Early activation of the p42/p44MAPK pathway mediates adenosine-induced nitric oxide production in human endothelial cells: a novel calcium-insensitive mechanism.

Abstract

Adenosine is released from the myocardium, endothelial cells, and skeletal muscle in ischemia and is an important regulator of coronary blood flow. We have already shown that acute (2 min) activation of A2a purinoceptors stimulates NO production in human fetal umbilical vein endothelial cells (1) and now report a key role for p42/p44 mitogen-activated protein kinases (p42/p44MAPK) in the regulation of the l-arginine-nitric oxide (NO) signaling pathway. Expression of mRNA for the A2a-, A2b-, and A3-adenosine receptor subtypes was abundant whereas A1-adenosine receptor mRNA levels were negligible. Activation of A2a purinoceptors by adenosine (10 microM) or the A2a receptor agonist CGS21680 (100 nM) resulted in an increase in l-arginine transport and NO release that was not mediated by changes in intracellular Ca2+, pH, or cAMP. Stimulation of endothelial cells with adenosine was associated with a membrane hyperpolarization and phosphorylation of p42/p44MAPK. l-NAME abolished the adenosine-induced hyperpolarization and stimulation of l-arginine transport whereas sodium nitroprusside activated an outward potassium current. Genistein (10 microM) and PD98059 (10 microM), an inhibitor of MAPK kinase 1/2 (MEK1/2), inhibited adenosine-stimulated l-arginine transport, NO production, and phosphorylation of p42/p44MAPK. We found no evidence for activation of eNOS via the serine/threonine kinase Akt/PKB (protein kinase B) in adenosine-stimulated cells. Our results provide the first evidence that adenosine stimulates the endothelial cell l-arginine-NO pathway in a Ca2+-insensitive manner involving p42/p44MAPK, with release of NO leading to a membrane hyperpolarization and activation of l-arginine transport.

Full text

Page 1

Early activation of the p42/p44MAPKpathway mediates
adenosine-induced nitric oxide production in human
endothelial cells: a novel calcium-insensitive
mechanism
AMANDA W. WYATT, JOERN R. STEINERT, CAROLINE P. D. WHEELER-JONES,*
ANTHONY J. MORGAN, DAVID SUGDEN,†JEREMY D. PEARSON, LUIS SOBREVIA,‡
AND GIOVANNI E. MANN1
Centre for Cardiovascular Biology and Medicine, GKT School of Biomedical Sciences, King’s College
London, Guy’s Campus, London SE1 1UL, UK; *Department of Veterinary Basic Science, Royal
Veterinary College, London NW1 0UT, UK;†Endocrinology and Reproduction Research Group,
GKT School of Biomedical Sciences, King’s College London, Guy’s Campus, London SE1 1UL, UK;
and‡Cellular and Molecular Physiology Laboratory, Faculty of Biological Sciences, University of
Concepcio ´n, Concepcio ´n, Chile
ABSTRACT
dium, endothelial cells, and skeletal muscle in ischemia
and is an important regulator of coronary blood flow.
We have already shown that acute (2 min) activation of
A2apurinoceptors stimulates NO production in human
fetal umbilical vein endothelial cells (1) and now report
a key role for p42/p44 mitogen-activated protein ki-
nases (p42/p44MAPK) in the regulation of the L-argi-
nine-nitric oxide (NO) signaling pathway. Expression
of mRNA for the A2a-, A2b-, and A3-adenosine receptor
subtypes was abundant whereas A1-adenosine receptor
mRNA levels were negligible. Activation of A2apurino-
ceptors by adenosine (10 ?M) or the A2areceptor
agonist CGS21680 (100 nM) resulted in an increase in
L-arginine transport and NO release that was not medi-
ated by changes in intracellular Ca2?, pH, or cAMP.
Stimulation of endothelial cells with adenosine was
associated with a membrane hyperpolarization and
phosphorylation of p42/p44MAPK. L-NAME abolished
the adenosine-induced hyperpolarization and stimula-
tion of L-arginine transport whereas sodium nitroprus-
side activated an outward potassium current. Genistein
(10 ?M) and PD98059 (10 ?M), an inhibitor of MAPK
kinase 1/2 (MEK1/2), inhibited adenosine-stimulated
L-arginine transport, NO production, and phosphoryla-
tion of p42/p44MAPK. We found no evidence for acti-
vation of eNOS via the serine/threonine kinase Akt/
PKB (protein kinase B) in adenosine-stimulated cells.
Our results provide the first evidence that adenosine
stimulates the endothelial cell L-arginine-NO pathway in
a Ca2?-insensitive manner involving p42/p44MAPK, with
release of NO leading to a membrane hyperpolariza-
tion and activation of
A. W., Steinert, J. R., Wheeler-Jones, C. P. D., Morgan,
A. J., Sugden, D., Pearson, J. D., Sobrevia, L., Mann,
G. E. Early activation of the p42/p44MAPKpathway
mediates adenosine-induced nitric oxide production in
Adenosine is released from the myocar-
L-arginine transport.—Wyatt,
human endothelial cells: a novel calcium-insensitive
mechanism. FASEB J. 16, 1584–1594 (2002)
Key Words: mitogen-activated protein kinases ? NO ? L-argi-
nine ? A2apurinoceptors
Adenosine is an endogenously produced purine nu-
cleoside generated from either the actions of 5? nucle-
otidases on AMP or specific hydrolases on S-adenosyl-
homocysteine (2). It is an important regulator of blood
flow and has been shown to decrease coronary resis-
tance under basal and restricted blood flow conditions
(3). In many cell types, adenosine elicits vasodilation
via direct stimulation of G-protein-coupled purinocep-
tors located on vascular smooth muscle cells by modu-
lating intracellular cAMP levels (2). Activation of A1
and A3receptor subtypes leads to inactivation of adeny-
lyl cyclase and a decrease in cAMP levels; activation of
A2aand A2breceptor subtypes enhances cAMP accumu-
lation, resulting in vasorelaxation (2). Although aden-
osine is classically regarded as an endothelium-inde-
pendent vasodilator, accumulating evidence implicates
endothelium-derived nitric oxide (NO) in adenosine-
mediated vasodilation (1, 4–9).
Our previous studies in human umbilical vein endo-
thelial cells (1) and studies by others in porcine carotid
artery endothelial cells (10) have established that acti-
vation of A2apurinoceptors by adenosine stimulates
NO release. Our experiments in umbilical vein endo-
thelial cells demonstrated that adenosine acutely stim-
ulates influx of l-arginine (substrate for eNOS) via a
sodium-independent y?transport system, whose activity
1Correspondence: Centre for Cardiovascular Biology and
Medicine, GKT School of Biomedical Sciences, King’s Col-
lege London, Guy’s Campus, London SE1 1UL, UK. E-mail
giovanni.mann@kcl.ac.uk
15840892-6638/02/0016-1584 © FASEB

