| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
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 Concepción, Concepción, Chile
1Correspondence: Centre for Cardiovascular Biology and Medicine, GKT School of Biomedical Sciences, King’s College London, Guy’s Campus, London SE1 1UL, UK. E-mail giovanni.mann{at}kcl.ac.uk
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
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 µM) 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 µM) and PD98059 (10 µM), 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.—Wyatt, 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/p44MAPK pathway mediates adenosine-induced nitric oxide production in human endothelial cells: a novel calcium-insensitive mechanism.
Key Words: mitogen-activated protein kinases • NO • L-arginine • A2a purinoceptors
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. It is an important regulator of blood flow and has been shown to decrease coronary resistance under basal and restricted blood flow conditions (3)
. In many cell types, adenosine elicits vasodilation via direct stimulation of G-protein-coupled purinoceptors located on vascular smooth muscle cells by modulating intracellular cAMP levels (2)
. Activation of A1 and A3 receptor subtypes leads to inactivation of adenylyl cyclase and a decrease in cAMP levels; activation of A2a and A2b receptor subtypes enhances cAMP accumulation, resulting in vasorelaxation (2)
. Although adenosine is classically regarded as an endothelium-independent vasodilator, accumulating evidence implicates endothelium-derived nitric oxide (NO) in adenosine-mediated vasodilation (1
, 4
5
6
7
8
9)
.
Our previous studies in human umbilical vein endothelial cells (1)
and studies by others in porcine carotid artery endothelial cells (10)
have established that activation of A2a purinoceptors by adenosine stimulates NO release. Our experiments in umbilical vein endothelial cells demonstrated that adenosine acutely stimulates influx of L-arginine (substrate for eNOS) via a sodium-independent y+ transport system, whose activity was enhanced in response to membrane hyperpolarization (1)
. Although adenosine has been reported to increase endothelial cell proliferation most likely via the activation of p42/p44MAPK (11
12
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 17ß-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 enzyme activation at basal levels of Ca2+ (17
, 18)
. Activation of eNOS in response to 17ß-estradiol has been reported to be mediated via p42/p44MAPK in a calcium-dependent manner (19)
. As recent evidence implicates Akt/PKB and p42/p44MAPK in 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 A2a purinoceptor-stimulated increases in L-arginine transport, NO production, and p42/44MAPK phosphorylation. A preliminary account of part of this work has been published in abstract form (21)
.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, HUVEC cultures were identified by their typical cobblestone morphology 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 µL 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)25 beads and cDNA was synthesized immediately. mRNA was added to oligo (dT)18 (1 µg) and random 10-MERS (1 µg); 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 µg/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 µL of cDNA (1:10 dilution), 100 µM of each deoxynucleoside 5'-triphosphate, 0.5 µM 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 µg/mL) staining. Adenosine receptor subtype PCR products were cut from agarose gels and purified using a Qiaquick gel extraction kit. The identity of each PCR product was confirmed by direct sequencing on an ABI automated sequencer.
Effects of A2a purinoceptor agonists and inhibitors of p42/p44MAPK and adenylyl cyclase on L-arginine transport
Confluent HUVEC in 96-well microtiter plates were preequilibrated 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 µM), and the phosphodiesterase inhibitor 3-isobutylmethylxanhine (IBMX, 1 µM), then stimulated for 2 min with adenosine (10 µM), the A2a purinoceptor agonist 2-p-(2-carboxyethyl) phenethylamino-5'-N-ethylcarboxamido-adenosine (CGS21680, 100 nM) or forskolin (1 µM). Basal and agonist-stimulated rates of L-[3H]arginine (100 µM) transport 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 nitrobenzylthioinosine (NBMPR, 10 µM, 30 min). In other experiments, cells were pretreated with the protein tyrosine kinase inhibitor genistein (10 µM, 30 min), its less active analog daidzein (10 µM, 30 min), an inhibitor of MEK1/2, PD98059 (2'-amino-3'-methoxyflavone, 10 µM, 30 min), or adenylyl cyclase SQ22536 (9-(tetrahydro-2-furanyl)-9H purin-6-amine, 100 µM, 30 min). In some experiments, cells were preincubated with the eNOS inhibitor NG-nitro-L-arginine-methyl ester (L-NAME, 100 µM, 30 min), calcium-free Krebs supplemented with ethylene glycol bis (2-aminoethylether)-N,N,N',N'-tetraacetic acid (EGTA, 1 mM), or exposed to 80 mM KCl to evoke a membrane depolarization.
