Rapid Stimulation of L-Arginine Transport by D-Glucose

Involves p42/44mapk and Nitric Oxide in Human Umbilical

Vein Endothelium
Carlos Flores, Susana Rojas, Claudio Aguayo, Jorge Parodi, Giovanni Mann, Jeremy D. Pearson,
Paola Casanello, Luis Sobrevia

Abstract—D-Glucose infusion and gestational diabetes induce vasodilatation in humans and increase L-arginine transport
and nitric oxide (NO) synthesis in human umbilical vein endothelial cells. High D-glucose (25 mmol/L, 2 minutes)
induced membrane hyperpolarization and an increase of L-arginine transport (Vmax 6.1⫾0.7 versus 4.4⫾0.1 pmol/␮g
protein per minute) with no change in transport affinity (Km 105⫾9 versus 111⫾16 ␮mol/L). L-[3H]Citrulline formation
and intracellular cGMP, but not intracellular Ca2⫹, were increased by high D-glucose. The effects of D-glucose were
mimicked by levcromakalim (ATP-sensitive K⫹ channel blocker), paralleled by p42/p44mapk and Ser1177– endothelial NO
synthase phosphorylation, inhibited by NG-nitro-L-arginine methyl ester (L-NAME; NO synthesis inhibitor), gliben-
clamide (ATP-sensitive K⫹ channel blocker), KT-5823 (protein kinase G inhibitor), PD-98059 (mitogen-activated
protein kinase kinase 1/2 inhibitor), and wortmannin (phosphatidylinositol 3-kinase inhibitor), but they were unaffected
by calphostin C (protein kinase C inhibitor). Elevated D-glucose did not alter superoxide dismutase activity. Our findings
demonstrate that the human fetal endothelial L-arginine/NO signaling pathway is rapidly activated by elevated D-glucose
via NO and p42/44mapk. This could be determinant in pathologies in which rapid fluctuations of plasma D-glucose may
occur and may underlie the reported vasodilatation in early stages of diabetes mellitus. (Circ Res. 2003;92:64-72.)
Key Words: humans 䡲 endothelium 䡲 glucose 䡲 arginine 䡲 nitric oxide

tivated protein (MAP) kinases (p42/44mapk).5,14,15 p42/44mapk

T he cationic amino acid L-arginine is the substrate for
nitric oxide (NO) synthesis via endothelial NO synthase
(eNOS)1 and is taken up primarily by the Na⫹-independent
activation may itself be dependent on PKC activation and NO
synthesis.5,14 However, the effect of short-term incubation
high-affinity (Km ⬇100 to 400 ␮mol/L) systems y⫹/CAT-1 with elevated D-glucose on the endothelial L-arginine/NO
and y⫹/CAT-2B (where CAT indicates cationic amino acid pathway has not been investigated.4,11,16,17
transporter) in human umbilical vein endothelial cells The present study shows that a 2-minute incubation with
(HUVECs).2,3 L-Arginine transport and NO synthesis (L- 25 mmol/L D-glucose increases L-arginine transport and NO
arginine/NO pathway) are increased in HUVECs from pa- synthesis in HUVECs. The underlying cellular mechanisms
tients with gestational diabetes.2 Interestingly, long-term involve phosphorylation of eNOS at Ser1177 via phosphatidyl-
incubation (24 hours) of HUVECs from normal pregnancies inositol 3-kinase (PI3-k) and activation of eNOS and p42/
with elevated D-glucose mimics the effect of gestational
p44mapk by D-glucose.18
diabetes on the L-arginine/NO pathway.4 In addition, elevated
D-glucose for 24 hours4,5 or 5 days6 increases eNOS gene
expression. A recent report shows that D-glucose infusion Materials and Methods
induces vasodilatation in humans,7 and in animal models, an
elevation of plasma D-glucose results in rapid (seconds to Cell Culture
minutes) vasodilatation.8 –10 Therefore, rapid fluctuations in Human umbilical vein endothelium was isolated (collagenase diges-
the D-glucose level are crucial in maintaining human fetal tion 0.25 mg/mL) and cultured (37°C, 5% CO2, confluent passage 2)
in medium 199 containing 5 mmol/L D-glucose, 10% newborn calf
endothelial function.2–5,11 serum, 10% fetal calf serum, 3.2 mmol/L L-glutamine, 100 ␮mol/L
D-Glucose activates protein kinase C (PKC), an enzyme L-arginine, and 100 U/mL penicillin-streptomycin (primary culture
involved with long-term stimulation of the L-arginine/NO medium).2– 4 Before an experiment (24 hours), the incubation me-
pathway,5,12–14 and (within 1 hour) p42 and p44 mitogen-ac- dium was changed to serum-free medium 199.

Original received July 26, 2002; revision received October 31, 2002; accepted November 11, 2002.
From the Cellular and Molecular Physiology Laboratory (C.F., S.R., C.A., J.P., P.C., L.S.), Department of Physiology, Faculty of Biological Sciences,
and the Department of Obstetrics and Gynaecology (P.C.), Faculty of Medicine, University of Concepción, Concepción, Chile, and King’s College
London (G.M., J.D.P.), Guy’s Campus, London, UK.
Correspondence to Dr L. Sobrevia, Cellular and Molecular Physiology Laboratory (CMPL), Department of Physiology, Faculty of Biological Sciences,
University of Concepción, PO Box 160-C, Concepción, Chile. E-mail lsobrev@udec.cl
© 2003 American Heart Association, Inc.
Circulation Research is available at http://www.circresaha.org DOI: 10.1161/01.RES.0000048197.78764.D6

