- Email: [email protected]
Characterization of pure human first-trimester cytotrophoblast cells in long-term culture: Growth pattern, markers, and hormone production
Connors et al.
12. Eldridge FL, Millhorn DE. Oscillation, gating and memory in the respiratory control system. In: Fishman AP, ed. Handbook of physiology. Washington, DC: American Physiological Society, 1986:93-114. 13. Phillipson EA, Duffin J, Caper JD. Critical dependence of respiratory rhythmicity on metabolic CO2 load. J Appl Physiol 1981 ;50:45-54. 14. Boddy K, Dawes GS, Fisher R, Pinter S, Robinson JS. Foetal respiratory moverpents, electrocortical and cardiovascular responses to hypoxemia and hypercapnia in sheep. J Physiol (Lond) 1974;243:599-618. 15. Nijhuis JG, Jonsgma HW, Crijns I, de Valk I, van den
April 1989 Am J Obstet Gynecol
Velden J. Effects of maternal glucose ingestion on human fetal breathing movements at weeks 24 and 28 of gestation. Early Hum Dev 1986;13:183-8. 16. Natale R, Patrick J, Richardson B. Effects of human maternal venous plasma glucose concentrations on fetal breatl~ing movements. AM J OBSTET GVNECOL 1978; 132:36-41. 17. Natale R, Nasello-Paterson C, Connors G. Patterns of fetal breathing activity in the human fetus at 24 to 28 weeks of gestation. AM J OBSTET GVNECOL 1988;158: 317-21.
Characterization of pure human first-trimester cytotrophoblast cells in long-term culture: Growth pattern, markers, and hormone production Simcha Yagel, MD," Robert F. Casper, MD: Wendy Powell, MSc: Ranjit S. Parhar, MS,. and Peeyush K. Lala, MD, PhD' London, Ontario, Canada Pure long-term cytotrophoblast cultures were established from human first-trimester placentas by growing chorionic villus explants without enzymatic digestion. Cytoplasmic human chorionic gonadotropin was detectable in all (100%) cells in culture when labeled with a polyclonal anti-human chorionic gonadotropin antibody and in 71% to'83% of cells labeled with a monoclonal anti-a-human chorionic gonadotropin antibody. Most of the cells expressed cytokeratin and surface Trop-1 and Trop-2 antigens (89% to 95%), but none expressed cytoplasmic vimentin or surface 6303 antigens. Study of the ultrastructure of the cells demonstrated epithelial morphologic features. The average doubling time of the trophoblast was 48 to 96 hours. Some of the lines have been continuously propagated for 8 months. They produced variable amounts of human chorionic gonadotropin (50 to 710 mlU/ml per 105 cells per 24 hours). The basal level of progesterone secreted by trophoblast (444.4 ± 32.4 pg/ml per 105 cells per 24 hours) doubled in the presence of pure human chorionic gonadotropin (100 ng/ml). They produced small amounts of 17J3-estradiol «20 pg/ml per 100 cells per 24 hours); human choriOniC gonadotropin had no effect on the estradiol production. Trophoblast-derived human chorionic gonadotropin acted as a growth factor because trophoblast proliferation (measured by uptake of thymidine labeled with tritium) was reduced by 60% in the presence of an anti-human chorionic gonadotropin antibody. Availability of pure, functionally competent human cytotrophoblast in long-term cultures is relevant for further studies in reproduction biology. (AM J OSSTET GVNECOL 1989;160:938-45.)
Key words: Trophoblast, trophoblast culture, human chorionic gonadotropin, progesterone
From the Department of Anatom1 and the Department of ObstetriCs and Gynecology/ Universl/.V Hospital, The University of We~tern Ontario. Supported by the Medical Research Counal of Canada, the National Cancer Institute of Canada, the Cancer Research Society, and Pharmasaence Inc., Montreal, Quebec. S. Yagel was a fellow of the Medical Research CounCil of Canada. Presented at the Forty{ourth Annual Meetzng of the Socuty of Obstetricians and Gynaecologis~ of Canada, Vancouver, Brll~h Columbza, Canada, June 21-25,1988, Reprint reques~: P. K. Lala, MD, PhD, Department of Anatomy, The University of Western Ontano, London, Ontario, Canada N6A 5C1. 938
A variety of techniques has been used to grow and maintain human first-trimester or term placental trophoblast cells in culture.'-6 Some cultures have been reported to retain trophoblastic markers even after several passages in vitro. The difficulty in obtaining pure trophoblast cultures can partly be explained by the fact that enzymatic digestion of the placenta, usually undertaken before culturing of mixed cell populations, results in inclusion of trophoblast cells, macrophages, endothelial cells, and fibroblasts. Kliman et al. reported j
Pure human cytotrophoblast in long-term culture
Volume 160 Number 4
939
Table I. Markers used for trophoblast characterization* Marker
Specificity
Cytokeratin Epithelial cell marker Vimentin Mesenchymally derived cell marker LCG hCG a-hCG a-Chain of hCG Trop-I Trop-2 63D3
Primary antibody
Polyclonal rabbit, I: 500
Secondary antIbody
Biotinylated goat antirabbit immunoglobulin Polyclonal goat, I: 500 Biotinylated rabbit antigoat immunoglobulin Biotinylated goat antirabbit Polyclonal rabbit, I: 100 Monoclonal mouse antihuman, Rhodamine-conjugated rabbit anti-mouse 1:100 Trophoblast antigen Monoclonal mouse antihuman, Biotinylated horse antimouse I: 100 Trophoblast antigen Monoclonal mouse antihuman, Biotinylated horse antimouse I: 100 Monoclonal mouse antihuman Biotinylated horse anti mouse Human monocyte and macrophage marker Hybridoma sup. (X 10 conc.)