Page 2

was enhanced in response to membrane hyperpolariza-
tion (1). Although adenosine has been reported to
increase endothelial cell proliferation most likely via
the activation of p42/p44MAPK(11–13), the signal
transduction pathway(s) mediating adenosine-induced
activation of eNOS have not yet been elucidated.
In endothelial cells, constitutive nitric oxide synthase
(eNOS) is classically activated by agonists that raise
intracellular Ca2?([Ca2?]i) levels such as histamine
and bradykinin. Recent evidence suggests that eNOS
can also be activated in a Ca2?-insensitive manner in
response to fluid shear stress (14, 15) and 17?-estradiol
(16). In response to fluid shear stress, eNOS can be
phosphorylated by the serine/threonine protein kinase
Akt (protein kinase B, PKB) in a phosphatidylinositol
3-kinase (PI3-kinase) -dependent manner, allowing en-
zyme activation at basal levels of Ca2?(17, 18). Activa-
tion of eNOS in response to 17?-estradiol has been
reported to be mediated via p42/p44MAPKin a calcium-
dependent manner (19). As recent evidence implicates
Akt/PKB and p42/p44MAPKin phosphorylation of
eNOS (20), we investigated the role of these kinases in
modulation of the endothelial l-arginine-NO pathway
by adenosine.
Acute stimulation of NO production by adenosine
occurred independent of changes in intracellular Ca2?,
pH, or cAMP but was associated with phosphorylation
of p42/p44MAPK. Inhibition of protein tyrosine kinases
and MAPK kinase (MEK1/2, upstream activator of
p42/44MAPK) abolished A2apurinoceptor-stimulated
increases in l-arginine transport, NO production,
and p42/44MAPKphosphorylation. A preliminary ac-
count of part of this work has been published in
abstract form (21).
MATERIALS AND METHODS
Endothelial cell culture
Endothelial cells were isolated from human umbilical veins
(HUVEC) by collagenase (0.5 mg mL?1) digestion and
cultured in medium 199 containing 10% (v/v) fetal calf
serum (FCS), 10% (v/v) newborn calf serum, 5 mM l-
glutamine, penicillin (100 units mL?1), streptomycin (100
units mL?1), porcine heparin (90 ?g mL?1), and endothelial
cell growth factor (20 ?g mL?1). As described (22), HUVEC
cultures were identified by their typical cobblestone morphol-
ogy and positive staining for von Willebrand factor (data not
shown) (23). All experiments were performed using passage
2 cells.
RT-PCR analysis
Confluent endothelial cells in 24-well microtiter plates were
washed twice with phosphate-buffered saline (PBS, 37°C) and
200 ?L of ice-cold lysis buffer containing 100 mM Tris-HCl,
pH 7.5, 500 mM LiCl, 10 mM EDTA, 1% LiDS, and 5 mM
DTT was added to disrupt the cells. Poly A?mRNA in each
lysate was isolated using magnetic oligo (dT)25beads and
cDNA was synthesized immediately. mRNA was added to oligo
(dT)18(1 ?g) and random 10-MERS (1 ?g); the mixture was
heated (70°C, 5 min) to remove secondary RNA structures
and cooled immediately on ice. DTT (20 mM), dATP, dCTP,
dTTP, and dGTP (0.5 mM), 40 U recombinant ribonuclease
inhibitor (Rnasin), and avian Moloney murine leukemia virus
reverse transcriptase (MMLV-RT) (10 U) were added. Tubes
were incubated for 1 h at 37°C, followed by 15 min at 42°C.
MMLV-RT was inactivated by heating (98°C, 3 min). The
cDNA generated was diluted 1:10 with tRNA (10 ?g/mL) and
stored at ?85°C. The adenosine receptor subtype primers
were A1-adenosine receptor subtype; sense primer: 5?-AAT
TGC TGT GGA CCG CTA CCT C-3?, antisense primer:
5?-CGA CAC CTT CTT GTT GAG CTG-3?; A2a-adenosine
receptor subtype: sense primer 5?-TTG ACC GCT ACA TTG
CCA TCC G-3? antisense primer 5?-GAA GAT CCG CAA ATA
GAC ACC-3?; A2b-adenosine receptor subtype, sense primer:
5?-ACC AAC TAC TTC CTG GTG TCC-3?; antisense primer:
5?-GCA GCT TTC ATT CGT GGT TCC-3?; A3-adenosine
receptor subtype: sense primer: 5?-ATC GCT GTG GAC CGA
TAC TTG-3?; antisense primer: 5?-AAT GCA CCT GTC TCT
TTG GAG 3?. Predicted sizes of PCR products were A1, 349
bp; A2a, 305 bp; A2b, 374 bp; A3, 353 bp. PCR tubes contained
1 ?L of cDNA (1:10 dilution), 100 ?M of each deoxynucleo-
side 5?-triphosphate, 0.5 ?M of appropriate forward and
reverse primer, 1.5 mM MgCl2, 50 mM KCl, 10 mM Tris-HCL
(pH 8.3), 0.5% glycerol, 0.1% triton X-100, and 1 U Taq DNA
polymerase. The thermal cycling conditions for all PCR
reaction were 94°C, 1 min, 60°C, 1 min, and 72°C, 2 min for
35–40 cycles. The PCR products were separated by agarose
gel electrophoresis (1.8% w/v) and bands were visualized by
ethidium bromide (0.5 ?g/mL) staining. Adenosine receptor
subtype PCR products were cut from agarose gels and puri-
fied using a Qiaquick gel extraction kit. The identity of each
PCR product was confirmed by direct sequencing on an ABI
automated sequencer.
Effects of A2apurinoceptor agonists and inhibitors of
p42/p44MAPKand adenylyl cyclase on L-arginine transport
Confluent HUVEC in 96-well microtiter plates were preequili-
brated for 30 min in Krebs buffer containing (mM): 131
NaCl, 5.6 KCl, 25 NaHCO3; 1 NaH2PO4, 2.5 CaCl2, 1 MgCl2,
5 D-glucose, 20 HEPES, pH 7.4 supplemented with l-arginine
(100 ?M), and the phosphodiesterase inhibitor 3-isobutyl-
methylxanhine (IBMX, 1 ?M), then stimulated for 2 min with
adenosine (10 ?M), the A2apurinoceptor agonist 2-p-(2-
carboxyethyl) phenethylamino-5?-N-ethylcarboxamido-aden-
osine (CGS21680, 100 nM) or forskolin (1 ?M). Basal and
agonist-stimulated rates of l-[3H]arginine (100 ?M) trans-
port were measured over the last 30 s of a 2 min incubation
period; as in our previous studies (1), we excluded the
potential influence of adenosine transport by preincubating
cells with the adenosine transport inhibitor nitrobenzylthio-
inosine (NBMPR, 10 ?M, 30 min). In other experiments, cells
were pretreated with the protein tyrosine kinase inhibitor
genistein (10 ?M, 30 min), its less active analog daidzein (10
?M, 30 min), an inhibitor of MEK1/2, PD98059 (2?-amino-
3?-methoxyflavone, 10 ?M, 30 min), or adenylyl cyclase
SQ22536 (9-(tetrahydro-2-furanyl)-9H purin-6-amine, 100
?M, 30 min). In some experiments, cells were preincubated
with the eNOS inhibitor NG-nitro-l-arginine-methyl ester
(l-NAME, 100 ?M, 30 min), calcium-free Krebs supplemented
with ethylene glycol bis (2-aminoethylether)-N,N,N?,N?-tet-
raacetic acid (EGTA, 1 mM), or exposed to 80 mM KCl to
evoke a membrane depolarization.
Single-cell measurement of [Ca2?]iand pHi
Intracellular calcium ([Ca2?]i) and pH were measured using
ratiometric fluorescent dyes. HUVEC were seeded onto glass
1585ACTIVATION OF THE L-ARGININE-NO PATHWAY BY ADENOSINE