64
 Flores et al D-Glucose Acutely Activates L-Arginine/NO Pathway 65

L-Arginine Transport Intracellular Ca2ⴙ

L-Arginine transport (1 ␮Ci/mL, 37°C, 1 minute) was determined in Cells on glass coverslips were loaded (30 minutes, 23°C) with the
cells preincubated (15 seconds to 5 minutes) with Krebs solution acetoxymethyl derivative of fluo 3 (5 ␮mol/L). Coverslips were
(mmol/L: NaCl 131, KCl 5.6, NaHCO3 25, NaH2PO4 1, HEPES 20, transferred to an experimental bath with Krebs solution containing 5
CaCl2 2.5, and MgCl2 1 [pH 7.4, 37°C]) containing 5 or 25 mmol/L or 25 mmol/L D-glucose, and Ca2⫹ was imaged using a Zeiss LSM
D-glucose, 25 mmol/L L-glucose, or 5 mmol/L D-glucose plus 410 confocal microscope.4
20 mmol/L D-mannitol (osmotic controls).2– 4 L-Arginine transport
was also determined in Krebs solution in which NaCl was replaced Western Blots
by equimolar concentrations of choline chloride2– 4 or in cells After pretreatment with 10 ␮mol/L PD-98059 (30 minutes), 100
incubated (30 minutes) with KCl (5.5 to 131 mmol/L), with NaCl ␮mol/L SNAP (2 to 5 minutes), or 30 nmol/L wortmannin (30
decreased equivalently, or with 131 mmol/L KCl for 2, 4, 10, 20, or minutes), the cells were incubated with 5 or 25 mmol/L D-glucose (2
30 minutes. In trans-stimulation experiments, cells were preincu- minutes). Cell protein extracts were probed with a primary poly-
bated (2 hours) with primary culture medium containing 10 mmol/L clonal mouse antiphosphorylated (1:1000) or nonphosphorylated
L-lysine. Cell-associated radioactivity and data analyses were per-
(1:1500) p44/p42mapk, rabbit anti-eNOS (1:2500) or anti-phosphory-
formed as described.2– 4 lated Ser1177– eNOS (1:2500) antibodies, and horseradish peroxidase–
conjugated goat secondary antibodies as described.3,5 Primary poly-
L-Arginine transport was assayed in cells preincubated (30 min-
clonal mouse anti-actin (1:2000) served as the internal control.
utes) with 100 ␮mol/L NG-nitro-L-arginine methyl ester (L-NAME,
Proteins were detected by enhanced chemiluminescence and quan-
eNOS inhibitor),2,3 10 ␮mol/L PD-98059 (MAP kinase kinase 1/2
tified by densitometry (Ultroscan XL enhanced laser densitometer,
[MEK1/2] inhibitor), 19 100 ␮ mol/L S-nitroso-N-acetyl- L , D - LKB Instruments).3,5
penicillamine (SNAP, NO donor), 1 ␮mol/L KT-5823 (protein
kinase G [PKG] inhibitor),20 100 nmol/L calphostin C (PKC inhib-
Semiquantitative PCR
itor),21 10 ␮mol/L glibenclamide (ATP-sensitive K⫹ [K⫹ATP] channel
Extracted mRNA (Dynal) was reversed-transcribed into cDNA using
blocker),22 1 ␮mol/L levcromakalim (K⫹ATP activator),23 or 30
oligo(dT18) plus random hexamers (10-mer) and M-MLV reverse
nmol/L wortmannin (PI3-k inhibitor).24 transcriptase (Promega) for 1 hour at 37°C.3 Polymerase chain
reactions (PCRs) were performed in 20-␮L samples (2 ␮L of 10⫻
Membrane Potential PCR buffer, 0.8 ␮L of 50 mmol/L Mg2⫹, 0.4 ␮L dNTPs, 13.6 ␮L
[3H]Tetraphenylphosphonium ([3H]TPP⫹) influx (46 nmol/L, 0.5 RNase-free H2O, 0.2 ␮L Taq DNA polymerase, and 0.5 ␮mol/L
␮Ci/mL, 15 to 120 seconds), a membrane potential–sensitive sequence-specific oligonucleotide primers for human CAT-1, CAT-
probe,2,3,25 was determined in Krebs solution containing 5 or 2A, or CAT-2B). Samples were incubated (4 minutes, 95°C),
25 mmol/L D-glucose, 100 ␮mol/L L-arginine, and 5.5 or followed by 35 cycles of 30 seconds at 95°C, 30 seconds at 57°C, 30
131 mmol/L KCl.2,3 Resting membrane potential (Em, whole-cell seconds at 72°C, and a final extension for 7 minutes at 72°C. ␤-Actin
patch clamp) was recorded using an EPC-7 amplifier (List Medical) expression was used as a reference value. Reverse transcription
as described.2 Em was measured for at least 1 minute in the presence (RT)-PCR products were sequenced in both directions by Taq
of 5 or 25 mmol/L D-glucose. Only recordings with ⬍0.1-mV dideoxy terminator cycle sequencing (automated DNA sequencer
variations were considered. Em was also determined in cells prein- 373A, Applied Biosystems).3
cubated for 2 hours with 10 mmol/L L-lysine. Oligonucleotide primers were as follows: hCAT-1 (sense)
5⬘-CCAGTACTTCCGACGAGTTAGA-3⬘, hCAT-1 (antisense)
5⬘-CATCCACACAGCAAACCGGACC-3⬘, hCAT-2A (sense) 5⬘-
PKC Activity TATCCCGATTTTTTTGCTGTGTGC-3⬘, hCAT-2A (antisense)
PKC activity (32P incorporation from [␥-32P]ATP into a synthetic 5⬘-TGCAGTCAACGTGGCAGCAACT-3⬘, hCAT-2B (sense) 5⬘-
PKC substrate peptide analogue5) was determined in cells exposed to TCCCAATGCCTCGTGTAATCTA-3⬘, hCAT-2B (antisense)
5 or 25 mmol/L D-glucose (2 minutes) after pretreatment with 100 5⬘-GCATGCTGAAGCCCTGTCTCTGC-3⬘, ␤-actin (sense) 5⬘-
nmol/L phorbol 12-myristate 13-acetate (5 minutes, PKC activator), AACCGCGAGAAGATGACCCAGATCATCTTT-3⬘, and ␤-actin
100 nmol/L 4␣-phorbol 12,13-didecanoate (5 minutes, less active (antisense) 5⬘-AGCAGCCGTGGCCATCTCTTGCTCGAAGTC-3⬘.
phorbol 12-myristate 13-acetate analogue), 100 nmol/L calphostin C Expected size products were as follows: hCAT-1, 450 bp; hCAT-2A,
(30 minutes), 10 ␮mol/L PD-98059 (30 minutes), 100 ␮mol/L 690 bp; hCAT-2B, 360 bp; and ␤-actin, 350 bp.
SNAP (5 minutes), 100 ␮mol/L L-NAME (30 minutes), or 30
nmol/L wortmannin (30 minutes). SOD Activity and ␣-Tocopherol Experiments
Cells were homogenized in buffer containing 50 mmol/L Tris-(hy-
cGMP Determination droxymethyl)-aminomethane, 100 mmol/L potassium chloride,
Cells preincubated (30 minutes) in Krebs solution (37°C) containing 0.02% Triton X-100, 100 mmol/L sodium pyrophosphate, and
L -arginine (100 ␮ mol/L) and 3-isobutyl-1-methylxanthine 100 mmol/L sodium fluoride (pH 7.4), which was supplemented with
(0.5 mmol/L, phosphodiesterase inhibitor),2,3 in the absence or trypsin inhibitors (4 mg/mL aprotinin, 1 mg/mL benzamidine, 5
presence of L-NAME (100 ␮mol/L), were exposed to 5 or ␮g/mL leupeptin, and 200 ␮mol/L sodium orthovanadate). Aliquots
25 mmol/L D-glucose for the last 2 minutes of the 30-minute (1 mg protein/mL) were incubated (25°C, 2 minutes) with potassium
phosphate buffer (50 mmol/L, pH 10.2) containing adrenochrome
incubation period with 3-isobutyl-1-methylxanthine. cGMP was
(200 ␮mol/L) and epinephrine (10 ␮mol/L), and absorbance was
determined in HCl-cell extracts by radioimmunoassay.2,3
measured at 480 nm. Superoxide dismutase (SOD) activity was
calculated from the inhibition curve for epinephrine auto-oxidation
L-Citrulline Assay versus protein concentration. Basal absorbance (100% activity) was
Cells were incubated with L-[3H]arginine (100 ␮mol/L, 4 ␮Ci/mL, the reaction in the absence of cell extracts.26 Cells were also
30 minutes, 37°C) in the absence or presence of L-NAME (100 preincubated (30 minutes) with ␣-tocopherol (500 ␮g/mL, 1%
␮mol/L). Cells were exposed to 5 or 25 mmol/L D-glucose for 30 ethanol).27
minutes or for the last 2, 4, 10, or 20 minutes of this incubation
period. Some experiments were performed in cells exposed only to Materials
L-[3H]arginine (4 ␮Ci/mL) and 5 or 25 mmol/L D-glucose. Formic Sera, agarose, and buffers were from GIBCO Life Technologies.
acid– digested cells (200 ␮L) were passed through a sodium ion form Collagenase type II (Clostridium histolyticum) was from Boehring-
of the cation ion-exchange resin Dowex 50W (50X8-200), and er-Mannheim, and Bradford protein reagent was from Bio-Rad
L-[3H]citrulline concentration was determined in the H2O eluate.3 Laboratories. L-NAME and SNAP were from Calbiochem. Ethidium
66 Circulation Research January 10/24, 2003