Trophoblast labeled (%) (range)
81-91
o 100 71-83 89-93 91-95
o
*AIl markers except a-hCG were identified with avidin-biotin-peroxidase complex method. a-hCG was identified with immunofluorescence method.
culturing of contamination-free functional cytotrophoblast cells obtained by Percoll gradient centrifugation of enzymatically dispersed term placenta. However, they did not attempt long-term maintenance of these cells by repeated passage. Establishment of longterm cultures of pure first-trimester human cytotrophoblast cells with functional integrity should provide reproducible in vitro systems for investigating endocrinologic and immunologic roles of these cells in pregnancy and the biologic basis of invasive behavior by cells. Our study was designed to achieve this objective. We show that by using the described methods, longterm replicating trophoblast populations were obtained that were free of macrophages, fibroblasts, and endothelial cells, as judged by morphologic features and numerous cell-specific markers. Moreover, these longterm monolayer cultures were proved to be functional: they produced a significant amount of human chorionic gonadotropin (hCG) and progesterone. hCG was found to be an important growth factor for the long-term cultured trophoblast cells. Material and methods
Primary trophoblast cultures. First-trimester placental tissues were obtained after elective termination of pregnancy performed by dilation and curettage at Victoria Hospital, London, Ontario. The tissues were rinsed thoroughly in cold phosphate-buffered saline solution. Areas rich in chorionic villi were selected to isolate the villi, which were minced into pieces of ::; 1 mm. Nonvillous materials inclusive of decidual tissue, blood clots, and membranes were carefully discarded. The fragments of villi were washed three times with minimal essential medium supplemented with streptomycin (20 mg/ml), penicillin (500 U /ml), and amphotericin B (25 ILg/ml). These fragments were then
cultured in RPMI 1640 (Grand Island Biological Co., Grand Island, N.Y.) medium containing 10% fetal calf serum and antibiotics (as stated above) in closed tissue culture flasks (Nunclon, 25 cm\ Nunc, Denmark) for 3 to 5 days until nonadherent cells could be removed and discarded. Cultures of adherent cells were expanded for 1 to 2 weeks in these flasks in fresh medium before their passage and long-term maintenance in vitro in similar flasks. For passages, cells were removed by a mild (0.25%) trypsin treatment, washed, and replanted. Transmission and scanning electron microscopy. For transmission electron microscopy and scanning electron microscopy preparations, trophoblasts were seeded on epithelium-free amniotic basement membrane, prepared as reported by Liotta et al. 7 The basement membrane served as a substratum for the cultured trophoblasts. Cells were fixed in 1% glutaraldehyde and 4% formalin for transmission electron microscopy. For scanning electron microscopy, the cells were fixed in 2.5% glutaraldehyde, 0.8% para formaldehyde, and 0.1 moUL cacodylate buffer, pH 7.3. Immunohistochemical staining of long-term trophoblast cultures. Cytotrophoblast cultures after 4 to 10 passages were characterized by immunohistochemical staining for numerous markers, identified by respective antibodies (Table I). These antibodies were against cytokeratin (Dakopatts, Santa Barbara, Calif.), vimentin (Poly Science Inc., Warrington, Penn.), hCG (polyclonal ab, ICN, Lisle, Ill.), a-chain hCG (monoclonal ab, ICN) and Trop-l and Trop-2 antigens (monoclonal antibodies, gift of L. A. Herzenberg, Stanford, Calif.).B The supernatant of the 63D3 hybridoma line (ATCC, Rockville, Md.) was the source of a monoclonal antibody against human macrophages." The cultures were fixed with 100% methanol for 10 minutes
940 Yagel et al.
April 1989 Am J Obstet Gynecol
Fig. 1. Trophoblast primary culture showing morphologically different cell populations apparently stemming from amorphous chorionic villus. (X 335.)
at - 20° C. Nonspecific immunoglobulin binding sites were masked by a 20-minute preincubation with 5% horse serum or goat serum or rabbit serum diluted in phosphate-buffered saline solution. The cultures were then rinsed once with phosphate-buffered saline solution -I % bovine serum albumin. Endogenous peroxidase activity was quenched by 20-minute incubation with 0.5% hydrogen peroxide in absolute methanol. All subsequent washings were with phosphate-buffered saline solution. Cells were treated with various primary antibodies at appropriate dilutions and then with biotinylated secondary antibodies (Dimension Lab, Mississauga, Ontario), the binding of which was visualized with an avidin-biotin-peroxidase complex detection method lO in all cases except a-hCG, which was detected by immunofluorescence. The avidin-biotin-peroxidase complex detection kit was purchased from Vector Laboratories, Burlingame, Calif. The details for detection of these markers are listed in Table I. Hormone content analysis. One hundred thousand trophoblast cells were incubated in I ml of complete media for 24 hours with or without the addition of pure hCG (CR-125, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, Md.). The supernatants were assayed for hCG when cells were incubated alone, and for progesterone or 1713-estradiol when cells were incubated with pure hCG. The level of hCG was assayed by the Amerlex-M luteinizing hormone radioimmunoassay kit (Amersham, Oakville, Ontario) according to the manufacturer's directions. The assay uses an antibody that cross reacts with the a-hCG chain by 22.9%. The concentration of progesterone and estradiol in the incubation media was measured by radio-
immunoassay. The progesterone and estradiol antibodies were raised in goats against 4-pregnen-3, 20-dione 11a-hemisuccinate: bovine serum albumin and 1,3,5( 10) -estratriene - 3, 1713 -diol- 6-one-6-(0-carboxymethyl)oxime: bovine serum albumin, respectively. The progesterone antibody used at a dilution of I: 7000 cross reacts with pregnenolone (1.2%), 20ahydroxyprogesteTOne (0.17%), 17a-hydroxyprogesterone (1.2%), and 2013-hydroxyprogesterone (0.8%) and with 1713-estradiol, dehydroepiandTOsterone, androstenedione, cortisol, and testosterone (0.1 %). The estradiol antibody, used at a dilution of I : 20,000, cross reacts with estrone (17%) and estriol (1.1 %) and with progesterone, cortisol, testosterone, and androstenedione «0.01 %). The progesterone and estradiol antisera were a gift from Dr. D. T. Armstrong (Medical Research Council Group in Reproductive Biology, University Hospital, London, Ontario). Progesterone was measured without chromatography after extraction from the incubation media with petroleum ether (boiling range 35° to 60° C, BDH Chemicals, Toronto). Methodologic losses of progesterone resulting from the extraction procedure were monitored by the recovery of 1,2,6, 73 H(N) progesterone (specific activity 90.1 Ci/mol) added to quadruplicate samples in each assay. 1713-Estradiol was measured after extraction from media with anhydrous diethyl ether (BDH Chemicals, Toronto). Methodologic losses of estradiol resulting from extraction were also monitored by addition of 2,4,6,7 3 H(N) 1713-estradiol (specific activity 99.0 Ci/mmol) to quadruplicate samples in each estradiol assay. The intraassay and interassay coefficients of variation for the progesterone assay were 3.9% and 2.6%, respectively. In the estradiol assay, the intraassay co-
Pure human cytotrophoblast in long-term culture
Volume 160 Number 4
941
Fig. 2. Cytotrophoblast in short-term culture showing four different types of cells (a to d). a, Cells resembling epithelial cells. (x 225.) b, Pleomorphic interdigitating cells resemble crazy paving. (X 225.) c, Multiangular cells. (x 450.) d, Closely packed spindle shaped cells. (X 225.)
efficient of variation was 6.5% and the interassay coefficient of variation was 7%. The sensitivity of the progesterone and estradiol assays was 4 and 6 pg per tube. respectively. The radioimmunoassay procedure was performed as described previously. I I. I ' Effects of anti-heG antibody on trophoblast proliferation. Trophoblast cell lines were cultured for 24 hours at a concentration of 15 X 103 cells per well. The cells were washed twice with phosphate-buffered saline solution and culture media were reconstituted. AntihCe antibody was then added to the wells at dilutions ranging from 1: 100 to I: 100.000. Normal rabbit serum at I : 100 dilution was added to the control wells. 'H-Thymidine (6.7 Ciimmol) at a concentration of 1 f.LCi per well was added to the same time point as the
anti-hCe antibody and the cultures were incubated for 48 hours . Proliferative response was evaluated from the beta counts of cells harvested at the end of the incubation period with a cell harvester (Titretek. Flow Laboratories. Rockville. Md.) .