Single-cell measurement of [Ca2+]i and pHi
Intracellular calcium ([Ca2+]i) and pH were measured using ratiometric fluorescent dyes. HUVEC were seeded onto glass no. 1 coverslips and cultured for 2–3 days before use. To measure single-cell [Ca2+]i changes, cells were loaded with 1 µM 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 µM 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 experimental runs, the same solution was used but with a lower BSA concentration (0.1%). Coverslips were mounted on a thermostatted stage (30 or 37°C) of an inverted epifluorescence microscope equipped with a 40x 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 (Improvision, Coventry UK). An image pair was captured every 1–3 s. [Ca2+]i and pHi are expressed qualitatively as the ratio of fluorescence at 350/380 nm or 490/440 nm, respectively. For single-cell photometric measurements of [Ca2+]i, excitation was via a 75W xenon arc lamp and fluorescence was monitored at >450 nm. For Fura-2 loaded cells, autofluorescence was estimated by addition of 2 mM Mn2+ plus 10 µM 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 µM L-arginine and rolipram (50 µM, cAMP-specific phosphodiesterase inhibitor). cAMP levels were determined under basal conditions and in cells challenged for 2 min with adenosine (10 µM), CGS21680 (100 nM) or forskolin (1 µM). 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 µM) and IBMX (1 µM). The preincubation medium was removed and cells were stimulated with adenosine (10 µM, 2 min) or the A2a receptor agonist CGS21680 (100 nM, 2 min). HCl (0.1M) cell extracts were stored at -20°C for radioimmunoassay of cGMP levels (22)
. Basal and A2a purinoceptor-stimulated cGMP levels were also measured in cells pretreated with inhibitors of eNOS (L-NAME, 100 µM, 30 min), adenylyl cyclase (SQ22536, 100 µM, 30 min), tyrosine kinases (genistein, 10 µM, 30 min), a less active analog of genistein (daidzein, 10 µM, 30 min), MEK1/2 (PD98059, 10 µM, 30 min), PI3-kinase (wortmannin, 20 nM, 30 min or LY294002, 10 µM, 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 (Biological, 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 generation, 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 µM) after pretreatment in the absence or presence of L-NAME (100 µM, 30 min) or challenged with the NO donor sodium nitroprusside (10 µM).
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 µM L-arginine in the absence or presence of a protein tyrosine kinase inhibitor, genistein (10 µM, 30 min) or the MEK1/2 inhibitors PD98059 (10 µM, 30 min) and U0126 (1,4-diamino-2,3-dicyano-1,4-bis[aminophenylthio]butadiene, 1 µM, 30 min). The concentrations of the structurally distinct MEK1/2 inhibitors used are consistent with our previous reports (27
, 28)
and have no inhibitory effect on the p38MAPK or JNK pathways (29
, 30)
. Cells were preincubated with a soluble guanylyl cyclase inhibitor 1H-[1,2,4] oxadiazolo [4,3-a] quinoxalin-1-one (ODQ, 10 µM, 30 min), L-NAME (100 µM, 30 min), ZM241385 (100 nM, 30 min, A2a purinoceptor antagonist), or the PI3-kinase inhibitor LY249002 (10 µM, 30 min). HUVEC were subsequently challenged with adenosine (10 µM, 2 min) and the reaction was stopped by washing monolayers with ice-cold PBS containing 200 µM 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/p44MAPK antibody raised against the dually phosphorylated proteins provides a good indicator of p42/p44MAPK activity, since we previously correlated p42/p44MAPK activity in HUVEC using an in gel kinase assay with dual phosphorylation of p42/p44MAPK (31)
. To ascertain equal protein loading, 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).
