Permissive role of nitric oxide in endothelin-induced migration of endothelial cells.

Endothelin (ET) synthesis is enhanced at sites of ischemia or in injured vessels. The purpose of this study was to explore the possibility of autocrine stimulation of endothelial cell migration by members of the endothelin family. Experiments with microvascular endothelial cell transmigration in a Boyden chemotactic apparatus showed that endothelins 1 and 3, as well as a selective agonist of ETB receptor IRL-1620, equipotently stimulated migration. Endothelial cell migration was unaffected by the blockade of ETA receptor, but it was inhibited by ETB receptor antagonism. Based on our previous demonstration of signaling from the occupied ETB receptor to constitutive nitric oxide (NO) synthase (Tsukahara, H., Ende, H., Magazine, H. I., Bahou, W. F., and Goligorsky, M. S. (1994) J. Biol. Chem. 269, 21778-21785), we next examined the contribution of ET-stimulated NO production to endothelial cell migration. In three independent cellular systems, 1) migration and wound healing by microvascular endothelial cells, 2) wound healing by Chinese hamster ovary cells stably expressing ETB receptor with or without endothelial NO synthase, and 3) application of antisense oligodeoxynucleotides targeting endothelial NO synthase in human umbilical vein endothelial cells, an absolute requirement for the functional NO synthase in cell migration has been demonstrated. These findings establish the permissive role of NO synthesis in endothelin-stimulated migration of endothelial cells.

Formation of new blood vessels, driven by the morphogenetic program and/or by the functional demand for increased blood supply, is initiated by the budding of endothelial cells off the microcirculatory bed. Tissue hypoxia represents a physiologic stimulus for angiogenesis, and several autocrine angiogenic mediators produced by endothelial cells have been recognized (1)(2)(3). Endothelin, one of such mediators, is constitutively produced by endothelial cells and its production is enhanced by hypoxia (4 -6). Two investigative teams have recently demonstrated that ET-1 1 and ET-3, acting via the ET B receptor, stimulate endothelial cell migration and proliferation (7,8).
Since we and others have previously demonstrated that occupancy of ET B receptors in endothelial cells is accompanied by the activation of constitutive endothelial NO synthase (9,10), we hypothesized that the observed motogenic effects of members of the ET family may in fact be mediated by the release of NO. Indeed, there is emerging evidence that NO release serves as a prerequisite for epithelial and endothelial cell motility (11)(12)(13). Leibovich et al. (13) have shown that production of angiogenic activity by activated monocytes (assayed by chemotaxis of endothelial cells and corneal angiogenesis) is absolutely dependent on L-arginine and NO synthase. These observations are in concert with findings reported by Ziche et al. (12) who detected the potentiation by sodium nitroprusside of the angiogenic effect of substance P in the rabbit cornea. Our own observations expand this function to the classical angiogenic signal, vascular endothelial growth factor. We demonstrated that endogenous NO production by the endothelial cells is a prerequisite for the motogenic and angiogenic effects of this factor. 2 In the present study, we used three different approaches (migration and wound healing by endothelial cells, wound healing by Chinese hamster ovary cells expressing ET B receptor with or without eNOS, and application of antisense oligodeoxynucleotides targeting eNOS) to provide evidence that the effect of ET-1 on cell migration is mediated via ET B receptor and requires functional enzymatic machinery for NO generation.
Functional Expression of ET B R and eNOS in CHO Cells-The generation and characterization of individual cell lines used in this study have been detailed previously (9). Briefly, CHO-ET B cell cultures stably integrated copies of the ET B R cDNA cloned, and the amplified insert encompassing the ET B open reading frame was digested with PstI, gel-purified, and ligated into the eukaryotic expression plasmid pMT2. This plasmid is driven by the adenovirus major late promoter and * This work was supported in part by National Institutes of Health Grants DK45695 and DK45462. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
contains the sequences encoding murine dihydrofolate reductase immediately downstream of the cDNA insert, thereby allowing for the establishment of stably integrated cell lines with greatly amplified copy numbers when propagated in the presence of methotrexate (17). Dihydrofolate reductase-deficient CHO cells (18) were propagated in nucleoside-free Iscove's modified Dulbecco's medium (Sigma), supplemented with 10% dialyzed fetal calf serum, and subsequently plated at a density of 5 ϫ 10 5 /100 mm 2 dish. Cells were transfected with 30 g of pMT2-ET B or pMT2-wild type using the calcium phosphate transfection method of Chen and Okayama (19). After 5 days, selection and amplification were initiated by supplementation of the media with 0.02 M methotrexate and cells were cloned by limiting dilution. Resistant colonies were subsequently expanded and simultaneously screened for stable integration by genomic hybridization using the 32 P-radiolabeled ET B cDNA as probe and for functional expression using fluorescence mapping with biotinylated highly specific ligand, IRL-1620, as previously detailed (9).