Results Morphologic characteristics of cytotrophoblasts. Typically after 5 to 12 days in the primary explant, growing cell populations appeared to be stemming from the villi that had eventually necrotized (Fig. 1). Occasionally more than one cell type constituted the trophoblast outgrowth around the villi. After 12 to 25 days. the cells morphologically resembled those reported by earlier investigators. 1·5 In agreement with the
942
Yagel et al.
Apnl1989 Am J Obstet G ynecol
Fig. 3. a, Trophoblast plated on amniotic basement me mbrane viewed by transmissio n electron microscopy. 8M, basement membrane; S, stroma; M. mitochondria; G, Golgi; j, junctio n. ( X 6438.) b, Scanning electron microscopy oftrophoblasts cultured o n amniotic basement membrane. ( x 1130.)
descriptions by Loke et al.," we observed four distinct types of morphologic patterns: the epithelial type (Fig. 2, a), the interdigitating cells, which resemble crazy paving (Fig. 2, b), the multiangular cells (Fig. 2, c), and the closely packed spindle-shaped cells (Fig. 2, d). However, only the multiangular cell type predominated in the long-term cultures. When cultures became confluent, some of the cells resembled fibroblasts (Fig. 2, d). Cultures usually showed signs of syncytia formation (under the light microscope) but the incidence of multinucleate cells never exceeded 10%. Transmission and scanning electron microscopy of trophoblasts. Trophoblast cells from long-term cultures were replated on amniotic basement membranes for morphologic studies. Proliferating trophoblasts grown on membranes for 3 days showed epithelial char-
acteristics. They exhibited large nuclei, prominent nucleoli, and cytoplasm filled with vesicles, Golgi, and mitochondria; occasional desmosomes have also been identified (Fig. 3, a). Scanning electron microscopy revealed fairly uniform shaped cells that formed multilayers with long cytoplasmic processes that crisscrossed (Fig. 3, b). Immunohistochemical staining. Cultures derived after four or more passages were subjected to immunolabeling to examine the cellular composition of longterm culture. The cells were positive for the following markers (Table I) : cytokeratin (Fig. 4, a), hCG (Fig. 4, b), a-hCG, and Trop-l and Trop-2. The cells were negative for the vimentin and 63D3 antigens. The latter antigen is a monocyte-macrophage marker. Human monocytes that served as a negative control for hCG,
Pure human cytotrophoblast in long-term culture
Volume 160 "'umber 4
943
Fig. 4. Long-term cultured trophoblast cells labeled with anticYlOkeratin (a) and anti-hCG antibody (b) with the avidin-blotin-peroxidase complex technique. (a, x 755: b, x 463.)
Table II. hCG production by long-term trophoblast culture* !tce level (mlUlml)
Trophoblall hnes
TR-19-1 TR-19-2 TR-30-4-1 TR-30-4-2 TR-30-4-3 TR-22-7-1 TR-22-7-2 TR-22-7-3 TR-22-7-4 TR-22-7-6
0 0 0
3 3 2 2 2 2 2
444.3 404.4 272.7 342.4 709.2 51.3 91.1 211.1 67.2 114.1
hCe level (rnlU I ml)
5 4 5 5 6 5 5 6 5 5
391.7 320.8 281.1 311.7 584.1 50.7 111.1 141.I 50.7 143.6
!tce level (rnlU I rnl)
9 9
361.1 611.6
*Level of hCG in complete media of 10' cells per milliliter cultured for 24 hours after subtraction of control level. Measurements were done in triplicate. SD < 5%.
o:-hCG, Trop-l, and Trop-2 were always negative for these markers. They were positive for 63D3 marker and vimentin. It should be mentioned that decidual cells from fl'esh non villous placental preparations or short-term cultures used in other studies were 63D3 negative and vimentin positive. Hormone production by cultured trophoblast cells heG. The level of hCG in the primary and long-term trophoblast cultures was highly variable . The concentration ranged between 50 and 710 m I U I mil 5 x lO' cells per 24 hours for the cell lines tested. The level was also dependent on the batch of fetal calf serum used to enrich the trophoblast media at the time of hCG assay (data not shown), despite the fact that all batches of fetal calf serum were devoid of hCG. It is interesting that the level ofhCG did not decline, at least up to the ninth passage of three lines tested (Table II).
Progesterone. Progesterone synthesis and secretion increased when pure hCG was added to the culture media (Fig. 5). Estrogen. Only small amounts «20 pg/ml) of estrogen were detected in the trophoblast culture media when androstenedione was added to the culture (Fig. 5). hCG did not enhance the production and secretion of estradiol by the trophoblast cells. As shown in Fig. 6, hCG was an important mitogen for the long-term cultured trophoblast cells. At the dilutions of anti-hCG antibody used, trophoblast thymidine uptake was decreased to about 40% of the control at the lowest (1: 100) dilution. Normal rabbit serum at a dilution of 1 :100 did not show any effect on trophoblast growth (97% to 105% of the control). Subcultures. The average doubling time of the trophoblasts varied between 48 and 96 hours for eighteen
944
Yagel et al.
April 1989 Am J Obstet Gynecol
120
1000
-I 100
0
0::: IZ
E ......
0 0
'" a.
(f)
LL
Q 0
a: 400 w f-
(f)
200
0
60
IZ W 0 0::: W
40
a...
__________1.----------------
oo.'