Page 3

no. 1 coverslips and cultured for 2–3 days before use. To
measure single-cell [Ca2?]ichanges, cells were loaded with 1
?M Fura-2/AM for 60 min at room temperature in 20% FCS,
80% DMEM, and fluorescence ratios measured using either
imaging (350/380 nm) or spectrophotometric (340/380 nm)
analysis (24, 25). To determine pHi, cells were loaded under
similar conditions but using 2 ?M BCECF/AM. After loading,
cells were maintained in a HEPES-buffered balanced salt
solution (HBSS) of the following composition (mM): 145
NaCl, 5 KCl, 1 MgCl2, 1 CaCl2, 10 HEPES, 5 glucose, 1%
(w/v) bovine serum albumin (BSA), pH 7.4. For experimen-
tal runs, the same solution was used but with a lower BSA
concentration (0.1%). Coverslips were mounted on a ther-
mostatted stage (30 or 37°C) of an inverted epifluorescence
microscope equipped with a 40? objective (N.A. 1.3) and
superfused with HBSS by gravity feed. For imaging, excitation
was via an LEP dual filter wheel system (Ludl, Hawthorne,
NY) equipped with various neutral density and interference
filters: 350 and 380 nm (Fura-2); 440 and 490 nm (BCECF).
Emission measured either ?400 nm (Fura-2) or ?505 nm
(BCECF), which was captured on a 14 bit cooled CCD camera
Hamamatsu C4880–80 controlled by Openlab software (Im-
provision, Coventry UK). An image pair was captured every
1–3 s. [Ca2?]iand pHiare expressed qualitatively as the ratio
of fluorescence at 350/380 nm or 490/440 nm, respectively.
For single-cell photometric measurements of [Ca2?]i, excita-
tion was via a 75W xenon arc lamp and fluorescence was
monitored at ?450 nm. For Fura-2 loaded cells, autofluores-
cence was estimated by addition of 2 mM Mn2?plus 10 ?M
histamine.
Measurement of intracellular cAMP levels
Confluent HUVEC in 24-well plates were incubated for 30
min with warmed (37°C) Krebs solution containing 100 ?M
l-arginine and rolipram (50 ?M, cAMP-specific phosphodi-
esterase inhibitor). cAMP levels were determined under basal
conditions and in cells challenged for 2 min with adenosine
(10 ?M), CGS21680 (100 nM) or forskolin (1 ?M). Cells were
extracted in 65% ethanol and dried before cAMP analysis
using an enzyme immunoassay kit.
Measurement of intracellular cGMP accumulation
As described previously (1, 22), basal and stimulated cGMP
accumulation in HUVEC monolayers was abolished by the
NOS inhibitor l-NAME, and thus cGMP levels were used as
an index of NO production. Cells were preincubated for 30
min in Krebs buffer containing l-arginine (100 ?M) and
IBMX (1 ?M). The preincubation medium was removed and
cells were stimulated with adenosine (10 ?M, 2 min) or the
A2areceptor agonist CGS21680 (100 nM, 2 min). HCl (0.1M)
cell extracts were stored at ?20°C for radioimmunoassay of
cGMP levels (22). Basal and A2apurinoceptor-stimulated
cGMP levels were also measured in cells pretreated with
inhibitors of eNOS (l-NAME, 100 ?M, 30 min), adenylyl
cyclase (SQ22536, 100 ?M, 30 min), tyrosine kinases
(genistein, 10 ?M, 30 min), a less active analog of genistein
(daidzein, 10 ?M, 30 min), MEK1/2 (PD98059, 10 ?M, 30
min), PI3-kinase (wortmannin, 20 nM, 30 min or LY294002,
10 ?M, 30 min). In some experiments, cGMP accumulation
was assayed in HUVEC exposed to nominally Ca2?-free Krebs
buffer or 80 mM KCl.
Whole-cell patch clamp analysis of K?currents
Ionic currents were measured using the whole-cell recording
mode of the patch clamp technique with a RK-400 (Biologi-
cal, France) amplifier (26). Recording pipettes (2–5 M?)
were filled with an intracellular solution containing (mM)
140 KCl, 2 MgCl2, 15 HEPES, and Nystatin (0.2 mg mL?1).
The bath solution contained (mM) 137 NaCl, 5.4 KCl, 2
CaCl2, 15 HEPES, 5 D-glucose, pH ? 7.4. Voltage pulses were
applied in steps of 20 mV (?100–100 mV) for 400 ms and the
holding potential was set to ?60 mV. Voltage pulse genera-
tion, data acquisition, and analysis were performed using
software written by J. Dempster (University of Strathclyde,
Glasgow, UK). All traces shown were leak subtracted. Current
recordings were obtained from HUVEC challenged with
adenosine (10 ?M) after pretreatment in the absence or
presence of l-NAME (100 ?M, 30 min) or challenged with
the NO donor sodium nitroprusside (10 ?M).
Immunoblotting
Confluent passage 2 HUVEC in 60 mm dishes were deprived
of serum for 12–16 h. Monolayers were washed twice with
Krebs buffer (37°C), then preincubated with Krebs buffer
containing 100 ?M l-arginine in the absence or presence of
a protein tyrosine kinase inhibitor, genistein (10 ?M, 30 min)
or the MEK1/2 inhibitors PD98059 (10 ?M, 30 min) and
U0126 (1,4-diamino-2,3-dicyano-1,4-bis[aminophenylthio]
butadiene, 1 ?M, 30 min). The concentrations of the struc-
turally distinct MEK1/2 inhibitors used are consistent with
our previous reports (27, 28) and have no inhibitory effect on
the p38MAPKor JNK pathways (29, 30). Cells were preincu-
bated with a soluble guanylyl cyclase inhibitor 1H-[1,2,4]
oxadiazolo [4,3-a] quinoxalin-1-one (ODQ, 10 ?M, 30 min),
l-NAME (100 ?M, 30 min), ZM241385 (100 nM, 30 min, A2a
purinoceptorantagonist),or
LY249002 (10 ?M, 30 min). HUVEC were subsequently
challenged with adenosine (10 ?M, 2 min) and the reaction
was stopped by washing monolayers with ice-cold PBS con-
taining 200 ?M sodium orthovanadate. Cells were lysed and
extracts were immunoblotted using a polyclonal antibody
raised against dually phosphorylated (threonine 183 and
tyrosine 185) p42/p44MAPK, total p42/p44MAPK, or serine-
phosphorylated Akt/PKB, with protein bands detected by
enhanced chemiluminescence (27). The polyclonal p42/
p44MAPKantibody raised against the dually phosphorylated
proteins provides a good indicator of p42/p44MAPKactivity,
since we previously correlated p42/p44MAPKactivity in
HUVEC using an in gel kinase assay with dual phosphoryla-
tion of p42/p44MAPK(31). To ascertain equal protein load-
ing, membranes were stained with Ponceau red (0.1% w/v in
5% acetic acid). The density of Western blots was analyzed
using ScnImage software (Scion Corporation, Frederick, MD).
thePI3-kinaseinhibitor
Materials
All cell culture reagents, collagenase type II from Clostridium
histolyticum, adenosine, histamine, IBMX, rolipram, wortman-
nin, ODQ, genistein, daidzein, SQ22536, CGS21680, N6-
cyclopentyladenosine (CPA), N-ethylcarboxamidoadenosine
(NECA), and (2-chloro-N6-(3-iodobenzyl)-5?-(N-methylcar-
bamoyl)adenosine (Cl-IB-MECA) were from Sigma (Poole,
Dorset, UK); A2apurinoceptor antagonist 4-(2-[7-amino-2-(2-
furyl)]1,2,4-triazolo[2,3-a] [1,3,5]triazin-5-ylamino]ethyl)phe-
nol (ZM241385) from AstraZeneca Pharmaceuticals (Maccles-
field, Cheshire, UK); PD98059, LY294002, and U0126 from
Calbiochem (Beeston, Nottingham, UK); Fura-2 acetoxy-
methyl ester and BCECF from Molecular Probes (Cambridge
Bioscience, UK). l-[2,3-3H]arginine (36.1 Ci mmol?1) 3?,5?
cyclic GMP-TME[tyrosine-125I], ECL reagents and the cAMP
EIA kit were purchased from Amersham plc (Buckingham-
shire, UK). Antibodies against the dually phosphorylated and
1586 Vol. 16October 2002WYATT ET AL.The FASEB Journal

Page 4

total p42/p44MAPKand all reagents for the RT-PCR were from
Promega (Southampton, UK); the anti-phosphoserine Akt/
PKB antibody and positive control were from New England
Biolabs (Hertfordshire, UK). The magnetic oligo (dT)25
beads (Dynabeads) were from Dynal (Wirral, UK) and the
Qiaquick gel extraction kit from Qiagen (Teddington, Mid-
dlesex, UK).
Statistics
Data are expressed as means ? se, where n indicates the
number of different umbilical vein endothelial cell cultures
with at least 2–6 replicate measurements per experimental
condition in each cell culture. Statistical analyses were carried
out using ordinary ANOVA where applicable and a two-tailed
Student’s t test for unpaired data, with P ? 0.05 considered
statistically significant.
RESULTS
RT-PCR analysis of adenosine receptor subtypes
in human fetal endothelial cells
We have previously reported that A2a- but not A1-
adenosine receptor agonists stimulate the l-argi-
nine-NO pathway in human umbilical vein endothelial
cells (1). To extend our pharmacological findings, we
have used RT-PCR analyses to identify adenosine recep-
tor subtypes in HUVEC. A2a-, A2b-, and A3-adenosine
receptor subtype mRNA was readily detected, whereas
expression of the A1-adenosine receptor subtype was
negligible (Fig. 1). Sequence analysis of the A2a-, A2b-,
and A3-adenosine receptor subtype PCR products
showed a 99–100% homology with known human
sequences (data not shown).
Effects of [Ca2?]oremoval or elevated K?on
L-arginine transport and cGMP production
Adenosine (10 ?M, 2 min) -stimulated l-arginine trans-
port [pmol (?g protein)?1
3.7?0.4, n?4, P?0.05] was significantly inhibited by
the NOS inhibitor l-NAME (2?0.9, n?4, P?0.05).
Adenosine-stimulated increases in cGMP levels [pmol
(106cells)?15 min?1: 2.2?0.2 vs. 4.8?0.1, n?4,
P?0.01] were mimicked by an A1/2purinoceptor ago-
nist NECA [100 nM, 2 min: 3.3?0.2 pmol (106cells)?1
5 min?1, P ? 0.05], unaffected by an A3agonist
Cl-IB-MECA [100 nM, 2 min: 2.5?0.2 pmol (106
cells)?1
5min?1], and
(2.0?0.06, n?4). As in our earlier study (1), pretreat-
ment of HUVEC with a selective A2apurinoceptor
antagonist (ZM241385) prevented both CGS21680
(data not shown) and NECA [2.4?0.04 pmol (106
cells)?15 min?1]-induced cGMP accumulation. Prein-
cubation of HUVEC in nominally Ca2?-free buffer had
no significant effect on basal, adenosine- or CGS21680-
stimulated rates of l-arginine transport or cGMP pro-
duction (Table 1). Depolarization of cells with 80 mM
K?reduced stimulated rates of l-arginine transport but
had no significant effect on cGMP levels (Table 1).
l-arginine transport and cGMP production were also
unaffected by NBMPR (data not shown), an inhibitor of
the equilibrative es nucleoside transporter (32, 33).
min?1: 1.45?0.4 vs.
abolished by
l-NAME
Effects of adenosine receptor agonists and histamine
on [Ca2?]iand pHi
Given that eNOS can be stimulated directly by Ca2?
(34), it was pertinent to examine the effects of adeno-
sine receptor activation on potential changes in
[Ca2?]i. Despite a significant stimulation of NO pro-
duction (Table 1), adenosine (10 ?M) and CGS21680
(100 nM) failed to stimulate a detectable increase in
[Ca2?]i, even though histamine evoked a characteristic
biphasic response in the same cells (Fig. 2A and inset).
Similar data were obtained in cells challenged initially
with histamine, followed by acute exposure to adeno-
sine or CGS21680 (data not shown). Careful examina-
tion of the 350/380 nm ratio images did not reveal any
highly localized [Ca2?]iincreases in response to aden-
osine that might have escaped detection when measur-
ing global cytosolic [Ca2?]i. Nevertheless, subcellular
Ca2?wave initiation loci were readily detected during
the early phase of a histamine-stimulated Ca2?spike.
Acute exposure of HUVEC to specific A1(CPA, 100
nM), A1/2(NECA, 100 nM), or A3(Cl-IB-MECA, 100
nM) purinoceptor agonists failed to elevate cytosolic
[Ca2?]i(Fig. 2B–D).
In addition to [Ca2?]i, changes in pHihave been
suggested to modulate eNOS activity, i.e., alkalinization
of the cytosol promotes NO release in some endothelial
cells (35). Although histamine (10 ?M) evoked a
transient increase in the 490/440 nm ratio (data not
shown) reflecting an increase in pHi, adenosine (10
?M) caused negligible changes in pHi(490/440 nm ?
Figure 1. RT-PCR analysis of adenosine receptor subtypes in
human umbilical vein endothelial cells. HUVEC mRNA was
isolated as described in Materials and Methods. cDNA was
generated from the mRNA and amplified by PCR using
specific adenosine receptor primers, with PCR products sep-
arated on a 1.8% agarose gel and visualized by ethidium
bromide staining. Images are representative of 3 different cell
cultures. C ? control (HUVEC mRNA not subjected to
reverse transcription); P ? plasmid containing the corre-
sponding adenosine receptor sequence; HUVEC ? cDNA.
1587 ACTIVATION OF THE L-ARGININE-NO PATHWAY BY ADENOSINE

