0021-972X/91/7301-0060$03.00/0 Journal of Clinical Endocrinology and Metabolism Copyright © 1991 by The Endocrine Society
Vol. 73, No. 1 Printed in U.S.A.
Stimulation by Human Chorionic Gonadotropin of Prostaglandin Synthesis by Early Human Placental Tissue R. A. NORTH*, R. WHITEHEAD, AND R. G. LARKINS Department of Medicine, University of Melbourne, Royal Melbourne Hospital (R.A.N., R.G.L.), and Ludwig Institute for Cancer Research (R. W.), Parkville, Victoria 3050, Australia
ABSTRACT. Successful establishment of pregnancy is dependent on inhibition of clotting and suppression of the maternal immune response at the feto-maternal interface. Early human placental production of prostacyclin (PGI2) and prostaglandin E2 (PGE2) may be important in this process. To examine the possible role of these PGs, we studied PGE and 6-ketoPGF]n (stable metabolite of PGI2) synthesis in human placental (9-17 weeks gestation) organ cultures, and monolayer cultures of purified trophoblasts. PGE2 appeared to be the major prostanoid formed. Other arachidonic acid metabolites identified in placental organ culture were 6-keto-PGFiOJ thromboxane B2, PGF2o, leukotriene B4, 5(S)-hydroxyeicosatetraenoic acid (5HETE), 12-HETE, and 15-HETE. The synthesis of PGE and
6-keto-PGFin altered with gestation and was maximal in the younger placentas. Arachidonic acid (33 (iM) stimulated and indomethacin (28 ^M) inhibited PG production. hCG, including physiological concentrations, stimulated PGE and 6-keto-PGFln synthesis in placental organ cultures. This effect was most striking in the 9-12 week placentas, compared to 15-17 week placentas. A similar hCG-induced stimulation of PGE production occurred in monolayer cultures of trophoblasts. The addition of TSH, FSH, and LH indicated that this effect was specific for hCG. These data suggest that hCG may have a biological role in the regulation of PG synthesis in early human placenta. (J Clin Endocrinol Metab 73: 60-70, 1991)
E
STABLISHMENT and maintenance of pregnancy involve two poorly understood events. Trophoblast cells invade the uterine wall and change the maternal spiral arteries into dilated vessels that terminate in the intervillous space. As pregnancy progresses, these structural changes to the spiral arteries enable the increased maternal blood flow to the feto-placental unit to be met. This is dependent on inhibition of clotting at the maternal-placental interface. The placenta and fetus are not rejected despite their antigenic differences from the mother. While unsuccessful implantation results in spontaneous abortion, it has been postulated that defective implantation and establishment of the maternal placental vascular bed are central in the pathogenesis of preeclampsia (1). One potential mechanism for inhibition of clotting during trophoblast invasion is prostacyclin (PGI2) production, as PGI2 is a potent inhibitor of platelet aggregation. Early human placental tissue has been shown to inhibit platelet aggregation (2) and reported to produce
PGI2, a potent inhibitor of platelet aggregation, in placental monolayer cultures (3). However, in term placentas the extent and biological significance of PGI2 synthesis are controversial (4, 5). Trophoblasts play an essential role in the recruitment and activation of suppressor cells in the decidua (6, 7). Placental cells synthesize a number of mediators with direct immunosuppressive effects, including arachidonic acid metabolites (8-10). In particular, prostaglandin E2 (PGE2) has been implicated in the immunosuppression at the feto-maternal interface (11-13). PGE2 inhibits Tlymphocyte proliferation, including inhibition of cytotoxic T-lymphocytes, and inhibits interleukin-2 (IL-2) production, thereby blocking the activation of lymphokine-activated cells and natural killer cells (13-15). These functions are likely to be important in suppression of maternal rejection of the invading fetal trophoblasts. PGE2 is the major prostanoid produced by term placentas (16, 17), but little is known of its production by early human placental tissue (18, 19). hCG is a major hormone produced in early pregnancy, and it has been reported to have immunosuppressive properties (20). This has been disputed by others who suggested that these findings resulted from either impurities in the hCG or the use of pharmacological doses
Received May 21, 1990. Address requests for reprints to: Dr. R. A. North, Department of Medicine, University of Melbourne, Royal Melbourne Hospital, Victoria 3050, Australia. * Postgraduate medical research scholar of the National Health and Medical Research Council of Australia at the time of these studies.
60
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 23:39 For personal use only. No other uses without permission. . All rights reserved.