80
20
1•
20
40
60
00
100
120
hCG(ng/ml)
Fig. 5. 10' Cytotrophoblast was incubated for 24 hours at 37° C in complete medium with indicated amount of hCG added. Concentrations of progesterone and estradiol were assayed as described under Material and methods. Net synthesis of hormones was calculated by subtracting their content in media at zero time. Values represent mean ± SE of quadruplicate samples. Progesterone (.); estradiol (e).
lines tested. Six cultures died after 4 months (six to eight passages). At the time of the manuscript preparation, the others were 5 to 8 months old (five to thirteen passages). Cells have been successfully recultured after storage in liquid nitrogen. Comment
This study demonstrates a simple technique for establishment of long-term first-trimester human trophoblast culture. The cultures maintained their trophoblastic purity even after many passages, as indicated by the presence of cytoplasmic hCG, cytokeratin, and surface Trop-l and Trop-2 antigens. Moreover, no cell was positive for vimentin, thus excluding fibroblastic and endothelial contamination. No cell was positive for 63D3, thus excluding macrophage contamination. The cells in culture were functionally active as demonstrated by hCG production and progesterone production. The latter was enhanced in the presence of pure hCG. Because up to 10% of the trophoblasts in culture showed signs of syncytial formation, the possibility of hormone production by syncytial cells alone is not ruled out. Estrogen was produced in small amounts by these cultures, even in the presence of androstenedione. This phenomenon might be explained by the lack of aromatase in first-trimester trophoblasts or by the need for specific stimulant(s) that may not have been available in these cultures. Human chorionic gonadotropin was found to be an important growth factor for the trophoblasts in culture, shown by up to
OL--L------~----~----~------~-
CONTROL
I: 100,000 I: 10,000
1'1,000
I : 100
ANTI-hCG ANTIBODY DILUTION Fig. 6. Effect of anti-hCG antibody (at dilutions indicated) on proliferation of 15 X 10' trophoblasts was tested. Counts were taken 48 hours after 'H-thymidine was added to the culture. Triplicate samples were assayed for each concentration point. Results represent mean ± SE for two experiments. TR-19-1 (.); TR-IO-2 (e); TR-30-4-2 (0).
60% suppression of 'H-thymidine uptake by trophoblasts incubated with anti-hCG antibody, suggesting an autocrine type regulation of trophoblast growth. However, even in the highest concentration, anti-hCG antibody did not completely prevent trophoblast proliferation, possibly indicating the presence of other trophoblast growth factors as described by others.'" l4 As mentioned earlier, all the markers and functions were tested at fourth or subsequent passages because long-term trophoblast culture was the primary goal of this study. Generally accepted morphologic characteristics alone are unreliable in defining cells in culture. Indeed, the morphology of cells changed dramatically from the primary to the long-term cultures without affecting their functional and antigenic characteristics. This study shows that the progenitor cells of the cytotrophoblast have extensive proliferative properties in vitro, although they are not immortal. Approximately one third of the trophoblast lines died after 4 months. This indicates a finite life span. However, some lines are still continuously growing after 8 months and 13 passages in culture. According to one study,' cells from term placentas have been maintained in continuous culture for up to I year. The major difference between the described method and those reported recently by others is that we deliberately did not use any enzymatic digestion of the chorionic tissues. Thus when undispersed villus fragments were plated as organ cultures on the dishes, there was a lesser chance for nontrophoblast cells to attach to the plastic and to prolif-
Pure human cytotrophoblast in long-term culture
Volume 160 Number 4
erate. We believe that this may explain the lack of macrophages, fibroblasts, and endothelial contamination in our established lines compared to other investigators. I -4 ,I5 The present technique is simple and reproducible and is currently being applied to studies of trophoblast function at the cellular and molecular levels, For example, we have shown that these cells are invasive in vitro and share key molecular mechanisms of invasion with highly metastatic tumor cells. I6
7.
8. 9. 10.
REFERENCES 1. Morgan DML, Toothill VJ, Landon MJ Long-term culture of human trophoblast cells. Br J Obstet Gynaecol 1985;92:84, 2, Loke YW, Butterworth BH, MargettsJJ, Burland K. Identification of cytotrophoblast colonies in cultures of human placental cells using monoclonal antibodies, Placenta 1986;7:221. 3. Schwab ME, Muller C, Schmid-Tannwald I. Fast and reliable culture method for cells from 8-10 week trophoblast tissue, Lancet 1984; 1: 1082, 4. Gustavii B, Edvall H, Mineur A, Heim S, Mandahl N, Kristoffersson U, Mitelman F. Trophoblast samples suitable for long-term culture. Acta Obstet Gynecol Scand 1985;64:661. 5. Kliman HJ, Nestler JE, Sermasi E, Sanger JM, StraussJF III. Purification, characterization, and in vitro differentiation of cytotrophoblasts from human term placentae, Endocrinology 1986; 118: 1567. 6, CulouscouJ-M, Remacle-Bonnet MM, Pommier G, Rance RJ, Depieds RC. Immunosuppressive properties of hu-
11.
12. 13.
14. 15. 16.
945
man placenta: study of supernatants from short-term syncytiotrophoblast cultures. J Reprod Immunol 1986;9:33. Liotta LA, Lee CW, Morakis DJ New method for preparing large surfaces of intact human basement membrane for tumor invasion studies. Cancer Lett 1980; 11:141. Lipinski M, Parks DR, Rouse RV, Herzenberg LA. Human trophoblast cell-surface antigens defined by monoclonal antibodies. Proc Natl Acad Sci USA 1981;78:5147. Ugolini V, Nunez G, Smith G, Stastny P, CapraJD. Initial characterization of monoclonal antibodies against human monocytes. Proc Natl Acad Sci USA 1980;77:6764. Hsu SM, Raine L, Fanger H. Use of avidin-biotin peroxidase complex (ABC) in immunoperoxidase techniques: a comparison between ABC and unlabeled antibody (PAP) procedures. J Histochem Cytochem 1981; 29:557. Patrick j, Challis j, Natale R, Richardson B. Circadian rhythms in maternal plasma cortisol, estrone, estradiol, and estriol at 34 to 35 weeks' gestation. AM j OBSTET GYNECOL 1979; 135:797. Leung RCK, Armstrong DT. Estrogen treatment of immature rats inhibits androgen production in vitro. Endocrinology 1979; 104: 1411. Goustin SA, Betsholtz C, Pfeifer-Ohlsson S, et al. Coexpression of the SIS and myc proto-oncogenes in developing human placenta suggests autocrine control of trophoblast growth. Cell 1985;41:301. Sen-Majumdar A, Murthy U, Das M. A new trophoblastderived growth factor from human placenta: purification and receptor identification. Biochemistry 1986;25:627. Stromberg K. The human placenta in cell and organ culture. Methods Cell Bioi 1980;21:227. Yagel S, Parhar RS, jeffrey jj, Lala PK. Normal nonmetastatic human trophoblast cells share in vitro invasive properties of malignant cells. j Cell Physiol1988; 136:455.