Page 5

ratio 0.06?0.004, n?98 cells). As expected, an ammo-
nium pulse (20 mM) caused a profound alkalinization
(490/440 nm ? ratio 0.57?0.01, n?98 cells, P?0.001),
followed by an acid rebound on ammonium with-
drawal.
Effects of adenosine, CGS21680, and forskolin
on intracellular cAMP levels
As activation of A2apurinoceptors has been shown to
increase intracellular cAMP levels in endothelial cells
(36), we investigated whether adenosine (10 ?M, 2
min) or CGS21680 (100 nM, 2 min) increased cAMP
levels in HUVEC. Although activation of adenylyl cy-
clase by forskolin (1 ?M, 2 min) increased cAMP
accumulation (% of control: 231?15, n?4, P?0.05),
treatment of cells with either adenosine or CGS21680
had no effect on intracellular cAMP levels (% of
control: 113?11, n?4 and 117?19, n?4, respectively).
Basal cAMP levels ranged between 7.5 and 14 pmol (106
cells)?15 min?1. Forskolin increased basal rates of
l-arginine transport without altering intracellular
cGMP levels (Fig. 3A). Pretreatment of endothelial
cells with the adenylyl cyclase inhibitor SQ22536 had
no effect on either adenosine-induced increases in
l-arginine transport or cGMP accumulation (Fig. 3B).
Effects of PI3-kinase inhibitors on adenosine-induced
stimulation of the L-arginine-NO pathway
As eNOS has been reported to be regulated in a
Ca2?-independent manner via PI3-kinase/Akt-medi-
ated phosphorylation (17, 18), we investigated whether
adenosine-stimulated NO production required activa-
tion of PI3-kinase and Akt/PKB. Inhibition of PI3-
kinase with wortmannin or LY294002 had no effect on
adenosine-mediated increases in cGMP accumulation
(Fig. 4A) or phosphorylation of Akt/PKB (Fig. 4B).
Figure 2. Activation of adeno-
sine receptors does not elevate
[Ca2?]i
in human umbilical
vein endothelial cells. HUVEC
were maintained in Ca2?-con-
taining solution and basal and
agonist-stimulated peak and pla-
teau [Ca2?]ilevels were moni-
tored in single cells using imag-
ing (350/380 nm fluorescence
ratio, A) or spectrophotometric
analysis(340/380
B–D). A) [Ca2?]iin cells chal-
lenged with adenosine (10 ?M)
and histamine (10 ?M); inset:
cells challenged with the A2a
purinoceptor agonist CGS21680
(100 nM) and histamine (10
?M). Images are pseudo-col-
ored to show ratio and are mod-
ulated by the brightness of the
sum of the 350 and 380 nm
images with a gamma correction
of 1.5 and filtered with a 3 ? 3
median filter. B–D) Cells challenged initially with either an A1(100 nM CPA), A1/2(100 nM NECA), or A3(100 nM Cl-IB-MECA)
purinoceptor agonist, then histamine (10 ?M). Data are representative of similar experiments in 3 different cell cultures.
nm ratio,
TABLE 1. Effect of extracellular calcium removal and membrane depolarization on adenosine- and CGS21680-induced L-arginine
transport and cGMP productiona
l-arginine transport (pmol (?g protein)?1min?1)
Intracellular cGMP (pmol (106cells)?15 min?1)
Normal
?[Ca2?]o
80 mM K?
Normal
?[Ca2?]o
80 mM K?
Control
Adenosine
CGS21680
1.8 ? 0.1
2.9 ? 0.2*
3.2 ? 0.1*
2.1 ? 0.1
3.1 ? 0.3*
3.4 ? 0.1*
1.7 ? 0.2
1.5 ? 0.04
2.1 ? 0.7
3.4 ? 0.4
6.1 ? 0.6**
6.4 ? 1.1**
2.5 ? 0.5
5.2 ? 0.6*
5.5 ? 0.6*
2.5 ? 0.4
4.5 ? 0.6*
5.5 ? 0.5*
aConfluent HUVEC were washed twice with Krebs (37°C) and incubated in Ca2?free Krebs buffer containing 100 ?M l-arginine and EGTA
(1 mM) or elevated KCl (80 mM) before stimulation with adenosine (10 ?M, 2 min) or CGS21680 (100 nM, 2 min). Transport of 100 ?M
l-[3H]arginine was determined for the last 30 s of a 2 min incubation period and intracellular cGMP levels were determined in HCl cell extracts
by radioimmunoassay. Values denote the means ? se of triplicate measurements in each of 4 different cell cultures; * P ? 0.05, ** P ? 0.01
vs. respective control.
1588 Vol. 16October 2002 WYATT ET AL.The FASEB Journal