bromide, Dowex (50WX8-400), and all other reagents were from

Sigma Chemical Co. L-[2,3-3H]Arginine (36.1 Ci/mmol), D-[1-
14
C]mannitol (49.3 mCi/mmol), [␥-32P]ATP, and [3H]TPP⫹ (37
Ci/mmol) were from NEN. 3⬘,5⬘-cGMP-TME was from ICN. Anti-
bodies were from Cell Signaling, New England Biolabs.

Statistical Analysis
Values are mean⫾SEM, where n indicates the number of different
cell cultures (4 to 8 replicates per experiment). Statistical analyses
were carried out on raw data using the Peritz F multiple-means
comparison test.28 A Student t test was applied for unpaired data, and
a value of P⬍0.05 was considered statistically significant.

Results
L-Arginine Transport
Elevated D-glucose (2 minutes), but not L-glucose or
D-mannitol, stimulated L-arginine transport (half-maximal
effect [K1/2] 13⫾2 mmol/L D-glucose) (Figure 1A). Basal
transport rates increased significantly after 30 seconds of
exposure to elevated D-glucose (K1/2 25⫾5 seconds), with
maximal rates achieved within 1 minute and sustained over 5
minutes (Figure 1B). Subsequent experiments were per-
formed using 25 mmol/L D-glucose for 2 minutes. D-Glu-
cose–stimulated L-arginine transport decreased to basal val-
ues within 5 minutes after reexposure of cells to 5 mmol/L
D-glucose (Figure 1B). RT-PCR analysis detected only
hCAT-1 and hCAT-2B mRNA in HUVECs (Figure 1C).
Elevated D-glucose had no effect on the nonsaturable
component (KD) of overall L-arginine transport but increased
Vmax, with no change in apparent Km (Figure 2A, Table 1).
Cell incubation with L-lysine (10 mmol/L, 2 hours) increased
6.5-fold the L-arginine transport in 5 mmol/L D-glucose
(Figure 2B). However, L-arginine transport was increased
2.9-fold in cells exposed to 25 mmol/L D-glucose (last 2
minutes of the 2-hour incubation with L-lysine) (Figure 2B).
D-Glucose stimulation and trans-stimulation by L-lysine of
⫹
L-arginine transport was unaltered in Na -free Krebs solution
(not shown).

TPPⴙ Influx and Membrane Potential

Elevated D-glucose increased TPP⫹ influx (1.8-fold) and
caused membrane hyperpolarization (Table 2). TPP⫹ influx
and L-arginine transport were inhibited by KCl-induced
membrane depolarization. Glibenclamide (K⫹ATP channel
blocker) blocked D-glucose–increased L-arginine transport
Figure 1. Effect of D-glucose on L-arginine transport and CAT
and changes in TPP⫹ influx and Em (Table 2). Levcromakalim expression. A, L-Arginine transport (100 ␮mol/L, 1 minute, 37°C)
(K⫹ATP channel activator) hyperpolarized the plasma mem- in HUVECs incubated (2 minutes) with increasing concentrations
brane and increased L-arginine transport and TPP⫹ influx only of D-glucose (䢇), L-glucose (䡩), or 5 mmol/L D-glucose plus 5 to
in 5 mmol/L D-glucose; effects were blocked by gliben- 20 mmol/L D-mannitol (▫). *P⬍0.05 and **P⬍0.04 vs 5 or
7.5 mmol/L D-glucose and corresponding values in L-glucose
clamide (Table 2). The effects of D-glucose were also blocked and D-glucose⫹D-mannitol. B, Time course of effect of
by wortmannin (not shown). Preloading cells with L-lysine D-glucose on L-arginine transport (as in panel A) in cells incu-
did not alter Em (⫺66.1⫾0.3 mV, P⬎0.05; n⫽29 cells) bated for 0 to 5 minutes in 5 mmol/L D-glucose (䡩), 25 mmol/L
D-glucose (䢇), or 5 mmol/L D-glucose⫹20 mmol/L D-mannitol (▫).
compared with control cells (⫺67.5⫾0.5 mV, P⬎0.05; n⫽45 High D-glucose– containing Krebs solution was replaced (arrows)
cells). L-Arginine transport was inhibited by KCl with half- by 5 mmol/L D-glucose, and cells were incubated for 10 min-
maximal inhibition at 12⫾2 and 17⫾4 mmol/L KCl for 5 and utes. *P⬍0.04 vs all other values. Values are mean⫾SEM
(n⫽13). C, RT-PCR for mRNA from cells in 5 mmol/L D-glucose.
25 mmol/L D-glucose, respectively. The time needed to mRNA was reversed-transcribed into cDNA, and PCR was per-
induce half-maximal inhibition of transport with 131 mmol/L formed for human CAT-1 (lane 2, 449 bp), CAT-2A (lane 3, 690
KCl was similar in cells in 5 mmol/L D-glucose (6.5⫾0.6 bp), or CAT-2B (lane 4, 357 bp). Lane 5 is ␤-actin (350 bp), and
minutes) compared with 25 mmol/L D-glucose (7.2⫾0.6 lane 1 is DNA ladder (100 to 2000 bp). Data are representative
of 14 cell cultures.
minutes).
 Flores et al D-Glucose Acutely Activates L-Arginine/NO Pathway 67

