Study of thromboxane and prostacyclin metabolism in an in vitro model of first -trimester human trophoblast Eleanor M. Diss, MD: Steven G. Gabbe, MD: Jay W. Moore, PhD,b and Douglas A. Kniss, PhD" Columbus, Ohio OBJECTIVES: The purpose of our study was to establish an in vitro tissue culture system to study eicosanoid metabolism in first-trimester trophoblastic tissue. Thromboxane A2, a potent vasoconstrictor, and prostacyclin, a potent vasodilator, were analyzed to evaluate their production in early pregnancy. STUDY DESIGN: Trophoblastic tissue was obtained via transabdominal chorionic villous sampling from 33 pregnanCies at 9 to 12 weeks' gestation for cytogenetic diagnosis. Initially, tissue obtained from the cytogenetics lab was morphologically consistent with villous core cells. Through altering cell density and passage, the cells became morphologically consistent with cytotrophoblasts. The cell lines were exposed to arachidonic acid (50 ILmol/L) and aspirin (1 to 100 ILmol/L) for 24 hours. Thromboxane B2 and 6-keto prostaglandin F2a were measured by radioimmunoassay. RESULTS: Villous core cells and cytotrophoblasts increased production of thromboxane A2 and prostacyclin in the presence of arachidonic acid (p < 0.002). The villous core cells produced more thromboxane A2 and prostacyclin than cytotrophoblasts (p < 0.02). A significant inhibition of both thromboxane A2 and prostacyclin production was seen in the presence of 100 ILmoi/L aspirin in both cell types (p < 0.05). CONCLUSIONS: This model may be useful for studying placental function in the first trimester because individual placental compartments can be evaluated in tissue culture. At the cellular level we were not able to detect a preferential decrease in thromboxane A2 production in the presence of aspirin (1 to 100 ILmoI/L). (AM J OaSTET GVNECOL 1992;167:1046-52.)
Key words: Trophoblast, prostacyclin, thromboxane, chorionic villous sampling Preeclampsia, a disorder complicating between 5% to 7% of all pregnancies, is characterized by diffuse vasospasm, enhanced platelet aggregation, endothelial cell injury, and abnormal placentation.'·5 Although the exact cause of preeclampsia has not been elucidated, arachidonic acid metabolites have been implicated as a possible factor.'·s The eicosanoids, prostacyclin (PGI 2) and thromboxane (Tx) A2, have been extensively studied as possible causes of preeclampsia.'·' PGI 2is a potent vasodilator and inhibitor of platelet aggregation, and TxA2 is a potent vasoconstrictor and platelet aggregator. 6 Studies have shown that placental PGI 2 and TxA2 levels are altered in patients with preeclampsia. 7. g In addition, the separate cell types of the human placenta have different production rates of PGI 2 and TxA2.1D Recent work by Nelson and WalshlD demonstrated that villous core cells have a sevenfold greater production of TxA2' measured as the stable metabolite TxB2' than From Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology: and Division of Cytogenetics, Department of Pathology,' The Ohio State University College of Medicine. Presented at the Twelfth Annual Meeting of the Society of Perinatal Obstetricians, Orlando, Florida, February 3-8, 1992. Reprint requests: Douglas A. Kniss, PhD, Department of Obstetrics and Gynecology, The Ohio State University, College of Medicine, 1654 Upham Dr., Means Hall, Columbus, OH 43210. 616139694
1046
trophoblastic cells. In contrast, these investigators found that the production of PGI 2, as measured as the stable metabolite 6-keto prostaglandin (PG) F,., was approximately the same in both the villous core and trophoblastic cells. Low-dose aspirin therapy has been used to prevent preeclampsia and fetal growth retardation."·" Prostacyclin has been demonstrated in trophoblastic tissues as early as 6 weeks' gestation and is known to increase throughout the first trimester. 14 Tissue obtained from first-trimester terminations in patients treated with high-dose aspirin showed significant decreases in PGI 2 levels.'5 Studies have also shown that aspirin preferentially inhibits TxA2 production in explanted chorionic villi and differentially affects TxA2 and PGI 2 production by the trophoblast and villous core compartments.'6 TxA 2 leveis have not been studied in firsttrimester placental tissue. Abnormal placentation has been documented in preeclamptic pregnancies ..· 5 Although normal implantation requires invasion of the cytotrophoblast into the myometrial segment of the spiral arteries,17 Khong et al. 5 found an absence of this process in patients with preeclampsia. Successful invasion of the trophoblast into the spiral arteries may be dependent on a delicate balance between TxA2 and PGI 2. Low-dose aspirin has