Page 6

Effect of genistein, PD98059, and U0126 on
adenosine-induced stimulation of L-arginine transport,
cGMP accumulation, and activation of p42/p44MAPK
Preincubation of HUVEC with genistein, a broad spec-
trum tyrosine kinase inhibitor, or the MEK1/2 inhibi-
tor PD98059 had no effect on basal rates of l-arginine
transport or cGMP accumulation (Table 2). In these
same experiments, adenosine (10 ?M, 2 min) evoked a
twofold increase in l-arginine transport and cGMP
accumulation, which was abolished in cells pretreated
with either genistein (10 ?M) or PD98059 (10 ?M) but
unaffected by daidzein (10 ?M), a less active analog of
genistein.
In parallel experiments, we found that adenosine (10
?M) caused a time-dependent (15 s–5 min) increase in
p42/p44MAPKphosphorylation (Fig. 5A), which was
attenuated by the selective A2areceptor antagonist
ZM241385 (Fig. 5B). The increase in p42/p44MAPK
phosphorylation in response to adenosine was not the
result of enhanced protein levels of p42/p44MAPK, since
total p42/p44MAPKprotein levels were unchanged (Fig.
5A). Our ability to detect basal phosphorylation of
p42/p44MAPKis consistent with our previous studies
with HUVEC in which we suggested that removal of
media and addition of agonist could result in shear
stress and subsequent phosphorylation of p42/p44MAPK
(27, 37). As shown in Fig. 6A–C, adenosine-induced
increases in p42/p44MAPKphosphorylation were abol-
ished by pretreatment of cells with genistein (10 ?M),
PD98059 (10 ?M), and U0126 (1 ?M). The inhibitory
effects of PD98059 and U0126 on p42/p44MAPKactiva-
tion and/or cGMP accumulation could not be attrib-
uted to actions on cyclooxygenases (38), since indo-
methacin (10 ?M) had no effect on adenosine-
stimulatedcGMPaccumulation
phosphorylation (data not shown). Treatment of cells
with LY294002 had no effect on adenosine-induced
orp42/p44MAPK
Figure 3. Role of adenylyl cyclase in adenosine and
CGS21680-stimulated l-arginine transport and intracellular
cGMP accumulation. A) Basal and agonist-stimulated l-argi-
nine transport and cGMP production were determined in
HUVEC challenged with adenosine (10 ?M, 2 min),
CGS21680 (100 nM, 2 min), or forskolin (1 ?M, 2 min).
l-arginine transport was measured over the final 30 s of a 2
min incubation period and cGMP production in HCl extracts
was determined by radioimmunoassay. B) HUVEC monolay-
ers were incubated with adenosine (10 ?M, 2 min) in the
absence or presence of an adenylyl cyclase inhibitor SQ22536
(100 ?M, 30 min) and l-arginine transport and cGMP
accumulation were determined. Basal l-arginine transport
and cGMP levels ranged between 1.3–2.3 pmol (?g pro-
tein)?1min?1and 8.1–9.8 pmol (106)?15 min?1, respec-
tively. Values denote the means ? se of triplicate measure-
ments in 4 different cell cultures, *P ? 0.05 vs. control.
Figure 4. Lack of involvement of PI3-kinase and Akt/PKB in
adenosine-stimulated cGMP production. A) HUVEC were
preincubated with the PI3-kinase inhibitors LY294002 (10
?M, 30 min) or wortmannin (20 nM, 30 min) before acute
exposure to adenosine (10 ?M, 2 min). Intracellular cGMP
levels were determined by radioimmunoassay. Values denote
the means ? se of triplicate measurements in 4 different cell
cultures. Basal cGMP levels ranged between 1 and 2.5 pmol
(106)?15 min?1. B) quiescent HUVEC were stimulated with
adenosine (10 ?M, 2 min); lysates were separated by SDS-
PAGE, transferred to membranes, and probed with an anti-
phosphoserine Akt/PKB antibody. A commercially available
positive control for phosphorylated Akt/PKB obtained from
NIH/3T3 cells treated with platelet-derived growth factor
(PDGF, 100 ?g mL?1, 10 min) was also analyzed. Blot is
representative of similar experiments in 5 different cell
cultures.
1589ACTIVATION OF THE L-ARGININE-NO PATHWAY BY ADENOSINE

Page 7

p42/p44MAPKphosphorylation, confirming that adeno-
sine does not signal via PI3-kinases in HUVEC (Fig. 7A).
Effect of eNOS inhibition on adenosine-induced
p42/p44MAPKphosphorylation
Although adenosine-induced stimulation of l-arginine
transport and cGMP accumulation was prevented by
l-NAME, inhibition of eNOS had no effect on aden-
osine-stimulated p42/p44MAPKphosphorylation (Fig.
7B). Moreover, pretreatment of cells with ODQ, an
inhibitor of soluble guanylyl cyclase, had only a mar-
ginal effect on p42/p44MAPKphosphorylation (Fig. 7C)
but, as expected, abolished adenosine-stimulated cGMP
accumulation (data not shown).
Effects of adenosine on whole-cell potassium currents
As adenosine-stimulated l-arginine transport is sensi-
tive to changes in membrane potential (1), we deter-
mined whether adenosine acutely modulates mem-
brane currents in HUVEC using the whole-cell patch
clamp technique. Adenosine (10 ?M) activated an
outward K?current (Fig. 8A), which was inhibited by
tetraethylammonium (10 mM; data not shown) and
l-NAME (Fig. 8B). Activation of this K?current by
adenosine was mimicked by the NO donor sodium
nitroprusside (Fig. 8C), suggesting that a NO-mediated
membrane hyperpolarization may account for the stim-
ulation of l-arginine transport in HUVEC challenged
with adenosine.
DISCUSSION
The present study provides the first evidence that acute
A2a purinoceptor activation stimulates the l-argi-
nine-NO pathway in fetal endothelial cells via Ca2?-,
pH-, and cAMP-independent mechanisms involving
phosphorylation of p42/p44MAPK. Release of NO (or a
downstream mediator) leads to a membrane hyperpo-
larization and stimulation of l-arginine influx, which in
HUVEC is mediated predominantly via the Na?-inde-
pendent cationic amino acid transport system y?desig-
nated CAT-1 (1, 39). Stimulation of eNOS activity in
response to A2apurinoceptor activation was not medi-
ated via PI3 kinase and Akt/PKB, as previously docu-
mented for shear stress and estradiol-mediated NO
production (17, 18, 40).
We have previously shown that adenosine-induced
activation of l-arginine transport and NO production
TABLE 2. Involvement of protein tyrosine kinases and
p42/p44MAPKin adenosine-induced activation
of the L-arginine–NO pathwaya
Condition
l-arginine
(pmol
(?g protein)?1
min1)
Intracellular cGMP
(pmol (106cells)?1
5min?1)
Control
Genistein (10 ?M)
Daidzein (10 ?M)
PD98059 (10 ?M)
Adenosine (10 ?M)
? Genistein (10 ?M)
? Daidzein (10 ?M)
? PD98059 (10 ?M)
2.2 ? 0.3
2.4 ? 0.6
2
? 0.4
2.6 ? 0.5
4.1 ? 0.2†
2.7 ? 0.4*
4.2 ? 0.2†
2.2 ? 0.3*
1.8 ? 0.3
2.4 ? 0.3
2.3 ? 0.2
2.4 ? 0.4
4.8 ? 0.06†
2
? 0.4*
5.1 ? 1.8†
2.2 ? 0.2*
aConfluent HUVEC were preincubated in the absence or pres-
ence of genistein (10 ?M, 30 min), daidzein (10 ?M, 30 min), or
PD98059 (10 ?M, 30 min) and stimulated with adenosine (10 ?M, 2
min). l-arginine transport was measured over the final 30 s of the
incubation period and cGMP levels were determined in HCl cell
extracts by radioimmunoassay. Data denote the means ? se of
duplicate measurements in each of 8 different cell cultures; † P ?
0.01 compared to control, * P ? 0.01 compared to adenosine alone.
Figure 5. Acute effects of adenosine on p42/p44MAPKphos-
phorylation is mediated by A2apurinoceptor activation. A)
Quiescent HUVEC were stimulated with adenosine (10 ?M)
for 15 s, 30 s, 1 min, 2 min, and 5 min and cell lysates were
separated by SDS-PAGE, transferred to membranes, and
probed using the p42/p44MAPKantibody against the dually
phosphorylated isoforms and an antibody against total p42/
p44MAPKprotein. Blot is representative of experiments in 3
different cell cultures. B) Quiescent HUVEC were stimulated
with adenosine (10 ?M, 2 min) after pretreatment of cells
with Krebs or ZM241385 (A2aantagonist, 100 nM, 30 min).
Cell lysates were immunoblotted for dually phosphorylated
p42/p44MAPKand equal protein loading was assessed by
Ponceau red staining (not shown). Analysis of the fold
changes in band density over control is illustrated below the
blot. Data denote the means ? se of density measurements
from 4 different cell cultures, *P ? 0.05 compared to control.
1590 Vol. 16 October 2002WYATT ET AL. The FASEB Journal