lation at Ser1177, cGMP, and L-[3H]citrulline formation. Sim-

ilar results were found in cells exposed to 25 mmol/L
D-glucose for the last 4, 10, or 20 minutes of the 30-minute
incubation period with L-[3H]arginine (not shown). Intracel-
lular Ca2⫹ in cells incubated with 25 mmol/L D-glucose for 2
minutes (42⫾7 nmol/L) was not statistically different
(P⬎0.05, n⫽125 cells) from values in cells incubated with
5 mmol/L D-glucose (35⫾5 nmol/L).
L-NAME also inhibited the effect of D-glucose on
⫹
L-arginine transport Vmax (Table 1), TPP influx, and Em
(Table 2); however, L-NAME did not alter TPP⫹ influx or Em
in 5 mmol/L D-glucose. Other experiments show that SNAP
(NO donor) induces TPP⫹ influx, L-arginine transport, and
membrane hyperpolarization only in 5 mmol/L D-glucose
(Table 2) and that dibutyryl cGMP (dbcGMP) increased
⫹
L-arginine transport (Figure 4A) and TPP influx (Figure 4B).
The effects of D-glucose and dbcGMP were blocked by
KT-5823, a PKG inhibitor (Table 2).

PKC and MAP Kinase Involvement

PKC activity was unaltered at up to 5 minutes of incubation
with high D-glucose (Figure 5A) or after the addition of
SNAP (not shown). Furthermore, calphostin C (PKC inhibi-
tor) had no effect on D-glucose–increased L-arginine transport
(Figure 5B), TPP⫹ influx, or NO synthesis (not shown).
However, longer incubation with elevated D-glucose (⬎10
Figure 2. Effect of D-glucose on kinetic parameters and trans- minutes) increased membrane PKC activity (not shown),
stimulation of L-arginine transport. A, Saturable L-arginine trans- confirming our previous observations in HUVECs.5 In con-
port (1 minute, 37°C) in HUVECs incubated (2 minutes) in trast, high D-glucose induced p42/p44mapk phosphorylation
5 mmol/L D-glucose (䡩) or 25 mmol/L D-glucose (䢇). B,
L-Arginine transport (100 ␮mol/L, 1 minute, 37°C) in cells prein-
(Figure 6A), an effect blocked by PD-98059 and mimicked
cubated (2 hours) in medium 199 in the absence (control) or by brief exposure (2 minutes) to SNAP (Figure 6B). Inter-
presence of 10 mmol/L L-lysine. Transport assays were per- estingly, the effect of D-glucose on p42/44mapk phosphoryla-
formed in cells exposed to 5 mmol/L D-glucose (open bars) or tion was blocked by wortmannin (Figure 6C). PD-98059 also
25 mmol/L D-glucose (filled bars) for the last 2 minutes of the
2-hour incubation period with L-lysine. Values are mean⫾SEM blocked the stimulatory effects of elevated D-glucose and
(n⫽16). *P⬍0.04 vs all other values. SNAP on TPP⫹ influx, L-arginine transport, and Em (Table 2).

NO Involvement SOD Activity and ␣-Tocopherol Effect

Elevated D-glucose increased eNOS phosphorylation at SOD activity in cells in 5 mmol/L D-glucose (5.5⫾0.6 U/mL)
Ser1177 (Figure 3A), L-[3H]citrulline (Figure 3B), and cGMP was not significantly altered (P⬎0.05, n⫽5) by 25 mmol/L
accumulation (Figure 3C). L-NAME inhibited the effect of D-glucose (6⫾1 U/mL). Extracellular SOD or ␣-tocopherol
3
D-glucose on cGMP and L-[ H]citrulline formation but did not did not block (P⬎0.05, n⫽4 to 8) the effect of D-glucose on
alter eNOS-Ser 1177
phosphorylation (not shown). However, L-arginine transport (5.6⫾0.3 and 5.8⫾0.6 pmol/␮g protein
wortmannin inhibited D-glucose–induced eNOS phosphory- per minute, respectively), L-citrulline formation (3.1⫾0.2 and

TABLE 1. D-Glucose Effect on L-Arginine Transport in HUVECs

Vmax /Km, pmol/␮g KD, pmol/␮g
Vmax, pmol/␮g Protein per Minute Protein per Minute
Km, ␮mol/L Protein per Minute per ␮mol/L per ␮mol/L
5 mmol/L D-Glucose 105⫾9 4.4⫾0.1 0.044⫾0.011 0.002⫾0.0007
PD-98059 123⫾12 4.2⫾0.2 0.034⫾0.012 0.003⫾0.0012
L-NAME 97⫾16 3.8⫾0.3 0.039⫾0.016 0.003⫾0.0009
25 mmol/L D-Glucose 111⫾16 6.1⫾0.7* 0.055⫾0.012* 0.003⫾0.0011
PD-98059 133⫾22 3.7⫾0.3† 0.028⫾0.009† 0.004⫾0.0012
L-NAME 125⫾22 3.6⫾0.5† 0.029⫾0.007† 0.003⫾0.0008
L-Arginine transport (100 ␮mol/L, 37°C) in cells preincubated (30 minutes) with 5 mmol/L D-glucose, in absence
or presence of 100 ␮mol/L PD-98059 or 100 ␮mol/L L-NAME, and exposed (2 minutes) to Krebs containing 5 or
25 mmol/L D-glucose. Values are mean⫾SEM, n⫽6.
*P⬍0.05, †P⬍0.05 vs 5 and 25 mmol/L D-glucose, respectively.
68 Circulation Research January 10/24, 2003