Volume 167 Number 4, Part 1
been shown to inhibit TxA2 production without significantly affecting PGI 2 production in term placentas and umbilical arteries of preeclamptic patients, 16, 18·21 Term-placenta tissue may not be representative of first-trimester placental function. However, the only method currently available to study the placenta in early gestation is to use tissue obtained after termination of pregnancy. We sought to develop a model to assess the effects of low-dose aspirin on TxA2 and PGI 2 production in first-trimester placental tissue obtained by chorionic villous sampling (CVS). Material and methods
Tissue acquisition. First-trimester chorionic villi were obtained from 33 patients for medically indicated cytogenetic studies by transabdominal CVS at 9 to 12 weeks' gestation under ultra so no graphic guidance. All patients gave informed consent for the procedure. Villous tissues were used for research purposes only after completion of the cytogenetic analysis. The tissue was rinsed with sterile Hanks' balanced salt solution (HBSS), and any nonvillous tissue was removed. The villi were first incubated in trypsin-ethylenediaminetetraacetic acid (0.05% trypsin; 0.53 mmoll L ethylenedraminetetraacetic acid, 4 sodium) for 1 hour at 37° C. The tissue was then transferred by Pasteur pipette to a second tube containing 10 mg/ml collagenase and incubated for 2 hours at 37° C with agitation to dissociate the cells of the mesenchymal core of the villi. The villous core cells were grown in 25 cm 2 tissue culture flasks with Chang's medium for amniocytes supplemented with penicillin-streptomycin (50 IJ.g/ml; 50 IJ.g/ml), Chang's supplement, L-glutamine (2 mmoIlL), and 10% fetal bovine serum to near confluence for cytogenetic analysis. The investigators were blind as to the identity of the patients. Cell cultures. After completion of the genetic studies the cells were subcultured for our investigations. Cells were rinsed with HBSS and then trypsinized with 0.25% trypsin for 10 minutes at 37° C, after which they were centrifuged at 600g for 5 minutes and an aliquot was removed for counting. A single-cell suspension was seeded into 25 cm 2 tissue culture flasks at 1 to 2 x 106 cells per flask, and cultures were grown in Ham Fl2lDulbecco's Modified Eagle Medium (DMEM) (1: 1) supplemented with 15% fetal bovine serum, 1 mmollL sodium pyruvate, 2 mmollL L-glutamine, and 50 IJ.g/ml gentamicin. The cells were maintained in a humidified incubator at 37° C with 5% carbon dioxide/95% air. After 5 to 7 days in vitro the cells reached confluent density, at which time they were subcultured at a split ratio of I: 2 in the same medium. After no more than three passages the cells were subcultured into 24-well plates (2.5 x 105 cells/well) and grown in serum-free FI2/DMEM + 0.1% bovine se-
First-trimester trophoblast in vitro
1047
rum albumin. At this stage the cells exhibited a fibroblastoid morphologic structure typical of undifferentiated mesenchymal cells of the villous core!2 These cells were referred to as villous core cells. In most cell preparations, after several serial passages a second cell type with an epithelioid phenotype became evident. These cells had features similar to cytotrophoblasts isolated from term placentas 23 and were referred to as cytotrophoblasts. The cytotrophoblasts were plated into 24-well plates (5 x 105 cells/well) in serum-free medium + 0.1 % bovine serum albumin. Eicosanoid studies. To initiate experiments the cells were seeded into 24-well plates as described and treated with aspirin (l to 100 IJ.mollL) in the presence or absence of 50 IJ.mollL arachidonic acid. Control cultures received only the vehicle (medium + 0.1 % bovine serum albumin). Cells were incubated for 24 hours at 37° C as described at which time the medium was removed and frozen at - 80° C until assayed. The cells were solubilized in 1 ml of 1 moliL sodium hydroxide and total protein was measured by the method of Bradford 24 with bovine serum albumin as standard. TxB2 and 6-ketoPGF 1o were measured by specific radioimmunoassay with antibodies generated in domestic chickens!5 The data are expressed as picograms of TxB2 or 6-ketoPGF,o produced per milligram of protein per 24 hours. Statistical analysis. Data were expressed as median values of TxB2 or 6-ketoPGF,o in picograms per milligram of protein because of interpatient variability. The data for multiple groups were analyzed by the nonparametric Friedman analysis of variance with the biomedical-designed