Materials
All cell culture reagents, collagenase type II from Clostridium histolyticum, adenosine, histamine, IBMX, rolipram, wortmannin, ODQ, genistein, daidzein, SQ22536, CGS21680, N6-cyclopentyladenosine (CPA), N-ethylcarboxamidoadenosine (NECA), and (2-chloro-N6-(3-iodobenzyl)-5'-(N-methylcarbamoyl)adenosine (Cl-IB-MECA) were from Sigma (Poole, Dorset, UK); A2a purinoceptor antagonist 4-(2-[7-amino-2-(2-furyl)]1,2,4-triazolo[2,3-a] [1,3,5]triazin-5-ylamino]ethyl)phenol (ZM241385) from AstraZeneca Pharmaceuticals (Macclesfield, Cheshire, UK); PD98059, LY294002, and U0126 from Calbiochem (Beeston, Nottingham, UK); Fura-2 acetoxymethyl 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 (Buckinghamshire, UK). Antibodies against the dually phosphorylated and total p42/p44MAPK and 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, Middlesex, 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 |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. To extend our pharmacological findings, we have used RT-PCR analyses to identify adenosine receptor 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+]o removal or elevated K+ on L-arginine transport and cGMP production
Adenosine (10 µM, 2 min) -stimulated L-arginine transport [pmol (µg protein)-1 min-1: 1.45±0.4 vs. 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 (106 cells)-1 5 min-1: 2.2±0.2 vs. 4.8±0.1, n=4, P<0.01] were mimicked by an A1/2 purinoceptor agonist NECA [100 nM, 2 min: 3.3±0.2 pmol (106 cells)-1 5 min-1, P < 0.05], unaffected by an A3 agonist Cl-IB-MECA [100 nM, 2 min: 2.5±0.2 pmol (106 cells)-1 5 min-1], and abolished by L-NAME (2.0±0.06, n=4). As in our earlier study (1)
, pretreatment of HUVEC with a selective A2a purinoceptor antagonist (ZM241385) prevented both CGS21680 (data not shown) and NECA [2.4±0.04 pmol (106 cells)-1 5 min-1]-induced cGMP accumulation. Preincubation of HUVEC in nominally Ca2+-free buffer had no significant effect on basal, adenosine- or CGS21680-stimulated rates of L-arginine transport or cGMP production (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)
.
|
Effects of adenosine receptor agonists and histamine on [Ca2+]i and pHi
Given that eNOS can be stimulated directly by Ca2+ (34)
, it was pertinent to examine the effects of adenosine receptor activation on potential changes in [Ca2+]i. Despite a significant stimulation of NO production (Table 1)
, adenosine (10 µM) 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. 2
A and inset). Similar data were obtained in cells challenged initially with histamine, followed by acute exposure to adenosine or CGS21680 (data not shown). Careful examination of the 350/380 nm ratio images did not reveal any highly localized [Ca2+]i increases in response to adenosine that might have escaped detection when measuring 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 pHi have been suggested to modulate eNOS activity, i.e., alkalinization of the cytosol promotes NO release in some endothelial cells (35)
. Although histamine (10 µM) evoked a transient increase in the 490/440 nm ratio (data not shown) reflecting an increase in pHi, adenosine (10 µM) caused negligible changes in pHi (490/440 nm 
ratio 0.06±0.004, n=98 cells). As expected, an ammonium 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 withdrawal.