Page 8

in HUVEC is mediated via A2apurinoceptors (1) and
have now established that this cell type possesses mRNA
for the A2a-, A2b-, and A3purinoceptor subtypes with
little or no detectable mRNA for the A1purinoceptor.
Our findings are supported by a recent study with
HUVEC reporting that mRNA levels for the A2apuri-
noceptor are 10-fold greater than the A2bpurinoceptor
(41). In agreement with our findings, the latter study
failed to detect mRNA for the A1purinoceptor. More-
over, A2a- and A2bpurinoceptors have been reported in
porcine and human coronary artery endothelial cells by
RT-PCR and Western blot analysis, establishing a correla-
tion between mRNA levels and protein expression (42).
In agreement with the present study and an earlier
report by Sobrevia et al. (1), a primary role for A2a
purinoceptors in adenosine-stimulated NO synthesis
has been documented in human iliac and porcine
carotid artery endothelial cells (10), guinea pig coro-
nary vasculature (4), hamster aorta (43), and porcine
coronary artery endothelial cells (44). Moreover, acti-
vation of A1purinoceptors markedly attenuates aden-
osine-stimulated NO production in iliac and carotid
artery endothelial cells (10). The only report implicat-
ing A1purinoceptors in adenosine-mediated NO pro-
duction in HUVEC used much higher adenosine con-
centrations (5 mM) (45) than the present study (10
?M), which may have led to activation of different
signaling pathways.
Classically, agonists that increase NO production in
endothelial cells do so via an elevation in [Ca2?]iand,
in the case of bradykinin, an intracellular alkalinization
Figure 6. Effect of genistein, PD98059, and U0126 on aden-
osine-induced activation of p42/p44MAPK. Quiescent endo-
thelial cells were stimulated with adenosine (Ado, 10 ?M, 2
min) in the absence or presence of genistein (10 ?M, 30 min;
A), PD98059 (10 ?M, 30 min; B), and U0126 (1 ?M, 30 min;
C). Cell lysates were analyzed by SDS-PAGE and blots were
probed using a p42/p44MAPKantibody raised against the
dually phosphorylated isoforms. Equal protein loading was
assessed by Ponceau red staining (data not shown). Blots in
panels A, B, and C are representative of experiments in 4, 6,
and 3 different cell cultures, respectively, and the fold
changes in band density over control for blots in panels A and
B are summarized below the respective blots. Data denote the
means ? se of density measurements from 4 and 6 different
cell cultures, *P ? 0.05 compared to control.
Figure 7. Characterization of adenosine-induced activation of
p42/p44MAPKphosphorylation. Quiescent cells were preincu-
bated with either a PI3-kinase inhibitor LY294002 (10 ?M, 30
min; A), l-NAME (100 ?M, 30 min; B), or ODQ (a soluble
guanylyl cyclase inhibitor, 10 ?M, 30 min; C), then chal-
lenged with adenosine (10 ?M, 2 min). Immunoblots were
probed using the p42/p44MAPKantibody raised against the
dually phosphorylated isoforms. Equal protein loading was
assessed by Ponceau red staining (not shown). Blots are
representative of experiments in 4 different cell cultures.
1591ACTIVATION OF THE L-ARGININE-NO PATHWAY BY ADENOSINE

Page 9

(35). However, adenosine- and CGS21680-mediated
NO synthesis in HUVEC occurred without measurable
changes in cytosolic Ca2?or pH, and chelation of
extracellular Ca2?had no effect on A2apurinoceptor-
stimulated l-arginine transport or NO production.
These findings agree with recent evidence that shear
stress (46) and 17?-estradiol-induced NO accumulation
in HUVEC occurs via a Ca2?-insensitive mechanism
(16). Recent reports have further established that
eNOS can be serine phosphorylated in response to
shear stress via Akt/PKB, which itself is regulated by
PI3-kinase (17, 18). The involvement of Akt/PKB and
PI3-kinase in A2apurinoceptor-mediated responses in
HUVEC seems unlikely, since we found that cGMP
accumulation was unaffected by the PI3-kinase inhibi-
tors LY294002 or wortmannin and Akt/PKB was not
serine phosphorylated in response to adenosine.
Activation of A2purinoceptors is known to increase
cAMP levels in vascular smooth muscle and endothelial
cells (2), and elevated cAMP levels have been reported
to stimulate endothelium-derived NO production (47).
Inthepresent study,however,
CGS21680-stimulated NO release was not accompanied
by an increase in intracellular cAMP levels. Moreover,
inhibition of adenylyl cyclase with SQ22536 had no
effect on A2apurinoceptor-stimulated l-arginine trans-
port or cGMP accumulation (see Fig. 3B). Our findings
agree with previous reports showing that cAMP levels in
HUVEC and coronary arterioles are not altered by
adenosine (36, 48). Direct activation of adenylyl cyclase
with forskolin increased cAMP levels and l-arginine
transport but had no effect on cGMP production (see
Fig. 3A). As cAMP activates large conductance K?
channels in endothelial cells (49), it is possible that
membrane hyperpolarization might explain the forsko-
lin-mediated increase in l-arginine transport. Forskolin
has been reported to potentiate agonist (bradykinin
and histamine) –stimulated but not basal cGMP accu-
mulation in endothelial cells (50), consistent with our
observation that forskolin did not alter basal cGMP
levels.
A role for protein tyrosine kinases in shear stress and
17?-estradiol-mediated activation of eNOS has been
reported (14, 51). In endothelial cells, acute stimula-
tion of NO production in response to estradiol can be
inhibited by herbimycin (16) and the MEK1/2 inhibi-
tor PD98059 (19), suggesting that protein tyrosine
kinases and p42/p44MAPKare involved in the signaling
pathway regulating NO production in endothelial cells.
Our study of HUVEC provides further evidence that
adenosine-stimulated activation of p42/p44MAPKis pre-
vented by genistein and the structurally distinct
MEK1/2 inhibitors PD98059 and U0126 (Fig. 6). Aden-
osine-stimulated phosphorylation of p42/p44MAPKwas
prevented by the selective A2a-adenosine receptor an-
tagonist ZM241385 but unaffected by nitrobenzylthio-
inosine, an inhibitor of the es nucleoside transporter
(data not shown), confirming that phosphorylation of
p42/p44MAPKis a result of A2apurinoceptor activation
rather than a consequence of adenosine transport. Our
findings are consistent with an earlier report that
stimulation of A2apurinoceptors in HUVEC leads to
activation of the p42/p44MAPKpathway and cell prolif-
eration (13). A2b-Adenosine receptors have been impli-
cated in the activation p42/p44MAPKin HEK-293 cells,
with activation mediated via Gq/11receptor coupling
dependent on activation of Ras (52). We can exclude a
role for A2breceptors in the activation of p42/p44MAPK
in HUVEC since 1) phosphorylation is reduced to basal
levels by the selective A2apurinoceptor antagonist
ZM241385 and activated by the A2apurinoceptor ago-
nist CGS21680 (100 nM, 2 min; data not shown), and 2)
receptor coupling to Gq/11leads to activation of phos-
pholipase C and subsequent release of store Ca2?,
adenosine-and
Figure 8. Effects of adenosine and sodium nitroprusside on
outward K?currents. Whole-cell currents were recorded from
single cells in the absence or presence of A) adenosine (10
?M), B) adenosine (10 ?M) ? l-NAME (100 ?M, pretreat-
ment for 30 min), or C) sodium nitroprusside (SNP, 10 ?M).
Voltage was applied in steps of 20 mV between ?100 mV and
?100 mV. Values denote the means ? se of current/voltage
relationships in 3–4 different cell cultures with 4 replicate
measurements per condition, *P ? 0.05 vs. corresponding
control or control ? l-NAME.
1592Vol. 16October 2002WYATT ET AL. The FASEB Journal