12. Eldridge FL, Millhorn DE. Oscillation, gating and memory in the respiratory control system. In: Fishman AP, ed. Handbook of physiology. Washington, DC: American Physiological Society, 1986:93-114. 13. Phillipson EA, Duffin J, Caper JD. Critical dependence of respiratory rhythmicity on metabolic CO2 load. J Appl Physiol 1981 ;50:45-54. 14. Boddy K, Dawes GS, Fisher R, Pinter S, Robinson JS. Foetal respiratory moverpents, electrocortical and cardiovascular responses to hypoxemia and hypercapnia in sheep. J Physiol (Lond) 1974;243:599-618. 15. Nijhuis JG, Jonsgma HW, Crijns I, de Valk I, van den
April 1989 Am J Obstet Gynecol
Velden J. Effects of maternal glucose ingestion on human fetal breathing movements at weeks 24 and 28 of gestation. Early Hum Dev 1986;13:183-8. 16. Natale R, Patrick J, Richardson B. Effects of human maternal venous plasma glucose concentrations on fetal breatl~ing movements. AM J OBSTET GVNECOL 1978; 132:36-41. 17. Natale R, Nasello-Paterson C, Connors G. Patterns of fetal breathing activity in the human fetus at 24 to 28 weeks of gestation. AM J OBSTET GVNECOL 1988;158: 317-21.
Characterization of pure human first-trimester cytotrophoblast cells in long-term culture: Growth pattern, markers, and hormone production Simcha Yagel, MD," Robert F. Casper, MD: Wendy Powell, MSc: Ranjit S. Parhar, MS,. and Peeyush K. Lala, MD, PhD' London, Ontario, Canada Pure long-term cytotrophoblast cultures were established from human first-trimester placentas by growing chorionic villus explants without enzymatic digestion. Cytoplasmic human chorionic gonadotropin was detectable in all (100%) cells in culture when labeled with a polyclonal anti-human chorionic gonadotropin antibody and in 71% to'83% of cells labeled with a monoclonal anti-a-human chorionic gonadotropin antibody. Most of the cells expressed cytokeratin and surface Trop-1 and Trop-2 antigens (89% to 95%), but none expressed cytoplasmic vimentin or surface 6303 antigens. Study of the ultrastructure of the cells demonstrated epithelial morphologic features. The average doubling time of the trophoblast was 48 to 96 hours. Some of the lines have been continuously propagated for 8 months. They produced variable amounts of human chorionic gonadotropin (50 to 710 mlU/ml per 105 cells per 24 hours). The basal level of progesterone secreted by trophoblast (444.4 ± 32.4 pg/ml per 105 cells per 24 hours) doubled in the presence of pure human chorionic gonadotropin (100 ng/ml). They produced small amounts of 17J3-estradiol «20 pg/ml per 100 cells per 24 hours); human choriOniC gonadotropin had no effect on the estradiol production. Trophoblast-derived human chorionic gonadotropin acted as a growth factor because trophoblast proliferation (measured by uptake of thymidine labeled with tritium) was reduced by 60% in the presence of an anti-human chorionic gonadotropin antibody. Availability of pure, functionally competent human cytotrophoblast in long-term cultures is relevant for further studies in reproduction biology. (AM J OSSTET GVNECOL 1989;160:938-45.)
Key words: Trophoblast, trophoblast culture, human chorionic gonadotropin, progesterone
From the Department of Anatom1 and the Department of ObstetriCs and Gynecology/ Universl/.V Hospital, The University of We~tern Ontario. Supported by the Medical Research Counal of Canada, the National Cancer Institute of Canada, the Cancer Research Society, and Pharmasaence Inc., Montreal, Quebec. S. Yagel was a fellow of the Medical Research CounCil of Canada. Presented at the Forty{ourth Annual Meetzng of the Socuty of Obstetricians and Gynaecologis~ of Canada, Vancouver, Brll~h Columbza, Canada, June 21-25,1988, Reprint reques~: P. K. Lala, MD, PhD, Department of Anatomy, The University of Western Ontano, London, Ontario, Canada N6A 5C1. 938
A variety of techniques has been used to grow and maintain human first-trimester or term placental trophoblast cells in culture.'-6 Some cultures have been reported to retain trophoblastic markers even after several passages in vitro. The difficulty in obtaining pure trophoblast cultures can partly be explained by the fact that enzymatic digestion of the placenta, usually undertaken before culturing of mixed cell populations, results in inclusion of trophoblast cells, macrophages, endothelial cells, and fibroblasts. Kliman et al. reported j
Pure human cytotrophoblast in long-term culture
Volume 160 Number 4
939
Table I. Markers used for trophoblast characterization* Marker
Specificity
Cytokeratin Epithelial cell marker Vimentin Mesenchymally derived cell marker LCG hCG a-hCG a-Chain of hCG Trop-I Trop-2 63D3
Primary antibody
Polyclonal rabbit, I: 500
Secondary antIbody
Biotinylated goat antirabbit immunoglobulin Polyclonal goat, I: 500 Biotinylated rabbit antigoat immunoglobulin Biotinylated goat antirabbit Polyclonal rabbit, I: 100 Monoclonal mouse antihuman, Rhodamine-conjugated rabbit anti-mouse 1:100 Trophoblast antigen Monoclonal mouse antihuman, Biotinylated horse antimouse I: 100 Trophoblast antigen Monoclonal mouse antihuman, Biotinylated horse antimouse I: 100 Monoclonal mouse antihuman Biotinylated horse anti mouse Human monocyte and macrophage marker Hybridoma sup. (X 10 conc.)