hCG STIMULATION OF PLACENTAL PG (21). Treatment of preovulatory ovarian follicles with hCG has been shown to be associated with increased PG production (22, 23). In view of the potential role of both PGE2 and PGI2 in the processes of implantation and supression of the maternal immune response, the aim of this study was to investigate early human placental prostanoid synthesis in the hope of gaining more understanding of the disturbance that may lead to placental ischemia and, thus, preeclampsia. As hCG production is maximal in early pregnancy, its role in the regulation of placental prostanoid production in vitro was investigated. Materials and Methods Placental tissue Human placental tissue from 9-17 weeks gestation was collected with sponge forceps at surgical terminations of pregnancies. The women were in good health, normotensive, and on no medication. In particular, they had not taken aspirin or other nonsteroidal antiinflammatory medication in the previous week. The duration of pregnancy was assessed by the date of the last menstrual period, estimation of uterine size when examined under anesthetic, and, in some cases, ultrasound examination. The study was approved by the Ethics Committee of the Royal Women's Hospital. Human placental organ cultures Basal PGE and 6-keto-PGFla production The placental tissue was placed in sterile phosphate-buffered saline and Dulbecco's formula without calcium or magnesium (Flow Laboratories, Rickmansworth, England) on ice and processed within 30 min. The tissue was washed three times in icecold PBS and stored at 4 C in medium 199 with Hanks' salts, 20 HIM HEPES buffer, 0.04% sodium bicarbonate, 2 mM L-glutamine, 60 Mg/mL penicillin, 100 iig/mL streptomycin (medium 199) and 5% fetal calf serum (FCS; Flow Laboratories). The placental tissue was dissected into 3- to 10-mg pieces of placental villi, and all decidua were removed. The dissected placenta was left in medium 199 at 4 C for 30 min before commencing an incubation. Each piece of tissue was incubated in 1 mL medium 199 with 5% FCS at 37 C in 5% CO2 in air for 6 h. At 6 h the medium was removed and stored at -20 C. Placental tissue was then incubated in fresh medium 199 with 5% FCS alone or with either 33 /xM arachidonic acid, 28 ixM indomethacin, 1 nM angiotensin-II, or 1 jiM bradykinin (Sigma Chemical Co., St. Louis, MO) for a further 12 h (6-18 h) under the same conditions. At the completion of an experiment, the medium was removed and stored at -20 C for estimation of PGE and 6-keto-PGFi« by RIA within 4 weeks of the incubation. The placental pieces were weighed wet after excess medium was absorbed onto filter paper. Tissue viability was established by light and electron microscopy after incubation. Forty-seven placentas (9-17 weeks gestation) were studied under basal conditions, and PGE was measured in the medium in
61
47 and 6-keto-PGFi« in 36 placentas. The effect of arachidonic acid and indomethacin was studied in 10 placentas from 9-16 weeks gestation, and angiotensin-II and bradykinin were studied in 4 placentas from 9-14 weeks gestation. In each placenta, tissue was incubated in quadruplicate in control and test groups. Mean PGE and 6-keto-PGFla values were calculated for each placenta, and the data from individual placentas were combined for statistical analysis. hCG effect on early human placental PGE and 6-keto-PGFla synthesis in organ culture Twenty-four placentas were established in organ culture as outlined. Six placentas were studied at each of the following gestational ages: 9-10, 11-12, 13-14, and 15-17 weeks. In a 12h incubation, the tissue was cultured in either medium 199 with 5% FCS alone or with hCG (Serono, Switzerland) at final concentrations of 103, 104, 105, 106, and 5 X 106 IU/L. These concentrations were chosen to include the range of serum hCG levels seen in early pregnancy, with peak levels at 10-12 weeks gestation on the order of 4.5 x 104 to 2 X 105 IU/L (24). Each placenta provided quadruplicate determinations in the control and experimental groups. PGE and 6-keto-PGFia levels were measured in the medium by RIA. In time-course studies, tissue was incubated in either medium 199 and 5% FCS alone or with 106 IU/L hCG. At 15 min and 1, 2,4, 6, and 12 h, 250 id medium were removed and replaced with the appropriate medium. The removed medium was stored at —20 C until assayed for PGE and 6-keto-PGFi«. In the calculation of the data, the PG levels at each time point were corrected for the dilution of the medium occurring with the repetitive removal and addition of 250 id medium. Endogenous production of hCG was measured in the medium of control (0 hCG) incubates in 18 placentas (9-17 weeks) with a hCG Maiaclone immunoradiometric assay kit (Serono Diagnostics). To assess the specificity of the response for hCG, placental tissue was incubated in an identical manner with LH (0.1, 1, 10, 100, and 1000 IU/L; Amersham International, Buckinghamshire, United Kingdom), highly purified human pituitary FSH (0.062, 0.62, 6.2, 62, and 620 IU/L; gift Dr. A. F. Parlow, National Pituitary Agency, NIH, Bethesda, MD; code no. AFP-739B), and TSH (100 and 200 mU/L; Armour Pharmaceuticals Co. Ltd., Eastbourne, Sussex, United Kingdom). LH and FSH were tested in two placentas (9-10 weeks) and TSH in 10 placentas (9-17 weeks). Purified trophoblast monolayer cultures The placental monolayer cell culture method was based on previous methods, which were modified to produce cultures of purified trophoblasts that remained viable for at least 2 weeks (25, 26). Human placental tissue (8-10 weeks gestation) was collected from termination of pregnancy, washed three times in sterile PBS at 4 C, and then placed in RPMI-1640 medium (Commonwealth Serum Laboratories, Melbourne, Australia) with penicillin (50 /^g/mL) and streptomycin (12 ixg/mL) on ice. The tissue was dissected in 4 C PBS into 2- to 5-mg pieces of villi, and decidual tissue was removed. The dissected placenta was allowed to settle for 5 min at 4 C, and excess PBS was removed. The trophoblast cells were separated after 20 min of
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 23:39 For personal use only. No other uses without permission. . All rights reserved.
