| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* INSERM U702; Hôpital Tenon; Université Pierre et Marie Curie, Paris, France;
Department of Biochemistry and Biophysics, UNC Lineberger Comprehensive Cancer Center, University of North Carolina School of Medicine, Chapel Hill, North Carolina, USA;
INSERM U637; Université Montpellier 1; CHU Arnaud de Villeneuve, France;
CNRS UMR 6188; Faculté de Médecine, Université d’Angers, Angers, France;
|| Division of Nephrology, Department of Medicine, Duke University and Durham Veterans Affairs Medical Center, Durham, North Carolina, USA; and
Université Pierre et Marie Curie, AP-HP; School of Medicine Saint-Antoine, Paris, France
2Correspondence. INSERM U702, Hôpital Tenon, Université Pierre et Marie Curie, 4 rue de la Chine, Paris 75020, France. E-mail: pierre-louis.tharaux{at}chusa.jussieu.fr
ABSTRACT
Endothelin-1 (ET-1), a powerful vasoconstrictor, is involved in vasospastic diseases such as coronary artery disease and subarachnoidal hemorrhage, as well as in renal and cardiovascular fibrotic remodeling. Transactivation of the epidermal growth factor receptor (EGFR) mediates ET-1 signaling in vascular smooth muscle cells (VSMCs) and isolated arteries. Moreover, EGFR is required for a full constrictive response to ET-1. However, the relevant mechanisms mediating EGFR transactivation in response to ET-1 have not been identified. The present study used isolated arteries and VSMCs to investigate the role of the EGFR ligand heparin binding-epidermal growth factor (HB-EGF) in ET-1-induced transactivation of EGFR, intracellular calcium mobilization, and VSMCs contraction. While baseline blood pressures were similar in HB-EGF-deficient and in wild-type littermate mice, the vasoconstrictor actions of ET-1 were attenuated in HB-EGF–/– animals. In isolated mouse carotid artery segments mounted in an arteriograph, ET-1 caused only a weak increase in isovolumetric tone in HB-EGF-deficient vessels, and this effect was mimicked by inhibition of EGFR tyrosine kinase or phosphoinositide 3-kinase (PI3K) in wild-type arteries with or without endothelium, indicating a specific role in VSMCs. EGFR or PI3K inhibitors had no effect on KCl-induced contraction, which was normal in HB-EGF-deficient mice. To confirm that the abnormal responses in HB-EGF-deficient mice were due to impaired EGFR signaling, we studied VSMCs from waved-2 (wa2) mice; these animals have a mutation causing a partial loss of function of EGFR tyrosine kinase activity. The ET-1-induced calcium peak was reduced by 30% in VSMCs from wa2 mice and from HB-EGF–/– mice. This effect was reproduced by preincubation of wild-type VSMCs with EGFR inhibitor AG1478 and PI3K inhibitors LY294002 and wortmannin. ProHB-EGF is bound to the cell membrane and released after cleavage by metalloproteinases; its action may contribute to effects of GPCR agonists on cell growth. Pretreatment of mouse VSMCs with batimastat, a metalloproteinase inhibitor, significantly attenuated ET-1-induced [Ca2+]i response in wild-type cells. Human proHB-EGF has been shown to be the endogenous receptor for Corynebacterium diphteriae toxin (DT). Mutated DT toxin (CRM197) is devoid of toxicity but it neutralizes HB-EGF binding to EGFR. Pretreatment of human VSMCs from internal mammary arteries with CRM197 significantly blunted ET-1-stimulated calcium transients. In conclusion, these findings suggest that the mechanism of ET-1-induced vasoconstriction involves HB-EGF-mediated transactivation of the EGFR. This functional cascade requires modulation of agonist-induced calcium transient by EGFR and PI3K with extremely fast kinetics, suggesting a novel paradigm for GPCR-mediated calcium signaling, which may offer future therapeutic targets.—Chansel, D., Ciroldi, M., Vandermeersch, S., Jackson, L. F., Gomez, A-M., Henrion, D., Lee, D. C., Coffman, T. M., Richard, S., Dussaule, J-C., Tharaux, P-L., Heparin binding EGF is necessary for vasospastic response to endothelin.
Key Words: EGF receptor • calcium • HB-EGF • transactivation • PI3-kinase • human
IN RECENT YEARS, A ROLE for crosstalk between G-protein-coupled receptors (GPCRs) and receptor tyrosine kinases (RTKs) in cellular responses has been well documented (1
2
3)
. This interaction allows GPCRs to take advantage of pathways downstream of RTKs to influence cell function. For example, GPCR signaling systems require epidermal growth factor receptor (EGFR) tyrosine kinase activation to trigger mitogen-activated protein kinase activation and transmission of mitogenic signals to the nucleus. Crosstalk between the EGFR and the endothelin-1 (ET-1) type A receptor (ETA, a GPCR) has been demonstrated in rat-1 cells and in vascular smooth muscle cells (VSMCs) (1)
. This cooperation between a GPCR and EGFR mediates cell growth (4)
and collagen I synthesis (5)
via an MAPK-dependent pathway. In addition, we recently reported that potent aortic contractile responses to ET-1 are suppressed and even reversed by inhibition of EGFR tyrosine kinase activity (5)
; however, the mechanism underlying the EGFR involvement in this critical contractile response to ET-1 remains unclear. There are at least two possible mechanisms that might be involved in this process. First, binding of ET-1 to the ET receptors in VSMC triggers phospholipase C-dependent production of 1,4,5-inositol trisphosphate (IP3), which binds to its receptor in sarcoplasmic reticulum to initiate Ca2+ release from this intracellular store (6
7
8
9)
. The resulting increase in [Ca2+]i may provoke EGFR transactivation via a PYK2 or Src-related tyrosine kinase (10)
, with subsequent tyrosine phosphorylation events promoting contraction by a mechanism such as increasing the Ca2+ sensitivity of the contractile apparatus through activation of the Rho-Rho kinase pathway (11)
. Alternatively, ET-1 binding to the ET receptors may cause transactivation of the EGFR, which then increases [Ca2+]i as might occur as a result of phospholipase C
activation (9)
. An important difference between these two sequences is that the first one involves [Ca2+]i-induced stimulation of EGFR-mediated events that favor contraction without further alteration of intracellular Ca2+ homeostasis, whereas the second case requires EGFR tyrosine kinase activity to achieve the full [Ca2+]i response to ET-1. To distinguish whether EFGR modulates the full [Ca2+]i response to ET-1, we designed experiments using pharmacological and genetic interventions to address the hypothesis that rapid activation of EGFR tyrosine kinase activity by heparin binding epidermal growth factor EGF (HB-EGF) contributes to ET-1–induced [Ca2+]i responses in VSMCs. Next, we tested the efficiency of pharmacological inhibitors of phosphoinositide 3-kinase (PI3K), a known downstream mediator of EGFR or ETA signaling, to alter ET-1-induced calcium transients and vasocontraction. Our studies also evaluated whether HB-EGF is required for ET-1-induced intracellular calcium signals in human VSMCs.
