Induction by Lysophosphatidylcholine, a Major Phospholipid Component of Atherogenic Lipoproteins, of Human Coronary Artery Smooth Muscle Cell Migration
Background—The objectives of the present study were (1) to determine whether lysophosphatidylcholine (lyso-PC), a prominent component of oxidatively modified LDL, induces migration of human coronary artery smooth muscle cells (SMCs) and, if so, to clarify the mechanism, and (2) to investigate the possible interactions of lyso-PC and platelet-derived growth factor (PDGF)-BB, endothelin-1 (ET-1), adrenomedullin (AM), or vitamin E on SMC migration by the Boyden’s chamber method.
Methods and Results—Lyso-PC induced SMC migration in a concentration-dependent manner between 10−6 and 5×10−5 mol/L. By contrast, phosphatidylcholine was without significant activity, and lysophosphatidylinositol and lysophosphatidylserine were much less effective than lyso-PC. Lyso-PC increased basic fibroblast growth factor (bFGF) production in a concentration-dependent manner between 10−6 and 5×10−5 mol/L in these cells. Furthermore, lyso-PC–induced SMC migration was inhibited by neutralizing antibody to bFGF but not by neutralizing antibody to transforming growth factor-β1. Lyso-PC–induced migration was significantly enhanced by PDGF-BB or ET-1 but was clearly inhibited by human AM and vitamin E.
Conclusions—These results indicate that (1) lyso-PC induces human coronary artery SMC migration at least in part through release of endogenous bFGF and (2) this lyso-PC–induced migration can be further induced by PDGF-BB and ET-1 and can be inhibited by human AM and vitamin E. Lyso-PC may recruit medial SMCs during the process of coronary atherosclerosis in part by releasing bFGF in concert with PDGF-BB or ET-1 in vascular tissues. This lyso-PC–induced SMC migration may be suppressed by AM and vitamin E under certain pathological conditions.
Elevated levels of LDL cholesterol are an important risk factor for coronary atherosclerosis and cardiovascular morbidity.1 However, oxidized LDL is shown to possess more atherogenic properties than native LDL.2 Actually, several lines of evidence have indicated that oxidatively modified LDL plays a key role in atherogenesis.3 4 In oxidized LDL particles, the content of lyso-PC, a polar phospholipid, is found to be dramatically increased.3 4 Actually, Parthasarathy and coworkers5 have demonstrated that as much as 40% of the phosphatidylcholine of LDL is converted to lyso-PC during oxidized modification. In addition, the concentration of lyso-PC is shown to be increased in atherosclerotic arterial lesions in animals fed an atherosclerotic diet.6
On the other hand, the migration of coronary artery medial SMCs into the intima is proposed to be an initial process of intimal thickening in coronary atherosclerotic lesions.7 8 Here, we show for the first time that lyso-PC, not phosphatidylcholine, potently induces migration of human coronary artery SMCs and that this lyso-PC–induced migration can be further induced by a potent mitogen, PDGF-BB, and a vasoconstrictive and growth-promoting peptide, ET-1.
Recently, Chai and coworkers9 demonstrated that the mechanism of enhanced DNA synthesis by oxidized LDL and lyso-PC at least in part involves an autocrine or paracrine action of release of endogenous bFGF. Therefore, we examined whether this lyso-PC–induced SMC migration is caused by the release of endogenous bFGF.
Finally, we show the interaction of lyso-PC and AM, a novel endothelium-derived vasorelaxant peptide,10 or vitamin E, a potent antioxidant, on the migration of human coronary artery SMCs.
SmBM and human coronary artery SMCs were purchased from Clonetics Corp. FCS, trypsin, and Versine were purchased from Gibco Laboratories. Synthetic human AM (1–52) and ET-1 were purchased from Peptide Institute. 8-Bromo-cAMP, IBMX, lyso-PC (palmitoyl), phosphatidylcholine (dipalmitoyl), lysophosphatidylinositol (palmitoyl), lysophosphatidylserine, and BSA were purchased from Sigma Chemical Co. Type I, II, III, or V collagen was purchased from Koken Inc. Flasks and multiwell plates were purchased from Becton Dickinson Co. The cAMP assay kit was purchased from Yamasa Shoyu Co, Ltd. Diff-Quick staining solution was purchased from Green-cross Corp. Forskolin was provided as a gift by Nihon Kayaku Co, Ltd. ETA antagonist, BQ-123, was provided as a gift from New Drug Discovery Research Laboratories, Tsukuba Research Institute, Banyu Pharmaceutical. Neutralizing antibody to bFGF (mouse monoclonal IgG1K) and human bFGF (recombinant) were purchased from Upstate Biotechnology, Inc. Neutralizing antibody to TGF-β1 (rabbit polyclonal antibody) was purchased from Santa Cruz Biotechnology, Inc.