62
NORTH, WHITEHEAD, AND LARKINS
trypsinization at 37 C with 100 mL 0.05% trypsin containing 0.02% EDTA (Flow Laboratories) The suspension was then sieved through a 150-Mm sieve (Endecotts Ltd., England), and the trypsin in the filtrate was neutralized with 10 mL FCS. The cell suspension was washed three times in RPMI with 10% FCS at 4 C. On the final wash the cells were resuspended in 20 mL PBS at 4 C, pipetted onto Ficoll-Paque (Pharmacia Fine Chemicals, Uppsala, Sweden), and centrifuged at 1500 rpm for 20 min at 4 C; the layer of cells at the interface was removed and washed twice with RPMI. Cell viability was assessed by eosin exclusion. The cells were diluted to 5 x 105 viable cells/ mL in RPMI-1640 with 10"5 M thioglycerol (Sigma), 25 U/L insulin (Commonwealth Serum Laboratories, Victoria, Australia), 1 Mg/mL hydrocortisone sodium succinate (Solu-Cortef, Upjohn, Kalamazoo, MI), 5 ng/mL mouse salivary gland epidermal growth factor (EGF; gift from Dr. R. Whitehead, Ludwig Cancer Research Institute, Melbourne, Australia), 50 Mg/mL penicillin, 12 /ig/mL streptomycin, and 10% FCS. One milliliter of cell suspension was added to 1-cm diameter Nunclon tissue culture wells (Nunc, Copenhagen, Denmark) lined with rat tail collagen. The medium was changed every third day. On days 3, 7, and 14 the cells were removed from the wells with collagenase type V (Sigma) and 0.025% trypsin with 0.01% EDTA and assessed for viability using 0.5% eosin. To identify the cell types, cultured cells were deposited on slides using a cytospin centrifuge and studied with immunofluorescence using antibodies directed against keratin (rabbit antikeratin, diluted 1:40, gift from Dr. R. Whitehead, Ludwig Cancer Research Institute, Melbourne, Australia, and antikeratin monoclonal LE 61, diluted 1:10, gift from Dr. B. Lane, Imperial Cancer Research Fund, London, England) and vimentin (antivimentin monoclonal 2H2, diluted 1:10, gift from Dr. I. Pedersen, Department of Pathology, Alfred Hospital, Melbourne, Australia). To further identify the cells in culture, immunoperoxidase staining was performed using antisera for hCG, human placental lactogen (hPL), and pregnancy-specific /3i-glycoprotein (SPi; Dako, Santa Barbara, CA). First trimester placental sections were used as positive controls. HLA class 1 markers were identified with a FACScanner, using mouse antiHLA class 1, ascites 1620 (gift from Prof. I. F. C. McKenzie, Research Centre for Cancer and Transplantation, Melbourne, Australia), diluted 1:100 as first antibody, and the human monocyte line U937 (Walter and Eliza Hall Institute, Melbourne, Australia) as a positive control. hCG effect on PGE and PGI2 production by first trimester placental cells in monolayer culture Placental monolayer cultures were established in RPMI with 10~5 M thioglycerol, 25 U/L regular insulin, 1 Mg/mL hydrocortisone, 50 jig/mL penicillin, 12 /zg/mL streptomycin, 5 ng/mL EGF, and 10% FCS for 7 days. For the 12 h before the addition of hCG, the cells were cultured in RPMI with 10"5 M thioglycerol, 25 U/L regular insulin, 1 /xg/mL hydrocortisone, 50 ng/ mL penicillin, 12 jug/mL streptomycin, and 10% FCS. EGF was excluded from the medium for this incubation, because it has been shown previously to stimulate synthesis of prostanoids (27). The cells were incubated for 12 h in 0.5 mL of either RPMI with 10% FCS alone or with hCG in final concentrations
JCE & M • 1991 Vol 73 • No 1
of 104, 105, 106, and 5 X 106 IU/L or indomethacin (28 Five incubations were performed in each of the experimental and control groups. The media were stored at —20 C for determination of PGE and 6-keto-PGFi0. The trophoblast cells were removed from the wells by sequential collagenase and trypsin treatment and counted using a hemocytometer. To assess the specificity of the effect of hCG for placental cells, the effect of hCG on prostanoid synthesis by bovine aortic endothelial cells was studied. The endothelial cells were cultured using methodology adapted from Gimbrone and co-workers (28). Endothelial cells were isolated with collagenase (Cooper Biomedical, Malvern, PA) digestion of the intimal layer of the thoracic aorta of freshly slaughtered cattle and grown in RPMI-1640 supplemented with FCS and containing minomycin, gentamicin, and fungizone. For the experimental protocol, the cells were passaged into 35-mm wells (Flow) and cultured in quadruplicate in either medium 199 with 5% FCS alone or with hCG in a final concentration of 1O4 to 5 x 106 IU/L for 12 h. The medium was assayed in duplicate for PGE and 6-keto-PGFia. The number of viable cells was estimated after 0.5% eosin exclusion, and results were expressed as nanograms per 106 viable cells/12 h. 6-keto PGFla and PGE RIA The PGs 6-keto PGF la (stable metabolite of prostacyclin) and PGE were measured by RIA according to the method of Nolan et al. (29). Rabbit antisera directed against 6-keto-PGFia and PGE2 (Seragen Research Products Boston, MA) were used at final dilutions of 1:140,000 and 1:21,000, respectively. The anti-PGE2 antibody cross-reacted 100% with PGEi, and therefore, this assay reflects total PGE. The standards 6-keto-PGFi« and PGE2 (Sigma) were diluted from 4-1000 pg/100 nL in appropriate media or in Tris-gelatin buffer, pH 7.4. 6-Keto[5,8,9,ll,12,14,15-N-3H]PGFla and [5,6,8,11,12,14,15-N-3H] PGE2 (Amersham International) were used as tracers. All samples and standards were assayed in duplicate. In the 6-ketoPGFia immunoassay, the intraassay coefficients of variation were 3.3% at 30 pg/100 nL (n = 6) and 6.4% at 90 pg/100 nL (n = 10). The interassay coefficients of variation were 10.5% at 15 pg/100 nh (n = 5) and 6.3% at 80 pg/100 nL (n = 5). In the PGE immunoassay, the intraassay coefficients of variation were 8.4% at 25 pg/100 nL (n = 10) and 11.2% at 100 pg/100 JUL (n = 10). The interassay coefficients of variation were 9.6% at 25 pg/100 ML (n = 7) and 11.7% at 100 pg/100 tih (n = 7). The limit of detection of the assay, defined as the amount of unlabeled PG that displaced 10% of the tracer bound in the zero standard, was 8 pg/100 /iL PGE and 4 pg/100 ML 6-ketoPGFlff. To validate the use of these RIAs in our culture system, the RIAs were performed after arachidonic acid metabolites in incubated medium were extracted with ethyl acetate, pH 3.5, and separated with high performance liquid chromatography (HPLC). In this system there was no apparent significant crossreactivity of the other eicosanoid products in the PGE RIA. However, at the levels of prostanoids measured in this system there was significant cross-reactivity in the 6-keto-PGFia immunoassay. The production of authentic 6-keto-PGFi« was supported by the presence of the major peak of immunoreactiv-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 23:39 For personal use only. No other uses without permission. . All rights reserved.