program statistical series 3S on the University mainframe computer. Materials. All culture media were purchased from GIBCO/BRL (Grand Island, N.Y.). Fetal bovine serum was purchased from Upstate Biotechnology, Inc. (Lake Placid, N.Y.). Materials for immunostaining were obtained from the following companies: (1) anti-~-hu man chorionic gonadotropin (hCG) and anti-a-hCG, Serono (Allentown, Pa.); (2) anti-human vimentin, Chemicon International (Temecula, Calif.); (3) monoclonal anti-cytokeratin peptide 8, Sigma Chemical Co. (St. Louis); (4) anti-human von Willebrand factor, Calbiochem Corp. (San Diego), and (5) anti-leukocyte common antigen, DAKO Corp. (Carpinteria, Calif.). Tritiated TxB2 and tritiated 6-keto PGF 10 were purchased from Amersham (Arlington Hts., 11.). Unlabeled prostaglandin standards, arachidonic acid, and aspirin, were obtained from Sigma Chemical Co. Immunocytochemistry. For immunostaining, cells were grown on 12 mm glass coverslips in standard growth medium. The cells were rinsed with Tris-buffered saline solution (pH 7.6) and fixed with zinc formalin for 30 minutes. After rinsing the cells with Tris-
1048
Diss et al.
October 1992 Am J Obstct Gynccol
Fig. 1. Villous core cells. Fibroblastoid morphologic structure typical of undifferentiated mesenchymal cells of villous core. (Phase contrast microscopy; x 205.)
buffered saline solution. they were permeabilized with methanol (26° C) and incubated with the appropriate dilution of the various antibodies for 2 hours at 37° C. After rinsing the cells with Tris-buffered saline solution + 1% newborn calf serum, they were incubated with biotinylated goat anti-rabbit immunoglobulin G for 1 hour at 37° C. The cells were rinsed again with Tris-buffered saline solution and incubated with streptavidin conjugated to horseradish peroxidase. The antigen was visualized by developing with aminoethylcarbazole (3-amino, 9-ethylcarbazole), 0.03% hydrogen peroxide, after which the cells were counterstained with hematoxylin. Results
Cytogenetic assessment of chorionic villous tissues. All karyotypes were normal with the exception of one trisomy 21. There was an equal distribution of males and females. Of the pregnancies completed, only one was complicated by preeclampsia at term. Morphologic characteristics of first-trimester chorionic villous cells. Villous core cells are illustrated in Fig. 1. These cells exhibited a fibroblastoid morphologic structure characteristic of undifferentiated mesenchymal cells. After an average of seven passages and low density plating at 5 x 105 cells per 25 cm 2 tissue culture flask, the primary cultures developed an epithelioid appearance characteristic of cytotrophoblasts (Fig. 2). Villous core cells were vimentin-positive, hCG-negative, and cytokeratin 8-negative. Cytotrophoblastic
cells were vimentin-negative, a-hCG-positive, f3-hCG weekly-positive, and cytokeratin 8-positive. To verify that villous core cells and cytotrophoblast cells were not contaminated with endothelial cells, immunostaining with anti-von Willebrand factor was performed. All cultures were negative for this marker. In addition, cultures were negative for the presence of macrophage and leukocyte markers (unpublished data). Eicosanoid studies. Cells derived from the villous core synthesized large amounts of the eicosanoids TxB2 and 6-ketoPGF'Q' However, cytotrophoblastic cell lines produced detectable levels of both eicosanoids only when incubated with arachidonic acid (50 j-lmoIlL). Therefore all cytotrophoblast data refer to cells incubated with arachidonic acid. Because of the wide variation in basal production of TxB2 and 6-ketoPGF,. between the various established cell lines, the data were presented as the median values in picograms of eicosanoid produced per milligrams of mg protein per 24 hours. Villous core cells. Villous core cell results include 13 different CVS samples. Fig. 3 shows representative experiments in which villous core cells were incubated in the presence or absence of arachidonic acid for 24 hours after which TxB2 and 6-ketoPGF'Q, respectively, were measured by radioimmunoassay. When villous core cells were incubated with arachidonic acid, there was a threefold to fourfold stimulation of eicosanoid production over control levels. The median values for control and arachidonic acid-treated villous core cells were 2645 pg TxB2/mg protein and 8000 pg TxB2/mg
First-trimester trophoblast in vitro
Volume 167 !';umber 4. Part 1
1049
Fig. 2. Cytotrophoblastic cells. Epithelioid morphologic structure of typical cytotrophoblastic cells. (Phase contrast microscopy x 205.)