Effects of adenosine, CGS21680, and forskolin on intracellular cAMP levels
As activation of A2a purinoceptors has been shown to increase intracellular cAMP levels in endothelial cells (36)
, we investigated whether adenosine (10 µM, 2 min) or CGS21680 (100 nM, 2 min) increased cAMP levels in HUVEC. Although activation of adenylyl cyclase by forskolin (1 µM, 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)-1 5 min-1. Forskolin increased basal rates of L-arginine transport without altering intracellular cGMP levels (Fig. 3
A). 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-mediated phosphorylation (17
, 18)
, we investigated whether adenosine-stimulated NO production required activation 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. 4
A) or phosphorylation of Akt/PKB (Fig. 4B
).
|
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 spectrum tyrosine kinase inhibitor, or the MEK1/2 inhibitor PD98059 had no effect on basal rates of L-arginine transport or cGMP accumulation (Table 2
). In these same experiments, adenosine (10 µM, 2 min) evoked a twofold increase in L-arginine transport and cGMP accumulation, which was abolished in cells pretreated with either genistein (10 µM) or PD98059 (10 µM) but unaffected by daidzein (10 µM), a less active analog of genistein.
|
In parallel experiments, we found that adenosine (10 µM) caused a time-dependent (15 s–5 min) increase in p42/p44MAPK phosphorylation (Fig. 5
A), which was attenuated by the selective A2a receptor 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/p44MAPK protein levels were unchanged (Fig. 5A
). Our ability to detect basal phosphorylation of p42/p44MAPK is 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. 6
A–C, adenosine-induced increases in p42/p44MAPK phosphorylation were abolished by pretreatment of cells with genistein (10 µM), PD98059 (10 µM), and U0126 (1 µM). The inhibitory effects of PD98059 and U0126 on p42/p44MAPK activation and/or cGMP accumulation could not be attributed to actions on cyclooxygenases (38)
, since indomethacin (10 µM) had no effect on adenosine-stimulated cGMP accumulation or p42/p44MAPK phosphorylation (data not shown). Treatment of cells with LY294002 had no effect on adenosine-induced p42/p44MAPK phosphorylation, confirming that adenosine does not signal via PI3-kinases in HUVEC (Fig. 7
A).
|
|
|
Effect of eNOS inhibition on adenosine-induced p42/p44MAPK phosphorylation
Although adenosine-induced stimulation of L-arginine transport and cGMP accumulation was prevented by L-NAME, inhibition of eNOS had no effect on adenosine-stimulated p42/p44MAPK phosphorylation (Fig. 7B
). Moreover, pretreatment of cells with ODQ, an inhibitor of soluble guanylyl cyclase, had only a marginal effect on p42/p44MAPK phosphorylation (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 sensitive to changes in membrane potential (1)
, we determined whether adenosine acutely modulates membrane currents in HUVEC using the whole-cell patch clamp technique. Adenosine (10 µM) activated an outward K+ current (Fig. 8
A), 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 stimulation of L-arginine transport in HUVEC challenged with adenosine.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 39)
. Stimulation of eNOS activity in response to A2a purinoceptor activation was not mediated via PI3 kinase and Akt/PKB, as previously documented 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 in HUVEC is mediated via A2a purinoceptors (1)
and have now established that this cell type possesses mRNA for the A2a-, A2b-, and A3 purinoceptor subtypes with little or no detectable mRNA for the A1 purinoceptor. Our findings are supported by a recent study with HUVEC reporting that mRNA levels for the A2a purinoceptor are 10-fold greater than the A2b purinoceptor (41)
. In agreement with our findings, the latter study failed to detect mRNA for the A1 purinoceptor. Moreover, A2a- and A2b purinoceptors have been reported in porcine and human coronary artery endothelial cells by RT-PCR and Western blot analysis, establishing a correlation 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 coronary vasculature (4)
, hamster aorta (43)
, and porcine coronary artery endothelial cells (44)
. Moreover, activation of A1 purinoceptors markedly attenuates adenosine-stimulated NO production in iliac and carotid artery endothelial cells (10)
. The only report implicating A1 purinoceptors in adenosine-mediated NO production in HUVEC used much higher adenosine concentrations (5 mM) (45)
than the present study (10 µM), 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+]i and, in the case of bradykinin, an intracellular alkalinization (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 A2a purinoceptor-stimulated L-arginine transport or NO production. These findings agree with recent evidence that shear stress (46)
and 17ß-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 A2a purinoceptor-mediated responses in HUVEC seems unlikely, since we found that cGMP accumulation was unaffected by the PI3-kinase inhibitors LY294002 or wortmannin and Akt/PKB was not serine phosphorylated in response to adenosine.