TABLE 2. D-Glucose Effect on L-Arginine Transport, [3H]TPPⴙ Influx, and Em in HUVECs

L-Arginine Transport, pmol/␮g [3H]TPP⫹ Influx, pmol/mg
Protein per Minute Protein per Minute Em, mV
5 mmol/L D-Glucose
Control 1.8⫾0.3 2.1⫾0.3 ⫺67.2⫾0.5
KCl 0.2⫾0.1* 0.1⫾0.04* ⫺8.1⫾0.5*
Glibenclamide 1.3⫾0.3 1.5⫾0.3 ⫺63.9⫾0.6
Levcromakalim 4.6⫾0.3* 5.6⫾0.3* ⫺74.2⫾0.3*
Levcromakalim⫹glibenclamide 1.6⫾0.2† 1.4⫾0.6† ⫺64.5⫾0.2†
PD-98059 2.1⫾0.1 1.7⫾0.2 ⫺64.2⫾0.5
L-NAME 1.6⫾0.1 1.5⫾0.3 ⫺65.1⫾0.3
SNAP 5.1⫾0.3* 3.9⫾0.4* ⫺76.2⫾1.0*
SNAP⫹PD-98059 2.1⫾0.3† 1.9⫾0.5† ⫺65.3⫾1.0†
KT-5823 1.9⫾0.5 1.9⫾0.3 ⫺66.8⫾0.9
25 mmol/L D-Glucose
Control 4.0⫾0.5* 3.7⫾0.3* ⫺77.1⫾0.2*
KCl 0.3⫾0.2*‡ 0.3⫾0.03*‡ ⫺6.2⫾0.5*‡
Glibenclamide 2.3⫾0.4‡ 2.4⫾0.4‡ ⫺62.7⫾0.5‡
Levcromakalim 4.3⫾0.4* 4.7⫾0.2* ⫺75.1⫾0.3*
Levcromakalim⫹glibenclamide 2.2⫾0.3†‡ 1.8⫾0.3†‡ ⫺61.9⫾0.3†‡
PD-98059 0.9⫾0.3‡ 1.7⫾0.1‡ ⫺69.9⫾1.0‡
L-NAME 1.4⫾0.1‡ 1.8⫾0.2‡ ⫺65.7⫾0.7‡
SNAP 4.1⫾0.4* 3.2⫾0.2* ⫺78.2⫾1.0*
SNAP⫹PD-98059 1.1⫾0.5†‡ 1.9⫾0.2†‡ ⫺65.7⫾1.0†‡
KT-5823 1.9⫾0.5‡ 1.9⫾0.3‡ ⫺65.3⫾0.5‡
⫹
L-Arginine transport (100 ␮mol/L), 关 H兴TPP influx (46 nmol/L), and Em (whole-cell patch clamp) in cells
3

preincubated (30 minutes) with 5 mmol/L D-glucose and 5.5 mmol/L (control) or 131 mmol/L KCl. Cells in 5.5 mmol/L
KCl were preincubated with 10 ␮mol/L glibenclamide (5 minutes), 1 ␮mol/L levcromakalim (5 minutes), 10 ␮mol/L
PD-98059 (30 minutes), 100 ␮mol/L L-NAME (30 minutes), 10 ␮mol/L KT-5823 (30 minutes), or 100 ␮mol/L SNAP
(2 minutes) and exposed (2 minutes) to 5 or 25 mmol/L D-glucose. Values are mean⫾SEM, n⫽17.
*P⬍0.05 vs control in 5 mmol/L D-glucose; †P⬍0.05 vs corresponding values in SNAP or levcromakalim; ‡P⬍0.05
vs control in 25 mmol/L D-glucose.

2.8⫾0.4 pmol/␮g protein per 30 minutes, respectively), TPP⫹ confirmed here, is Na⫹ independent and inhibited by mem-
influx (4.4⫾0.2 and 4.1⫾0.3 pmol/mg protein per minute, brane depolarization.2,3,32 Because of their similar kinetic
respectively), or changes in Em (⫺76⫾0.3 and ⫺78⫾0.3 mV, properties, CAT-1 and CAT-2B are hard to distinguish at the
respectively). functional level.31 Our results show that both high-affinity
hCAT-1 (Km ⬇100 to 200 ␮mol/L) and hCAT-2B (Km ⬇200
Discussion to 400 ␮mol/L), but not the low-affinity hCAT-2A trans-
The present study establishes that D-glucose induces a rapid porter (Km ⬇2 to 5 mmol/L), are present in HUVECs,
concentration-dependent stimulation of L-arginine transport confirming previous reports.3,33 CAT-1 is more sensitive than
in HUVECs. This effect requires NO synthesis associated CAT-2B to trans-stimulation by cationic amino acids.3,31,34
with increased phosphorylation of eNOS at Ser1177 and acti- When we preloaded HUVECs with L-lysine, L-arginine trans-
vation of p42/p44mapk and PI3-k, and it is independent of PKC port was increased by ⬇7-fold in 5 mmol/L D-glucose.
and intracellular Ca2⫹ changes. These findings provide the However, the L-lysine trans-stimulatory effect was less ef-
first evidence that short-term hyperglycemia activates the fective (⬇3-fold) in cells exposed for 2 minutes to 25 mmol/L
L -arginine/NO signaling pathway in human fetal D-glucose. Because L-arginine transport is trans-stimulated
endothelium. by 9.8-fold or 1.8-fold in Xenopus oocytes injected with
⫹
L-Arginine transport is mediated by systems y /CATs,2,3,12 hCAT-1 or hCAT-2B mRNA, respectively, 34 trans-
y⫹L,29,30 and b0,⫹12 in HUVECs, with the first likely predom- stimulation in HUVECs in 5 mmol/L D-glucose may be
inating at the physiological concentration of extracellular preferentially mediated by hCAT-1. The reduced trans-
⫹
L-arginine. The cDNAs for four potential human y trans- stimulation of transport in high D-glucose may result from a
porters (hCAT-1, hCAT-2B, hCAT-2A, and hCAT-4) have state of maximal activity of L-arginine transporters already
been sequenced.31 L-Arginine transport in HUVECs occurs induced by L-lysine; therefore, high D-glucose could not
with relatively high affinity (Km ⬇80 to 100 ␮mol/L) and, as further increase L-arginine transport. In addition, the possi-
 Flores et al D-Glucose Acutely Activates L-Arginine/NO Pathway 69

Figure 4. cGMP involvement in the effect of D-glucose on

L-arginine transport and [3H]TPP⫹ influx. HUVECs preincubated
(30 minutes) with 5 mmol/L D-glucose in the absence (control) or
presence of KT-5823 or dbcGMP were exposed (2 minutes) to
5 mmol/L D-glucose (open bars) or 25 mmol/L D-glucose (filled
bars) in the presence of KT-5823 and/or dbcGMP, and
L-arginine transport (100 ␮mol/L, 1 minute, 37°C) (A) and
[3H]TPP⫹ influx (46 nmol/L, 1 minute, 37°C) (B) were determined.
Values are mean⫾SEM (n⫽12). *P⬍0.05 vs all other values.