Culture of Human Coronary Artery SMCs
Human coronary artery SMCs were cultured in SmBM containing human epidermal growth factor (0.5 ng/mL), human FGF (2 ng/mL), insulin (5 μg/mL), 5% FCS, 50 μg/mL gentamicin sulfate, and 50 μg/mL amphotericin-B. Cells were identified as SMCs according to their morphological and growth characteristics.11 12 Cultures were maintained at 37°C with atmospheric air and 5% CO2. Cells were subcultured after treatment with 0.25% trypsin and 0.02% EDTA. Subconfluent SMCs between the fourth and eighth passages were used for the experiments.
Migration of SMCs was assayed by a modification of the Boyden’s chamber method using microchemotaxis chambers (Neuro Probe Inc) and polycarbonate filters (Nucleopore Corp), as previously reported.11 12 In all experiments, collagen-coated filters were used. Briefly, the membranes were treated with 0.5N acetic acid and then incubated for 48 to 72 hours at 25°C in a collagen solution (100 μg/mL type I collagen in 0.5N acetic acid). They were then air-dried.11 12 Type I collagen was used to coat the filters in the chemotaxis chambers, because the chemotactic effect of lyso-PC was greater when type I collagen was used as a substrate than when type II, type III, or type V collagen was used (type I>type III>type V>type II).
Cultured SMCs were trypsinized and suspended at a concentration of ≈5.0×105 cells/mL in SmBM. The cell number was counted with an electronic cell counter (model ZB1, Coulter Electronics). A volume of 200 μL of SMC suspension was placed in the upper chamber, and 40 μL of medium (0.4% BSA containing 10−7, 5×10−7, 10−6, 5×10−6, 10−5, 5×10−5, 10−4, or 5×10−4 mol/L lyso-PC) was placed in the lower chamber. The chamber was incubated at 37°C under 5% CO2 in air for 6 hours.
Migration activity is calculated as the mean number of migrated cells observed in four high-power fields, as previously reported.11 Four different wells were studied at the same day on the same cell preparation. This procedure was performed in four independent experiments.
In addition, we examined the effects of lysophospholipids structurally similar to lyso-PC, such as phosphatidylcholine, lysophosphatidylinositol, and lysophosphatidylserine on human coronary artery SMC migration.
In experiments to determine the effect of endogenous bFGF or TGF-β1 on lyso-PC–induced SMC migration, two concentrations (1 and 10 μg/mL) of neutralizing antibodies to bFGF and TGF-β1 were added to the upper chamber 30 minutes before the addition of lyso-PC to the lower chamber. Neutralizing antibody to bFGF used here reacts with bovine, human, rat, and mouse bFGF and does not cross-react with bovine acidic FGF. The neutralizing antibody to TGF-β1 used here reacts with mouse, rat, and human TGF-β1 and does not cross-react with either TGF-β2 or TGF-β3.
Effects of PDGF-BB or ET-1 on lyso-PC–induced SMC migration were also examined. Two concentrations of PDGF-BB (0.5 and 2 ng/mL) or ET-1 (10−8 and 10−7 mol/L) were added to the lower chamber with 5×10−6 mol/L lyso-PC. In an experiment to determine the effect of the ETA receptor antagonist BQ-123 on ET-1–induced migration, BQ-123 and ET-1 were added to the lower chamber with lyso-PC.
In experiments to determine the effects of AM or vitamin E on lyso-PC–induced SMC migration, various concentrations (10−9, 10−8, 10−7, and 10−6 mol/L) of human AM or 5×10−5 mol/L vitamin E were added to the lower chamber in addition to lyso-PC.
In separate experiments to determine the effects of 8-bromo-cAMP and forskolin on lyso-PC–induced SMC migration, these cAMP-elevating agents were added to the lower chamber in addition to 5×10−5 mol/L lyso-PC.