protein (p < 0.0002), respectively, and 2788 pg 6ketoPGF'a/mg protein and 10,926 pg 6-ketoPGF,a/mg protein (p < 0.0002), respectively. When villous core cells were incubated with aspirin (l to 100 J.LmoIlL), there was a dose-dependent decrease in TxB2 biosynthesis (l00 J.LmollL, 92% reduction; 10 J.LmoIlL, 82% reduction; 1 J.LmoIlL, 21 % reduction; Fig. 2, a). In cells incubated with arachidonic acid the degree of aspirin-induced inhibition of TxB2 synthesis was less pronounced (l00 J.,tmoIlL, 56% reduction; 10 J.,tmoIlL. 16% reduction; 1 J.LmoIlL, 15% reduction; Fig. 3, b). A similar decrease in 6-ketoPGF'a biosynthesis was seen in villous core cells incubated with aspirin (l00 J.,tmoIlL. 90% reduction; 10 J.,tmoIlL, 86% reduction; 1 J.LmoIlL, 50% reduction; Fig. 3, c). In cells incubated with arachidonic acid, the degree of aspirininduced inhibition of6-ketoPGF'a synthesis was also less pronounced (l00 J.,tmoIlL. 56% reduction; 10 J.l.moIlL, 37% reduction; 1 J.,tmoIlL, 17% reduction; Fig. 3, d). These reductions were statistically significant at 100 J.LmollL and 10 J.,tmollL concentrations of aspirin (p < 0.05). In cultures incubated with arachidonic acid, only 100 J.,tmollL aspirin caused a significant reduction of TxB. and 6-ketoPGF,a production (p < 0.05). CytotrophobIastic cells. The cytotrophoblastic cells were derived from 27 different CVS samples. Data are presented for cultures supplemented with arachidonic acid (Fig. 4). The median production of TxB2 was 1276 pg/mg protein; the median production of6-ketoPGF'a was 560 pg/mg protein. Data for aspirin-induced inhibition of TxB2 and 6-ketoPGF'a production are as
follows: TxB2 (l00 J.,tmoIlL. 56% reduction; 10 J.,tmoIlL, 40% reduction; 1 J.,tmoIlL, 20% reduction) and 6-ketoPGF,a (l00 J.,tmoIlL. 57% reduction; 10 J.,tmoIlL. 41 % reduction; 1 J.,tmoIlL. 16% reduction). This reduction was significant in cells incubated with 100 J.,tmollL and 10 J.,tmollL concentrations of aspirin (p < 0.05). Experiments were conducted in both villous core cells and cytotrophoblastic cells in seven CVS samples within the same cell line. All cultures had been supplemented with arachidonic acid. The median production ofTxB e was 2071 pg I mg protein in the cytotrophoblastic cells and 8000 pg I mg protein in the villous core cells (p < 0.02). The median production of 6-ketoPGF'a was 933 pg/mg protein in cells with villous core morphologic characteristics and 6577 pg/mg protein in the cytotrophoblastic cells (p < 0.009) (Fig. 5). Fig. 6 illustrates the TxB zI6-ketoPGF,a ratios in the villous core cells and cytotrophoblastic cells in the presence of arachidonic acid. There were no significant differences in the median ratios between the aspirintreated and control groups. The patient with preeclampsia had the highest TxB 2 /6-ketoPGF'a ratio of all of the cytotrophoblastic cell lines. This elevated ratio did not improve in the presence of aspirin (1 to 100 J.,tmoIlL).
Comment Previous studies 111 term placentas have demonstrated abnormal TxA./PGI 2 ratios in patients with preeclampsia,"' 9 placentas from whom produce more TxAz
1050
(a)
Diss et al.
Am
Thromboxane Produc:tion: Villous
..
12000
'i
i
c... ~
&.
,:x ~ .....
.e
9000 6000
3000
0
(b)
...c
~
4 & ~ ~ .....
...
...1..
•t
Control
Cells
*p
2000 1000 0
(b)
'0 CI>
()
,....
