Role of the Lipoxygenase Pathway in Angiotensin II–Induced Vasoconstriction in the Human Placenta
We have previously shown that the vasopressor effect of angiotensin II (Ang II) is inhibited by lipoxygenase (LO) blockers. To elucidate the specific mechanism involved, we studied the relationship between the contractile effect of Ang II and LO products in a human placental preparation. In perfused placental cotyledons, Ang II (boluses of 1 to 10 μg) increased perfusion pressure and 12-hydroxyeicosatetraenoic acid (12-HETE) release. The LO blockers phenidone and n-propyl gallate reduced the maximal Ang II–induced increment in pressure from 26±3 to 16±3 and 18±4 mm Hg, respectively (P<.05 for both). Ang II alone (10 μg) increased 12-HETE release from 8.9±3.6 to 37.6±0.4 ng/min, and this rise was entirely blocked by both phenidone and n-propyl gallate. Pressure increase generated by an increase in flow rate had no effect on 12-HETE formation. In deendothelialized umbilical artery segments, Ang II (10–7 mol/L) increased 12-HETE formation by 47±5% (n=20). In cultured umbilical artery smooth muscle cells, Ang II increased 12-HETE formation from 48.1±7.2 to 75.1+15.3 ng/mg protein, and this effect was also blocked by the specific LO inhibitor baicalein (10−6 mol/L). These results provide evidence that the vasopressor effect of Ang II is functionally coupled to 12-LO activation, which apparently takes place in arterial smooth muscle cells.
Angiotensin II–mediated vasoconstriction is involved in both the physiological maintenance of arterial pressure and the pathogenesis of various forms of experimental and human hypertension. The direct vascular effects of Ang II depend on early activation of phospholipase C to induce the hydrolysis of phosphatidylinositol diphosphate into inositol 1,4,5-trisphosphate, which releases calcium from intracellular storage sites and DG, which activates calcium- and phospholipid-sensitive protein kinase C.1 DG can also be generated by direct phospholipase C–mediated hydrolysis of phosphatidylinositol2 or from phosphatidylcholine via a specific phospholipase D.3 Metabolism of DG by DG lipase leads to the release of arachidonic acid from the Sn 2 position. Indeed, Ang II–induced increase in DG in glomerulosa cells is followed by the release of arachidonic acid,4 presumably via activation of DG lipase.5 Arachidonic acid thus formed may be rapidly oxidized via several enzyme systems, including the LO pathway, which has been shown to be present in target cells for Ang II action.6 7 8 9 Furthermore, Ang II was previously shown to activate 12-LO in glomerulosa cells6 7 8 and in porcine vascular smooth muscle cells.9 Activation of the 12-LO pathway appears mandatory for Ang II–dependent steroidogenesis, as indicated by reports that LO inhibition blocks the aldosterone response to Ang II and that this effect can be overcome by the addition of 12-HETE.6 7
Observations that LO blockers attenuate Ang II–induced vasoconstriction10 and lower arterial pressure in renovascular hypertensive rats11 indirectly suggest a role for LO activation in the vasopressor effects of Ang II. However, evidence that Ang II–induced vasoconstriction is functionally coupled to activation of a distinct LO enzyme has thus far not been presented. Recently, McQueen et al12 have reported that the human placental vascular system is Ang II responsive, such that arterial pressure in placental cotyledons perfused with Ang II rises in a dose-dependent fashion reminiscent of Ang II's effects in the human systemic circulation. In the present study, we examined the relationship between the vasopressor effect of Ang II and LO activity in the human placental circulation.
Umbilical Artery Segments
Approval for the use of human placental tissue was obtained from the human subject committee of our medical center. Umbilical cords were collected from normal pregnancies and placed in ice-cold saline. The arteries were dissected from the umbilical cord and isolated, and the adventitia were separated from the media and removed mechanically. The arteries were then cut longitudinally and divided into 5-cm-long segments. The exposed endothelial surface was scraped gently with a coverslip to remove the endothelial lining, and the remaining lamina media was equilibrated for 2 hours in a Krebs' bicarbonate buffer containing (mmol/L) NaHCO3 25.6, NaCl 122.1, KCl 5.16, CaCl2-2H2O 1.4, MgSO4 1.2, and EDTA 0.03, along with 2 g/L glucose. The medial segments were then resuspended in the same buffer solution and incubated at 37°C for 10 minutes in a shaking incubator under O2 with the LO inhibitor baicalein (10–6 mol/L) or its vehicle (dimethyl sulfoxide, 50 μL in 2 mL). Ang II (l0–7 mol/L) was then added, and the incubation was carried on for 10 additional minutes. The incubation was terminated by adding 2 mL of ice-cold ethanol, followed by removal of the tissue. Samples were stored at −80°C and subsequently homogenized, extracted, and assayed for HETE concentrations. The results are expressed per tissue protein.