Vmax for L-arginine transport without altering the apparent Km.

As noted above, L-arginine transport is sensitive to changes in
extracellular K⫹ and Em. Because high D-glucose induced
membrane hyperpolarization, stimulation of L-arginine trans-
port could result from changes in Em. The stimulatory effect
Figure 3. Effect of D-glucose on eNOS activity. A, Immunoblot of D-glucose on TPP⫹ influx and L-arginine transport was
for total eNOS and phosphorylated eNOS at Ser1177 (eNOSⵑP- blocked by glibenclamide, a K⫹ATP channel blocker. In addi-
Ser1177) in HUVECs preincubated (30 minutes) with 5 mmol/L
D-glucose in the absence or presence of wortmannin and then tion, levcromakalim (K⫹ATP activator)23 mimics D-glucose–
exposed (2 minutes) to 5 or 25 mmol/L D-glucose (see Materials induced changes in Em, L-arginine transport, and TPP⫹ influx.
and Methods). Top panel shows densitometry ratios. Data are K⫹ATP channels are expressed in the endothelium35 and are
representative of 6 cell cultures. B, L-[3H]Citrulline formation
from L-[3H]arginine (4 ␮Ci/mL, 37°C) in cells preincubated (30 activated by D-glucose36; thus, the effects of D-glucose may
minutes) with Krebs solution in the absence (control) or pres- involve changes in the activity of glibenclamide-sensitive
ence of L-NAME and exposed for the last 2 minutes to Krebs K⫹ATP channels.
solution containing 5 mmol/L D-glucose (open bars) or
D-Glucose (24 hours) also increases eNOS expression4 and
25 mmol/L D-glucose (filled bars). C, Intracellular cGMP under
same experimental conditions as in panel B. Values are activity2 in HUVECs. In the present study, eNOS activity was
mean⫾SEM (n⫽17). *P⬍0.05 vs all other values. increased in HUVECs exposed for 1 to 5 minutes to high
D-glucose, an effect associated with increased phosphoryla-
bility that L-lysine–stimulated L-arginine transport was due to tion of eNOS at Ser1177, a residue known to be associated with
membrane hyperpolarization is unlikely because Em was eNOS activation.37 The rapid eNOS stimulatory effect of
unaltered in L-lysine–preloaded cells. D-glucose was not further increased by longer incubations
Long-term incubation (24 hours) of HUVECs with high periods with D-glucose, which could be due to a maximal and
D-glucose increases Vmax for L-arginine transport.4 The pres- sustained activation of eNOS acutely induced by elevated
ent study shows that acute (2-minute) D-glucose increases D-glucose. Because D-glucose did not alter basal intracellular
70 Circulation Research January 10/24, 2003

Figure 5. Lack of effect of calphostin C in the effect of

D-glucose on L-arginine transport. A, PKC activity in cytosolic
(open symbols) and membrane (filled symbols) fractions from
HUVECs preincubated (30 minutes) with Krebs solution without
(䡩, 䢇) or with 100 nmol/L calphostin C (▫, 䡲). Cells were incu-
bated (30 seconds) in 25 mmol/L D-glucose in the absence or
presence of calphostin C. B, Time-course effect of 25 mmol/L
D-glucose on L-arginine transport (100 ␮mol/L, 1 minute, 37°C)
(as in panel A) in Krebs solution without (䡩) or with 100 nmol/L
calphostin C (䢇). Values are mean⫾SEM (n⫽18). *P⬍0.05 vs all
other values.

Ca2⫹, it is likely that rapid eNOS activation by D-glucose is

Ca2⫹ independent, supporting recent observations of Ca2⫹- Figure 6. NO and PI3-k involvement in effect of D-glucose on
independent eNOS activation in HUVECs.37,38 D-Glucose– p42/44mapk phosphorylation. A, Immunoblot for phosphorylated
p44mapk (p44ⵑP) and p42mapk (p42ⵑP) and nonphosphorylated
induced phosphorylation of eNOS at Ser1177, L-arginine trans- p44mapk (p44) or p42mapk (p42) in HUVECs preincubated (30 min-
port, and TPP⫹ influx were blocked by wortmannin, utes) with 5 mmol/L D-glucose in the absence or presence of
suggesting that the PI3-k pathway could be involved in the PD-98059 and then exposed (2 minutes) to 5 or 25 mmol/L
D-glucose. B, Effect of SNAP on p42/p42mapk in HUVECs in
effects of D-glucose. L-Arginine transport could be determi-
5 mmol/L D-glucose. C, Effect of wortmannin on p42/p42mapk in
nant for eNOS activity11; however, the possibility that the HUVECs (as in panel A). Data are representative of similar
D-glucose–induced increase of L-citrulline production was results in 17 cell cultures.
due to elevated L-arginine transport seems unlikely, inasmuch
as D-glucose–induced NO synthesis was unaltered in the synthase.2 NO causes membrane hyperpolarization in the
absence of extracellular L-arginine. endothelium,3,36 and NO (from SNAP) has been shown to
L-NAME blocked D-glucose–increased L-arginine trans- cause comparable increases in TPP⫹ influx and L-arginine
port and NO synthesis in HUVECs. This inhibitor does not transport and to cause membrane hyperpolarization to those
alter basal L-arginine transport in the endothelium2– 4,39; thus, caused by D-glucose, although SNAP treatment did not
NO most likely mediates changes in L-arginine transport, as further enhance the effects of high D-glucose, in HUVECs.
suggested in bovine aortic endothelium.40 This result is These findings support the hypothesis that NO acutely mod-
similar to that found in HUVECs from patients with gesta- ulates L-arginine transport by a mechanism that involves
tional diabetes; L-arginine transport in these cells is increased membrane hyperpolarization.
concomitantly with membrane hyperpolarization and NO NO-altered K⫹ channel activity may occur by both indirect
synthesis, and this increase is inhibited by blocking NO mechanisms via cGMP and direct NO action on channels.36
 Flores et al D-Glucose Acutely Activates L-Arginine/NO Pathway 71