For double transfectants, CHO-ET B cells were retransfected with the cDNA for endothelial cell-derived NOS (20) (kindly supplied by Dr. K. D. Bloch) using the eukaryotic expression plasmid pMEP4 (Invitrogen, San Diego, CA), which contains an inherent hygromycin B resistance gene and an inducible metallothionein promoter upstream from the NOS cloning site. Stable transfectants (CHO-ET B NOS) were selected using culture medium supplemented with 300 g/ml hygromycin B (1000 units/mg, Sigma). Resistant colonies were ring cloned, propagated, and functionally evaluated for NO expression using an NOselective electrode, as detailed previously (9). NOS expression was maximized by supplementation of the medium with 50 M ZnCl 2 for 5 h prior to functional measurements.
Cell Migration Assay-Migration assay was performed according to the previously described technique (22), with minor modifications, in a Boyden chemotactic apparatus (Neuroprobe, Cabin John, MD). RMVEC were lifted with 0.05% trypsin, 0.53 mM EDTA (Life Technologies, Inc.) and washed, and 10 6 cells/ml suspended in 25 l of M199 with 0.1% bovine serum albumin were added to the lower chamber of a Boyden apparatus. Polycarbonate filters with 8-m pores (Poretics Corporation, Livermore, CA) were coated with 10 g/ml ProNectin F (Protein Polymer Technologies, San Diego, CA), washed twice with phosphate-buffered saline, and positioned above the wells of the lower chemotactic chamber which contained cells. The top half of the chamber was reattached, and the chamber was incubated in an inverted position at 37°C in 95% air ϩ 5% CO 2 for 2 h to allow a uniform cell attachment to the filter. The test agents, as specified under "Results," or a vehicle suspended in 50 l of M199 with 0.1% bovine serum albumin were added to upper chambers. The chambers were wrapped with Parafilm and incubated for an additional 6 h in an upright position. After incubation, the filter was removed from the apparatus, and cells were fixed with methanol and stained with Dff-Quick (Baxter, Miami, FL). The number of migrated cells on the upper surface of the filter was counted in six randomly chosen fields under ϫ 400 magnification and averaged. All experiments were performed in quadruplicate, and each experiment was repeated at least three times.
Monitoring of NO Release-RMVEC or HUVEC grown in gelatincoated 35-mm dishes were preincubated in Krebs-Ringer buffer of the following composition (in mM): NaCl 136, KCl 5.4, CaCl 2 1.8, MgSO 4 0.8, NaH 2 PO 4 1.0, HEPES 10, and 1 g/liter glucose, adjusted to pH 7.4, which was supplemented with 7.5 units/ml superoxide dismutase (Sigma). NO release was monitored with an NO-selective microprobe (Inter Medical Co., Nagoya, Japan), as previously detailed (11). Tip diameter of the probe (25 m) permitted the use of a micromanipulator (Zeiss-Eppendorff) attached to the stage of an inverted microscope (Nikon Diaphot) and enclosed in a Faraday's chamber to position the sensor approximately 5 m above the cell surface. Calibration of the electrochemical sensor was performed using different concentrations of a nitrosothiol NO donor S-nitroso-N-acetyl-DL-penicillamine, as previously detailed (23).