MATERIALS AND METHODS
Materials
The following reagents were used: AG1478 and LY294002 (Calbiochem, San Diego, CA, USA), ET-1, BQ123, BQ788, sarafotoxin 6c, wortmannin, and CRM197 (Sigma, St. Louis, MO, USA). Bosentan was a generous gift from Dr. Martine Clozel (Actelion Ltd., Basel, Switzerland). None of the pharmacological agents exhibited autofluorescence at the excitation and emission wavelengths used for Fura-2 acetoxymethyl ester (AM) and for fluo-4 AM (Calbiochem). Batimastat was obtained from British Biotech, antiphospho-EGFR (Tyr-1068); pan anti-EGFR antibodies were purchased from Cell Signaling (Danvers, MA, USA) (2234).
Animals
Waved-2 (wa2) mice, which contain a point mutation in EGFR that reduces receptor tyrosine kinase activity by > 90%, were studied (12)
. These mice have a mild phenotype (wavy coat, curly whiskers, and runted stature) and normally developed vessels and kidneys (5
,12
13
14)
. They were bred on the original genetic background purchased from the Jackson Laboratory (Bar Harbor, ME, USA).
HB-EGF-deficient mice were studied. This strain has recently been generated and described in detail by L. Jackson et al. (15)
. All animals studied were between 8 and 14 wk old. The protocol followed the French Department of Agriculture and the National Institute of Health guidelines for animal care and protection. For both strains, all experiments used age- and sex-matched wild-type and genetically altered animals with similar genetic background (C57Bl/6 x Sv129/J mixed background).
Systolic blood pressure measurements in conscious mice
Systolic blood pressures were measured in groups of HB-EGF+/+ (n=8), HB-EGF–/– (n=10), wild-type (n=20), and wa2 (n=20) mice by using a computerized tail-cuff system (Visitech Systems, Cary, NC, USA) as described previously (16)
. Before the study was initiated, mice were adapted to the apparatus for at least 5 days. Systolic blood pressure was then measured for 10 consecutive days at the same time.
Intra arterial measurements of mean blood pressure in anesthetized mice
The procedure was adapted from previously described method (17)
. On the day of the study, animals were anesthetized with isoflurane and a flexible plastic catheter (0.020ID; Braintree Scientific, Braintree, MA, USA) was placed in the carotid artery to monitor arterial pressure. After completion of all tail-cuff measurements, intra-arterial BPs were determined on the mouse groups described above. Surgeries were performed in a uniform fashion by a single investigator (P-L T) between 9 AM and 2 PM. Mice were anesthetized with isoflurane (1.2%) and placed on an operating surface maintained at 38°C. A midline incision was made in the neck. With care taken to avoid the vagus nerve and carotid sinus, the left carotid artery was isolated below the level of the bifurcation and was tied off distally with 5.0 silk suture, and a vascular clamp was applied proximally. Approximately 3.5 mm of beveled Microrenathane catheter tubing (0.025 inch optical density (OD), 0.01 inch ID; Braintree Scientific) was inserted into the vessel so that its tip was approximately at the junction of the aorta and the carotid. The catheter was then firmly sutured in place. The catheter was flushed with heparinized (20 U/ml) PBS. A second catheter was placed in the jugular vein to infuse ET-1. Intra-arterial blood pressure was recorded continuously through the carotid catheter with a pressure transducer and Windaq acquisition and playback software (Dataq Instruments, Akron, OH, USA). Beginning with an equilibration period of 5 min, blood pressure was recorded continuously for the duration of the experiment at 50 recordings/s. Data collected from each animal were compressed to five recordings per second, and the mean arterial blood pressure (MAP) recordings from animals in each experimental group were integrated and averaged. Immediately after the equilibration period, mice received a bolus injection of 0.9% sodium chloride solution (NS) in a volume equal to 0.2% of their body weight. This is the volume used for each of the injections throughout the remainder of the experiment. At 20 min intervals thereafter, 3 nmol·kg–1 of ET-1 (Sigma Aldrich) was injected intravenously and MAP was monitored for an additional 30 min.
For each mouse, the mean arterial pressure and heart rate were defined as the average mean and rate values for all waveforms obtained during the recording session. Requirements for inclusion of data were a pulsatile waveform, minimum heart rate of 400 beats per minute.