Umbilical Artery Smooth Muscle Cells
Umbilical cords were collected shortly after delivery, squeezed gently to remove blood, and placed in ice-cold sterile dishes containing 50 mL of DMEM (Biological Industries), glucose (1 g), streptomycin (20 mg/mL), penicillin (20 000 U/mL), and nystatin (1250 U/mL). Arteries were dissected from the cord. The adventitia were removed mechanically, and the remaining arterial tissue was chopped into tiny slices (1 to 3 mm in width). Ten to 15 arterial segments per dish were subsequently placed in Corning gelatin-coated dishes (35 mm) and tightened to the gelatin layer. One milliliter of Earle's medium 199, containing l-glutamine, penicillin/streptomycin/nystatin (as above), and 20% fetal calf serum, was added, and the dishes were transferred to a 95% air/5% CO2 incubator at 37°C. Four hours later, an additional 1 mL of medium was added. Cell migration was usually detected within 5 to 7 days. Cells were fed twice a week, and at confluence they were trypsinized and transferred to Petri dishes containing no gelatin. Experiments were carried out on cells derived from the second to fifth passages. Cells were randomly screened for smooth muscle actin expression by immunofluorescent staining using mouse anti-human smooth muscle actin (Dako) visualized by fluorescein-conjugated goat anti-mouse immunoglobulin G (Caitag Laboratories).
Before the experiments, cells were incubated for 24 hours in serum-free medium 199. Cells were then preincubated with baicalein (10–5 mol/L) or its vehicle for 10 minutes. Ang II (10–7 mol/L) was subsequently added, and the incubation was carried out for 10 additional minutes and terminated by adding an equal volume of ice-cold ethanol. Cells were scraped off the dishes mechanically, and the mixture was aspirated and frozen at −80°C until assayed.
Placentas were taken immediately post partum from women with uncomplicated pregnancies, who delivered between the 37th and 41st week. All placentas were products of spontaneous vaginal deliveries. Placentas from cesarian sections were excluded, since these procedures, per se, apparently result in LO activation.13 In each experiment, the artery and vein of a single cotyledon were cannulated.
The cotyledon was perfused with Krebs' bicarbonate buffer at pH 7.4 via an arterial cannula at a fixed flow rate of 5 mL/min with a peristaltic infusion pump 202 (Medix Intl), resulting in a basal pressure of 110 mm Hg in the system. During these procedures, the placenta was kept in the buffer at 37°C in a water bath and aerated with a 95% O2/5% CO2 mixture. The placenta was allowed to equilibrate for at least 50 minutes before the initiation of the experiments, at which time all blood was washed out, and basal HETE outflow was stabilized.
Pressure was measured continuously by a Harvard transducer and registered by a recorder (Soltec). Before each experiment with Ang II, a baseline sample of 10 mL was collected from the venous outflow. Five minutes after each bolus dose of Ang II (1 to 10 μg, Hypertensin, CIBA), another outflow sample of 10 mL was taken. This time course for Ang II effects was determined in a set of preliminary experiments. Samples were placed on ice and in the dark, centrifuged for 15 minutes at 750g at 4°C, and frozen at –80°C until assayed.
In all samples, proteins were first precipitated out with 80% ethanol (or 15% ethanol for the placental outflow). In samples originating in cells or arterial segments, the supernatant was then transferred and dried under nitrogen. The precipitate was redissolved in a solution containing 1.5 mL methanol-NaOH (0.2N) and 150 μL H2O. Tubes were vigorously shaken and then blown with nitrogen at room temperature in the dark for 45 minutes. Distilled water was then added to yield a 15% methanol NaOH/H2O solution. The samples were acidified to pH 3.0 with 0.1N HCl and subsequently transferred to 1000-mg C18 Accubond columns (J&W Scientific) that had been prewashed successively in 5 mL methanol and 5 mL water. The elution of HETEs was performed using a modification of the method of Powell.14 The column was successively washed with 15% ethanol in water (3 mL), water (5 mL), and petroleum ether (3 mL), and the compounds of interest were eluted with ethyl acetate (4 μL). All solvents were of HPLC grade (Merck). The extracted samples in ethyl acetate were dried under nitrogen, dissolved in 500 μL ethanol, filtered, dried again, and redissolved in 80 μL of 100% methanol. Based on preliminary experiments with tritiated 5-, 12-, and 15-HETE, the mean recoveries for these compounds were 87%, 89%, and 87%, respectively. Since our samples contained no 5-HETE, the recovery for 12- and 15-HETE was calculated for each sample through adding known amounts of 5-HETE (Biomol) before extraction.