Trophoblast labeled (%) (range)
81-91
o 100 71-83 89-93 91-95
o
*AIl markers except a-hCG were identified with avidin-biotin-peroxidase complex method. a-hCG was identified with immunofluorescence method.
culturing of contamination-free functional cytotrophoblast cells obtained by Percoll gradient centrifugation of enzymatically dispersed term placenta. However, they did not attempt long-term maintenance of these cells by repeated passage. Establishment of longterm cultures of pure first-trimester human cytotrophoblast cells with functional integrity should provide reproducible in vitro systems for investigating endocrinologic and immunologic roles of these cells in pregnancy and the biologic basis of invasive behavior by cells. Our study was designed to achieve this objective. We show that by using the described methods, longterm replicating trophoblast populations were obtained that were free of macrophages, fibroblasts, and endothelial cells, as judged by morphologic features and numerous cell-specific markers. Moreover, these longterm monolayer cultures were proved to be functional: they produced a significant amount of human chorionic gonadotropin (hCG) and progesterone. hCG was found to be an important growth factor for the long-term cultured trophoblast cells. Material and methods
Primary trophoblast cultures. First-trimester placental tissues were obtained after elective termination of pregnancy performed by dilation and curettage at Victoria Hospital, London, Ontario. The tissues were rinsed thoroughly in cold phosphate-buffered saline solution. Areas rich in chorionic villi were selected to isolate the villi, which were minced into pieces of ::; 1 mm. Nonvillous materials inclusive of decidual tissue, blood clots, and membranes were carefully discarded. The fragments of villi were washed three times with minimal essential medium supplemented with streptomycin (20 mg/ml), penicillin (500 U /ml), and amphotericin B (25 ILg/ml). These fragments were then
cultured in RPMI 1640 (Grand Island Biological Co., Grand Island, N.Y.) medium containing 10% fetal calf serum and antibiotics (as stated above) in closed tissue culture flasks (Nunclon, 25 cm\ Nunc, Denmark) for 3 to 5 days until nonadherent cells could be removed and discarded. Cultures of adherent cells were expanded for 1 to 2 weeks in these flasks in fresh medium before their passage and long-term maintenance in vitro in similar flasks. For passages, cells were removed by a mild (0.25%) trypsin treatment, washed, and replanted. Transmission and scanning electron microscopy. For transmission electron microscopy and scanning electron microscopy preparations, trophoblasts were seeded on epithelium-free amniotic basement membrane, prepared as reported by Liotta et al. 7 The basement membrane served as a substratum for the cultured trophoblasts. Cells were fixed in 1% glutaraldehyde and 4% formalin for transmission electron microscopy. For scanning electron microscopy, the cells were fixed in 2.5% glutaraldehyde, 0.8% para formaldehyde, and 0.1 moUL cacodylate buffer, pH 7.3. Immunohistochemical staining of long-term trophoblast cultures. Cytotrophoblast cultures after 4 to 10 passages were characterized by immunohistochemical staining for numerous markers, identified by respective antibodies (Table I). These antibodies were against cytokeratin (Dakopatts, Santa Barbara, Calif.), vimentin (Poly Science Inc., Warrington, Penn.), hCG (polyclonal ab, ICN, Lisle, Ill.), a-chain hCG (monoclonal ab, ICN) and Trop-l and Trop-2 antigens (monoclonal antibodies, gift of L. A. Herzenberg, Stanford, Calif.).B The supernatant of the 63D3 hybridoma line (ATCC, Rockville, Md.) was the source of a monoclonal antibody against human macrophages." The cultures were fixed with 100% methanol for 10 minutes
940 Yagel et al.
April 1989 Am J Obstet Gynecol
Fig. 1. Trophoblast primary culture showing morphologically different cell populations apparently stemming from amorphous chorionic villus. (X 335.)
at - 20° C. Nonspecific immunoglobulin binding sites were masked by a 20-minute preincubation with 5% horse serum or goat serum or rabbit serum diluted in phosphate-buffered saline solution. The cultures were then rinsed once with phosphate-buffered saline solution -I % bovine serum albumin. Endogenous peroxidase activity was quenched by 20-minute incubation with 0.5% hydrogen peroxide in absolute methanol. All subsequent washings were with phosphate-buffered saline solution. Cells were treated with various primary antibodies at appropriate dilutions and then with biotinylated secondary antibodies (Dimension Lab, Mississauga, Ontario), the binding of which was visualized with an avidin-biotin-peroxidase complex detection method lO in all cases except a-hCG, which was detected by immunofluorescence. The avidin-biotin-peroxidase complex detection kit was purchased from Vector Laboratories, Burlingame, Calif. The details for detection of these markers are listed in Table I. Hormone content analysis. One hundred thousand trophoblast cells were incubated in I ml of complete media for 24 hours with or without the addition of pure hCG (CR-125, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, Md.). The supernatants were assayed for hCG when cells were incubated alone, and for progesterone or 1713-estradiol when cells were incubated with pure hCG. The level of hCG was assayed by the Amerlex-M luteinizing hormone radioimmunoassay kit (Amersham, Oakville, Ontario) according to the manufacturer's directions. The assay uses an antibody that cross reacts with the a-hCG chain by 22.9%. The concentration of progesterone and estradiol in the incubation media was measured by radio-
immunoassay. The progesterone and estradiol antibodies were raised in goats against 4-pregnen-3, 20-dione 11a-hemisuccinate: bovine serum albumin and 1,3,5( 10) -estratriene - 3, 1713 -diol- 6-one-6-(0-carboxymethyl)oxime: bovine serum albumin, respectively. The progesterone antibody used at a dilution of I: 7000 cross reacts with pregnenolone (1.2%), 20ahydroxyprogesteTOne (0.17%), 17a-hydroxyprogesterone (1.2%), and 2013-hydroxyprogesterone (0.8%) and with 1713-estradiol, dehydroepiandTOsterone, androstenedione, cortisol, and testosterone (0.1 %). The estradiol antibody, used at a dilution of I : 20,000, cross reacts with estrone (17%) and estriol (1.1 %) and with progesterone, cortisol, testosterone, and androstenedione «0.01 %). The progesterone and estradiol antisera were a gift from Dr. D. T. Armstrong (Medical Research Council Group in Reproductive Biology, University Hospital, London, Ontario). Progesterone was measured without chromatography after extraction from the incubation media with petroleum ether (boiling range 35° to 60° C, BDH Chemicals, Toronto). Methodologic losses of progesterone resulting from the extraction procedure were monitored by the recovery of 1,2,6, 73 H(N) progesterone (specific activity 90.1 Ci/mol) added to quadruplicate samples in each assay. 1713-Estradiol was measured after extraction from media with anhydrous diethyl ether (BDH Chemicals, Toronto). Methodologic losses of estradiol resulting from extraction were also monitored by addition of 2,4,6,7 3 H(N) 1713-estradiol (specific activity 99.0 Ci/mmol) to quadruplicate samples in each estradiol assay. The intraassay and interassay coefficients of variation for the progesterone assay were 3.9% and 2.6%, respectively. In the estradiol assay, the intraassay co-
Pure human cytotrophoblast in long-term culture
Volume 160 Number 4
941
Fig. 2. Cytotrophoblast in short-term culture showing four different types of cells (a to d). a, Cells resembling epithelial cells. (x 225.) b, Pleomorphic interdigitating cells resemble crazy paving. (X 225.) c, Multiangular cells. (x 450.) d, Closely packed spindle shaped cells. (X 225.)
efficient of variation was 6.5% and the interassay coefficient of variation was 7%. The sensitivity of the progesterone and estradiol assays was 4 and 6 pg per tube. respectively. The radioimmunoassay procedure was performed as described previously. I I. I ' Effects of anti-heG antibody on trophoblast proliferation. Trophoblast cell lines were cultured for 24 hours at a concentration of 15 X 103 cells per well. The cells were washed twice with phosphate-buffered saline solution and culture media were reconstituted. AntihCe antibody was then added to the wells at dilutions ranging from 1: 100 to I: 100.000. Normal rabbit serum at I : 100 dilution was added to the control wells. 'H-Thymidine (6.7 Ciimmol) at a concentration of 1 f.LCi per well was added to the same time point as the
anti-hCe antibody and the cultures were incubated for 48 hours . Proliferative response was evaluated from the beta counts of cells harvested at the end of the incubation period with a cell harvester (Titretek. Flow Laboratories. Rockville. Md.) .