Measurements of vasoconstriction
Carotid artery rings (2 mm length, 150–200 µm) were obtained from HB-EGF-deficient mice, and their respective littermate controls, and placed in physiological salt solution (PSS) on ice (pH 7.4) of the following composition, in mmol/l: NaCl, 140; KCl, 4.7; MgS047H20, 1.17; NaHC03, 5; KH2P04, 1.10, CaC12, 1 .O; HEPES, 50; and glucose (Glc), 5, pH 7.4. Carotid artery segments were mounted in the jaws of a wire myograph (Kent Scientific, Litchfield, CT, USA) to measure isometric tension. Isometric force generation was measured using described methods (5
,18)
. Once mounted, the PSS was warmed to 37°C and gassed with 95%/5% air/CO2, and the vessel segments were stretched to a predetermined length that was equivalent to an internal diameter of 200–250 µm and allowed to equilibrate for 30 min. Endothelium vitality was verified by measuring endothelium-dependent relaxation induced by acetylcholine (Ach) in precontracted vessels as described previously (18
, 19)
. After equilibration, the vessels were induced to contract with 125 mmol/l KCl until reproducible contractile responses were obtained (3–4 times). The effects of EGFR antagonism with AG1478 (500 nmol/l) or PI3K inhibitor LY2940052 (10 µmol/l) on ET-1-induced contraction were compared in HB-EGF+/+ and HB-EGF–/– carotids by adding the respective antagonists 2 min before ET-1. In addition, contraction to calcium was tested by adding cumulative concentrations of calcium in the bath of the arteries in the presence or absence of EGFR and PI3K inhibitors. The bath solution was then a calcium-free solution containing 125 mmol/l KCl. The contraction response to ET-1 in HB-EGF–/– and +/+ carotid artery segments was investigated at doses of ET-1 between 10–1 and 10–4 nmol/l in a noncumulative manner since the response to ET-1 was sustained and long-lasting. The magnitude of contraction was expressed as the percentage of the maximum force response to KCl (20)
or in millinewton (mN) (5)
.
Similarly, mouse aorta segments (that did not exhibit significant qualitative differences with carotids when challenged with ET-1) were studied with ET-1 (10 nmol/l) as described (5)
. After 2 min, PI3K inhibitor LY2940052 (10 µmol/l) or vehicle (DMSO) was added. Continuous recording of isometric force generation for the next 10 min was performed using previously described methods (5)
.
Cell culture
Mouse aortic smooth muscle cells were isolated from collagenase-treated aortas of male wild-type mice and of wa2 mice as described previously (21)
. They were grown in RPMI 1640 medium supplemented with 20% fetal calf serum (BioWest, Nuaillé, France) and used between passages 2 and 3. All cells were cultured in a humidified atmosphere containing 5% CO2 and growth arrested by 24 h of serum deprivation before experiments.
Human VSMCs from internal mammary arteries
Fragments of the internal mammary artery that would otherwise have been discarded were obtained from patients undergoing coronary artery bypass graft at CHU Arnaud de Villeneuve, Montpellier, France. The Institutional Ethical Committee of the Hospital approved the study. The cells were cultured according to a procedure described in detail before for human coronary VSMCs (22
, 23)
. In brief, after enzymatic dispersion, cells were maintained at 37°C in an air-CO2 incubator with Dubelcco’s minimum essential medium and Ham’s F-10 (1:1 v/v; Eurobio, Courtaboeuf, France) containing 5% human serum and 5% myoclone super plus (fetal bovine; Life Technologies, Gaithersburg, MD, USA). After reaching confluence, cells were trypsinized. Subcultured cells, used between passages 2 and 3, were plated in disposable borosilicate recording chambers (Lab-tek; Polylabo) at 500,000 cells/well. Cultures were subsubconfluent at the time of recording.
Dye loading and measurement of free [Ca2+]i
In suspended cells, a ratiometric method that is well adapted to quantitative assessment was used
Subconfluent cells were loaded for 40 min at 37°C with 5 µmol/l Fura-2/AM dissolved in phosphate buffer at pH 7,4 (135 mmol/l NaCl; 1 mmol/l Na2HPO4; 5 mmol/l KCl; 0. 5 mmol/l Mg2SO4; 1.8 mmol/l CaCl2; 10 mmol/l Glc; 10 mmol/l HEPES) supplemented with 1 mg/ml of BSA. After being washed, the cells were trypsinized, resuspended in the same buffer and transferred into a quartz cuvette under constant stirring at 37°C, and fluorescence was continuously monitored by a spectrofluorometer Quanta Master 1 (Photon Technology International, Birmingham, NJ, USA) before and after the addition of the agents to be tested. Fluorescence was measured at 510 nm after alternating excitation at 340 nm and 380 nm, and the 340/380 ratio was determined as a measure of intracellular free calcium as described previously (24)
. Both conditions were obtained by calibration with 0.02% Triton x-100 for the maximum value and 10 mmol/l ethylene glycol-bis-aminoethylether-N,N,N',N'-tetra-acetic acid (EGTA) for the minimum value. Felix software 1.1 program was used for [Ca2+]i calculation.
In adherent cells, confocal microscopy analysis, most suitable for imaging Ca2+ dynamics, was used
The vascular myocytes were plated in disposable borosilicate recording chambers (Lab-tek, Rochester, NY, USA) at 500,000 cells/well. Cultures were always subconfluent, with most cells being physically isolated. Cells were maintained at 37°C in an air-CO2 incubator. Twenty-four hours prior to the experiments, cells were incubated in the same culture media but without serum. At the time of the experiment, cells were loaded with the membrane permeant Ca2+-sensitive dye fluo-4 AM (5 µmol/l) from a stock fluo-4 AM solution dissolved in DMSO plus 20% Pluronic acid, as explained earlier (25)
during 15 min at 37°C in a humidified air atmosphere. Then cells were rinsed with a normal Tyrode solution (140 mmol/l NaCl; 4 mmol/l KCl; 1.1 mmol/l MgCl2; 10 mmol/l HEPES; 10 mmol/l Glc; 1.8 mmol/l CaCl2; pH=7.4 with NaOH) and 20 extra minutes were allowed to de-esterify the dye. Cells were superfused with the Tyrode solution or with the same solution supplemented with ET-1 during images acquisition. Images were recorded at 5 s intervals with a confocal microscope (Zeiss LSM 510) fitted with an argon ion laser (488 nm). Emission was collected at >505 nm. Images were corrected for the background fluorescence. Then the fluorescence values (F) were normalized to the basal fluorescence (before ET-1 application), F0, to obtain the fluorescence ratio (F/F0). Transient variations of [Ca2+]i were measured at their maximal amplitude.