Human coronary artery SMCs were cultured in 35-mm culture dishes, treated with various concentrations of lyso-PC for 6 hours, and then harvested by first rinsing the cultures in Dulbecco’s PBS and then overlaying them with 0.2 mL of Dulbecco’s PBS containing 2×10−3 mol/L EDTA, pH 8.0, 2×10−3 mol/L N-ethylmaleimide, and 10−3 mol/L PMSF, as previously reported.13 After cells were disrupted by three cycles of freezing and thawing followed by sonication for 1 minute, the homogenate was centrifuged at 25 000g for 30 minutes. The supernatant was stored at −80°C until bFGF quantification was performed with a solid-phase ELISA (Quantikine FGF Basic Immunoassay Kit; R&D Systems, Inc). The average recovery rate of bFGF in cell culture medium was 89%. The minimal detectable dose of bFGF was 1 pg/mL. The interassay variation was 8.7%, and the intra-assay variation was 4.6%.
After preincubation, the cell monolayers were washed twice with PBS and then stimulated for 30 minutes with different concentrations of human AM dissolved in SmBM that contained 5×10−4 mol/L IBMX. cAMP levels were determined by radioimmunoassay done with the cAMP assay kit, as previously described.14 15
Calculations and Analysis
The statistical significance of differences in the results was evaluated with an unpaired ANOVA, and P values were calculated by Scheffé’s method.16 All values were expressed as mean±SD.
Effect of Lyso-PC on SMC Migration
Figure 1⇓ shows concentration-dependent effects of lyso-PC (10−7, 5×10−7, 10−6, 5×10−6, 10−5, and 5×10−5 mol/L) on SMC migration for 6 hours of incubation. Lyso-PC significantly induced migration in a concentration-dependent manner between 10−6 and 5×10−5 mol/L. By contrast, the migration-stimulatory effects of the higher range of concentrations (10−4 and 5×10−4 mol/L) of lyso-PC were clearly weak or none (Figure 1⇓).
Micrographs of human coronary artery SMCs taken after a 6-hour migration period clearly show that lyso-PC, not phosphatidylcholine, induced migration in a concentration-dependent manner and without causing apparent morphological injury (Figure 2⇓).
Table 1⇓ shows the effects of other lysophospholipids structurally similar to lyso-PC on SMC migration. When these phospholipids were tested at two comparable concentrations, phosphatidylcholine was without significant activity, and lysophosphatidylinositol and lysophosphatidylserine appear to be much less effective than lyso-PC.
Effects of Neutralizing Antibody to bFGF on Lyso-PC–Induced SMC Migration
To clarify the role of endogenous bFGF in lyso-PC–induced SMC migration, cells were treated with antibody that neutralize bFGF. The results of these experiments are shown in Figure 3⇓. bFGF slightly but significantly stimulates SMC migration in a concentration-dependent manner (Figure 3A⇓). The addition of neutralizing antibody to bFGF to the upper chamber clearly abolished bFGF-induced SMC migration (Figure 3A⇓). The addition of 1 μg/mL of neutralizing antibody to bFGF to the upper chamber significantly inhibited lyso-PC–induced migration. When 10 μg/mL of antibody was added, the effect was even greater (Figure 3B⇓). By contrast, neither 1 μg/mL nor 10 μg/mL of neutralizing antibody to TGF-β1 to the upper chamber affected lyso-PC–induced migration (Figure 3B⇓).
Furthermore, we examined the effect of lyso-PC on bFGF production in cultured human coronary artery SMCs. Lyso-PC significantly increased bFGF production in a concentration-dependent manner between 10−6 and 5×10−5 mol/L (Figure 4⇓). bFGF concentration caused by 5×10−5 mol/L were ≈9-fold greater than nonstimulated controls. These results suggest that lyso-PC induces SMC migration at least in part through release of endogenous bFGF.
To further clarify the mechanism by which lyso-PC induces human coronary artery SMC migration, we examined the ability of vitamin E, a potent antioxidant, to inhibit lyso-PC–induced chemotaxis. Vitamin E (5×10−5 mol/L) significantly inhibited SMC migration induced by 5×10−5 mol/L lyso-PC, whereas vitamin E could not inhibit the nonstimulated and bFGF-induced SMC migration (Figure 5⇓). This suggested that the vitamin E–inhibitable step in the lyso-PC–induced migration was before stimulation by bFGF.