0
0..
0
c..
to
w:~
0..
0 +' CI>
~
I
oc
i
+ ""i
Control
" t
.
,
" "
t + " f
Cl.
"" " "
+'
2000
"
1500
"
c..Q. 0>
E
1000
"0> Q.
500 0
"
" t "
+I Control
w:~
Cl. 0 +' Q)
~
I
o£J
OVillous Core
,....
~ Cytotroph .
C
7500
'rl Q)
+' 0 L
Q.
5000
(J)
E
"(J)
Q.
**
2500
'-
0
a
-6
10000
::J
U 0
***
(\j
-5 -4 Log [ASA] M
1051
TxA2 and PGI 2 Production Villous Core Cells vs. Cytotrophoblasts
u Q) u
'-
""
Prostacyclin Production: Cytotrophoblasts Arachidonic Acid Stimulation
....CI>C
::J '0
i A
"0>
Q. '-'
""
*p
Key words: Trophoblast, prostacyclin, thromboxane, chorionic villous sampling Preeclampsia, a disorder complicating between 5% to 7% of all pregnancies, is characterized by diffuse vasospasm, enhanced platelet aggregation, endothelial cell injury, and abnormal placentation.'·5 Although the exact cause of preeclampsia has not been elucidated, arachidonic acid metabolites have been implicated as a possible factor.'·s The eicosanoids, prostacyclin (PGI 2) and thromboxane (Tx) A2, have been extensively studied as possible causes of preeclampsia.'·' PGI 2is a potent vasodilator and inhibitor of platelet aggregation, and TxA2 is a potent vasoconstrictor and platelet aggregator. 6 Studies have shown that placental PGI 2 and TxA2 levels are altered in patients with preeclampsia. 7. g In addition, the separate cell types of the human placenta have different production rates of PGI 2 and TxA2.1D Recent work by Nelson and WalshlD demonstrated that villous core cells have a sevenfold greater production of TxA2' measured as the stable metabolite TxB2' than From Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology: and Division of Cytogenetics, Department of Pathology,' The Ohio State University College of Medicine. Presented at the Twelfth Annual Meeting of the Society of Perinatal Obstetricians, Orlando, Florida, February 3-8, 1992. Reprint requests: Douglas A. Kniss, PhD, Department of Obstetrics and Gynecology, The Ohio State University, College of Medicine, 1654 Upham Dr., Means Hall, Columbus, OH 43210. 616139694
1046
trophoblastic cells. In contrast, these investigators found that the production of PGI 2, as measured as the stable metabolite 6-keto prostaglandin (PG) F,., was approximately the same in both the villous core and trophoblastic cells. Low-dose aspirin therapy has been used to prevent preeclampsia and fetal growth retardation."·" Prostacyclin has been demonstrated in trophoblastic tissues as early as 6 weeks' gestation and is known to increase throughout the first trimester. 14 Tissue obtained from first-trimester terminations in patients treated with high-dose aspirin showed significant decreases in PGI 2 levels.'5 Studies have also shown that aspirin preferentially inhibits TxA2 production in explanted chorionic villi and differentially affects TxA2 and PGI 2 production by the trophoblast and villous core compartments.'6 TxA 2 leveis have not been studied in firsttrimester placental tissue. Abnormal placentation has been documented in preeclamptic pregnancies ..· 5 Although normal implantation requires invasion of the cytotrophoblast into the myometrial segment of the spiral arteries,17 Khong et al. 5 found an absence of this process in patients with preeclampsia. Successful invasion of the trophoblast into the spiral arteries may be dependent on a delicate balance between TxA2 and PGI 2. Low-dose aspirin has
Volume 167 Number 4, Part 1
been shown to inhibit TxA2 production without significantly affecting PGI 2 production in term placentas and umbilical arteries of preeclamptic patients, 16, 18·21 Term-placenta tissue may not be representative of first-trimester placental function. However, the only method currently available to study the placenta in early gestation is to use tissue obtained after termination of pregnancy. We sought to develop a model to assess the effects of low-dose aspirin on TxA2 and PGI 2 production in first-trimester placental tissue obtained by chorionic villous sampling (CVS). Material and methods