Results Morphologic characteristics of cytotrophoblasts. Typically after 5 to 12 days in the primary explant, growing cell populations appeared to be stemming from the villi that had eventually necrotized (Fig. 1). Occasionally more than one cell type constituted the trophoblast outgrowth around the villi. After 12 to 25 days. the cells morphologically resembled those reported by earlier investigators. 1·5 In agreement with the
942
Yagel et al.
Apnl1989 Am J Obstet G ynecol
Fig. 3. a, Trophoblast plated on amniotic basement me mbrane viewed by transmissio n electron microscopy. 8M, basement membrane; S, stroma; M. mitochondria; G, Golgi; j, junctio n. ( X 6438.) b, Scanning electron microscopy oftrophoblasts cultured o n amniotic basement membrane. ( x 1130.)
descriptions by Loke et al.," we observed four distinct types of morphologic patterns: the epithelial type (Fig. 2, a), the interdigitating cells, which resemble crazy paving (Fig. 2, b), the multiangular cells (Fig. 2, c), and the closely packed spindle-shaped cells (Fig. 2, d). However, only the multiangular cell type predominated in the long-term cultures. When cultures became confluent, some of the cells resembled fibroblasts (Fig. 2, d). Cultures usually showed signs of syncytia formation (under the light microscope) but the incidence of multinucleate cells never exceeded 10%. Transmission and scanning electron microscopy of trophoblasts. Trophoblast cells from long-term cultures were replated on amniotic basement membranes for morphologic studies. Proliferating trophoblasts grown on membranes for 3 days showed epithelial char-
acteristics. They exhibited large nuclei, prominent nucleoli, and cytoplasm filled with vesicles, Golgi, and mitochondria; occasional desmosomes have also been identified (Fig. 3, a). Scanning electron microscopy revealed fairly uniform shaped cells that formed multilayers with long cytoplasmic processes that crisscrossed (Fig. 3, b). Immunohistochemical staining. Cultures derived after four or more passages were subjected to immunolabeling to examine the cellular composition of longterm culture. The cells were positive for the following markers (Table I) : cytokeratin (Fig. 4, a), hCG (Fig. 4, b), a-hCG, and Trop-l and Trop-2. The cells were negative for the vimentin and 63D3 antigens. The latter antigen is a monocyte-macrophage marker. Human monocytes that served as a negative control for hCG,
Pure human cytotrophoblast in long-term culture
Volume 160 "'umber 4
943
Fig. 4. Long-term cultured trophoblast cells labeled with anticYlOkeratin (a) and anti-hCG antibody (b) with the avidin-blotin-peroxidase complex technique. (a, x 755: b, x 463.)
Table II. hCG production by long-term trophoblast culture* !tce level (mlUlml)
Trophoblall hnes
TR-19-1 TR-19-2 TR-30-4-1 TR-30-4-2 TR-30-4-3 TR-22-7-1 TR-22-7-2 TR-22-7-3 TR-22-7-4 TR-22-7-6
0 0 0
3 3 2 2 2 2 2
444.3 404.4 272.7 342.4 709.2 51.3 91.1 211.1 67.2 114.1
hCe level (rnlU I ml)
5 4 5 5 6 5 5 6 5 5
391.7 320.8 281.1 311.7 584.1 50.7 111.1 141.I 50.7 143.6
!tce level (rnlU I rnl)
9 9
361.1 611.6
*Level of hCG in complete media of 10' cells per milliliter cultured for 24 hours after subtraction of control level. Measurements were done in triplicate. SD < 5%.
o:-hCG, Trop-l, and Trop-2 were always negative for these markers. They were positive for 63D3 marker and vimentin. It should be mentioned that decidual cells from fl'esh non villous placental preparations or short-term cultures used in other studies were 63D3 negative and vimentin positive. Hormone production by cultured trophoblast cells heG. The level of hCG in the primary and long-term trophoblast cultures was highly variable . The concentration ranged between 50 and 710 m I U I mil 5 x lO' cells per 24 hours for the cell lines tested. The level was also dependent on the batch of fetal calf serum used to enrich the trophoblast media at the time of hCG assay (data not shown), despite the fact that all batches of fetal calf serum were devoid of hCG. It is interesting that the level ofhCG did not decline, at least up to the ninth passage of three lines tested (Table II).
Progesterone. Progesterone synthesis and secretion increased when pure hCG was added to the culture media (Fig. 5). Estrogen. Only small amounts «20 pg/ml) of estrogen were detected in the trophoblast culture media when androstenedione was added to the culture (Fig. 5). hCG did not enhance the production and secretion of estradiol by the trophoblast cells. As shown in Fig. 6, hCG was an important mitogen for the long-term cultured trophoblast cells. At the dilutions of anti-hCG antibody used, trophoblast thymidine uptake was decreased to about 40% of the control at the lowest (1: 100) dilution. Normal rabbit serum at a dilution of 1 :100 did not show any effect on trophoblast growth (97% to 105% of the control). Subcultures. The average doubling time of the trophoblasts varied between 48 and 96 hours for eighteen
944
Yagel et al.
April 1989 Am J Obstet Gynecol
120
1000
-I 100
0
0::: IZ
E ......
0 0
'" a.
(f)
LL
Q 0
a: 400 w f-
(f)
200
0
60
IZ W 0 0::: W
40
a...
__________1.----------------
oo.'