hCG STIMULATION OF PLACENTAL PG ity in the position of the 6-keto-PGFla standard. Most of the cross-reacting material eluted in the position of PGE2, where it was found that there was a contribution of 0.7-1.9% of the value of radioimmunoassayable PGE in this peak to the measured 6-keto-PGFlCT value. 6-Keto-PGFln values were, therefore, corrected by subtracting 1.3% of the measured PGE value in each experiment. HPLC characterization of arachidonic acid metabolites by early human placenta Ten- to 15-mg pieces of 10-12 week human placental tissue were incubated for 12 h with 1 mL of either medium 199 with 6.4 ng [14C]arachidonic acid (SA, 60.1 mCi/mmol; Amersham International) or medium 199 with 6.4 /ug [14C]arachidonic acid/ mL and 5000 IU hCG/mL. In each placenta, six incubations were performed with hCG and without hCG. At the conclusion of the incubation, the medium was stored at -20 C and extracted twice with acidified ethyl acetate within 2 weeks. The recovery of added standard 6-keto-PGFi« after the extraction procedure was 95 ± 4% (n = 4); for added standard PGE2 the recovery was 96 ± 3 % (n = 4). In the incubated samples, the recovery of total 14C disintegrations per min after extraction was 69 ± 9% (n = 12). HPLC was performed by adapting the method of Eling et al. (30). Radiolabeled standards were obtained from Amersham International. Standards employed were 6-keto[5,8,9,ll,12,14,15-N-3H]PGFla, [5,6,8,9,11,12,14,15-N-3H] thromboxane B2 ([3H]TXB2), [9-3H]PGF2«, [5,6,8,11,12,14,15N-3H]PGE2, [5,6,8,9,ll,12,14,15-N-3H]leukotriene B4 ([3H] LTB4), [14,15-N-3H]LTC4, [14,15-N-3H]LTD4, 5(S)-hydroxy[5,6,8,9,11,12,14,15-N-3H]eicosatetraenoic acid ([3H]5-HETE), [5,6,8,9,11,12,14,15-N-3H]12-HETE, [5,6,8,9,11,12,14,15^-^] 15-HETE, and [l-14C]aracidonic acid. Separation of the eicosanoid products was performed on a Waters HPLC instrument (Waters Associates, Inc., Milford, MA) fitted with model 510 and M-45 high pressure pumps, model 680 Automated Gradient Controller, Waters Intelligent Sample Processor model 710B, and a Radial Compression Separation System RCM-100 containing a 5-mm X 10-cm Radial Pak-A (C-18) cartridge. The column was eluted at 3 mL/min, and the eluate was collected in fractions every 0.3 min with a LKB 2111 Multirac fraction collector (LKB, Bromma, Sweden). The standards and samples were dissolved and injected in 22% acetonitrile in ultrapure water (pH 3.5). The column was equilibrated for 10 min under initial conditions of 22% acetonitrile in water at 3 mL/min. After injection of the sample, the column was eluted isocratically with 22% acetonitrile for 10 min and then 24% acetonitrile for 14 min. Using linear gradient 6 on the Waters solvent programmer, the acetonitrile concentration was increased from 24% to 50% over 13 min, held at 50% for 11 min, and then run from 50-80% acetonitrile over 4 min, using gradient 6. The column was run for 6 min at 80% acetonitrile, purged with 100% acetonitrile for 10 min, and then reequilibrated with 22% acetonitrile in water for 15 min. The fractions collected were mixed with 15 mL toluene-teric scintillant (7:3), and the radioactivity was measured on a /3-scintillation spectrometer.
63
Statistical analysis Results have been expressed as the mean ± SE. Statistical analysis of the results was performed using the t test for unpaired observations, Wilcoxon signed rank test, and multiple linear regression techniques using the statistical package GLIM (31). GLIM enables interactive analysis of generalized linear models under a unified frame work. This methodology assumes that the variation in the continuous dependent variable, about the fitted mean(s), is normally distributed. This does not imply that the marginal distribution itself be normally distributed. Results were considered significant at the 0.05 level. Results HPLC characterization of arachidonic acid metabolites by early human placenta Typical HPLC profiles of arachidonic acid metabolites formed by first trimester placental tissue in organ culture with and without hCG are shown in Fig. 1. PGE2 appeared to be the major prostanoid formed in both the presence and absence of hCG. There was increased [14C] PGE2 formed in the incubations with hCG (0.017 ± 0.003% of total dpm/mg wet wt placenta) compared to incubations without hCG (0.006 ± 0.001% of total dpm/ mg wet wt placenta; P < 0.05). A 6-keto-PGFi« peak was evident in the stimulated incubation (hCG present), but was barely discernible as a separate peak in the incubations with no hCG. Peaks corresponding to the positions of TXB2, PGF2a, LTB4, 5-HETE, 12-HETE, and 15HETE standards were also seen. Other unidentified peaks eluted at 28, 31, and 36 min.
Early human placental PGE and 6-keto-PGFia production in organ culture: effect of incubation period and gestation Placental tissue from 9-17 weeks gestation produced significantly more PGE than 6-keto-PGFi« (Table 1). The hourly rate of production of each prostanoid was significantly greater in the second 12-h incubation compared to the initial 6-h incubation. PGE increased from 8.7 ± 0.9 pg/h-mg wet wt in the 6-h incubation to 61.3 ± 6.0 pg/h-mg wet wt in the 12-h incubation, and 6keto-PGF la from 4.7 ± 0.5 to 11.2 ± 2.2 pg/h-mg wet wt (P < 0.01). Time-course experiments indicated that the production rate over a 12-h culture period was not linear, with the major proportion of the prostanoid being formed from the sixth to the twelfth hour (Fig. 2). The relationship between placental gestation and PGE and 6-keto-PGFia synthesis is shown in Table 2. In the initial 6-h incubation PGE production was maximal between 11 and 13 weeks gestation and then decreased, following a quadratic relationship with gestation (P
Vol. 73, No. 1 Printed in U.S.A.