Tissue acquisition. First-trimester chorionic villi were obtained from 33 patients for medically indicated cytogenetic studies by transabdominal CVS at 9 to 12 weeks' gestation under ultra so no graphic guidance. All patients gave informed consent for the procedure. Villous tissues were used for research purposes only after completion of the cytogenetic analysis. The tissue was rinsed with sterile Hanks' balanced salt solution (HBSS), and any nonvillous tissue was removed. The villi were first incubated in trypsin-ethylenediaminetetraacetic acid (0.05% trypsin; 0.53 mmoll L ethylenedraminetetraacetic acid, 4 sodium) for 1 hour at 37° C. The tissue was then transferred by Pasteur pipette to a second tube containing 10 mg/ml collagenase and incubated for 2 hours at 37° C with agitation to dissociate the cells of the mesenchymal core of the villi. The villous core cells were grown in 25 cm 2 tissue culture flasks with Chang's medium for amniocytes supplemented with penicillin-streptomycin (50 IJ.g/ml; 50 IJ.g/ml), Chang's supplement, L-glutamine (2 mmoIlL), and 10% fetal bovine serum to near confluence for cytogenetic analysis. The investigators were blind as to the identity of the patients. Cell cultures. After completion of the genetic studies the cells were subcultured for our investigations. Cells were rinsed with HBSS and then trypsinized with 0.25% trypsin for 10 minutes at 37° C, after which they were centrifuged at 600g for 5 minutes and an aliquot was removed for counting. A single-cell suspension was seeded into 25 cm 2 tissue culture flasks at 1 to 2 x 106 cells per flask, and cultures were grown in Ham Fl2lDulbecco's Modified Eagle Medium (DMEM) (1: 1) supplemented with 15% fetal bovine serum, 1 mmollL sodium pyruvate, 2 mmollL L-glutamine, and 50 IJ.g/ml gentamicin. The cells were maintained in a humidified incubator at 37° C with 5% carbon dioxide/95% air. After 5 to 7 days in vitro the cells reached confluent density, at which time they were subcultured at a split ratio of I: 2 in the same medium. After no more than three passages the cells were subcultured into 24-well plates (2.5 x 105 cells/well) and grown in serum-free FI2/DMEM + 0.1% bovine se-
First-trimester trophoblast in vitro
1047
rum albumin. At this stage the cells exhibited a fibroblastoid morphologic structure typical of undifferentiated mesenchymal cells of the villous core!2 These cells were referred to as villous core cells. In most cell preparations, after several serial passages a second cell type with an epithelioid phenotype became evident. These cells had features similar to cytotrophoblasts isolated from term placentas 23 and were referred to as cytotrophoblasts. The cytotrophoblasts were plated into 24-well plates (5 x 105 cells/well) in serum-free medium + 0.1 % bovine serum albumin. Eicosanoid studies. To initiate experiments the cells were seeded into 24-well plates as described and treated with aspirin (l to 100 IJ.mollL) in the presence or absence of 50 IJ.mollL arachidonic acid. Control cultures received only the vehicle (medium + 0.1 % bovine serum albumin). Cells were incubated for 24 hours at 37° C as described at which time the medium was removed and frozen at - 80° C until assayed. The cells were solubilized in 1 ml of 1 moliL sodium hydroxide and total protein was measured by the method of Bradford 24 with bovine serum albumin as standard. TxB2 and 6-ketoPGF 1o were measured by specific radioimmunoassay with antibodies generated in domestic chickens!5 The data are expressed as picograms of TxB2 or 6-ketoPGF,o produced per milligram of protein per 24 hours. Statistical analysis. Data were expressed as median values of TxB2 or 6-ketoPGF,o in picograms per milligram of protein because of interpatient variability. The data for multiple groups were analyzed by the nonparametric Friedman analysis of variance with the biomedical-designed program statistical series 3S on the University mainframe computer. Materials. All culture media were purchased from GIBCO/BRL (Grand Island, N.Y.). Fetal bovine serum was purchased from Upstate Biotechnology, Inc. (Lake Placid, N.Y.). Materials for immunostaining were obtained from the following companies: (1) anti-~-hu man chorionic gonadotropin (hCG) and anti-a-hCG, Serono (Allentown, Pa.); (2) anti-human vimentin, Chemicon International (Temecula, Calif.); (3) monoclonal anti-cytokeratin peptide 8, Sigma Chemical Co. (St. Louis); (4) anti-human von Willebrand factor, Calbiochem Corp. (San Diego), and (5) anti-leukocyte common antigen, DAKO Corp. (Carpinteria, Calif.). Tritiated TxB2 and tritiated 6-keto PGF 10 were purchased from Amersham (Arlington Hts., 11.). Unlabeled prostaglandin standards, arachidonic acid, and aspirin, were obtained from Sigma Chemical Co. Immunocytochemistry. For immunostaining, cells were grown on 12 mm glass coverslips in standard growth medium. The cells were rinsed with Tris-buffered saline solution (pH 7.6) and fixed with zinc formalin for 30 minutes. After rinsing the cells with Tris-
1048
Diss et al.
