, did not affect antiproliferative effects dysfunction with different early stages of coronary atherosclero-
of NO-generating compounds. Hence, cGMP appears sis. Circulation 1991;83:391 – 401.
 R. Heller et al. / Atherosclerosis 144 (1999) 49–57 57

[8] Garg UC, Hassid A. Nitric oxide-generating vasodilators and control of growth and apoptosis of endothelial cells. Am J
8-bromo-cyclic guanosine monophosphate inhibit mitogenesis Physiol 1997;272:H760– 8.
and proliferation of cultured rat vascular smooth muscle cells. J [25] Ziche M, Morbidelli L, Masini E, Amerini S, Granger HJ,
Clin Invest 1989;83:1774–7. Maggi CA, Geppetti P, Ledda F. Nitric oxide mediates angio-
[9] Dubey RK. Vasodilator-derived nitric oxide inhibits fetal calf genesis in vivo and endothelial growth and migration in vitro
serum- and angiotensin-II-induced growth of renal arteriolar promoted by substance P. J Clin Invest 1994;94:2036 –44.
smooth muscle cells. J Pharmacol Exp Ther 1994;269:402 – 8. [26] Kröncke K-D, Fehsel K, Kolb-Bachofen V. Inducible nitric
[10] Kullo IJ, Schwartz RS, Pompili VJ, Tsutsui M, Milstien S, oxide synthase and its product nitric oxide, a small molecule
Fitzpatrick LA, Katusic ZS, O’Brien T. Expression and function with complex biological activities. Biol Chem Hoppe-Seyler
of recombinant endothelial NO synthase in coronary artery 1995;376:327 – 43.
smooth muscle cells. Arterioscler Thromb Vasc Biol [27] Garthwaite J, Southam E, Boulton CL, Nielsen EB, Schmidt K,
1997;17:2405 – 12. Mayer B. Potent and selective inhibition of nitric oxide-sensitive
[11] Kojda G, Noack E. Effects of pentaerythrityl-tetranitrate and guanylyl cyclase by 1H-[1,2,4]oxadiazolo[4,3,-a]quinoxalin-1-
isosorbide-5-mononitrate in experimental atherosclerosis. Agents one. Mol Pharmacol 1995;48:184 – 8.
Actions Suppl 1995;45:201–6. [28] Ormerod MG. Analysis of DNA — general methods. In:
[12] von der Leyen HE, Gibbons GH, Morishita R, Lewis NP, Ormerod MG, editor. Flow Cytometry. A Practical Approach.
Zhang L, Nakajima M, Kaneda Y, Cooke JP, Dzau VJ. Gene Oxford: IRL Press, 1994:119 – 35.
therapy inhibiting neointimal vascular lesion: in vivo transfer of [29] Ghigo D, Alessio P, Foco A, Bussolino F, Costamagna C, Heller
endothelial cell nitric oxide synthase gene. Proc Natl Acad Sci R, Garbarino G, Pescarmona GP, Bosia A. Nitric oxide synthe-
USA 1995;92:1137 – 41. sis is impaired in glutathione-depleted human umbilical vein
[13] Yu S-M, Hung LM, Lin CC. cGMP-elevating agents suppress endothelial cells. Am J Physiol 1993;265:C728– 32.
proliferation of vascular smooth muscle cells by inhibiting the [30] Lowry OH, Rosenbrough AL, Farr RJ, Randall RJ. Protein
activation of epidermal growth factor signaling pathway. Circu- measurement with folin phenol reagent. J Biol Chem
lation 1997;95:1269 –77. 1951;193:265 – 75.
[14] Abell TJ, Richards AM, Ikram H, Espinier EA, Yandle T. Atrial [31] Verbeuren TJ, Bonhomme E, Laubie M, Simonet S. Evidence
natriuretic factor inhibits proliferation of vascular smooth mus- for induction of nonendothelial NO synthase in aortas of choles-
cle cells stimulated by platelet-derived growth factor. Biochem terol-fed rabbits. J Cardiovasc Pharmacol 1993;21:841 –5.
Biophys Res Commun 1989;160:1392–6. [32] Yan Z-Q, Yokota T, Zhang W, Hansson GK. Expression of
[15] Garg U, Hassid A. Inhibition of rat mesangial cell mitogenesis inducible nitric oxide synthase inhibits platelet adhesion and
by nitric oxide-generating vasodilators. Am J Physiol restores blood flow in the injured artery. Circ Res 1996;79:38–
1989;257:60F – 6F. 44.
