003 1-3998/89/2604-0330$02.00/0 PEDlATRIC RESEARCH Vol. 26, No. 4, 1989 Copyright 0 1989 International Pediatric Research Foundation, Inc. Printed in U.S. A. Oxygen-Related Prostaglandin Synthesis in Ductus Arteriosus and Other Vascular Cells MARLENE RABINOVITCH, NANCY BOUDREAU, GEORGE VELLA, FLAVIO COCEANI, AND PETER M. OLLEY Divisions of Cardiology, Pathology and Neurosciences, Research Institute, The Hospital for Sick Children and the Departments of Pediatrics, Pathology and Physiology, University of Toronto, Toronto, Ontario, Canada M5GIX8 ABSTRACT. We compared oxygen-related prostaglandin natal constriction of the vessel wall (2). The identity of the synthesis in fetal lamb ductus arteriosus (DA) pulmonary constrictor remains unknown, although its formation from a artery (PA) and aorta endothelial and smooth muscle cells. cytochrome P-450-dependent monooxygenase pathway is likely We measured basal synthesis of 6-keto-PGR, and PGE2, (3). We have previously established and characterized endothelial the response to calcium ionophore (A23187), a nonspecific and smooth muscle cell cultures from the DA (4). When com- stimulus of prostaglandin production, as well as the re- pared with cells from the Ao and PA, DA cells are similar in sponse to oxygen, a perinatal stimulus, monitoring both appearance and growth, but show both qualitative and quanti- the effects of hyperoxia (95% 02) and hypoxia (2% 02). tative differences in protein synthesis (4). The purpose of this In addition, we established whether differences observed study was to establish whether oxygen-induced production of in fetal lamb PA cells related to oxygen tension were also prostaglandins is also different in DA cells. The results of these observed in newborn central and microvessel PA cells. Our experiments could therefore serve as the background for future results indicate that DA endothelial cells increase 6-keto- work related to the mechanism of oxygen-induced release of PGF,, in response to ionophore (p < 0.05). With hyper- vasoactive substances. We compared basal DA endothelial and oxia, DA endothelial cells increase PGE2 synthesis and smooth muscle synthesis of PGE2 and 6-keto-PGF,,, the stable DA smooth muscle cells increase 6-keto-PGK, (p < 0.05 metabolite of prostacyclin, with that of PA and Ao cells, as well and 0.02, respectively). Aorta smooth muscle cells increase as the response to a nonspecific stimulus, the calcium ionophore 6-keto-PGF,, in response to ionophore and hyperoxia (p (A23 187), and a perinatal stimulus, oxygen. We also assessed < 0.003 and 0.05, respectively). PA endothelial and smooth whether changes in prostaglandin production in the fetal PA muscle cells have higher levels of basal prostaglandin were also reflected in newborn central and microvessel PA cells. synthesis when compared with DA and aorta. In response to ionophore, increased 6-keto-PGP1, is observed in both PA endothelial and smooth muscle cells (p < 0.02 and MATERIALS AND METHODS 0.0004, respectively), and PGE2 is increased in PA smooth Materials. DPBS, medium 199 (contains 5 mM CaC12) with muscle cells (p < 0.003). Hypoxia, however, decreases PA 25 mM HEPES buffer and glutamine, antibiotic-antimycotic smooth muscle production of both 6-keto-PGS, and PGE2 solution (penicillin, streptomycin, fungizone), and FCS were (p < 0.02 and 0.01, respectively). Similar observations purchased from GIBCO, Burlington, Ontario. Culture dishes (35 were made in newborn lamb central and microvessel PA mm) were Nunclon, from Nunc, Denmark, flasks and 15-mL cells. Thus, the hyperoxia-induced increase in prostaglan- tubes were from Costar, Cambridge, MA. Microcamer beads din synthesis in DA (endothelium and smooth muscle) and (Cytodex III), were from Pharmacia, Uppsala, Sweden. Antibody aorta (smooth muscle) and the hypoxia-induced decrease to human factor VIII was obtained from Behring Diagnostics in prostaglandin synthesis in PA (smooth muscle) seem to Inc. Branchburg, NJ and ovine factor VIII antibody was kindly be phenotypic properties of the cells in culture. (Pediatr. supplied by Dr. M. Andrew, McMaster University, Hamilton, Res 26: 330-335,1989) Ontario, Canada. Angiotensin converting enzyme substrate, [3H] benzoyl-phe-ala-pro came from Ventrex Laboratories, Portland, Abbreviations ME. Radiolabeled compounds, [5,6,8,9,11,12,14,15,-'H(N)]ar- achidonic acid, (83.6 Ci/mmol) and [5,6,8,11,12,14,15-3H(N)] DA, ductus arteriosus PGE2 (185 Ci/mmol) and [5,8,9,1 1,12,14,15,-3H(N)] 6-keto- Ao, aorta PGF,, (157 Ci/mmol) were obtained from New England Nu- PA, pulmonary artery clear, Boston, MA. Unlabeled thromboxane B2, prostaglandin DPBS, Dulbecco's PBS D2, PGF2,, were from Upjohn, Kalamazoo, MI and 5-HETE and 15-HETE, from Biomol, Philadelphia, PA. Calcium iono- phore A23 187 was obtained from Sigma Chemical Co., St. Louis, MO and scintillation fluid (ACS)-from Amersham, Oakville, Normal closure