October 1992 Am J Obstct Gynccol
Fig. 1. Villous core cells. Fibroblastoid morphologic structure typical of undifferentiated mesenchymal cells of villous core. (Phase contrast microscopy; x 205.)
buffered saline solution. they were permeabilized with methanol (26° C) and incubated with the appropriate dilution of the various antibodies for 2 hours at 37° C. After rinsing the cells with Tris-buffered saline solution + 1% newborn calf serum, they were incubated with biotinylated goat anti-rabbit immunoglobulin G for 1 hour at 37° C. The cells were rinsed again with Tris-buffered saline solution and incubated with streptavidin conjugated to horseradish peroxidase. The antigen was visualized by developing with aminoethylcarbazole (3-amino, 9-ethylcarbazole), 0.03% hydrogen peroxide, after which the cells were counterstained with hematoxylin. Results
Cytogenetic assessment of chorionic villous tissues. All karyotypes were normal with the exception of one trisomy 21. There was an equal distribution of males and females. Of the pregnancies completed, only one was complicated by preeclampsia at term. Morphologic characteristics of first-trimester chorionic villous cells. Villous core cells are illustrated in Fig. 1. These cells exhibited a fibroblastoid morphologic structure characteristic of undifferentiated mesenchymal cells. After an average of seven passages and low density plating at 5 x 105 cells per 25 cm 2 tissue culture flask, the primary cultures developed an epithelioid appearance characteristic of cytotrophoblasts (Fig. 2). Villous core cells were vimentin-positive, hCG-negative, and cytokeratin 8-negative. Cytotrophoblastic
cells were vimentin-negative, a-hCG-positive, f3-hCG weekly-positive, and cytokeratin 8-positive. To verify that villous core cells and cytotrophoblast cells were not contaminated with endothelial cells, immunostaining with anti-von Willebrand factor was performed. All cultures were negative for this marker. In addition, cultures were negative for the presence of macrophage and leukocyte markers (unpublished data). Eicosanoid studies. Cells derived from the villous core synthesized large amounts of the eicosanoids TxB2 and 6-ketoPGF'Q' However, cytotrophoblastic cell lines produced detectable levels of both eicosanoids only when incubated with arachidonic acid (50 j-lmoIlL). Therefore all cytotrophoblast data refer to cells incubated with arachidonic acid. Because of the wide variation in basal production of TxB2 and 6-ketoPGF,. between the various established cell lines, the data were presented as the median values in picograms of eicosanoid produced per milligrams of mg protein per 24 hours. Villous core cells. Villous core cell results include 13 different CVS samples. Fig. 3 shows representative experiments in which villous core cells were incubated in the presence or absence of arachidonic acid for 24 hours after which TxB2 and 6-ketoPGF'Q, respectively, were measured by radioimmunoassay. When villous core cells were incubated with arachidonic acid, there was a threefold to fourfold stimulation of eicosanoid production over control levels. The median values for control and arachidonic acid-treated villous core cells were 2645 pg TxB2/mg protein and 8000 pg TxB2/mg
First-trimester trophoblast in vitro
Volume 167 !';umber 4. Part 1
1049
Fig. 2. Cytotrophoblastic cells. Epithelioid morphologic structure of typical cytotrophoblastic cells. (Phase contrast microscopy x 205.)