[16] Garg UC, Hassid A. Nitric oxide decreases cytosolic free calcium [33] Sobey CG, Brooks RM, Heistad DD. Evidence that expression
in Balb/c 3T3 fibroblasts by a cyclic GMP-independent mecha- of inducible nitric oxide synthase in response to endotoxin is
nism. Biochem Biophys Res Commun 1990;171:474–9. augmented in atherosclerotic rabbits. Circ Res 1995;77:536–43.
[17] Denham S, Rowland IJ. Inhibition of the reactive proliferation [34] Kreye VAW. Sodium nitroprusside. In: Scriabine A, editor.
of lymphocytes by activated macrophages: the role of nitric Pharmacology of Antihypertensive Drugs. New York: Raven
oxide. Clin Exp Immunol 1992;87:157–62. Press, 1980:373 – 96.
[18] Ziche M, Morbidelli L, Masini E, Granger H, Geppetti P, Ledda [35] Chasseaud LF, Down WH, Grundy RK. Concentrations of the
F. Nitric oxide promotes DNA synthesis and cyclic GMP forma- vasodilator isosorbide dinitrate and its metabolites in the blood
tion in endothelial cells from postcapillary venules. Biochem of human subjects. Eur J Clin Pharmacol 1975;8:157 – 60.
Biophys Res Commun 1993;192:1198–203. [36] Abshagen UW. Pharmacokinetics of isosorbide mononitrate.
[19] Yang W, Ando J, Korenaga R, Toyo-oka T, Kamiya A. Exoge- Am J Cardiol 1992;70:61G – 6G.
nous nitric oxide inhibits proliferation of cultured vascular en- [37] Morbidelli L, Chang C-H, Douglas JG, Granger HJ, Ledda F,
dothelial cells. Biochem Biophys Res Commun 1994;203:1160 – 7. Ziche M. Nitric oxide mediates mitogenic effect of VEGF on
[20] Arnal J-F, Yamin J, Dockery S, Harrison DG. Regulation of coronary venular endothelium. Am J Physiol 1996;270:H411–5.
endothelial nitric oxide synthase mRNA, protein, and activity [38] Okamura T, Yoshida K, Toda N. Suppression by methylene
during cell growth. Am J Physiol 1994;267:C1381–8. blue of prostaglandin I2 synthesis in isolated dog renal arteries.
[21] Papapetropoulos A, Desai KM, Rudic RD, Mayer B, Zhang R, J Pharmacol Exp Ther 1990;254:198 – 203.
Ruiz-Torres M, Garcia-Cardena G, Madri JA, Sessa WC. Nitric [39] Wolin MS, Cherry PD, Rodenburg JM, Messina EJ, Kaley G.
oxide synthase inhibitors attenuate transforming-growth-factor- Methylene blue inhibits vasodilation of skeletal muscle arterioles
b1-stimulated capillary organization in vitro. Am J Pathol to acetylcholine and nitric oxide via the extracellular generation
1997;150:1835– 44. of superoxide anion. J Pharmacol Exp Ther 1990;254:872–6.
[22] Gooch KJ, Dangler CA, Frangos JA. Exogenous, basal, and [40] Mayer B, Brunner F, Schmidt K. Inhibition of nitric oxide
flow-induced nitric oxide production and endothelial cell prolif- synthesis by methylene blue. Biochem Pharmacol 1993;45:367–
eration. J Cell Physiol 1997;171:252–8. 74.
[23] Sarkar R, Webb RC, Stanley JC. Nitric oxide inhibition of [41] Kwon NS, Stuehr DJ, Nathan CF. Inhibition of tumor cell
endothelial cell mitogenesis and proliferation. Surgery ribonucleotide reductase by macrophage-derived nitric oxide. J
1995;118:274 – 9. Exp Med 1991;174:761 – 7.
[24] Lopez-Farre A, Sanchez de Miguel L, Caramelo C, Gomez-Ma- [42] Lepoivre M, Flaman J-M, Henry Y. Early loss of the tyrosyl
cias J, Garcia R, Mosquera JR, de Frutos T, Millas I, Rivas F, radical in ribonucleotide reductase of adenocarcinoma cells pro-
Echezarreta G, Casado S. Role of nitric oxide in autocrine ducing nitric oxide. J Biol Chem 1992;267:22994– 3000.