Activation of A2 purinoceptors 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)
. In the present study, however, adenosine- and 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 A2a purinoceptor-stimulated L-arginine transport 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 forskolin-mediated increase in L-arginine transport. Forskolin has been reported to potentiate agonist (bradykinin and histamine) –stimulated but not basal cGMP accumulation 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 17ß-estradiol-mediated activation of eNOS has been reported (14
, 51)
. In endothelial cells, acute stimulation of NO production in response to estradiol can be inhibited by herbimycin (16)
and the MEK1/2 inhibitor PD98059 (19)
, suggesting that protein tyrosine kinases and p42/p44MAPK are involved in the signaling pathway regulating NO production in endothelial cells. Our study of HUVEC provides further evidence that adenosine-stimulated activation of p42/p44MAPK is prevented by genistein and the structurally distinct MEK1/2 inhibitors PD98059 and U0126 (Fig. 6)
. Adenosine-stimulated phosphorylation of p42/p44MAPK was prevented by the selective A2a-adenosine receptor antagonist ZM241385 but unaffected by nitrobenzylthioinosine, an inhibitor of the es nucleoside transporter (data not shown), confirming that phosphorylation of p42/p44MAPK is a result of A2a purinoceptor activation rather than a consequence of adenosine transport. Our findings are consistent with an earlier report that stimulation of A2a purinoceptors in HUVEC leads to activation of the p42/p44MAPK pathway and cell proliferation (13)
. A2b-Adenosine receptors have been implicated in the activation p42/p44MAPK in HEK-293 cells, with activation mediated via Gq/11 receptor coupling dependent on activation of Ras (52)
. We can exclude a role for A2b receptors in the activation of p42/p44MAPK in HUVEC since 1) phosphorylation is reduced to basal levels by the selective A2a purinoceptor antagonist ZM241385 and activated by the A2a purinoceptor agonist CGS21680 (100 nM, 2 min; data not shown), and 2) receptor coupling to Gq/11 leads to activation of phospholipase C and subsequent release of store Ca2+, 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 intracellular cGMP levels that in turn lead to the phosphorylation of p42/p44MAPK (53)
. However, in our experiments adenosine-stimulated NO production required activation of p42/p44MAPK. Since adenosine-stimulated p42/p44MAPK phosphorylation was not inhibited by L-NAME and ODQ only partially attenuated the response, this suggests that p42/p44MAPK phosphorylation precedes NO release and cGMP accumulation. These findings implicate p42/p44MAPK as upstream regulators of NO production in response to A2a purinoceptor activation in HUVEC. The observation that ODQ led to a slight decrease in p42/p44MAPK phosphorylation 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/p44MAPK phosphorylation. 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 increased 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 membrane potential (22
, 55
56
57)
and there is evidence for the activation of ion channels by NO (58)
. In the present study, adenosine activated an outward K+ current, which was abolished by L-NAME and mimicked by a NO donor (Fig. 8)
. These findings suggest that NO generated after stimulation of A2a purinoceptors activates this current. Alternatively, as reported in HEK293 cells, cGMP-activated protein kinase may lead to phosphorylation of a large conductance BKCa channel (59)
, leading to membrane hyperpolarization and an increased driving force for L-arginine entry.
In conclusion, we have shown that acute activation of A2a purinoceptors in fetal endothelial cells stimulates L-arginine transport and cGMP accumulation independent of changes in cytosolic Ca2+, pH, or cAMP, but involving protein tyrosine kinases and p42/p44MAPK phosphorylation. The resulting increase in NO production (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.
|
ACKNOWLEDGMENTS |
|---|
Received for publication December 21, 2001.
Revision received May 30, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