Effects of PDGF-BB and ET-1 on Lyso-PC–Induced SMC Migration
Figure 6⇓ shows the effects of PDGF-BB and ET-1 on lyso-PC–induced SMC migration. Lyso-PC–induced migration was significantly enhanced by PDGF-BB. Lyso-PC–induced migration was also enhanced by ET-1. It is of note that ET-1 alone cannot induce SMC migration at the concentrations used in this study (Figure 6B⇓).
Table 2⇓ shows the effect of the ETA receptor antagonist BQ-123 on the enhancement by ET-1 of lyso-PC–induced migration. BQ-123 clearly inhibited the enhancement by ET-1. This indicates that the enhancement by ET-1 of lyso-PC–induced migration is probably mediated through ETA receptors. By contrast, BQ-123 had no significant effect on lyso-PC–induced migration in the absence of ET-1.
Effect of AM on Lyso-PC–Induced SMC Migration
Figure 7A⇓ shows effects of various concentrations (10−9, 10−8, 10−7, and 10−6 mol/L) of human AM on 5×10−6 mol/L or 5×10−5 mol/L lyso-PC–induced SMC migration. Human AM significantly inhibited 5×10−6 mol/L lyso-PC–induced migration at concentrations of 10−7 and 10−6 mol/L. Human AM also inhibited 5×10−5 mol/L lyso-PC–induced migration in a concentration-dependent manner between 10−8 and 10−6 mol/L.
The effects of AM on cellular cAMP levels in cells treated with 5×10−6 or 5×10−5 mol/L lyso-PC are shown in Figure 7B⇑. In parallel with the inhibition by AM of lyso-PC–induced SMC migration, cellular cAMP increased after treatment with AM (Figure 7A⇑ and 7B⇑).
Effects of 8-Bromo-cAMP and Forskolin on Lyso-PC–Induced SMC Migration
To elucidate whether the inhibitory effect of AM on lyso-PC–induced SMC migration is causally linked to the increase in cellular cAMP, we examined the effect of 8-bromo-cAMP, a cAMP analogue, on lyso-PC (5×10−5 mol/L)–induced SMC migration. Inhibition of lyso-PC–induced SMC migration by AM could be reproduced by this analogue at concentrations of 10−6 and 10−4 mol/L (Figure 8A⇓).
Furthermore, the effect of forskolin, an activator of adenylate cyclase, on lyso-PC (5×10−5 mol/L)–induced SMC migration was examined. The addition of forskolin also reduced lyso-PC–induced SMC migration at concentrations of 10−7 and 10−5 mol/L (Figure 8B⇑).
In the present study, we show that lyso-PC, a prominent component of oxidized LDL,3 5 induces migration of human coronary artery SMCs. To the best of our knowledge, this is the first demonstration concerning the relationship between lyso-PC and SMC migration. This may support the finding shown by Autio and coworkers17 and subsequently by our group12 that oxidized LDL stimulates the migration of cultured rat or human SMCs, because phosphatidylcholine of LDL is found to be converted to lyso-PC during oxidized modification.5 Furthermore, both lyso-PC and oxidized LDL are shown to stimulate SMC proliferation.9 18 Several lines of evidence suggest that oxidized LDL19 and its major phospholipid component, lyso-PC, are present in the intima. Therefore, these observations raise the hypothesis that the conversion of phosphatidylcholine into lyso-PC in the intima may render it chemotactic for medial SMCs, thus contributing to plaque formation and atherogenesis in the coronary artery.
The specific mechanisms of action of lyso-PC as a chemotactic factor for human coronary artery SMCs remains to be established, but the present findings at least allow us to say something about the possibilities. bFGF was a weak chemoattractant for human coronary artery SMCs and neutralizing antibody to bFGF-inhibited lyso-PC–induced migration of human coronary artery SMCs. Actually, lyso-PC increased bFGF production in these cells. These results raise the possibility that lyso-PC stimulates human coronary artery SMC migration at least in part through release of bFGF, which is similar to the mechanism involved in stimulation of SMC proliferation by lyso-PC.9 Furthermore, we showed that vitamin E significantly inhibited lyso-PC–induced SMC migration but could not inhibit the bFGF-induced SMC migration. It is therefore likely that the vitamin E–inhibitable step in the lyso-PC–induced SMC migration is before stimulation by bFGF.