protein (p < 0.0002), respectively, and 2788 pg 6ketoPGF'a/mg protein and 10,926 pg 6-ketoPGF,a/mg protein (p < 0.0002), respectively. When villous core cells were incubated with aspirin (l to 100 J.LmoIlL), there was a dose-dependent decrease in TxB2 biosynthesis (l00 J.LmollL, 92% reduction; 10 J.LmoIlL, 82% reduction; 1 J.LmoIlL, 21 % reduction; Fig. 2, a). In cells incubated with arachidonic acid the degree of aspirin-induced inhibition of TxB2 synthesis was less pronounced (l00 J.,tmoIlL, 56% reduction; 10 J.,tmoIlL. 16% reduction; 1 J.LmoIlL, 15% reduction; Fig. 3, b). A similar decrease in 6-ketoPGF'a biosynthesis was seen in villous core cells incubated with aspirin (l00 J.,tmoIlL. 90% reduction; 10 J.,tmoIlL, 86% reduction; 1 J.LmoIlL, 50% reduction; Fig. 3, c). In cells incubated with arachidonic acid, the degree of aspirininduced inhibition of6-ketoPGF'a synthesis was also less pronounced (l00 J.,tmoIlL. 56% reduction; 10 J.l.moIlL, 37% reduction; 1 J.,tmoIlL, 17% reduction; Fig. 3, d). These reductions were statistically significant at 100 J.LmollL and 10 J.,tmollL concentrations of aspirin (p < 0.05). In cultures incubated with arachidonic acid, only 100 J.,tmollL aspirin caused a significant reduction of TxB. and 6-ketoPGF,a production (p < 0.05). CytotrophobIastic cells. The cytotrophoblastic cells were derived from 27 different CVS samples. Data are presented for cultures supplemented with arachidonic acid (Fig. 4). The median production of TxB2 was 1276 pg/mg protein; the median production of6-ketoPGF'a was 560 pg/mg protein. Data for aspirin-induced inhibition of TxB2 and 6-ketoPGF'a production are as
follows: TxB2 (l00 J.,tmoIlL. 56% reduction; 10 J.,tmoIlL, 40% reduction; 1 J.,tmoIlL, 20% reduction) and 6-ketoPGF,a (l00 J.,tmoIlL. 57% reduction; 10 J.,tmoIlL. 41 % reduction; 1 J.,tmoIlL. 16% reduction). This reduction was significant in cells incubated with 100 J.,tmollL and 10 J.,tmollL concentrations of aspirin (p < 0.05). Experiments were conducted in both villous core cells and cytotrophoblastic cells in seven CVS samples within the same cell line. All cultures had been supplemented with arachidonic acid. The median production ofTxB e was 2071 pg I mg protein in the cytotrophoblastic cells and 8000 pg I mg protein in the villous core cells (p < 0.02). The median production of 6-ketoPGF'a was 933 pg/mg protein in cells with villous core morphologic characteristics and 6577 pg/mg protein in the cytotrophoblastic cells (p < 0.009) (Fig. 5). Fig. 6 illustrates the TxB zI6-ketoPGF,a ratios in the villous core cells and cytotrophoblastic cells in the presence of arachidonic acid. There were no significant differences in the median ratios between the aspirintreated and control groups. The patient with preeclampsia had the highest TxB 2 /6-ketoPGF'a ratio of all of the cytotrophoblastic cell lines. This elevated ratio did not improve in the presence of aspirin (1 to 100 J.,tmoIlL).
Comment Previous studies 111 term placentas have demonstrated abnormal TxA./PGI 2 ratios in patients with preeclampsia,"' 9 placentas from whom produce more TxAz
1050
(a)
Diss et al.
Am
Thromboxane Produc:tion: Villous
..
12000
'i
i
c... ~
&.
,:x ~ .....
.e
9000 6000
3000
0
(b)
...c
~
4 & ~ ~ .....
...
...1..
•t
Control
Cells
*p
2000 1000 0
(b)
'0 CI>
()
,....
0
0..
0
c..
to
w:~
0..
0 +' CI>
~
I
oc
i
+ ""i
Control
" t
.
,
" "
t + " f
Cl.
"" " "
+'
2000
"
1500
"
c..Q. 0>
E
1000
"0> Q.
500 0
"
" t "
+I Control
w:~
Cl. 0 +' Q)
~
I
o£J
OVillous Core
,....
~ Cytotroph .
C
7500
'rl Q)
+' 0 L
Q.
5000
(J)
E
"(J)
Q.
**
2500
'-
0
a
-6
10000
::J
U 0
***
(\j
-5 -4 Log [ASA] M
1051
TxA2 and PGI 2 Production Villous Core Cells vs. Cytotrophoblasts
u Q) u
'-
""
Prostacyclin Production: Cytotrophoblasts Arachidonic Acid Stimulation
....CI>C
::J '0
i A
"0>
Q. '-'
""
*p