80
20
1•
20
40
60
00
100
120
hCG(ng/ml)
Fig. 5. 10' Cytotrophoblast was incubated for 24 hours at 37° C in complete medium with indicated amount of hCG added. Concentrations of progesterone and estradiol were assayed as described under Material and methods. Net synthesis of hormones was calculated by subtracting their content in media at zero time. Values represent mean ± SE of quadruplicate samples. Progesterone (.); estradiol (e).
lines tested. Six cultures died after 4 months (six to eight passages). At the time of the manuscript preparation, the others were 5 to 8 months old (five to thirteen passages). Cells have been successfully recultured after storage in liquid nitrogen. Comment
This study demonstrates a simple technique for establishment of long-term first-trimester human trophoblast culture. The cultures maintained their trophoblastic purity even after many passages, as indicated by the presence of cytoplasmic hCG, cytokeratin, and surface Trop-l and Trop-2 antigens. Moreover, no cell was positive for vimentin, thus excluding fibroblastic and endothelial contamination. No cell was positive for 63D3, thus excluding macrophage contamination. The cells in culture were functionally active as demonstrated by hCG production and progesterone production. The latter was enhanced in the presence of pure hCG. Because up to 10% of the trophoblasts in culture showed signs of syncytial formation, the possibility of hormone production by syncytial cells alone is not ruled out. Estrogen was produced in small amounts by these cultures, even in the presence of androstenedione. This phenomenon might be explained by the lack of aromatase in first-trimester trophoblasts or by the need for specific stimulant(s) that may not have been available in these cultures. Human chorionic gonadotropin was found to be an important growth factor for the trophoblasts in culture, shown by up to
OL--L------~----~----~------~-
CONTROL
I: 100,000 I: 10,000
1'1,000
I : 100
ANTI-hCG ANTIBODY DILUTION Fig. 6. Effect of anti-hCG antibody (at dilutions indicated) on proliferation of 15 X 10' trophoblasts was tested. Counts were taken 48 hours after 'H-thymidine was added to the culture. Triplicate samples were assayed for each concentration point. Results represent mean ± SE for two experiments. TR-19-1 (.); TR-IO-2 (e); TR-30-4-2 (0).
60% suppression of 'H-thymidine uptake by trophoblasts incubated with anti-hCG antibody, suggesting an autocrine type regulation of trophoblast growth. However, even in the highest concentration, anti-hCG antibody did not completely prevent trophoblast proliferation, possibly indicating the presence of other trophoblast growth factors as described by others.'" l4 As mentioned earlier, all the markers and functions were tested at fourth or subsequent passages because long-term trophoblast culture was the primary goal of this study. Generally accepted morphologic characteristics alone are unreliable in defining cells in culture. Indeed, the morphology of cells changed dramatically from the primary to the long-term cultures without affecting their functional and antigenic characteristics. This study shows that the progenitor cells of the cytotrophoblast have extensive proliferative properties in vitro, although they are not immortal. Approximately one third of the trophoblast lines died after 4 months. This indicates a finite life span. However, some lines are still continuously growing after 8 months and 13 passages in culture. According to one study,' cells from term placentas have been maintained in continuous culture for up to I year. The major difference between the described method and those reported recently by others is that we deliberately did not use any enzymatic digestion of the chorionic tissues. Thus when undispersed villus fragments were plated as organ cultures on the dishes, there was a lesser chance for nontrophoblast cells to attach to the plastic and to prolif-
Pure human cytotrophoblast in long-term culture
Volume 160 Number 4
erate. We believe that this may explain the lack of macrophages, fibroblasts, and endothelial contamination in our established lines compared to other investigators. I -4 ,I5 The present technique is simple and reproducible and is currently being applied to studies of trophoblast function at the cellular and molecular levels, For example, we have shown that these cells are invasive in vitro and share key molecular mechanisms of invasion with highly metastatic tumor cells. I6
7.
8. 9. 10.
REFERENCES 1. Morgan DML, Toothill VJ, Landon MJ Long-term culture of human trophoblast cells. Br J Obstet Gynaecol 1985;92:84, 2, Loke YW, Butterworth BH, MargettsJJ, Burland K. Identification of cytotrophoblast colonies in cultures of human placental cells using monoclonal antibodies, Placenta 1986;7:221. 3. Schwab ME, Muller C, Schmid-Tannwald I. Fast and reliable culture method for cells from 8-10 week trophoblast tissue, Lancet 1984; 1: 1082, 4. Gustavii B, Edvall H, Mineur A, Heim S, Mandahl N, Kristoffersson U, Mitelman F. Trophoblast samples suitable for long-term culture. Acta Obstet Gynecol Scand 1985;64:661. 5. Kliman HJ, Nestler JE, Sermasi E, Sanger JM, StraussJF III. Purification, characterization, and in vitro differentiation of cytotrophoblasts from human term placentae, Endocrinology 1986; 118: 1567. 6, CulouscouJ-M, Remacle-Bonnet MM, Pommier G, Rance RJ, Depieds RC. Immunosuppressive properties of hu-
11.
12. 13.
14. 15. 16.
945
man placenta: study of supernatants from short-term syncytiotrophoblast cultures. J Reprod Immunol 1986;9:33. Liotta LA, Lee CW, Morakis DJ New method for preparing large surfaces of intact human basement membrane for tumor invasion studies. Cancer Lett 1980; 11:141. Lipinski M, Parks DR, Rouse RV, Herzenberg LA. Human trophoblast cell-surface antigens defined by monoclonal antibodies. Proc Natl Acad Sci USA 1981;78:5147. Ugolini V, Nunez G, Smith G, Stastny P, CapraJD. Initial characterization of monoclonal antibodies against human monocytes. Proc Natl Acad Sci USA 1980;77:6764. Hsu SM, Raine L, Fanger H. Use of avidin-biotin peroxidase complex (ABC) in immunoperoxidase techniques: a comparison between ABC and unlabeled antibody (PAP) procedures. J Histochem Cytochem 1981; 29:557. Patrick j, Challis j, Natale R, Richardson B. Circadian rhythms in maternal plasma cortisol, estrone, estradiol, and estriol at 34 to 35 weeks' gestation. AM j OBSTET GYNECOL 1979; 135:797. Leung RCK, Armstrong DT. Estrogen treatment of immature rats inhibits androgen production in vitro. Endocrinology 1979; 104: 1411. Goustin SA, Betsholtz C, Pfeifer-Ohlsson S, et al. Coexpression of the SIS and myc proto-oncogenes in developing human placenta suggests autocrine control of trophoblast growth. Cell 1985;41:301. Sen-Majumdar A, Murthy U, Das M. A new trophoblastderived growth factor from human placenta: purification and receptor identification. Biochemistry 1986;25:627. Stromberg K. The human placenta in cell and organ culture. Methods Cell Bioi 1980;21:227. Yagel S, Parhar RS, jeffrey jj, Lala PK. Normal nonmetastatic human trophoblast cells share in vitro invasive properties of malignant cells. j Cell Physiol1988; 136:455.