Next, we showed that lyso-PC–induced migration could be enhanced by a potent mitogen, PDGF-BB, and the endothelium-derived vasoconstrictive and growth-promoting peptide ET-1. It is of note that ET-1 significantly enhanced lyso-PC–induced migration even though ET-1 alone had no effects on SMC migration, which means the synergism of ET-1 with lyso-PC. Actually, a low concentration of exogenous ET-1 is found to potentiate the vasoconstrictive or mitogenic action of norepinephrine,20 serotonin,21 or PDGF.22 In addition, ET-induced enhancement of SMC migration was clearly inhibited by the ETA receptor antagonist BQ-123. It is therefore likely that this naturally occurring phospholipid may function in combination with PDGF or ET-1 (possibly mediated by ETA receptors) and may contribute to the intimal thickening in coronary atherosclerotic lesions.23 24 Conversely, phosphatidylcholine was without significant activity, and lysophosphatidylinositol and lysophosphatidylserine were much less effective than lyso-PC. However, the precise structural specificity of lyso-PC that has such an effect on SMC migration remains to be clarified at this time.
Then, we showed that human AM strongly inhibits lyso-PC–induced migration of human coronary artery SMCs in a concentration-dependent manner. Actually, 5×10−5 mol/L lyso-PC–induced SMC migration was significantly inhibited by human AM at concentrations of 10−8 to 10−6 mol/L. Although human AM (1–52) is the major circulating form of AM, the normal plasma concentrations (≈10−11 to 10−10 mol/L) are much lower than those of synthetic human AM that inhibited SMC migration in our study.25 26 However, local levels of AM in coronary artery tissues may be much higher than plasma concentrations of AM, because it has been demonstrated that a considerable amount of AM is synthesized in and secreted from vascular endothelial cells.27 Taking the matter into account, our results suggest that AM, by acting locally as a paracrine agent, inhibits the migration of human coronary artery SMCs. Recently, we showed that AM had a modest antiproliferative effect in cultured rat aortic SMCs.28 Consequently, these findings may raise the possibility that human AM in local vascular tissues antagonizes the development of intimal thickening during the process of coronary atherosclerosis. However, this experiment was done on cultured SMCs. Therefore, any extrapolation from the present experiment on cultured SMCs to in vivo conditions should be carefully performed.
Finally, we have shown that the inhibitory effect of AM on lyso-PC–induced SMC migration is causally linked to the increase in the cellular cAMP. Actually, human AM increased cAMP levels, and this effect paralleled the migration-inhibitory effect of AM. The cAMP-elevating agents 8-bromo-cAMP and forskolin significantly inhibited lyso-PC–induced migration. These results suggest that AM inhibits lyso-PC–induced migration, at least in part, through a cAMP-dependent process.
In conclusion, our results indicate that lyso-PC, a polar phospholipid that accumulates in oxidized LDL and in atherosclerotic arterial wall,3 4 can induce migration of human coronary artery SMCs, possibly in combination with PDGF or ET-1, and that this lyso-PC–induced SMC migration is at least in part caused by the release of endogenous bFGF. The conversion of phosphatidylcholine of LDL into lyso-PC during oxidized modification in atherosclerotic arterial wall may contribute to coronary atherogenesis, especially to the initial step of this disorder. Therefore, interventions that prevent this conversion in vivo may be predicted to act to reduce coronary artery medial SMCs into the intima in various pathophysiological settings. Furthermore, our results indicate that AM, a novel endothelium-derived relaxing factor, and vitamin E, an antioxidant, inhibit lyso-PC–induced human coronary artery SMC migration. However, further studies will be necessary to clarify the exact cellular mechanisms responsible for cellular actions of lyso-PC and the interaction of this phospholipid and other vasoactive substances.
Selected Abbreviations and Acronyms
|bFGF||=||basic fibroblast growth factor|
|ETA||=||endothelin A–type receptor|
|FGF||=||fibroblast growth factor|
|PDGF||=||platelet-derived growth factor|
|SmBM||=||smooth muscle cell basal medium|
|SMC||=||smooth muscle cell|
|TGF-β1||=||transforming growth factor-β1|
This work was supported by a grant-in-aid for scientific research from the Ministry of Education, Science, and Culture, Japan. The authors gratefully acknowledge the technical assistance of Atsumi Ohnishi, Yuka Inoshita, and Miyuki Kitakaze of the Division of Hypertension and Atherosclerosis, The First Department of Internal Medicine, Osaka City University Medical School.