EGFR phosphorylation on ET-1 or HB-EGF addition
Cultured aortic smooth muscle cells were serum starved for 24 h and stimulated with ET-1 (10 nmol/l) or HB-EGF (10 µg/l) in RPMI medium. After various periods (ranging from 15 s to 10 min), culture plates were put on ice and medium was quickly removed and replaced by ice-cold PBS. Cell were subsequently scraped in Phosphosafe® extraction buffer (Novagen, Madison, WI, USA) and homogenized with a dounce on ice. The tubes were spun, the supernatant isolated, and protein was assayed (Bradford method) for each sample. Seven micrograms of each sample were loaded in duplicate (with sample and reducing buffers (Invitrogen, Carlsbad, CA, USA) and run on in two separate NuPAGE 4/12% gels (Invitrogen), transferred to supported PVDF membranes, and subsequently exposed to either primary antiphopho-EGFR (Tyr-1068) or primary anti-EGFR antibody (Ab) (Cell Signaling) followed by secondary anti-rabbit IgG HRP-coupled Ab (Amersham, Arlington Heights, IL, USA) in 5% PBS-Tween buffers and detected by enhanced chemiluminescence (ECL)-plus on autoradiography films (Fuji). Densitometric analysis was used to quantitate protein levels and phosphorylated EGFR signal was normalized to total form.
PI3Kinase activity assay
PI3K activity was assessed with FACETM cell-based ELISA kit (Active Motif, Carlsbad, CA, USA) following the recommended procedure.
Statistical analysis
Results are expressed as mean ± SE. Comparisons between groups were analyzed by Mann-Whitney U test (2-sided) or ANOVA for experiments with more than two subgroups, followed by protected least significance difference Fisher’s test. Probability values of P < 0.05 were considered statistically significant. All analyses were performed with Statview software (SAS Institute Inc., Cary, NC, USA).
RESULTS
Roles of HB-EGF on baseline blood pressure and in ET-1-mediated rise of blood pressure
In the basal state on a 0.4% NaCl diet, systolic blood pressure as measured by tail cuff manometry was normal and comparable in conscious 8- to 10-wk-old HB-EGF–/– animals compared to homozygous HB-EGF+/+ controls (99±7 vs. 105±8 mmHg, respectively). The equivalence of blood pressures was also documented when intra-arterial pressures were measured directly in anesthetized animals (103±3 vs. 104±4 mmHg). In wild-type animals, acute infusion of ET-1 (0.3 µg/kg) provoked a marked pressor response. However, this response was significantly attenuated in HB-EGF–/– mice (maximum increase in mean blood pressure 10 min after ET-1 infusion (+12.5±5.8 vs. 1.7±5 mmHg, P<0.05) (Fig. 1
).
|
Selective alteration of vasocontractile response of HB-EGF-deficient arteries
In carotid artery segments from wild-type (WT) mice, ET-1 induced a robust contraction. However, the dose-response curve to ET-1 was markedly shifted to the right in HB-EGF–/– vessels (Fig. 1)
. EC50 were 42 ± 5 and 265 ± 56 nmol/l in HB-EGF+/+ and HB-EGF–/– carotids, respectively (P<0.01). Even at considerably high ET-1 concentrations (0.1–10 µmol/l), HB-EGF–/– carotids could not achieve full contraction. The maximal tone induced by ET-1 was 3-fold higher in HB-EGF+/+ than in HB-EGF–/– carotids (3.20±0.48 vs. 1.05±0.21 mN, respectively, P<0.01). By contrast, KCl produced robust contraction of similar magnitude in segments arteries from HB-EGF+/+ and –/– mice (not shown) and, more generally, on EGFR pharmacological blockade with tyrphostin AG1478 (Fig. 2
).
|
The endothelium does not affect the contractile action of HB-EGF
Since endothelin receptors on the endothelium may modulate the vasoconstrictive action of ET-1, we compared the effect of endothelial denudation on ET-1-stimulated HB-EGF–/– and control carotids. As depicted in Fig. 3
, the endothelium blunted the rise in isovolumetric tone at 0.1 nmol/l [ET-1] only in both groups (P<0.05). Removal of the endothelium did not influence the depressed response observed in HB-EGF–/– arteries, suggesting that the source of HB-EGF and its actions are mostly restricted to VSMCs.
|
Influence of HB-EGF deficiency on the kinetics of the vasoconstrictive response
ET-1-induced vasoconstrictive response was markedly diminished in carotid artery segments lacking HB-EGF for each time point within the first 10 min after addition of ET-1 (10 nmol/l) (P<0.05 for each comparison from 2 to 10 min, P<0.01 for global comparison with 2-way ANOVA, n=8 per group). Within the first 5 min of ET-1 stimulation, HB-EGF+/+ carotids exhibited a sustained isovolumetric tone that was 3- to 4-fold higher than HB-EGF–/– arteries (32±7 vs. 12±9% at 1 min and 73±14 vs. 17±11% of maximal contraction at 5 min for controls and HB-EGF-deficient carotids, respectively) (Fig. 4
).
|
Role of EGFR and PI3Kinase in ET-1-mediated vasoconstriction
HB-EGF binds to EGFR that transduces numerous cellular signaling cascades (26
, 27)
. Therefore, we tested whether acute EGFR pharmacological blockade would impair vasocontraction to ET-1 and confirmed previous data obtained in mouse aorta (5)
(Fig. 3)
. This result ruled out the involvement of a developmental action of congenital HB-EGF deficiency or EGFR loss of activity in ET-1 response in the HB-EGF–/– and waved-2 strains. Consistently, HB-EGF-deficient carotids and aortas were of normal size and structure (not shown).