The HPLC determination was modified from the method described by Estra.15 Extracted samples were separated by reverse-phase chromatography using a C18-column (Shandon Scientific), an integrator, and an isocratic pump (Milton Roy). The mobile phase was of 80% methanol, 20% water, 0.1% acetic acid, and 0.05% triethylene at 1 mL/min. Effluent was monitored at 237 nm with a Milton Roy UV detector. With this procedure, the detection limit of the HPLC is 0.5 ng of 12- and 15-HETE. For each set of experiments, representative data points were also validated by radioimmunoassay for 12- and 15-HETE (Advanced Magnetics) as previously described,6 8 10 and the overall correlation between the HPLC and radioimmunoassay was 0.84. All values are expressed as mean±SEM. Statistical assessment was made using analysis of variance, and Student's t test was applied for specific comparisons. Statistical significance was assumed for values of P<.05.
Effect of Ang II on LO Products in Umbilical Artery Segments
Both 12- and 15-HETE but not 5-HETE were detected by HPLC tracing in samples extracted from either the medium or the arterial tissue itself (Fig 1⇓). However, 15-HETE production was significantly reduced by the cyclooxygenase inhibitor indomethacin (10–6 mol/L, 42±10% of basal generation, n=5, P<.05), whereas 12-HETE production tended to be higher in the presence of indomethacin (147±32%, n=5, P=NS). Ang II increased the amount of 12-HETE released by deendothelialized umbilical artery segments by ≈50%, and this increase was blocked by the specific LO inhibitor baicalein (Table⇓). 12-HETE release by arterial tissue not subjected to deendothelialization was quantitatively similar to that seen in deendothelialized segments, thus suggesting that basal and Ang II–induced 12-HETE release predominantly reflect processes within the lamina media.
Effect of Ang II on LO Products in Umbilical Artery Smooth Muscle Cells
12-HETE, but not 15- or 5-HETE, was detected by HPLC tracing of extracts from the cellular material or the medium of cultured umbilical artery smooth muscle cells. Ang II increased cell-associated 12-HETE production from 48.1±7.2 to 75.1±15.3 ng/mg protein (n=7, P<.01), and this effect was not seen in the presence of the LO inhibitor baicalein16 (10–6 mol/L, 26.7±6.4 ng/mg protein). Ang II also increased the amount of 12-HETE released to the incubation medium, although both basal and Ang II–stimulated 12-HETE release were one order of magnitude less than the cell-associated 12-HETE concentrations (data not shown).
Effects of Ang II in the Placental Circulation
Ang II increased both arterial pressure and 12-HETE release in the placental system. The dose-related effect of Ang II on pressure and 12-HETE output is depicted in Fig 2⇓. As shown, stepwise increases in Ang II bolus from 1 to 10 μg elicited incremental changes in placental pressure and in 12-HETE release. The increase in 12-HETE output was transient and reversible. This was evidenced by a commensurate decline in 12-HETE release when a 1 μg Ang II bolus followed a 10-fold larger dose of Ang II (10 μg Ang II, 13.5±2 ng/min; 1 μg Ang II, 4.6±3.1 ng/min; P<.05). In this system, norepinephrine (1 to 2000 μg) did not elicit a detectable change in pressure (basal pressure, 118±6 mm Hg; 2000 μg norepinephrine, 117±7 mm Hg) or 12-HETE release (basal, 14.1±2.3 ng/mL; norepinephrine, 13.6±2.3 ng/mL). Nevertheless, to address the possibility that the rise in 12-HETE discharge rate was secondary to an increase in pressure per se rather than a response to Ang II, perfusion pressure was raised by 20 to 25 mm Hg by increasing the flow rate from 5 to 7 mL/min (Fig 3⇓). The attained rise in pressure had no effect on 12-HETE release (expressed in terms of nanograms per minute).