- Received November 17, 1997.
- Revision received January 20, 1998.
- Accepted February 4, 1998.
- Copyright © 1998 by American Heart Association
Schwartz SM, Campbell GR, Campbell JH. Replication of smooth muscle cells in vascular disease. Circ Res. 1986;58:427–444.
Witztum JL, Steinberg D. Role of oxidized low density lipoprotein in atherogenesis. J Clin Invest. 1991;88:1785–1792.
Parthasarathy S, Steinbrecher UP, Barnett J, Witztum JL, Steinberg D. Essential role of phospholipase A2 activity in endothelial cell-induced modification of low density lipoprotein. Proc Natl Acad Sci U S A. 1985;82:3000–3004.
Portman OW, Alexander M. Lysophosphatidylcholine concentration and metabolism in aortic intima plus inner media: effect of nutritionally induced atherosclerosis. J Lipid Res. 1969;10:158–165.
Clowes AW, Schwartz SM. Significance of quiescent smooth muscle migration in the injured rat carotid artery. Circ Res. 1985;56:139–145.
Chai YC, Howe PH, DiCorleto PE, Chisolm GM. Oxidized low density lipoprotein and lysophosphatidylcholine stimulate cell cycle entry in vascular smooth muscle cells. J Biol Chem. 1996;271:17791–17797.
Kohno M, Yokokawa K, Kano H, Yasunari K, Minami M, Hanehira T, Yoshikawa J. Adrenomedullin is a potent inhibitor of angiotensin II–induced migration of human coronary artery smooth muscle cells. Hypertension. 1997;29:1309–1313.
Kohno M, Yokokawa K, Yasunari K, Kano H, Minami M, Ueda M, Yoshikawa J. Effect of natriuretic peptide family on the oxidized LDL-induced migration of human coronary artery smooth muscle cells. Circ Res. 1997;81:585–590.
Bilato C, Pauly RR, Melillo G, Monticone R, Gorelick-Feldman D, Gluzband YA, Sollott SJ, Ziman B, Lakatta EG, Crow MT. Intracellular signaling pathways required for rat vascular smooth muscle cell migration. J Clin Invest. 1995;96:1905–1915.
Kohno M, Kano H, Horio T, Yokokawa K, Yasunari K, Takeda T. Inhibition of endothelin production by adrenomedullin in the vascular smooth muscle cells. Hypertension. 1995;25:1185–1190.
Horio T, Kohno M, Yasunari K, Yokokawa K, Minami M, Takeda T. Adrenomedullin as a novel antimigration factor of vascular smooth muscle cells. Circ Res. 1995;77:660–664.
Wallenstein S, Zucker CL, Fleiss JL. Some statistical methods useful in circulation research. Circ Res. 1980;47:1–9.
Chatterjee S. Role of oxidized human plasma low density lipoproteins in atherosclerosis: effects on smooth muscle cell proliferation. Mol Cell Biol. 1992;111:143–147.
Ylä-Herttuala S, Palinski W, Rosenfeld ME, Parthasarathy S, Carew TE, Butler S, Witztum JL, Steinberg D. Evidence for the presence of oxidatively modified low density lipoprotein in atherosclerotic lesions of rabbit and man. J Clin Invest. 1989;84:1086–1095.
Yang Z, Richard V, von Segesser L, Bauer E, Stulz P, Turina M, Lüscher TF. Threshold concentrations of endothelin-1 potentiate contractions to norepinephrine and serotonin in human arteries: a new mechanism of vasospasm. Circulation. 1990;82:188–195.
Jawien A, Bowen-Pope DF, Lindner V, Schwartz SM, Clowes AW. Platelet-derived growth factor promotes smooth muscle migration and intimal thickening in a rat model of balloon angioplasty. J Clin Invest. 1992;89:507–511.
Kohno M, Hanehira T, Kano H, Horio T, Yokokawa K, Ikeda M, Minami M, Yasunari K, Yoshikawa J. Plasma adrenomedullin concentrations in essential hypertension. Hypertension. 1996;27:102–107.