Page 10

which was not observed in our single-cell fluorescence
experiments (Fig. 2).
V EGF has been shown to stimulate NO production
in endothelial cells, resulting in an increase in intracel-
lular cGMP levels that in turn lead to the phosphoryla-
tion of p42/p44MAPK(53). However, in our experi-
ments adenosine-stimulated NO production required
activation of p42/p44MAPK. Since adenosine-stimulated
p42/p44MAPKphosphorylation was not inhibited by
l-NAME and ODQ only partially attenuated the re-
sponse, this suggests that p42/p44MAPKphosphoryla-
tion precedes NO release and cGMP accumulation.
These findings implicate p42/p44MAPKas upstream
regulators of NO production in response to A2apuri-
noceptor activation in HUVEC. The observation that
ODQ led to a slight decrease in p42/p44MAPKphos-
phorylation may stem from nonspecific effects of this
inhibitor since l-NAME, which reduces NO production
and cGMP levels, had no effect on adenosine-induced
p42/p44MAPKphosphorylation. Incubation of cells with
l-NAME before adenosine stimulation abolished the
up-regulation of l-arginine transport, suggesting that
NO (or a downstream mediator) may cause the in-
creased transport rate in HUVEC. Our observation is
consistent with the report that acute exposure of bovine
aortic endothelial cells to NO donors stimulates
l-arginine transport via the cationic amino acid
transporter CAT-1 (54).
l-arginine transport is sensitive to changes in mem-
brane potential (22, 55–57) and there is evidence for
the activation of ion channels by NO (58). In the
present study, adenosine activated an outward K?cur-
rent, which was abolished by l-NAME and mimicked by
a NO donor (Fig. 8). These findings suggest that NO
generated after stimulation of A2apurinoceptors acti-
vates this current. Alternatively, as reported in HEK293
cells, cGMP-activated protein kinase may lead to phos-
phorylation of a large conductance BKCachannel (59),
leading to membrane hyperpolarization and an in-
creased driving force for l-arginine entry.
In conclusion, we have shown that acute activation of
A2apurinoceptors in fetal endothelial cells stimulates
l-arginine transport and cGMP accumulation indepen-
dent of changes in cytosolic Ca2?, pH, or cAMP, but
involving protein tyrosine kinases and p42/p44MAPK
phosphorylation. The resulting increase in NO produc-
tion (or a downstream protein) then leads to increases
in l-arginine transport, most likely via a NO-induced
activation of an outward K?current and membrane
hyperpolarization.
This work was supported by the Medical Research Council,
Wellcome Trust (040727/Z/94; 052953/Z/97), British Heart
Foundation, Fondo Nacional de Desarrollo Cientifico y Tec-
nologico (1000354 and 7000354, Chile), and Direccion de Investi-
gacion,UniversidaddeConcepcion(DIUC201.084.003–1,Chile).
We are grateful to Dr. Rebecca Houliston for her assistance in
the initial immunoblotting experiments and to Dr. Albert
Ferro for his advice in measuring cAMP levels, and the
midwives of the Guy’s Hospital Maternity Ward for collecting
the umbilical cords.
REFERENCES
1. Sobrevia, L., Yudilevich, D. L., and Mann, G. E. (1997) Activation
of A2-purinoceptors by adenosine stimulates l-arginine transport
(system y?) and nitric oxide synthesis in human fetal endothelial
cells. J. Physiol. (London) 499, 135–140
Olsson, R. A., and Pearson, J. D. (1990) Cardiovascular purinocep-
tors. Physiol. Rev. 70, 761–845
Parratt, J. R., and Kane, K. A. (1994) KATPchannels in ischaemic
preconditioning. Cardiovasc. Res. 28, 783–787
Vials, A., and Burnstock, G. (1993) A2-purinoceptor-mediated
relaxation in the guinea-pig coronary vasculature: a role for nitric
oxide. Br. J. Pharmacol. 109, 424–429
Steinhorn, R. H., Morin, F. C., Van Wylen, D. G. L., Gugino, S. F.,
Giese, E. C., and Russell, J. A. (1994) Endothelium dependent
relaxations to adenosine in juvenile rabbit pulmonary arteries and
veins. J. Physiol. (London) 266, 2001–2006
Martin, P. L., and Potts, A. A. (1994) The endothelium of the rat
renal artery plays an obligatory role in A2adenosine receptor-
mediated relaxation induced by 5?-N-ethylcarboxamidoadenosine
and N6-cyclopentyladenosine. J. Pharmacol. Exp. Ther. 270, 893–899
Smits, P., Williams, S. B., Lipson, D. E., Banitt, P., Rongen, G. A.,
and Creager, M. A. (1995) Endothelial release of nitric-oxide
contributes to the vasodilator effect of adenosine in humans.
Circulation 92, 2135–2141
Abebe, W., Hussain, T., Olanrewaju, H., and Mustafa, S. J. (1995)
Role of nitric oxide in adenosine receptor-mediated relaxation of
porcine coronary artery. Am. J. Physiol. 269, H1672–H1678
Skinner, M. R., and Marshall, J. M. (1996) Studies on the roles of
ATP, adenosine and nitric-oxide in mediating muscle vasodilata-
tion induced in the rat by acute systemic hypoxia. J. Physiol.
(London) 495, 553–560
Li, J. M., Fenton, R. A., Wheeler, H. B., Powell, C. C., Peyton, B. D.,
Cutler, B. S., and Dobson, J. G. (1998) Adenosine A2areceptors
increase arterial endothelial cell nitric oxide. J. Surg. Res. 80,
357–364
Ethier, M. F., Chander, V., and Dobson, J. G. (1993) Adenosine
stimulates proliferation of human endothelial cells in culture.
Am. J. Physiol. 265, H131–H138
Sexl, V., Mancusi, G., Baumgartner-Parzer, S., Schutz, W., and
Freissmuth, M. (1995) Stimulation of human umbilical vein endo-
thelialcellproliferationbyA2-adenosineandbeta2-adrenoceptors.
Br. J. Pharmacol. 114, 1577–1586
Sexl, V., Mancusi, G., Holler, C., Gloria-Maercker, E., Schutz, W.,
and Freissmuth, M. (1997) Stimulation of the mitogen-activated
protein kinase via the A2A-adenosine receptor in primary human
endothelial cells. J. Biol. Chem. 272, 5792–5799
Ayajiki, K., Kindermann, M., Hecker, M., Fleming, I., and Busse, R.
(1996) Intracellular pH and tyrosine phosphorylation but not
calcium determine shear-stress induced nitric oxide production in
native endothelial cells. Circ. Res. 78, 750–758
Fleming, I., Bauersachs, J., Fisslthaler, B., and Busse, R. (1998)
Ca2?-independent activation of the endothelial nitric oxide syn-
thase in response to tyrosine phosphatase inhibitors and fluid
shear stress. Circ. Res. 82, 686–695
Caulin-Glaser, T., Garcia-Cardena, G., Sarrel, P., Sessa, W. C., and
Bender, J. R. (1997) 17 Beta-estradiol regulation of human endo-
thelial cell basal nitric oxide release, independent of cytosolic Ca2?
mobilization. Circ. Res. 81, 885–892
Dimmeler, S., Fleming, I., Fisslthaler, B., Hermann, C., Busse, R.,
and Zeiher, A. M. (1999) Activation of nitric oxide synthase in
endothelial cells by Akt-dependent phosphorylation. Nature (Lon-
don) 399, 601–605
Fulton, D., Gratton, J. P., McCabe, T. J., Fontana, J., Fujio, Y.,
Walsh, K., Franke, T. F., Papapetropoulos, A., and Sessa, W. C.
(1999) Regulation of endothelium-derived nitric oxide production
by the protein kinase Akt. Nature (London) 399, 597–601
Chen, Z., Yuhanna, I. S., GalchevaGargova, Z., Karas, R. H.,
Mendelsohn, R. E., and Shaul, P. W. (1999) Estrogen receptor
alpha mediates the nongenomic activation of endothelial nitric
oxide synthase by estrogen. J. Clin. Invest. 103, 401–406
Hisamoto, K., Ohmichi, M., Kanda, Y., Adachi, K., Nishio, Y.,
Hayakawa, J., Mabuchi, S., Takahashi, K., Tasaka, K., Miyamoto, Y.,
Taniguchi, N., and Murata, Y. (2001) Induction of endothelial
nitric oxide synthase phosphorylation by the raloxifene analog
LY117018 is differentially mediated by Akt and extracellular signal-
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
1593ACTIVATION OF THE L-ARGININE-NO PATHWAY BY ADENOSINE