Next, preincubation of carotids with PI3K inhibitor LY294002 for 2 min prior to ET-1 addition prevented ET-1-induced rise in isovolumetric tone in normal but not in HB-EGF-deficient arteries within a wide range of ET-1 concentrations (P<0.01 vs. controls) (Fig. 3)
. This result indicates that PI3 kinase is critical for the full normal response to ET-1. Similar results were reproduced when a time course study was performed with LY294002 for 2 min prior to ET-1 addition. Again, PI3K inhibition prevented ET-1-induced rise in isovolumetric tone after 1 min and for up to 5 min (P<0.05, n=3 per condition) (Fig. 4)
whereas LY294002 had no effect on KCl-induced contraction (not shown). Next, reversal of ET-1-induced vascular contraction was tested using LY294002. Two minutes after stimulation of wild-type mouse aorta rings with ET-1 (0.1 µmol/l), isovolumetric tone reached a plateau (76±12 vs. 16±0.8% of KCl-induced contraction at 2 min) that progressively recovered to baseline 3 min after inhibition of PI3K with LY294002 (10 µmol/l) (0.48±0.04% of KCl-induced contraction after T0+5 min). By contrast, aorta incubated with ET-1 and vehicle alone, which still displayed a tonic force plateau after 5 min (80±0.04% of KCl-induced contraction, P<0.05 for ET-1 and DMSO alone vs. ET-1 and LY 294002 treated groups, respectively, n=6 per condition) (Fig. 4)
.
Role of EGFR transactivation on the ET-1-induced increase in [Ca2+]i in mouse aortic vascular smooth muscle cells (VSMCs)
Comparison of ET-1 [Ca2+]i response in VSMC from wild-type and EGFR kinase-deficient wa2 mice
Basal [Ca2+]i in suspended VSMCs from wa2 and from wild-type mice was the same: 278 ± 15 (n=37) vs. 275 ± 8 nmol/l (n=81). In both cases [Ca2+]i release was immediate after stimulation by 100 nmol/l ET-1 (time to peak<3 s). The [Ca2+]i response was biphasic, with a peak followed by a subsequent sustained phase. In VSMCs from wa2 mice, the peak value was significantly reduced: +373 ± 27 (n=37) vs. +487 ± 28 nmol/l above baseline (n=45) in wild-type cells (P<0.01) (Fig. 5
).
|
To analyze the same individual cells before and during ET-1 application (paired measurements), we used adherent VSMC and a confocal microscopy approach. Figure 6
A shows that the wild-type VSMCs (top) have a stronger fast response to ET-1 than the wa2 VSMCs (bottom). Figure 6B
illustrates the response to ET-1 of a single wt VSMC and a single wa2 VSMC over time and it is clear that wa2 VSMC response to ET-1 is slower and depressed. Data are summarized in Fig. 6C
and show that the maximum increase in [Ca2+]i in response to ET-1 is reduced in wa2 VSMCs both at 1 nmol/l (F/F0: 1.24±0.06 in 25 wt vs. 1.10±0.02 in 12 wa2 cells, P<0.05) and at 30 nmol/l (which induced maximal response in both groups) (F/F0: 2.99±0.27 in 19 wt vs. 1.20±0.08 in 14 wa2 cells, P<0.0001, n=3–5 dishes for controls and 3–5 dishes for wa2 (number of littermate animals used for each cell isolation) for each concentration studied). This is consistent with data obtained in suspended VSMCs.
|
Although an apparent sustained response could be observed in some individual cells such as the one shown in Fig. 6
, this is not significant at the population level. We did not observe any strain difference in the values for the sustained phase at 180–600 s after application of 100 nmol/l ET-1 (for example, F/F0: 1.36±0.05 vs. 1.48±0.12 at 2 min after the onset of the calcium transient in 31 wa2 and 37 normal responsive VSCMs, respectively, P>0.05).
Part of the overall decrease in [Ca2+]i response observed in wa2 VSMCs was due to the fact that fewer wa2 cells responded at low ET-1 concentrations (32% vs. 53% at 1 nmol/l ET-1, 57% vs. 100% at 30 nmol/l, 86% vs. 80% at 100 nmol/l ET-1 in wa2 and controls, respectively, P<0.01), suggesting an involvement of EGFR in the dose response relationship. We checked that the viability of the cultured VSMCs were similar (>98% in both groups by the trypan blue exclusion method). Moreover, the responses in individual wa2 VSMCs exhibiting a rise in calcium transient were still much smaller (P<0.01).
HB-EGF deficiency is associated with impaired [Ca2+]i response to ET-1 in VSMCs
To verify that endogenous HB-EGF mediates ET-1-induced [Ca2+]i response in VSMCs, we assessed the response of HB-EGF–/– and control VSMCs derived from mouse aortas (Fig. 6D
). Whereas basal [Ca2+]i in VSMCs from HB-EGF–/– and from wild-type mice was the same (290±15 vs. 270±12 nmol/l, respectively, P>0.05), HB-EGF deficiency was associated with significant lower ET-1-stimulated increase in [Ca2+]i (peak minus baseline values: 322±39 vs. 475±22 nmol/l in HB-EGF–/– and HB-EGF+/+ VSMCs, respectively, P<0.05, n=13 and 7 measures per group).