Our results show that the D-glucose increases in L-arginine fellowships. We thank the midwives of the Hospital Regional-
transport and TPP⫹ influx were mimicked by dbcGMP and Concepción (Chile) labor ward for the supply of umbilical cords and
Isabel Jara for secretarial assistance.
blocked by the PKG inhibitor KT-5823. In addition, D-
glucose–induced membrane hyperpolarization was also
blocked by KT-5823. Thus, modulation of ion channel
References
1. Moncada S, Palmer RMJ, Higgs EA. Nitric oxide: physiology, patho-
activity (and hence, L-arginine transport) could be due to the physiology and pharmacology. Pharmacol Rev. 1991;43:109 –142.
activation of PKG downstream from NO synthesis. 2. Sobrevia L, Cesare P, Yudilevich DL, Mann GE. Diabetes-induced acti-
PKC activity is increased in subjects with diabetes mellitus vation of system y⫹ and nitric oxide synthase in human endothelial cells
association with membrane hyperpolarization. J Physiol (Lond). 1995;
or in endothelium chronically exposed to high 489:183–192.
D-glucose.12,16,17 Activation of diacylglycerol/phorbol ester– 3. Casanello P, Sobrevia L. Intrauterine growth retardation is associated
sensitive PKC isozymes activates eNOS and the NO- with reduced activity and expression of the cationic amino acid transport
dependent increased p42/p44 mapk phosphorylation in systems y⫹/hCAT-1 and y⫹/hCAT-2B, and lower activity of nitric oxide
synthase in human umbilical vein endothelial cells. Circ Res. 2002;91:
HUVECs exposed for 24 hours to high D-glucose.5 However, 127–134.
25 mmol/L D-glucose for 1 to 5 minutes did not alter PKC 4. Sobrevia L, Nadal A, Yudilevich DL, Mann GE. Activation of L-arginine
activity in this cell type, suggesting that the rapid D-glucose transport (system y⫹) and nitric oxide synthase by elevated glucose and
effect on L-arginine transport was PKC independent. Because insulin in human endothelial cells. J Physiol (Lond). 1996;490:775–781.
mapk 5. Montecinos VP, Aguayo C, Flores C, Wyatt AW, Pearson JD, Mann GE,
D-glucose induces a rapid (2-minute) p42/p44 phosphor- Sobrevia L. Regulation of adenosine transport by D-glucose in human
mapk
ylation and because inhibition of the p42/p44 phosphory- fetal endothelial cells: involvement of nitric oxide, protein kinase C and
lation by PD-98059 also inhibits the D-glucose increase in mitogen-activated protein kinase. J Physiol (Lond). 2000;529:777–790.
6. Cosentino F, Hishikawa K, Katusic ZS, Lüscher TF. High glucose
TPP⫹ influx and L-arginine transport, it is likely that p42/
increases nitric oxide synthase expression and superoxide anion gen-
p44mapk activation is involved in this pathway. Activation of eration in human aortic endothelial cells. Circulation. 1997;96:25–28.
p42/44mapk requires PI3-k activity in HUVECs.41 Our results 7. Beckman JA, Goldfine AB, Gordon MB, Garrett LA, Creager MA.
show that D-glucose–induced p42/44mapk phosphorylation is Inhibition of protein kinase C␤ prevents impaired endothelium-dependent
vasodilation caused by hyperglycemia in humans. Circ Res. 2002;90:
blocked by wortmannin, suggesting that the D-glucose effect 107–111.
requires PI3-k activity in HUVECs. SNAP-increased p42/ 8. Duckrow RB. Decreased cerebral blood flow during acute hypergly-
p44mapk phosphorylation and L-arginine transport were caemia. Brain Res. 1995;703:145–150.
blocked by PD-98059, complementing results showing that 9. Wascher TC, Bachernegg M, Kickenweiz A, Stark G, Stark U, Toplak H,
Graier WF. Involvement of the L-arginine-nitric oxide pathway in
NO, via cGMP, causes rapid p42/p44mapk phosphorylation in hyperglycaemia-induced coronary artery dysfunction of isolated guinea
the endothelium.4,42 Because D-glucose–induced and NO- pig hearts. Eur J Clin Invest. 1996;26:707–712.
induced membrane hyperpolarization are blocked by PD- 10. Cipolla MJ, Porter JM, Osol G. High glucose concentration dilate cerebral
98059, p42/p44mapk activation could modulate ion channel arteries and diminish myogenic tone through an endothelial mechanism.
Stroke. 1997;28:405– 411.
activity and L-arginine transport in HUVECs. 11. Sobrevia L, Mann GE. Dysfunction of the nitric oxide signalling pathway
Elevated D-glucose leads to overproduction of oxygen- in diabetes and hyperglycaemia. Exp Physiol. 1997;82:1–30.
derived free radicals in several cell types.11,16,17 We found 12. Pan M, Wasa M, Lind DS, Gertler J Abbott W, Souba WW. TNF-
that SOD activity in HUVECs was unaltered by 25 mmol/L stimulated arginine transport by human vascular endothelium requires
activation of protein kinase C. Ann Surg. 1995;221:590 – 601.
D-glucose and that SOD or ␣-tocopherol did not block the 13. Racké K, Hey C, Mössner J, Hammermann R, Stichnote C, Wessler I.
effects of D-glucose, suggesting that short-term incubation Activation of L-arginine transport by protein kinase C in rabbit, rat and
with elevated D-glucose would not generate enough oxygen- mouse alveolar macrophages. J Physiol (Lond). 1998;551:813– 825.
14. Haneda M, Araki S, Togawa M, Sugimoto T, Isono M, Kikkawa R.
derived free radicals to induce changes in the L-arginine/NO
Mitogen-activated protein kinase cascade is activated in glomeruli of
pathway in HUVECs. diabetic rats and glomerular mesangial cells cultured under high glucose
The present study has established that high D-glucose conditions. Diabetes. 1997;46:847– 853.
rapidly activates the L-arginine/NO pathway in HUVECs. 15. Natarajan R, Scott S, Bai W, Yernerni K, Nadler J. Angiotensin II
signalling in vascular smooth muscle cells under high glucose conditions.
The effect of D-glucose involves PI3-k– dependent, but PKC- Hypertension. 1999;33:378 –384.
independent and intracellular Ca2⫹–independent, eNOS and 16. Pieper GM. Review of alterations in endothelial nitric oxide production in
p42/44mapk activation. These results complement previous diabetes. Hypertension. 1998;31:1047–1060.
observations in animal models in which elevated plasma 17. De Vriese AS, Verbeuren TJ, Van de Voorde J, Lameire NH, Vanhoutte
PM. Endothelial dysfunction in diabetes. Br J Pharmacol. 2000;130:
D-glucose results in a rapid (seconds to minutes) vasodilata-
963–974.
tion8 –10 and observations of D-glucose–induced vasodilatation 18. Flores C, Sobrevia L. Acute effect of D-glucose on L-arginine transport
in humans.7 Local NO synthesis could be one mechanism by and nitric oxide synthesis in human fetal endothelium. J Physiol (Lond).
which rapid alterations in plasma D-glucose result in vasodi- 2000;523:104P. Abstract.
19. Kolch W. Meaningful relationships: the regulation of the Ras/Raf/
latation and may have important implications in diabetic MEK/ERK pathway by protein interactions. Biochem J. 2000;351:
patients, in whom plasma D-glucose concentrations may 289 –305.
change rapidly.11,12,16,17 20. Grider JR. Interplay of VIP and nitric oxide in regulation of the
descending relaxation phase of peristalsis. Am J Physiol. 1993;264:
G334 –G340.
Acknowledgments 21. Kobayashi E, Ando K, Nakano H, Iida T, Ohno H, Morimoto M, Tamaoki
This study was supported by Fondo Nacional de Ciencia y Tec- T. Calphostins (UCN-1028), novel and specific inhibitors of protein
nología (FONDECYT 1000354 and 7000354), Universidad de Con- kinase C, I: fermentation, isolation, physico-chemical properties and
cepción (DIUC 201.084.003-1) (Chile), and The Wellcome Trust biological activities. J Antibiot (Tokyo). 1989;42:1470 –1474.
(United Kingdom). P. Casanello, J. Parodi, and C. Aguayo hold Beca 22. Holzmann S, Kukovetz WR, Braida C, Poch G. Pharmacological inter-
Docente Universidad de Concepción and CONICYT-PhD (Chile) action experiments differentiate between glibenclamide-sensitive K⫹
72 Circulation Research January 10/24, 2003