Stimulation by Human Chorionic Gonadotropin of Prostaglandin Synthesis by Early Human Placental Tissue R. A. NORTH*, R. WHITEHEAD, AND R. G. LARKINS Department of Medicine, University of Melbourne, Royal Melbourne Hospital (R.A.N., R.G.L.), and Ludwig Institute for Cancer Research (R. W.), Parkville, Victoria 3050, Australia
ABSTRACT. Successful establishment of pregnancy is dependent on inhibition of clotting and suppression of the maternal immune response at the feto-maternal interface. Early human placental production of prostacyclin (PGI2) and prostaglandin E2 (PGE2) may be important in this process. To examine the possible role of these PGs, we studied PGE and 6-ketoPGF]n (stable metabolite of PGI2) synthesis in human placental (9-17 weeks gestation) organ cultures, and monolayer cultures of purified trophoblasts. PGE2 appeared to be the major prostanoid formed. Other arachidonic acid metabolites identified in placental organ culture were 6-keto-PGFiOJ thromboxane B2, PGF2o, leukotriene B4, 5(S)-hydroxyeicosatetraenoic acid (5HETE), 12-HETE, and 15-HETE. The synthesis of PGE and
6-keto-PGFin altered with gestation and was maximal in the younger placentas. Arachidonic acid (33 (iM) stimulated and indomethacin (28 ^M) inhibited PG production. hCG, including physiological concentrations, stimulated PGE and 6-keto-PGFln synthesis in placental organ cultures. This effect was most striking in the 9-12 week placentas, compared to 15-17 week placentas. A similar hCG-induced stimulation of PGE production occurred in monolayer cultures of trophoblasts. The addition of TSH, FSH, and LH indicated that this effect was specific for hCG. These data suggest that hCG may have a biological role in the regulation of PG synthesis in early human placenta. (J Clin Endocrinol Metab 73: 60-70, 1991)
E
STABLISHMENT and maintenance of pregnancy involve two poorly understood events. Trophoblast cells invade the uterine wall and change the maternal spiral arteries into dilated vessels that terminate in the intervillous space. As pregnancy progresses, these structural changes to the spiral arteries enable the increased maternal blood flow to the feto-placental unit to be met. This is dependent on inhibition of clotting at the maternal-placental interface. The placenta and fetus are not rejected despite their antigenic differences from the mother. While unsuccessful implantation results in spontaneous abortion, it has been postulated that defective implantation and establishment of the maternal placental vascular bed are central in the pathogenesis of preeclampsia (1). One potential mechanism for inhibition of clotting during trophoblast invasion is prostacyclin (PGI2) production, as PGI2 is a potent inhibitor of platelet aggregation. Early human placental tissue has been shown to inhibit platelet aggregation (2) and reported to produce
PGI2, a potent inhibitor of platelet aggregation, in placental monolayer cultures (3). However, in term placentas the extent and biological significance of PGI2 synthesis are controversial (4, 5). Trophoblasts play an essential role in the recruitment and activation of suppressor cells in the decidua (6, 7). Placental cells synthesize a number of mediators with direct immunosuppressive effects, including arachidonic acid metabolites (8-10). In particular, prostaglandin E2 (PGE2) has been implicated in the immunosuppression at the feto-maternal interface (11-13). PGE2 inhibits Tlymphocyte proliferation, including inhibition of cytotoxic T-lymphocytes, and inhibits interleukin-2 (IL-2) production, thereby blocking the activation of lymphokine-activated cells and natural killer cells (13-15). These functions are likely to be important in suppression of maternal rejection of the invading fetal trophoblasts. PGE2 is the major prostanoid produced by term placentas (16, 17), but little is known of its production by early human placental tissue (18, 19). hCG is a major hormone produced in early pregnancy, and it has been reported to have immunosuppressive properties (20). This has been disputed by others who suggested that these findings resulted from either impurities in the hCG or the use of pharmacological doses
Received May 21, 1990. Address requests for reprints to: Dr. R. A. North, Department of Medicine, University of Melbourne, Royal Melbourne Hospital, Victoria 3050, Australia. * Postgraduate medical research scholar of the National Health and Medical Research Council of Australia at the time of these studies.
60
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 23:39 For personal use only. No other uses without permission. . All rights reserved.