Effects of modulation of ETA and ETB receptors on ET-1-induced increase in [Ca2+]i
Addition of ETA selective antagonist BQ 123 fully prevented ET-1-induced increase in [Ca2+]i both in normal and in EGFR defective wa2 VSMCs (inhibition of [Ca2+]i release by more than 98% in both groups, P<0.0001). This result was mimicked with the mixed ET receptors blocker bosentan (Fig. 7
). By contrast, ETB antagonism with BQ 788 was ineffective in both strains (80±14 and 87±12% of ET-1 alone-stimulated [Ca2+]i responses in control and wa2 VSMCs, respectively, P>0.05). Consistently, selective ETB agonist sarafotoxin 6c was devoid of any action on [Ca2+]i. Altogether, these results suggest that both strains share a similar ETA-dependent exclusive pathway linking ET-1 to calcium signaling in VSMCs.
|
HB-EGF is not sufficient to promote calcium transients in VSMCs
To test if EGFR activity is sufficient to modulate [Ca2+]i in VSMCs, HB-EGF (10 ng/ml) was added in normal mouse VSMCs. EGFR phosphorylation was markedly enhanced 10 min later under the same conditions (125±12% and 293±45% of baseline on average on ET-1 and HB-EGF, respectively (P<0.05 and P<0.001, respectively, vs. control) (Fig. 8
). HB-EGF alone (10 ng/ml) failed to stimulate [Ca2+]i in normal VSMCs (peak values: 0 nmol/l). However, when HB-EGF was added 5 min after ET-1, we observed a modest but significant [Ca2+]i transient (peak minus baseline values: 514±128 and 90±14 nmol/l after ET-1 alone and 5 min later after HB-EGF, respectively (P<0.001 for HB-EGF after ET-1 vs. HB-EGF alone) (Fig. 8)
. These results indicate that HB-EGF availability modulates [Ca2+]i response once some specific required signal has been induced by activation of ETA receptors.
|
Effect of AG1478 on the ET-1-induced increase in [Ca2+]i in VSMC
To ascertain the role of EGFR transactivation in ET-1 [Ca2+]i increase, we used AG1478, a specific EGFR inhibitor. Wild-type VSMCs were incubated for 10 min, with or without 250 nmol/l AG1478, before stimulation with 100 nmol/l ET-1. We verified that AG1478 had no effect by itself on baseline [Ca2+]i. Figure 9
illustrates the marked inhibition (47%) of ET-1-induced peak above baseline value (182±20 vs. 341±32 nmol/l for AG1478 incubated and control cells, respectively; P<0.01, n=6 per group).
|
Role of PI3-kinase on the ET-1-induced increase in [Ca2+]i in VSMC
To study the role of PI3K in ET-1 stimulated [Ca2+] increase, we used wortmannin and LY 294002, two unrelated specific inhibitors of PI3K. The cells were incubated with various concentrations of both drugs for 10 min before stimulation with 100 nmol/l ET-1. Both inhibitors had no action on basal [Ca2+]i. By contrast, addition of each of these two inhibitors prior to ET-1 blunted [Ca2+]i peak value in a concentration-dependent manner. The maximal inhibition (50% of control, P<0.05) was obtained with 1 µmol/l wortmannin (314±54 vs. 626±40 nmol/l with and without wortmannin, respectively, n=8) and nearly the same inhibition (44% of control, P<0.01) was obtained with 50 µmol/l LY 294002 (382±36 with LY 294002 vs. 673±92 nmol/l for controls, respectively) (Fig. 9)
.
To test the role of intact EGFR on the PI3K-mediated ET-1-induced increase in [Ca2+]i, we performed the same [Ca2+]i measurement on wa2 VSMCs. Both drugs failed to inhibit ET-1-induced increase in [Ca2+]i in these cells (429±21 vs. 462±25 nmol/l for wa2 VSMCs+50 µmol/l LY 294002 or vehicle, respectively, and 417±72 vs. 412±90 nmol/l for wa2 VSMCs+1 µmol/l wortmannin or vehicle, respectively; n=8, NS, P>0.05) (Fig. 9)
.
Next, the effect on ET-1-induced calcium transient of concomitant inhibition of EGFR and PI3K, respectively, with AG1478 and LY 294002, was compared to the effect of each compound alone. Inhibition of the two pathways showed no additive effect and inhibited ET-1-induced [Ca2+]i peak by 44% (P<0.05) (Fig. 9)
.
Interaction between the EGFR transactivated by ET-1 and PI3K activation
We examined whether modulation of PI3K activity would affect EGFR phosphorylation after stimulation with ET-1. We also symmetrically verified the influence of EGFR activation on PI3K activity at baseline and upon addition of ET-1.
EGFR phosphorylation was increased by ET-1 in freshly isolated aortas (127±20% vs. 100±2% in controls, P<0.05, n=4 per condition) (Fig. 10
). Blockade of PI3K with wortmannin for 10 min or for up to 3 h prior to stimulation did not alter pEGFR/EGFR cell expression, indicating that although ET-1-induced transactivation of EGFR, this action was not mediated by PI3K activation.
|
EGF induced a robust 3-fold increase in pAKT/AKT cell content (not shown). In comparison, ET-1 (0.1 µmol/l) induced a 148 ± 16% increase in relative AKT phosphorylation compared with control (P<0.05, n=4 per condition) (Fig. 10)
. Of note, preincubation with EGFR inhibitor AG1478 prevented the ET-1-induced stimulation of PI3K (110±11% of baseline after 10 min, P<0.05 vs. ET-1+ vehicle), suggesting that ET-1-dependent PI3K activity is mediated by EGFR transactivation.