Page 11

regulated protein kinase in vascular endothelial cells. J. Biol. Chem.
276, 47642–47649
Wyatt, A. W., Sobrevia, L., Wheeler-Jones, C., Houliston, R. A., and
Mann, G. E. (1998) Involvement of protein tyrosine kinases in
stimulation by adenosine of the l-arginine-nitric oxide pathway in
humanumbilicalveinendothelialcells.J.Physiol.(London)506,33P
Sobrevia, L., Cesare, P., Yudilevich, D. L., and Mann, G. E. (1995)
Diabetes-induced activation of system y?and nitric oxide synthase
in human endothelial cells: association with membrane hyperpo-
larization. J. Physiol. (London) 489, 183–192
Jaffe,E.A.,Nachman,R.L.,Becker,C.G.,andMinick,C.R.(1973)
Culture of human endothelial cells derived from umbilical veins.
Identification by morphologic and immunologic criteria. J. Clin.
Invest. 52, 2745–2756
Jacob, R. (1990) Agonist-stimulated divalent cation entry into
single cultured human umbilical vein endothelial cells. J. Physiol.
(London) 421, 55–77
Jacob, R. (1991) Calcium oscillations in endothelial cells. Cell
Calcium 12, 127–134
Ruehlmann, D. O., Steinert, J. R., Valverde, M. A., Jacob, R., and
Mann, G. E. (1998) Environmental estrogenic pollutants induce
acute vascular relaxation by inhibiting l-type Ca2?channels in
smooth muscle cells. FASEB J. 12, 613–619
Wheeler-Jones,C.P.,May,M.J.,Houliston,R.A.,andPearson,J.D.
(1996) Inhibition of MAP kinase kinase (MEK) blocks endothelial
PGI2release but has no effect on von Willebrand factor secretion
or E-selectin expression. FEBS Lett. 388, 180–184
Wheeler-Jones, C., AbuGhazaleh, R., Cospedal, R., Houliston,
R. A., Martin, J., and Zachary, I. (1997) Vascular endothelial
growth factor stimulates prostacyclin production and activation of
cytosolic phospholipase A2in endothelial cells via p42/p44 mito-
gen-activated protein kinase. FEBS Lett. 420, 28–32
Alessi,D.R.,Cuenda,A.,Cohen,P.,Dudley,D.T.,andSaltiel,A.R.
(1995) PD 98059 is a specific inhibitor of the activation of
mitogen-activated protein kinase kinase in vitro and in vivo. J. Biol.
Chem. 270, 27489–27494
Favata, M. F., Horiuchi, K. Y., Manos, E. J., Daulerio, A. J., Stradley,
D. A., Feeser, W. S., Van Dyk, D. E., Pitts, W. J., Earl, R. A., Hobbs,
F., Copeland, R. A., Magolda, R. L., Scherle, P. A., and Trzaskos,
J. M. (1998) Identification of a novel inhibitor of mitogen-activated
protein kinase kinase. J. Biol. Chem. 273, 18623–18632
May, M. J., Wheeler-Jones, C., Houliston, R. A., and Pearson, J. D.
(1998) Activation of p42(mapk) in human umbilical vein endothe-
lial cells by interleukin-1 alpha and tumor necrosis factor-alpha.
Am. J. Physiol. 43, C789–C798
Sobrevia, L., Jarvis, S. M., and Yudilevich, D. L. (1994) Adenosine
transport in cultured human umbilical vein endothelial cells is
reduced in diabetes. Am. J. Physiol. 267, C39–C47
Griffiths, M., Beaumont, N., Yao, S. Y., Sundaram, M., Boumah,
C. E., Davies, A., Kwong, F. Y., Coe, I., Cass, C. E., Young, J. D., and
Baldwin, S. A. (1997) Cloning of a human nucleoside transporter
implicated in the cellular uptake of adenosine and chemothera-
peutic drugs. Nature Med. 3, 89–93
Lin, S., Fagan, K. A., Li, K. X., Shaul, P. W., Cooper, D. M., and
Rodman, D. M. (2000) Sustained endothelial nitric-oxide synthase
activation requires capacitative Ca2?entry. J. Biol. Chem. 275,
17979–17985
Fleming, I., Hecker, M., and Busse, R. (1994) Intracellular alkalin-
ization induced by bradykinin sustains activation of the constitutive
nitric oxide synthase in endothelial cells. Circ. Res. 74, 1220–1226
Richard, L. F., Dahms, T. E., and Webster, R. O. (1998) Adenosine
prevents permeability increase in oxidant-injured endothelial
monolayers. Am. J. Physiol. 43, H35–H42
Tseng, H., Peterson, T. E., and Berk, B. C. (1995) Fluid shear stress
stimulates mitogen-activated protein kinase in endothelial cells.
Circ. Res. 77, 869–878
Borsch-Haubold,A.G.,Pasquet,S.,andWatson,S.P.(1998)Direct
inhibition of cyclooxygenase-1 and -2 by the kinase inhibitors
SB203580 and PD98059-SB203580 also inhibits thromboxane syn-
thase. J. Biol. Chem. 273, 28766–28772
Sala, R., Rotoli, B. M., Colla, E., Visigalli, R., Parolari, A., Bussolati,
O., Gazzola, G. C., and Dall’Asta, V. (2002) Two-way arginine
transport in human endothelial cells: TNF-alpha stimulation is
restricted to system y?. Am. J. Physiol. 282, C134–C143
Haynes, M. P., Sinha, D., Russell, K. S., Collinge, M., Fulton, D.,
Morales-Ruiz, M., Sessa, W. C., and Bender, J. R. (2000) Membrane
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
estrogen receptor engagement activates endothelial nitric oxide
synthase via the PI3-kinase-Akt pathway in human endothelial cells.
Circ. Res. 87, 677–682
Feoktistov, I., Goldstein, A. E., Ryzhov, S., Zeng, D., Belardinelli, L.,
Voyno-Yasenetskaya, T., and Biaggioni, I. (2002) Differential ex-
pression of adenosine receptors in human endothelial cells: role of
A2breceptorsinangiogenicfactorregulation.Circ.Res.90,531–538
Olanrewaju, H. A., Qin, W., Feoktistov, I., Scemama, J. L., and
Mustafa, S. J. (2000) Adenosine A2aand A2breceptors in cultured
humanandporcinecoronaryarteryendothelialcells.Am.J.Physiol.
279, H650–H656
Prentice, D. J., and Hourani, S. M. (2000) Characterisation of
adenosine receptors mediating relaxation in hamster isolated
aorta. Naunyn-Schmiedebergs Arch. Pharmacol. 362, 427–434
Olanrewaju,H.A.,andMustafa,S.J.(2000)AdenosineA2aandA2b
receptors mediated nitric oxide production in coronary artery
endothelial cells. Gen. Pharmacol. 35, 171–177
Rolleston, S., and Marshall, J. M. (2000) The ability of adenosine to
generate nitric oxide in human umbilical vein endothelial cells.
J. Physiol. (London) 523, 151P
Fleming, I., Bauersachs, J., Fisslthaler, B., and Busse, R. (1998)
Ca2?-independent activation of the endothelial nitric oxide syn-
thase in response to tyrosine phosphatase inhibitors and fluid
shear stress. Circ. Res. 82, 686–695
Ferro, A., Queen, L. R., Priest, R. M., Xu, B., Ritter, J. M., Poston,
L.,andWard,J.P.(1999)Activationofnitricoxidesynthasebybeta
2-adrenoceptors in human umbilical vein endothelium in vitro.
Br. J. Pharmacol. 126, 1872–1880
Hein, T., and Kuo, L. (1998) cAMP-independent dilation of
coronary arterioles to adenosine: role of nitric oxide, G-proteins,
and KATPchannels. FASEB J. 12, 76–76
Graier, W. F., Kukovetz, W. R., and Groschner, K. (1993) Cyclic
AMP enhances agonist-induced Ca2?entry into endothelial cells
by activation of potassium channels and membrane hyperpolariza-
tion. Biochem. J. 291, 263–267
Graier, W. F., Groschner, K., Schmidt, K., and Kukovetz, W. R.
(1992) Increases in endothelial cyclic AMP levels amplify agonist-
induced formation of endothelium-derived relaxing factor
(EDRF). Biochem. J. 288, 345–349
Fleming, I., and Busse, R. (1999) Signal transduction of eNOS
activation. Cardiovasc. Res. 43, 532–541
Gao, Z. H., Chen, T. S., Weber, M. J., and Linden, J. (1999) A2b
adenosine and P2Y2receptors stimulate mitogen-activated protein
kinase in human embryonic kidney-293 cells—cross-talk between
cyclic AMP and protein kinase C pathways. J. Biol. Chem. 274,
5972–5980
Parenti, A., Morbidelli, L., Cui, X. L., Douglas, J. G., Hood, J. D.,
Granger, H. J., Ledda, F., and Ziche, M. (1998) Nitric oxide is an
upstream signal of vascular endothelial growth factor-induced
extracellular signal-regulated kinase(1/2) activation in postcapil-
lary endothelium. J. Biol. Chem. 273, 4220–4226
Ogonowski, A. A., Kaesemeyer, W. H., Jin, L., Ganapathy, V.,
Leibach, F. H., and Caldwell, R. W. (2000) Effects of NO donors
and synthase agonists on endothelial cell uptake of l-Arginine and
superoxide production. Am. J. Physiol. 278, C136–C143
Bussolati, O., Sala, R., Astorri, A., Rotoli, B. M., Dall’Asta, V., and
Gazzola, G. C. (1993) Characterization of amino acid transport in
human endothelial cells. Am. J. Physiol. 265, C1006–C1014
Kavanaugh, M. P. (1993) Voltage dependence of facilitated argi-
nine flux mediated by the system y?basic amino acid transporter.
Biochemistry 32, 5781–5785
Bogle, R. G., Baydoun, A. R., Pearson, J. D., and Mann, G. E.
(1996) Regulation of l-arginine transport and nitric oxide release
in superfused porcine aortic endothelial cells. J. Physiol. (London)
490, 229–241
Duszky, M., and Radomski, M. W. (2000) The role of nitric oxide
intheregulationofionchannelsinairwayepithelium:implications
for diseases of the lung. Free Radic. Res. 33, 449–459
Fukao, M., Mason, H. S., Britton, F. C., Kenyon, J. L., Horowitz, B.,
and Keef, K. D. (1999) Cyclic GMP-dependent protein kinase
activates cloned BKCachannels expressed in mammalian cells by
direct phosphorylation at serine 1072. J. Biol. Chem. 274, 10927–
10935
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
Received for publication December 21, 2001.
Revised for publication May 30, 2002.
1594Vol. 16October 2002 WYATT ET AL.The FASEB Journal