hCG STIMULATION OF PLACENTAL PG (21). Treatment of preovulatory ovarian follicles with hCG has been shown to be associated with increased PG production (22, 23). In view of the potential role of both PGE2 and PGI2 in the processes of implantation and supression of the maternal immune response, the aim of this study was to investigate early human placental prostanoid synthesis in the hope of gaining more understanding of the disturbance that may lead to placental ischemia and, thus, preeclampsia. As hCG production is maximal in early pregnancy, its role in the regulation of placental prostanoid production in vitro was investigated. Materials and Methods Placental tissue Human placental tissue from 9-17 weeks gestation was collected with sponge forceps at surgical terminations of pregnancies. The women were in good health, normotensive, and on no medication. In particular, they had not taken aspirin or other nonsteroidal antiinflammatory medication in the previous week. The duration of pregnancy was assessed by the date of the last menstrual period, estimation of uterine size when examined under anesthetic, and, in some cases, ultrasound examination. The study was approved by the Ethics Committee of the Royal Women's Hospital. Human placental organ cultures Basal PGE and 6-keto-PGFla production The placental tissue was placed in sterile phosphate-buffered saline and Dulbecco's formula without calcium or magnesium (Flow Laboratories, Rickmansworth, England) on ice and processed within 30 min. The tissue was washed three times in icecold PBS and stored at 4 C in medium 199 with Hanks' salts, 20 HIM HEPES buffer, 0.04% sodium bicarbonate, 2 mM L-glutamine, 60 Mg/mL penicillin, 100 iig/mL streptomycin (medium 199) and 5% fetal calf serum (FCS; Flow Laboratories). The placental tissue was dissected into 3- to 10-mg pieces of placental villi, and all decidua were removed. The dissected placenta was left in medium 199 at 4 C for 30 min before commencing an incubation. Each piece of tissue was incubated in 1 mL medium 199 with 5% FCS at 37 C in 5% CO2 in air for 6 h. At 6 h the medium was removed and stored at -20 C. Placental tissue was then incubated in fresh medium 199 with 5% FCS alone or with either 33 /xM arachidonic acid, 28 ixM indomethacin, 1 nM angiotensin-II, or 1 jiM bradykinin (Sigma Chemical Co., St. Louis, MO) for a further 12 h (6-18 h) under the same conditions. At the completion of an experiment, the medium was removed and stored at -20 C for estimation of PGE and 6-keto-PGFi« by RIA within 4 weeks of the incubation. The placental pieces were weighed wet after excess medium was absorbed onto filter paper. Tissue viability was established by light and electron microscopy after incubation. Forty-seven placentas (9-17 weeks gestation) were studied under basal conditions, and PGE was measured in the medium in
61
47 and 6-keto-PGFi« in 36 placentas. The effect of arachidonic acid and indomethacin was studied in 10 placentas from 9-16 weeks gestation, and angiotensin-II and bradykinin were studied in 4 placentas from 9-14 weeks gestation. In each placenta, tissue was incubated in quadruplicate in control and test groups. Mean PGE and 6-keto-PGFla values were calculated for each placenta, and the data from individual placentas were combined for statistical analysis. hCG effect on early human placental PGE and 6-keto-PGFla synthesis in organ culture Twenty-four placentas were established in organ culture as outlined. Six placentas were studied at each of the following gestational ages: 9-10, 11-12, 13-14, and 15-17 weeks. In a 12h incubation, the tissue was cultured in either medium 199 with 5% FCS alone or with hCG (Serono, Switzerland) at final concentrations of 103, 104, 105, 106, and 5 X 106 IU/L. These concentrations were chosen to include the range of serum hCG levels seen in early pregnancy, with peak levels at 10-12 weeks gestation on the order of 4.5 x 104 to 2 X 105 IU/L (24). Each placenta provided quadruplicate determinations in the control and experimental groups. PGE and 6-keto-PGFia levels were measured in the medium by RIA. In time-course studies, tissue was incubated in either medium 199 and 5% FCS alone or with 106 IU/L hCG. At 15 min and 1, 2,4, 6, and 12 h, 250 id medium were removed and replaced with the appropriate medium. The removed medium was stored at —20 C until assayed for PGE and 6-keto-PGFi«. In the calculation of the data, the PG levels at each time point were corrected for the dilution of the medium occurring with the repetitive removal and addition of 250 id medium. Endogenous production of hCG was measured in the medium of control (0 hCG) incubates in 18 placentas (9-17 weeks) with a hCG Maiaclone immunoradiometric assay kit (Serono Diagnostics). To assess the specificity of the response for hCG, placental tissue was incubated in an identical manner with LH (0.1, 1, 10, 100, and 1000 IU/L; Amersham International, Buckinghamshire, United Kingdom), highly purified human pituitary FSH (0.062, 0.62, 6.2, 62, and 620 IU/L; gift Dr. A. F. Parlow, National Pituitary Agency, NIH, Bethesda, MD; code no. AFP-739B), and TSH (100 and 200 mU/L; Armour Pharmaceuticals Co. Ltd., Eastbourne, Sussex, United Kingdom). LH and FSH were tested in two placentas (9-10 weeks) and TSH in 10 placentas (9-17 weeks). Purified trophoblast monolayer cultures The placental monolayer cell culture method was based on previous methods, which were modified to produce cultures of purified trophoblasts that remained viable for at least 2 weeks (25, 26). Human placental tissue (8-10 weeks gestation) was collected from termination of pregnancy, washed three times in sterile PBS at 4 C, and then placed in RPMI-1640 medium (Commonwealth Serum Laboratories, Melbourne, Australia) with penicillin (50 /^g/mL) and streptomycin (12 ixg/mL) on ice. The tissue was dissected in 4 C PBS into 2- to 5-mg pieces of villi, and decidual tissue was removed. The dissected placenta was allowed to settle for 5 min at 4 C, and excess PBS was removed. The trophoblast cells were separated after 20 min of
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 23:39 For personal use only. No other uses without permission. . All rights reserved.