ET-1-induced increase in [Ca2+]i in VSMC requires metalloproteinase activity
To investigate whether extracellular shedding of EGFR ligands leads to an ET-1-induced [Ca2+]i increase, we used a broad-spectrum matrix metalloproteinase inhibitor: batimastat (3)
. Preincubation of normal VSMCs with 5 µmol/l batimastat for 10 min induced a significant 40% decrease in ET-1 stimulated [Ca2+]i peak (+286±56 vs. +522±92 nmol/l above baseline for batimastat-treated and control cells, respectively, P<0.05, n=9) (Fig. 11
).
|
Neutralization of HB-EGF impairs ET-1-mediated [Ca2+]I in human VSMCs
Human proHB-EGF has been shown to be the endogenous receptor for Corynebacterium diphtheriae toxin (DT) (28)
. DT toxin has been mutated (CRM197) to be devoid of toxicity and is able to neutralize HB-EGF binding to EGFR; it has been a useful tool to demonstrate the role of HB-EGF in mediating GPCR agonist transactivation of EGFR (3
, 29)
. VSMCs from human internal mammary arteries were cultured and incubated with CRM 197 for 1 h before the experiments. In this condition, human VSMCs had a significantly lower Ca2+ response to ET-1 than controls (30 nmol/l ET-1 increased the fluorescence by 1.28±0.03-fold in 189 human mammary VSMC in control conditions and 1.15±0.01-fold in 157 human mammary VSMC treated with 10 mg/ml CRM197, P<0.001) (Fig. 11)
, suggesting a role of HB-EGF in mediating ET-1-induced Ca2+ signaling.
DISCUSSION
Through a combination of approaches, including the use of mouse models lacking either EGFR tyrosine kinase activity or HB-EGF as well as hemodynamic studies, ex vivo measurement of arterial isovolumetric tone, and [Ca2+]i assessment in mouse and human VSMCs, we observed that HB-EGF did not influence the control of blood pressure in physiological conditions but was required for ET-1-induced vasoconstriction and an increase in blood pressure in vivo. We previously demonstrated that pharmacologic blockade of the EGFR prevented or reversed ET-1-induced vasoconstriction of mouse aorta (5)
. However, we were not able to identify the mechanism involving EGFR tyrosine kinase. We now demonstrate that HB-EGF is the major ligand for EGFR, which mediates ET-1-induced a pressive response in vivo and vasoconstriction in freshly isolated mouse arteries. HB-EGF is likely to act through a metalloproteinase-dependent process since batimastat mimicked the effect of EGFR tyrosine kinase inhibition. Moreover, EGFR tyrosine kinase activity participates in the ET-1-stimulated rise in [Ca2+]i in mouse and human VSMCs. These data agree with the findings of Carmines et al. that pharmacological blockade of EGFR modulated Ca2+ influx in isolated afferent arterioles from rat kidney (30)
and with recent work suggesting a role for HB-EGF-EGFR pathway in sustained adrenergic vasoconstriction in rat mesenteric arteries (31)
. However, in this latter study, EGFR blockade with high concentration of AG1478 (10 µmol/l) had no significant effect on phenylephrine-induced contraction of rat mesenteric arteries during the first minute of incubation, although inhibition became significant after 2 min (31)
. Although [Ca2+]i signaling was not examined in this study, early [Ca2+]i signaling may not have been directly affected by alpha1ß-adrenoreceptor-induced EGFR transactivation in intact arteries. Alternative actions of EGFR on the contractile machinery may include activation of the Rho-Rho kinase pathway with subsequent calcium sensitization (32)
or delayed and sustained calcium extracellular entry through L-type voltage gated Ca2+ channels (33
, 34)
. Moreover, differences in animal species and vascular beds may also account for the variety of functional links between vasoconstrictive GPCRs and EGFR.
We further report that functional transactivation of EGFR was extremely rapid, since the [Ca2+]i transient peak was blunted by 30–40% within 5 s after the addition of ET-1 in either wa2 cells or AG1478-treated cells. This indicates rapid GPCR-EGFR functional crosstalk, which may have limited our ability to measure EGFR phosphorylation within the first 15 s of the initial [Ca2+]i mobilization (a technical challenge thus far unmet to our knowledge). However, EGFR autophosphorylation was observed 2 to 10 min after GPCR stimulation, as described by others. Thus, we report here a critical and very early involvement of EGFR in the ET-1-induced rise in [Ca2+]i with subsequent modulation of vascular tone. This confers on the HB-EGF-EGFR pathway a physiological role that is distinct from that described in previous studies in which the alpha1 receptor agonists phenylephrine and thrombin were both shown to induce subacute EGFR autophosphorylation in vitro and to display similar requirements for HB-EGF in order to promote protein synthesis in rat and mouse VSMCs after several hours (35
, 36)
, an in vitro trophic effect that was mimicked by HB-EGF itself (27)
. Further studies are needed to identify tyrosine residues that are rapidly phosphorylated (and probably dephosphorylated) upon ET-1 stimulation and to determine whether acute and subacute involvement of EGFR transactivation are functionally linked. Such a hypothesis may explain how transactivation of EGFR on vasoactive GPCR stimulation induces a variety of distinct cellular effects, with a variable connection between vasoconstriction and cell growth promotion. Alternatively, one may hypothesize different kinetics for EGFR transactivation, a process that apparently involves a combination of various EGFR ligands and metalloproteinases from the MMP and ADAM families to mediate cell growth, cell invasion, and migration (26
, 31
, 37
38
39
40)
. The data presented herein identify HB-EGF as the critical mediator of ET-1-induced vasoconstriction, since HB-EGF-deficient carotids were virtually insensitive to ET-1 and similar results were obtained with HB-EGF-deficient aorta. Moreover, the lack of a pressive effect of systemic infusion of ET-1 in HB-EGF-deficient animals suggests a direct or indirect interaction between ET receptors and HB-EGF release within the vasculature. Meanwhile, the lack of effect of ET-1 on mean blood pressure in HB-EGF-deficient mice might be attributed in part to their defective cardiac valvulogenesis and associated valve regurgitation (15)
. Since endothelial denudation did not restore normal response to ET-1 in HB-EGF –/– arteries nor alter the [Ca2+]i response in normal arteries, the possible role of endothelial ET receptors or of release of HB-EGF from endothelial cells is unlikely. The general mechanism for HB-EGF action on ET-1 stimulation is likely to involve acute cleavage of pro-HB-EGF and release of an active fragment since batimastat, an inhibitor of a wide range of metalloproteinases, mimicked the effect of EGFR blockade. Consistently, down-regulation of pro-HB-EGF in human VSMCs from internal mammary arteries with CRM 197 demonstrated a requirement for pro-HB-EGF in the ET-1-mediated full calcium response.