of the DA is dependent upon two processes, Ontario, Canada. the prenatal development of intima1 cushions (1) and the post- Culturing and characterizing vascular cells. Fetal Rambouillet lambs were delivered by cesarean section on d 135-138 of Received January 18, 1989; accepted May 25, 1989. gestation (term 145 d), and the great vessels were removed as Correspondence and reprints Marlene Rabinovitch, M.D., Department of Car- previously described (3, 4). The DA, PA, and Ao were separated, diology, The Hospital for Sick Children, 555 University Avenue, Toronto, Ontario, Canada M5G 1x8. opened, and rinsed in DPBS supplemented with ca2+and Mg2+ Supported by the Medical Research Council of Canada Grant MT 857. M.R. is (2.5 and 4.1 mM, and 3% antibiotic-antim~cotic a Research Associate of the Heart and Stroke Foundation of Ontario. solution. Endothelial cells were subsequently harvested by scrap- 330 PROSTAGLANDINS IN DUCTUS ARTERIOSUS CELLS 33 1 ing the luminal surface of the vessel with a no. 11 scalpel blade from at least three different harvests but each flask was used as (4, 5) and immersing the blade in a 35-mm culture dish contain- its own control. All cells were incubated with 1 pCi [3H]arachi- ing medium 199 (contains 5 mM CaC12), 20% FCS, and 1% donic acid as described above, and prostaglandin synthesis was antibiotic-antimycotic solution. The culture medium was not assessed during 1 h of exposure to room air. Fresh serum-free supplemented with growth factors. Cells were then grown to medium was added to the flask and repeat assessment was made confluence and passaged nonenzymatically 1:2. Growth rates during one hour of hypoxia (2% 02, 5% C02, 93% N2). Then, and attachment efficiencies have been previously published (4). 0.5 h equilibration in room air intervened after which, fresh Differences were apparent comparing endothelial and smooth serum-free medium with calcium ionophore A23 187 (final con- muscle cells, but not in relation to vascular site, i.e. DA, PA, or centration 5 pM) was added. Prostaglandin synthesis was reas- Ao. sessed in room air only to ascertain that, despite the change in After removing the adventitia from the vessels, explants of the precursor pool, the cells could respond to a prostaglandin-pro- muscular media were cultured in 25-mL flasks using the same ducing stimulus. medium as the endothelial cells, but with 10% FCS added. In all experiments, cells were checked after the exposure to Smooth muscle cells were propagated from the explants (6) and ionophore, hypoxia, or hyperoxia to verify that they had main- also passaged 1:2 nonenzymatically at confluence. tained an intact cobblestone phenotype and had no signs of Rambouillet lambs, 3-5 days of age, were used as the source detachment or injury from the 5-min bubbling. None was ob- of newborn PA central and microvessel endothelial and smooth served. The culture medium was then removed, and the cells muscle cells. Microvessel endothelial cells were harvested by a from each flask were subsequently scraped in DPBS, aspirated previously described perfusion technique using microcarrier up and down in a syringe via an 18-gauge needle to break up the beads, 40-140 pm (7, 8). The cells that attached to the beads clumps and counted (model HZ counter, Coutler Electronics, were largely endothelial, but some smooth muscle cells were also Hialeah, FL) so that results could be referred quantitatively to present that formed colonies that were later subcultured. Endo- the constituent cells. The culture medium removed was acidified thelial and smooth muscle cells from the central PA were har- with 1 N HC1 to pH 3 and passed through a C18Sep-Pak followed vested and cultured following the technique described for fetal by 2 x 5 mL DPBS washes. The cartridges were then rinsed with cells. 6 mL H20,6 mL hexane and the prostaglandins were eluted off Endothelial cells in culture were characterized by phase con- the Sep-Pak with 12 mL acetone. Recovery was 86-90%. The trast light microscopy as forming a typical "cobblestone," contact acetone fraction was evaporated under a stream of N2 and inhibited monolayer (4, 5, 7). At passage 2, positive immunoflu- samples were reconstituted in 250 pL acetonitrile. orescence was demonstrated using both a human and ovine Prostaglandins were resolved using an HPLC system from antibody to factor VIII (4, 5). Furthermore, levels of angiotensin Waters, Milford, MA which included two model 5 10 pumps and converting enzyme activity (9) were accordant with published a CISNova-Pak column (15.0 x 3.5 cm). A flow rate of 1.5 mL/ values (4). Newborn central PA endothelial cells were similar to min with an isocratic gradient system of 0.1 % H3P0JCH3CN both newborn microvessel PA and fetal central PA cells in terms (3:l by vol) (pH 3.5) increasing linearly to 100% CH3CN over of converting enzyme activity, but they took approximately 1 18 min was used.