Wound Healing Assays-Electrode fabrication and the design of electric cell-substrate impedance sensor, a device that detects with high precision the electrical resistance of cells cultured inside small wells on the surface of a miniature gold electrode, have been reported previously (24,25). Electrodes were precoated with 10 g/ml fibronectin (Collaborative Biomedical Products, Bedford, MA). Endothelial cells at the density of 2 ϫ 10 5 were resuspended in Basal Medium (Life Technologies, Inc.) supplemented with 0.1% bovine serum albumin, 250 M L-arginine (Sigma), and 100 units/ml penicillin 100 g/ml streptomycin, and seeded on electrodes. To generate "wounds" on the surface of microelectrodes, confluent endothelial monolayers were electropermeabilized with direct current from a 6-V battery generating a local electrical current of 0.12 mA. This procedure resulted in a loss of cells from the surface of the microelectrode, while all surrounding cells remained intact. The rate of repopulation of the microelectrode with migrating cells was monitored for 20 h as changes in resistance and capacitance. Alternatively, uniform wounds were generated using a laser beam. CHO-ET B and CHO-ET B NOS cell monolayers grown on glass coverslips for 2 days were exposed to a wounding procedure as previously detailed (11). Wounds were produced using a single pulse generated by a YAG laser (Coherent 7900, Palo Alto, CA) with energy of 6.5-7.5 mJ. Cell monolayers were arranged perpendicularly to and in the focal plane of the beam (using a micromanipulator, these parameters were kept constant between different experiments). After wounding, the medium was immediately changed to Dulbecco's modified Eagle's medium containing 0.1% bovine serum albumin supplemented with 100 units/ml penicillin and 100 g/ml streptomycin. After the wounding, images were recorded immediately and 3 and 6 h later and stored for analysis (Universal Imaging Corp., West Chester, PA). The surface area of wounds in the consecutive images was measured, the size of the initial wound was expressed as 100%, and the timed progression of wound healing was expressed as fractions of the initial size.
Antisense . These sequences included the 5Ј-untranslated region of eNOS cDNA and initiation codon. Thus designed sequences showed no homology with other known mammalian sequences deposited in the GenBank data base, as screened using a Blast program (21). All the S-ODNs were purified using oligonucleotide purification cartridges, dried down, resuspended in Tris-EDTA (10 mM Tris, pH 7.4, 1 mM EDTA, pH 8.0), and quantified spectrophotometrically.
Prior to migration assays, HUVEC were incubated with 10 M S-ODNs for 12 h under a serum-free condition. During typical experiments in a Boyden chemotactic chamber (see above), S-ODNs were present in the medium. In parallel experiments, companion HUVEC treated with 10 M S-ODNs in the serum-free medium were seeded on gelatin-coated coverslips, allowed to attach, and used to verify their responsiveness to ET-1 (1 nM) by monitoring NO release or were stained with monoclonal antibodies directed against eNOS (Transduction Laboratories, Lexington, KY), according to the previously described protocol (11).

Motogenic Effect of Endothelin Is Mediated via ET B Receptor
and Is NO-dependent-With regard to NO release and locomotive responses, RMVEC displayed sensitivity to ET-1 within the physiologic concentration range (Fig. 1). ET-1 stimulated NO production by RMVEC, and L-NAME blunted NO release in a dose-dependent manner, with 2 mM L-NAME resulting in complete inhibition of NO release (Fig. 1A). Transwell migration of RMVEC, studied in a modified Boyden chamber, showed the similar exquisite sensitivity to ET-1. A statistically significant almost 50% increase in the number of migrated cells was detected with ET-1 concentrations of 10 pM, and it almost doubled with further elevation of ET-1 concentration to 1 nM (Fig. 1B). Inhibition of NO synthase with 2 mM L-NAME completely abrogated the ET-1-induced cell migration. Under the conditions of NOS inhibition, endothelial cell migration could be restored in the presence of 8-bromo-cyclic GMP (100 M), as shown in Fig. 2.