Little is known about how EGFR can modulate calcium signaling on transactivation by GPCRs. Whereas EGF or HB-EGF alone did provoke phosphorylation of EGFR without inducing any rise in [Ca2+]i, addition of HB-EGF after ET-1 induced a mild but significant [Ca2+]i, transient. This result indicates that EGFR can interfere with calcium signaling once GPCR (ETA) activation has occurred and that HB-EGF availability is limiting for an ETA-induced response. Another significant finding of our study is that PI3K activity is also required for full ET-1-mediated vasoconstrictive action and [Ca2+]i signaling. When, in the present study, both PI3K and EGFR were pharmacologically inhibited, no additive inhibition of the ET-1 stimulated [Ca2+]i peak was observed. In addition, LY294002 and wortmaninn had no effect in EGFR tyrosine kinase-deficient wa2 VSMCs. Although samples from the literature suggest that PI3K is activated downstream of EGFR (41
42
43)
, we cannot exclude the possibility that ET receptors may also activate PI3K independently and in parallel. Northcott et al. (44)
found that arteries from rat models of hypertension develop a spontaneous tone that is inhibited by LY294002 or wortmannin whereas arteries from normal rats do not develop such a spontaneous tone. These authors suggested that PI3K may activate L-type Ca2+ channels as confirmed in vitro (33
, 34)
. Indeed, Gß proteins have been suggested to interact with PI3K and L-type calcium channels and to activate them on GPCR agonists (34)
.
Our data lead us to support this hypothesis, namely that HB-EGF-mediated stimulation of EGFR is a required intermediate between ETA receptors and the activation and stimulation of [Ca2+]i peak and contraction; if correct, this would be a novel paradigm in mouse and human VSMCs (Fig. 12
). PI3K activity is also a requirement for ET-1 actions; this pathway seems at least to be downstream to EGFR autophosphorylation, since we observed that ET-1-stimulated PI3K activity was blunted by AG1478 in cultured VSMCs. Conversely, EGFR phosphorylation, although stimulated by ET-1, was not influenced by PI3K inhibitors.
|
ET-1 was shown to activate two types of Ca2+-permeable nonselective cation channels (NSCC-1 and NSCC-2) along with store-operated Ca2+ channels (SOCC) that allow extracellular Ca2+ influx in rat VSMC (45)
. In CHO cells transfected with ETA receptor, NSCC-2 and SOCC are stimulated by ET-1 via a PI3K-dependent cascade (46)
. However, these authors implicated ET-1-mediated activation of these calcium channels prior to phosphorylation of EGFR in rabbit internal carotid artery VSMCs (47)
, an observation that does not fit with the present findings in mouse arteries and human internal mammary arteries cells. Studies that are beyond the scope of the present one are needed to elucidate which calcium pathway is affected by HB-EGF-EGFR transactivation and to what extent PI3K activation is induced with such a fast kinetic.
EGF or HB-EGF alone, even at concentrations that induced a strong phosphorylation of EGFR, failed to induce any rise in [Ca2+]i or in arterial tone of normal mouse arteries. We suggest that concomitant signaling cascades from ETA and EGFR are both required for [Ca2+]i peak in VSMCs and vasoconstriction. In this regard, our findings may explain interesting data by Watts and colleagues indicating that EGF failed to increase tone in aorta isolated from normal rats but enhanced aortic tone from DOCA-salt-treated hypertensive rats (48)
, a condition where accumulation of ET-1 in the vascular wall has been demonstrated (49)
. A synergistic effect of EFGR and ET receptors to promote vasoconstriction may occur in other circumstances where excess of ET-1 is found in the vascular wall such as NO deficiency (50)
, mineralosteroid intoxication (51)
, salt-sensitive hypertension (52
, 53)
, and atherosclerosis (54)
.
In conclusion, our results identify important new steps in the signal transduction pathway mediating ET-1 induction of Ca2+ signaling and vasoconstriction. Moreover, the strong requirement for HB-EGF and EGFR for early Ca2+ signaling in mouse and human VSMCs may constitute an amplifying loop with subsequent release and synthesis of HB-EGF, since [Ca2+]i increase is known to stimulate metalloproteinase cleavage and secretion of HB-EGF (55
, 56)
. Furthermore, our results may suggest a mechanism whereby injury causes prevailing levels of ET-1 in the vascular wall to become strongly vasoconstrictive and trophic (50
, 54)
(Fig. 12)
. That is, hypertension, nephrosclerosis, or atherosclerosis increase levels of key intermediates of the endothelin-HB-EGF pathway, i.e., metalloproteinases (57)
, HB-EGF (58
, 59)
, and PI3K (60)
.
ACKNOWLEDGMENTS
We thank Claude Kitou, Laetitia Breton, and Caroline Martin for excellent animal care and Guillermo Salazar for excellent technical assistance. Dr. Pierre-Louis Tharaux thanks the Simone and Cino del Duca Foundation and the Eli Lilly International Fellowship Foundation. The work was financially supported in part by NIH grant CA43793 and INSERM. We are deeply indebted to Richard D. Bukoski, who kindly helped us to perform the first myographic experiments of this study (Cardiovascular Disease Research Program, North Carolina Central University, Durham, NC, USA), and to Dr. Anne Virsolvy for providing us with human VSM cells. The authors have no conflicts of interest to disclose.
FOOTNOTES
1 These authors contributed equally to this work. 
Received for publication October 21, 2005. Accepted for publication April 17, 2006.
REFERENCES
[Ca2+]i: difference in [Ca2+]i between peak and baseline) (*P<0.01 vs. controls, n=37–45 per group).