62
NORTH, WHITEHEAD, AND LARKINS
trypsinization at 37 C with 100 mL 0.05% trypsin containing 0.02% EDTA (Flow Laboratories) The suspension was then sieved through a 150-Mm sieve (Endecotts Ltd., England), and the trypsin in the filtrate was neutralized with 10 mL FCS. The cell suspension was washed three times in RPMI with 10% FCS at 4 C. On the final wash the cells were resuspended in 20 mL PBS at 4 C, pipetted onto Ficoll-Paque (Pharmacia Fine Chemicals, Uppsala, Sweden), and centrifuged at 1500 rpm for 20 min at 4 C; the layer of cells at the interface was removed and washed twice with RPMI. Cell viability was assessed by eosin exclusion. The cells were diluted to 5 x 105 viable cells/ mL in RPMI-1640 with 10"5 M thioglycerol (Sigma), 25 U/L insulin (Commonwealth Serum Laboratories, Victoria, Australia), 1 Mg/mL hydrocortisone sodium succinate (Solu-Cortef, Upjohn, Kalamazoo, MI), 5 ng/mL mouse salivary gland epidermal growth factor (EGF; gift from Dr. R. Whitehead, Ludwig Cancer Research Institute, Melbourne, Australia), 50 Mg/mL penicillin, 12 /ig/mL streptomycin, and 10% FCS. One milliliter of cell suspension was added to 1-cm diameter Nunclon tissue culture wells (Nunc, Copenhagen, Denmark) lined with rat tail collagen. The medium was changed every third day. On days 3, 7, and 14 the cells were removed from the wells with collagenase type V (Sigma) and 0.025% trypsin with 0.01% EDTA and assessed for viability using 0.5% eosin. To identify the cell types, cultured cells were deposited on slides using a cytospin centrifuge and studied with immunofluorescence using antibodies directed against keratin (rabbit antikeratin, diluted 1:40, gift from Dr. R. Whitehead, Ludwig Cancer Research Institute, Melbourne, Australia, and antikeratin monoclonal LE 61, diluted 1:10, gift from Dr. B. Lane, Imperial Cancer Research Fund, London, England) and vimentin (antivimentin monoclonal 2H2, diluted 1:10, gift from Dr. I. Pedersen, Department of Pathology, Alfred Hospital, Melbourne, Australia). To further identify the cells in culture, immunoperoxidase staining was performed using antisera for hCG, human placental lactogen (hPL), and pregnancy-specific /3i-glycoprotein (SPi; Dako, Santa Barbara, CA). First trimester placental sections were used as positive controls. HLA class 1 markers were identified with a FACScanner, using mouse antiHLA class 1, ascites 1620 (gift from Prof. I. F. C. McKenzie, Research Centre for Cancer and Transplantation, Melbourne, Australia), diluted 1:100 as first antibody, and the human monocyte line U937 (Walter and Eliza Hall Institute, Melbourne, Australia) as a positive control. hCG effect on PGE and PGI2 production by first trimester placental cells in monolayer culture Placental monolayer cultures were established in RPMI with 10~5 M thioglycerol, 25 U/L regular insulin, 1 Mg/mL hydrocortisone, 50 jig/mL penicillin, 12 /zg/mL streptomycin, 5 ng/mL EGF, and 10% FCS for 7 days. For the 12 h before the addition of hCG, the cells were cultured in RPMI with 10"5 M thioglycerol, 25 U/L regular insulin, 1 /xg/mL hydrocortisone, 50 ng/ mL penicillin, 12 jug/mL streptomycin, and 10% FCS. EGF was excluded from the medium for this incubation, because it has been shown previously to stimulate synthesis of prostanoids (27). The cells were incubated for 12 h in 0.5 mL of either RPMI with 10% FCS alone or with hCG in final concentrations
JCE & M • 1991 Vol 73 • No 1
of 104, 105, 106, and 5 X 106 IU/L or indomethacin (28 Five incubations were performed in each of the experimental and control groups. The media were stored at —20 C for determination of PGE and 6-keto-PGFi0. The trophoblast cells were removed from the wells by sequential collagenase and trypsin treatment and counted using a hemocytometer. To assess the specificity of the effect of hCG for placental cells, the effect of hCG on prostanoid synthesis by bovine aortic endothelial cells was studied. The endothelial cells were cultured using methodology adapted from Gimbrone and co-workers (28). Endothelial cells were isolated with collagenase (Cooper Biomedical, Malvern, PA) digestion of the intimal layer of the thoracic aorta of freshly slaughtered cattle and grown in RPMI-1640 supplemented with FCS and containing minomycin, gentamicin, and fungizone. For the experimental protocol, the cells were passaged into 35-mm wells (Flow) and cultured in quadruplicate in either medium 199 with 5% FCS alone or with hCG in a final concentration of 1O4 to 5 x 106 IU/L for 12 h. The medium was assayed in duplicate for PGE and 6-keto-PGFia. The number of viable cells was estimated after 0.5% eosin exclusion, and results were expressed as nanograms per 106 viable cells/12 h. 6-keto PGFla and PGE RIA The PGs 6-keto PGF la (stable metabolite of prostacyclin) and PGE were measured by RIA according to the method of Nolan et al. (29). Rabbit antisera directed against 6-keto-PGFia and PGE2 (Seragen Research Products Boston, MA) were used at final dilutions of 1:140,000 and 1:21,000, respectively. The anti-PGE2 antibody cross-reacted 100% with PGEi, and therefore, this assay reflects total PGE. The standards 6-keto-PGFi« and PGE2 (Sigma) were diluted from 4-1000 pg/100 nL in appropriate media or in Tris-gelatin buffer, pH 7.4. 6-Keto[5,8,9,ll,12,14,15-N-3H]PGFla and [5,6,8,11,12,14,15-N-3H] PGE2 (Amersham International) were used as tracers. All samples and standards were assayed in duplicate. In the 6-ketoPGFia immunoassay, the intraassay coefficients of variation were 3.3% at 30 pg/100 nL (n = 6) and 6.4% at 90 pg/100 nL (n = 10). The interassay coefficients of variation were 10.5% at 15 pg/100 nh (n = 5) and 6.3% at 80 pg/100 nL (n = 5). In the PGE immunoassay, the intraassay coefficients of variation were 8.4% at 25 pg/100 nL (n = 10) and 11.2% at 100 pg/100 JUL (n = 10). The interassay coefficients of variation were 9.6% at 25 pg/100 ML (n = 7) and 11.7% at 100 pg/100 tih (n = 7). The limit of detection of the assay, defined as the amount of unlabeled PG that displaced 10% of the tracer bound in the zero standard, was 8 pg/100 /iL PGE and 4 pg/100 ML 6-ketoPGFlff. To validate the use of these RIAs in our culture system, the RIAs were performed after arachidonic acid metabolites in incubated medium were extracted with ethyl acetate, pH 3.5, and separated with high performance liquid chromatography (HPLC). In this system there was no apparent significant crossreactivity of the other eicosanoid products in the PGE RIA. However, at the levels of prostanoids measured in this system there was significant cross-reactivity in the 6-keto-PGFia immunoassay. The production of authentic 6-keto-PGFi« was supported by the presence of the major peak of immunoreactiv-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 23:39 For personal use only. No other uses without permission. . All rights reserved.