To elucidate the type of ET receptor(s) involved in the observed stimulation of RMVEC migration, transwell migration experiments were performed with different endothelins and in the presence of selective inhibitors of their receptors. As summarized in Fig. 3, ET-1 and ET-3 were equipotent in stimulating endothelial cell migration. A selective ET B R agonist, IRL-1620 exhibited a similar stimulatory effect, suggesting that the observed phenomena were mediated via ET B R. This conclusion is further supported by the observation that a selective ET A receptor antagonist BQ123 did not affect ET-1-induced endothelial cell transmigration, whereas a selective ET B R antagonist BQ788 completely abrogated the stimulatory action of ET-1. These findings were in concert with ET-1-induced NO release. While BQ123 did not interfere with NO production by endothelial cells stimulated with 1 nM ET-1, pretreatment with BQ788 abrogated NO release in response to ET-1 but not to bradykinin (Fig. 4). It is therefore concluded that ET-1 induces endothelial cell migration via the ET B R and that the previously demonstrated coupling of this receptor to eNOS (9) is critical for NO production and cell migration.
Effects of Endothelin on the Rate of Endothelial Wound Healing-The rate of endothelial wound healing was examined by electroporation-induced RMVEC denudation from the surface of a gold miniature electrode, as previously detailed (25). ET-1 accelerated wound healing at 10 h by 66.0 Ϯ 3.2% (n ϭ 5) compared with the control medium (40.4 Ϯ 2.0% (n ϭ 5), p Ͻ 0.05) (Fig. 5). The rate of restitution of endothelial integrity was equally enhanced by IRL-1620 and by ET-3, and this effect was abolished by co-application of ET B receptor antagonist BQ788, further implicating ET B receptor in the observed phenomena (Fig. 5, A and B). The addition of L-NAME to ET-1-or ET-3-stimulated endothelial cells virtually abrogated this response (Fig. 5, C and D).
Effects of Endothelin on Wound Healing in CHO Cells-To further test the above premise that NO production is required for ET-1-induced RMVEC migration, experiments were performed in a model system of CHO cells expressing ET B R and eNOS, as verified and detailed previously (9). ET-1 was ineffective in stimulating wound healing in CHO cells transfected with ET B R alone (Fig. 6). However, co-expression of both ET B R  and eNOS in CHO cells imparted on them the migratory response to ET-1 and resulted in acceleration of wound healing. The data further confirm the permissive role of NO production in ET-1-induced migration.
Effects of Endothelin in the Presence of Antisense Oligodeoxynucleotides Targeting eNOS-In the next series of experiments, a selective knockdown of eNOS was performed using antisense S-ODNs. To avoid any possible species-specific variations in the sequence, experiments were performed in HUVEC using an antisense construct directed against the initiation codon of human eNOS cDNA. Cells were treated with either antisense, sense, or scrambled S-ODNs, as detailed under "Materials and Methods." The expression of immunodetectable eNOS in endothelial cells subjected to S-ODNs is presented in Fig. 7, confirming the adequacy of treatments. Functional analysis of endothelial cells subjected to S-ODNs was accomplished by monitoring NO release. Fig. 8A depicts typical tracings of ET-1-induced NO release. Both sense and scrambled S-ODNtreated cells responded to ET-1 with increased NO release, similar to that observed in intact control cells. In contrast, antisense S-ODN-treated HUVEC failed to produce NO in response to ET-1. These data confirm the validity of the utilized S-ODN constructs. To study the effects of ET-1 on motility of S-ODN-treated HUVEC, experiments were performed in a modified Boyden apparatus. As summarized in Fig. 8B, pretreatment of HUVEC with antisense S-ODN resulted in a dramatic deceleration of transmigration, as compared with intact control cells and cells exposed to sense or scrambled constructs.

DISCUSSION
In this study, two different endothelial cell preparations, derived from rats and humans, showed a consistent acceleration of motility in response to members of the endothelin family. Through application of specific agonists and antagonists of known endothelin receptors, ET A and ET B , it was possible to implicate ET B receptor in the observed responses. However, these effects of ET were mediated by endogenous NO production, as confirmed by the results of three independent experimental approaches. In the first series of experiments, L-NAME prevented ET-1-induced endothelial cell transwell migration and wound healing. In the second series, CHO-ET B NOS cell, a model cell system devoid of complexity inherent to endothelial cells, showed NO dependence of migration induced by ET-1. Finally, the application of eNOS isoform-selective antisense S-ODNs, but not sense or scrambled constructs, to endothelial cells suppressed their migratory responsiveness to ET-1. From these observations, we derive a conclusion that stimulated NO production serves a permissive role in ET-induced acceleration of endothelial cell motility and wound healing.