channels and cyclic GMP as components of vasodilation by nicorandil. 33. Irie K, Tsukahara F, Fujii E, Uchida Y, Yoshioka T, He WR, Shitashige
Eur J Pharmacol. 1992;29:1–7. M, Murota S, Muraki T. Cationic amino acid transporter-2 mRNA
23. Katnik C, Adams DJ. Characterization of ATP-sensitive potassium induction by tumor necrosis factor-␣ in vascular endothelial cells. Eur
channels in freshly dissociated rabbit aortic endothelial cells. Am J J Pharmacol. 1997;339:289 –293.
Physiol. 1997;272:H2507–H2511. 34. Closs EI, Graf P, Habermeier A, Föstermann U. Interference of L-arginine
24. Isenovic E, Muniyappa R, Milivojevic N, Rao Y, Sowers JR. Role of analogues with L-arginine transport mediated by the y⫹ carrier hCAT-2B.
PI3-kinase in isoproterenol and IGF-1 induced ecNOS activity. Biochem Nitric Oxide. 1997;1:65–73.
Biophys Res Commun. 2001;285:954 –958. 35. Nilius B, Droogmans G. Ion channels and their functional role in vascular
25. Schilling WP. Effect of membrane potential on cytosolic calcium of endothelium. Physiol Rev. 2001;81:1415–1459.
bovine aortic endothelial cells. Am J Physiol. 1989;257:H778 –H784. 36. Sobey CG. Potassium channel function in vascular disease. Arterioscler
26. Misra HP. Adrenochrome assay. In: Greenwald RA, ed. Handbook of Thromb Vasc Biol. 2001;21:28 –38.
Methods for Oxygen Radical Research. Boca Raton, Fla: CRC Press; 37. Dimmeler S, Fleming I, Fisslthaler B, Hermann C, Busse R, Zeiher AM.
1989:237–241. Activation of nitric oxide synthase in endothelial cells by Akt-dependent
phosphorylation. Nature. 1999;399:601– 605.
27. Ganz MB, Seftel A. Glucose-induced changes in protein kinase C and
38. Shaul PW. Regulation of endothelial nitric oxide synthase: location,
nitric oxide are prevented by vitamin E. Am J Physiol. 2000;278:
location, location. Annu Rev Physiol. 2002;64:749 –774.
E146 –E152.
39. Bogle RG, Moncada S, Pearson JD, Mann GE. Identification of inhibitors
28. Harper JF. Peritz’ F test: BASIC program of a robust multiple comparison
of nitric oxide synthase that do not interact with the endothelial cell
test for statistical analysis of all differences among group means. Comput
L-arginine transporter. Br J Pharmacol. 1992;105:768 –770.
Biol Med. 1984;14:437– 445. 40. Ogonowski AA, Kaesemeyer WH, Jin L, Ganapathy V, Leibach FH,
29. Arancibia Y, Sobrevia L. Transport of L-arginine mediated by system y⫹L Caldwell RW. Effects of NO donors and synthase agonists on endothelial
in human fetal endothelial cells. J Physiol (Lond). 1999;517:47P. cell uptake of L-Arg and superoxide production. Am J Physiol. 2000;278:
Abstract. C136 –C146.
30. Sala R, Rotoli BM, Colla E, Visigalli R, Parolari A, Bussolati O, Gazzola 41. Lal BK, Varma S, Pappas PJ, Hobson RW II, Duran WN. VEGF
GC, Dall’Asta V. Two-way arginine transport in human endothelial cells: increases permeability of the endothelial cell monolayer by activation of
TNF-␣ stimulation is restricted to system y⫹. Am J Physiol. 2002;282: PKB/akt, endothelial nitric-oxide synthase, and MAP kinase pathways.
C134 –C143. Microvasc Res. 2001;62:252–262.
31. Devés R, Boyd CAR. Transporters for cationic amino acids in animal 42. Parenti A, Morbidelli L, Cui XL, Douglas JG, Hood JD, Granger HJ,
cells: discovery, structure and function. Physiol Rev. 1998;78:487–545. Ledda F, Ziche M. Nitric oxide is an upstream signal of vascular endo-
32. Bussolati O, Sala R, Astorri A, Rotoli BM, Dall’Asta V, Gazzola GC. thelial growth factor-induced extracellular signal-regulated kinase 1/2
Characterization of amino acid transport in human endothelial cells. Am J activation in postcapillary endothelium. J Biol Chem. 1998;273:
Physiol. 1993;265:C1006 –C1014. 4220 – 4226.