Two structurally unrelated LO blockers, phenidone and n-propyl gallate17 inhibited the Ang II–induced increment in 12-HETE output and concomitantly attenuated the vasopressor effect of Ang II (Fig 4⇓). Typical HPLC tracings of basal, Ang II–stimulated, and Ang II–stimulated phenidone-treated placental effluents are represented in Fig 5⇓. The insolubility of more specific LO blockers, (ie, esculetin, baicalein, and cinnamyl-3,4-dihydroxy α-cyanocinnamate) in aqueous solution made direct bolus injections or even slow perfusion of these agents through the placental circulation impossible, causing clogging of the perfusion system. The reduction in the vasopressor effect of Ang II did not, however, depend on the cyclooxygenase-blocking properties of the dual LO-cyclooxygenase inhibitor phenidone, since indomethacin did not attenuate the pressor response to Ang II (Ang II, 25±3 mm Hg; Ang II+indomethacin, 28±4 mm Hg). In fact, in some experiments (though not in the overall analysis), indomethacin augmented the rise in pressure in response to Ang II. Indomethacin also increased 12-HETE concentration in the venous outflow, presumably secondary to generalized shunting from the cyclooxygenase to the LO pathway in the placenta.
The present study provides evidence that Ang II leads to increased 12-HETE release in human arterial smooth muscle cells and that this effect is linked to Ang II–dependent vasoconstriction in the human placental circulation. These findings are consistent with our previous observations that LO blockers reduce the vasopressor effect of Ang II in the intact rat and impair the contractile response to Ang II of rat femoral artery rings.10 The results are also complementary to a previous report that LO blockers attenuate, but do not entirely inhibit, the intracellular calcium signal elicited by Ang II in cultured rat aortic smooth muscle cells.18 Although Ang II likewise increased 15-HETE concentration in arterial segments, its production was at least in part due to cyclooxygenase activity. Furthermore, 15-HETE was not detected in cultured vascular smooth muscle cells or in the placental perfusate. Overall, it appears from these findings that 12-HETE rather than 15-HETE is the LO product linked to Ang II–driven vasoconstriction.
The present study demonstrates that Ang II–induced vasoconstriction is associated with a marked increase in the “venous” efflux of 12-HETE in the placental circulation. The precise source for this Ang II–dependent increase in 12-HETE production cannot be determined directly in the placental perfusion studies. However, the arterial vessels in this system are the most likely target tissue for Ang II's effects. Indeed, in isolated umbilical arteries supplying the placenta, we have localized this effect to the smooth muscle cells derived from the lamina media of the artery. The precise intracellular mechanism through which 12-HETE may be involved in Ang II–dependent vasoconstriction is currently unknown. 12-HETE may be required for the generation of Ang II–induced intracellular calcium transients.18 19 Recent evidence also suggests that LO products can activate mitogen-activated protein kinases.20 Finally, high-affinity binding sites for 12-HETE have been identified in certain cells, and activation of these putative receptors may have a role in the regulation of protein kinase.21 These effects may explain the requirement for intracellular 12-HETE production in Ang II–driven cellular processes.
A similar relative rise in 12-HETE release in response to Ang II was seen in medial segments of the umbilical artery and in cultured umbilical smooth muscle cells. The relative Ang II–dependent rise in 12-HETE output was, however, much larger in the “intact” placental circulation than in the medial portion of the artery. Several mechanisms may account for this difference. First, the vasopressor effect of Ang II takes place mainly in the smooth muscle cells of the small resistance arterioles rather than in the large umbilical artery used for the medial segments and culture studies. Smooth muscle cells in the resistance arterioles may display a much more pronounced 12-HETE response to Ang II. Second, the presence of additional cellular elements in the intact placenta may facilitate the effect of Ang II on arterial smooth muscle via paracrine interactions. Finally, some of the increment observed in the placental circulation may originate in cell types other than vascular smooth muscle cells.
Of note is the finding that Ang II–induced 12-LO activation leads to the release of 12-HETE, presumably formed within the vascular smooth muscle cells, to the extracellular space, raising its concentration in the placental circulation. This may have pathophysiological sequels, since 12-HETE has been shown to directly promote vascular smooth muscle cell migration,22 enhance endothelial cell growth and replication,23 and inhibit renin secretion.24
Recently, Natarajan and colleagues9 25 have shown that two different Ang II–responsive cell types, ie, porcine vascular smooth muscle cells9 and human adrenal glomerulosa cells,25 express the porcine leukocyte form of 12-LO. The precise molecular form of 12-LO in human umbilical smooth muscle cells awaits characterization. Nevertheless, the present study is the first demonstration of 12-LO activity linked to a specific hormone-regulated arterial function in humans.