hCG STIMULATION OF PLACENTAL PG ity in the position of the 6-keto-PGFla standard. Most of the cross-reacting material eluted in the position of PGE2, where it was found that there was a contribution of 0.7-1.9% of the value of radioimmunoassayable PGE in this peak to the measured 6-keto-PGFlCT value. 6-Keto-PGFln values were, therefore, corrected by subtracting 1.3% of the measured PGE value in each experiment. HPLC characterization of arachidonic acid metabolites by early human placenta Ten- to 15-mg pieces of 10-12 week human placental tissue were incubated for 12 h with 1 mL of either medium 199 with 6.4 ng [14C]arachidonic acid (SA, 60.1 mCi/mmol; Amersham International) or medium 199 with 6.4 /ug [14C]arachidonic acid/ mL and 5000 IU hCG/mL. In each placenta, six incubations were performed with hCG and without hCG. At the conclusion of the incubation, the medium was stored at -20 C and extracted twice with acidified ethyl acetate within 2 weeks. The recovery of added standard 6-keto-PGFi« after the extraction procedure was 95 ± 4% (n = 4); for added standard PGE2 the recovery was 96 ± 3 % (n = 4). In the incubated samples, the recovery of total 14C disintegrations per min after extraction was 69 ± 9% (n = 12). HPLC was performed by adapting the method of Eling et al. (30). Radiolabeled standards were obtained from Amersham International. Standards employed were 6-keto[5,8,9,ll,12,14,15-N-3H]PGFla, [5,6,8,9,11,12,14,15-N-3H] thromboxane B2 ([3H]TXB2), [9-3H]PGF2«, [5,6,8,11,12,14,15N-3H]PGE2, [5,6,8,9,ll,12,14,15-N-3H]leukotriene B4 ([3H] LTB4), [14,15-N-3H]LTC4, [14,15-N-3H]LTD4, 5(S)-hydroxy[5,6,8,9,11,12,14,15-N-3H]eicosatetraenoic acid ([3H]5-HETE), [5,6,8,9,11,12,14,15-N-3H]12-HETE, [5,6,8,9,11,12,14,15^-^] 15-HETE, and [l-14C]aracidonic acid. Separation of the eicosanoid products was performed on a Waters HPLC instrument (Waters Associates, Inc., Milford, MA) fitted with model 510 and M-45 high pressure pumps, model 680 Automated Gradient Controller, Waters Intelligent Sample Processor model 710B, and a Radial Compression Separation System RCM-100 containing a 5-mm X 10-cm Radial Pak-A (C-18) cartridge. The column was eluted at 3 mL/min, and the eluate was collected in fractions every 0.3 min with a LKB 2111 Multirac fraction collector (LKB, Bromma, Sweden). The standards and samples were dissolved and injected in 22% acetonitrile in ultrapure water (pH 3.5). The column was equilibrated for 10 min under initial conditions of 22% acetonitrile in water at 3 mL/min. After injection of the sample, the column was eluted isocratically with 22% acetonitrile for 10 min and then 24% acetonitrile for 14 min. Using linear gradient 6 on the Waters solvent programmer, the acetonitrile concentration was increased from 24% to 50% over 13 min, held at 50% for 11 min, and then run from 50-80% acetonitrile over 4 min, using gradient 6. The column was run for 6 min at 80% acetonitrile, purged with 100% acetonitrile for 10 min, and then reequilibrated with 22% acetonitrile in water for 15 min. The fractions collected were mixed with 15 mL toluene-teric scintillant (7:3), and the radioactivity was measured on a /3-scintillation spectrometer.
63
Statistical analysis Results have been expressed as the mean ± SE. Statistical analysis of the results was performed using the t test for unpaired observations, Wilcoxon signed rank test, and multiple linear regression techniques using the statistical package GLIM (31). GLIM enables interactive analysis of generalized linear models under a unified frame work. This methodology assumes that the variation in the continuous dependent variable, about the fitted mean(s), is normally distributed. This does not imply that the marginal distribution itself be normally distributed. Results were considered significant at the 0.05 level. Results HPLC characterization of arachidonic acid metabolites by early human placenta Typical HPLC profiles of arachidonic acid metabolites formed by first trimester placental tissue in organ culture with and without hCG are shown in Fig. 1. PGE2 appeared to be the major prostanoid formed in both the presence and absence of hCG. There was increased [14C] PGE2 formed in the incubations with hCG (0.017 ± 0.003% of total dpm/mg wet wt placenta) compared to incubations without hCG (0.006 ± 0.001% of total dpm/ mg wet wt placenta; P < 0.05). A 6-keto-PGFi« peak was evident in the stimulated incubation (hCG present), but was barely discernible as a separate peak in the incubations with no hCG. Peaks corresponding to the positions of TXB2, PGF2a, LTB4, 5-HETE, 12-HETE, and 15HETE standards were also seen. Other unidentified peaks eluted at 28, 31, and 36 min.
Early human placental PGE and 6-keto-PGFia production in organ culture: effect of incubation period and gestation Placental tissue from 9-17 weeks gestation produced significantly more PGE than 6-keto-PGFi« (Table 1). The hourly rate of production of each prostanoid was significantly greater in the second 12-h incubation compared to the initial 6-h incubation. PGE increased from 8.7 ± 0.9 pg/h-mg wet wt in the 6-h incubation to 61.3 ± 6.0 pg/h-mg wet wt in the 12-h incubation, and 6keto-PGF la from 4.7 ± 0.5 to 11.2 ± 2.2 pg/h-mg wet wt (P < 0.01). Time-course experiments indicated that the production rate over a 12-h culture period was not linear, with the major proportion of the prostanoid being formed from the sixth to the twelfth hour (Fig. 2). The relationship between placental gestation and PGE and 6-keto-PGFia synthesis is shown in Table 2. In the initial 6-h incubation PGE production was maximal between 11 and 13 weeks gestation and then decreased, following a quadratic relationship with gestation (P