ET B receptor in endothelial cells serves a unique purpose for autocrine feedback regulation of several physiologic functions by the produced ET-1. The application of anti-ET ␥-globulin to cultured endothelial cells inhibits the rate of proliferation (26,27). Moreover, this receptor is implicated in ET-induced vasodilation (10). We have recently demonstrated that this nonisopeptide-selective receptor is functionally coupled to eNOS and coordinates the release of NO from endothelial cells (9). FIG. 5. Endothelial wound healing as monitored with an electrical impedance sensor. Microvascular endothelial cells were cultured to confluency on the surface of fibronectin-coated wells especially manufactured for measurements of resistance (24). At time 0, cells growing on the surface of a miniature electrode were destroyed by electroporation (25) (as detailed under "Materials and Methods"), and the rate of restitution of monolayer integrity, as judged by the restoration of the electrical impedance, was monitored for 20 h. Tracings are representative of 3-4 replicative experiments. In all these experiments, initial resistance of endothelial cells was within the range of 9,000 -10,000 ohms/cm 2 . To simplify the comparison, ordinates represent normalized resistance. Note that ET-1, ET-3, and IRL-1620 were equipotent in accelerating endothelial wound healing and that a selective ET B receptor antagonist BQ788 reversed the effect of ET-1 (A, B, and D). Administration of 2 mM L-NAME to the culture medium inhibited wound healing despite the continuous presence of 1 nM ET-1 or ET-3 in the medium (C and D).
Stably transfected CHO-ET B NOS cells were successfully employed in those studies to reveal, in a simplified model system, that receptor-enzyme coupling involves the tyrosine kinaseand calcium/calmodulin-dependent pathways (9). Hence, the studies established a physiologically meaningful communication and coordination between endothelium-derived vasoconstrictor and vasodilator systems that control the vascular tone. Other investigators have also viewed ET-1 generated by endothelial cells as an autocrine motility factor (7,8). However, results presented herein implicate for the first time the integral complex system of endothelium-derived vasoactive agonists, endothelin and NO, in endothelial cell motility and implicate this system in vascular remodeling.
We have previously demonstrated that NO production is a prerequisite for epithelial cell migration guided by several motogens, including hepatocyte growth factor and epidermal growth factor (11). The establishment of head-to-tail gradients of NO synthase in locomoting, but not in stationary, epithelial cells was construed as an important regulator of locomotion driven by these guidance cues (11). More recently, we have established that NO interferes with processes of endothelial cell adhesion to matrix proteins and participates in locomotion initiated by vascular endothelium growth factor. 2 These findings provided an experimental basis for a hypothesis ascribing to NO the function of a modulator of cell-matrix adhesion. In this vein, ET-induced NO production by endothelial cells can be considered as a particular example of a more generalized phenomenon of NO-regulated cell motility.
The described NO production serving as a prerequisite for endothelial cell locomotion in response to activation of ET B receptor may explain a host of pathophysiologic observations on inadequate angiogenesis despite enhanced generation of ET-1. In hypercholesterolemic pigs and in humans with atherosclerotic lesions, ET-1 generation is augmented (28 -30). The initial size of each wound was expressed as 100%, and the rate of wound healing was assessed as a decrease in wound size (panel A depicts representative images of wound healing). Note that ET-1 (1 nM) accelerated wound healing (* denotes p Ͻ 0.05 versus control, n ϭ 5-7 separate experiments) only in CHO-ET B R/NOS cells, and this effect was inhibited by L-NAME (panel B). Wild type CHO cells did not respond to ET-1 (not shown).  The reason for the insufficient angiogenesis toward ischemic sites could be conceptually explained by the fact that NO production by the endothelium is suppressed under these conditions (31)(32)(33). Similarly, endothelial wound healing after balloon angioplasty is retarded, neointimal formation by proliferating smooth muscle cells is enhanced, and endothelin receptor antagonist SB 209670 protects angioplastic vessels against neointimal formation (14). Interpretation of these observations may rest on the above presented findings on the permissive role of NO in endothelial cell migration and wound healing.