Selected Abbreviations and Acronyms
|Ang II||=||angiotensin II|
|HPLC||=||high-performance liquid chromatography|
The authors wish to thank Rivka Morris for her expert assistance in the preparation of this manuscript. We thank the staff of the delivery room of the Tel Aviv Sourasky Medical Center, and especially Zwia Gonen, for their help in obtaining material for this study.
Reprint requests to N. Stern, MD, Institute of Endocrinology, Elias-Sourasky-Tel-Aviv Medical Center, 6 Weizman St, Tel-Aviv, 64239, Israel.
- Received March 4, 1996.
- Revision received April 24, 1996.
- Revision received September 30, 1996.
Griendling KK, Berk BL, Ganz P, Gimbrone MA, Alexander W. Angiotensin II stimulation of vascular smooth muscle phosphoinositide metabolism. Hypertension. 1987;9(suppl III):III-181-III-185.
Majerus PW, Connoly T, Dechmyn H, Ross T, Ishii H, Bansal V, Wilson D. The metabolism of phosphoinositide-derived messenger molecules. Science.. 1986;234:1519-1523.
Nadler J, Natarajan R, Stern N. Specific action of the lipoxygenase pathway in mediating angiotensin II induced aldosterone synthesis in isolated adrenal glomerulosa cells. J Clin Invest.. 1987;80:1763-1769.
Natarajan R, Stern N, Hseuh W, Do Y, Nadler J. Role of the lipoxygenase pathway in angiotensin II-mediated aldosterone biosynthesis in human adrenal glomerulosa cells. J Clin Endocrinol Metab.. 1988;67:581-591.
Natarajan R, Gu JL, Rossi J, Gonzales N, Lanting L, Xu L, Nadler J. Elevated glucose and angiotensin II increase 12-lipoxygenase activity and expression in porcine aortic smooth muscle cells. Proc Natl Acad Sci U S A.. 1993;90:4947-4951.
Stern N, Golub M, Noazwa K, Berger M, Knoll E, Yanagawa N, Natarajan R, Nadler JL, Tuck ML. Selective inhibition of angiotensin II-mediated vasoconstriction by lipoxygenase blockade. Am J Physiol.. 1989;257:H434-H443.
Nozawa K, Tuck ML, Eggena P, Golub M, Nadler JL, Natarajan R, Stern N. Inhibition of the lipoxygenase pathway reduces blood pressure in renovascular hypertensive rats. Am J Physiol.. 1990;259:H1774-H1780.
McQueen J, Jardine A, Kingdom J, Templeton A, Whittle MJ, Connell JM. Interaction of angiotensin II and atrial natriuretic peptide in the human fetoplacental unit. Am J Hypertens. 1990;3(pt 1):641-644.
Nakadate T, Yamamoto S, Aizu E, Kato R. Effects of flavonoids and antioxidants on 12-0-tetradecanoyl-phorbol-13-acetate-caused epidermal ornithine decarboxylase induction and tumor promotion in relation to lipoxygenase inhibition by these compounds. Jpn J Cancer Res.. 1984;75:214-222.
Saito F, Hori MT, Berger M, Chang CT, Golub M, Stern N, Tuck ML. 12-Lipoxygenase products modulate calcium signals in vascular smooth muscle cells. Hypertension.. 1992;20:138-143.
Rao G, Baas AS, Glasgow WC, Eling TE, Runge MS, Alexander RW. Activation of MAP kinases by AA and its metabolites in VSMC. J Biol Chem.. 1994;269:32586-32591.
Lin B, Khahn WA, Hannum YA, Timar J, Taylor JD, Lundy S, Butovich I, Honn KV. 12 (s)-Hydroxyeicosatetraenoic acid regulation of protein kinase C-α in melanoma cells; role of receptor mediated hydrolysis of inositol phospholipids. Proc Natl Acad Sci U S A.. 1995;92:9323-9327.
Setty BNY, Graeber JE, Stuart MJ. The mitogenic effect of 15- and 12-hydroxyeicosatetraenoic acid on endothelial cells may be mediated via diacylglycerol kinase inhibition. J Biol Chem.. 1987;262:17613-17622.
Antonopillai I, Nadler J, Robin E, Horton R. The inhibitory role of 12- and 15-lipoxygenase products on renin release. Hypertension.. 1987;10:61-66.