IN VITRO Volume15, No. 2, 1979 All rightsreserved 9
D I F F E R E N T I A T I O N IN H U M A N A M N I O T I C F L U I D C E L L C U L T U R E S : CHORIONIC GONADOTROPIN PRODUCTION 1 ROBERT E. PRIEST, ~JEAN H. PRIEST, JESSIE F. MOINUDDIN, ANDDEMETRIOS S. SGOUTAS
Department of Pathology and Laboratory Medicine and Department of Pediatrics, Division of Medical Genetics, Emory University, Atlanta, Georgia30322
SUMMARY Two of the distinguishable cell classes subcultured from human amniotic fluid were examined for their capability to produce human chorionic gonadotropin (hCG) as determined by radioimmunoassay. The class that predominates in most cultures used for prenatal genetic diagnosis, previously termed AF (for amniotic fluidl, secretes hCG into the culture medium. Dermal fibroblasts do not, nor does another type of cultured cell from amniotic fluid, previously termed F because of a resemblance to fibroblasts. Primary AF cultures produce more hCG than do subcultures. Evidence that this hormone is intact hCG is provided by its immunoreactivity with antisera raised against the/~-subunit and against the intact molecule of hCG. Furthermore, a dose-response curve for hormone in culture medium is parallel to that of highly purified intact hCG. It is postulated that AF cultures are derived from fetal membranes and retain properties of trophoblast.
Key words: human amniotic fluid; cell cultures; hCG (human chorionic gonadotropink radioimmunoassay; trophoblast. INTRODUCTION Interpretation of some prenatal genetic diagnoses is difficult because we lack information about origins and characteristics of the different cells in human amniotic-fluid cultures from second-trimester pregnancies. For instance, evaluation of chromosomal mosaicism in amniotic-fluid cultures is a problem when one does not know with certainty if the cells originate from the fetus, from extraembryonic fetal membranes, or from both (1, 2). Another problem of interpretation concerns studies of enzymes and proteins that are tissue-specific, for example, inherited disorders of collagen (3). Three distinguishable cell classes cultured from amniotic fluid were termed E for epithelial, AF for amniotic fluid, and F for fibroblast by Hoehn et al. (4). Only the AF and F cell classes can be subcuhured routinely by trypsin, and are the subject of this report. Since the AF class predomi~Research supported by Grand HD 11379 from the National Institutes of Health. 2To whom request for reprints should be addressed at the Department of Pathology and Laboratory Medicine, Emory University, Atlanta, Georgia 30322.
nates in primary and early subcultures of secondtrimester amniotic fluids, these cells presumably also predominate in most cultures used for prenatal genetic diagnosis. AF subcultures show cellular pleomorphism and multinucleation, a loose network appearance at confluency, and growth potential less than half that of human diploid fibroblasts {4, 5). On the other hand, F subcultures resemble typical fibroblast cultures from dermis and other connective tissues in every respect so far studied. The cells have a spindle shape and form parallel arrays at confluency. There is no tendency to become multinucleated and the growth potential is like that of other human diploid fetal fibroblasts {4, 5). Other differences between these two cell classes reported recently from this laboratory include: (a) production of a glycoprotein specific for epithelial basement membrane by AF but not F cultures {6); (b) extracellular material with ultrastructural characteristics of Type I {connective tissue) collagen fibers in F but not AF cultures {5); and (c) biochemical evidence for Type I collagen in F cultures and for a basement membrane collagen in AF cultures (3). 142
143
CHORIONIC GONADOTROPIN PRODUCTION The information accumulating about F cultures strongly indicates their connective-tissue origin, whereas a source for AF cultures that would account for all characteristics described thus far is trophoblast. In these studies we provide further evidence for this extraembryonic fetal origin of the predominant cell type used for prenatal genetic diagnosis. We show that A F cells produce human chorionic gonadotropin lhCG) in standard culture conditions and without inducing agents. MATERIALS AND METHODS
Cell culture. The routine establishment and maintenance of amniotic-fluid cultures have been described previously (3, 7). Culture conditions and descriptions for individual experiments are included in t h e legends for Figs. 1 and 2, and Table 1. Dulbecco and Vogt ID&V) medium was prepared in this laboratory from a standard recipe, filter-sterilized with 0.22-tam pore size, and stored at 5 ~ C not longer than 3 months. Glutamine, fetal bovine serum, and antibiotics (penicillin and streptomycin) were stored at - 1 5 ~ C and added to complete medium just before use with cells. Primary cultures were grown in 25% and subcultures in 15% fetal bovine serum. Antibiotics were included routinely in medium of primary cultures but not in subcultures. Microbial growth was monitored by a change from the normal clear appearance of culture medium, unusual pH or cell growth, and increase in chromosomal breaks and rearrangements in a diploid culture. Mycoplasma assay was run, as indicated, by the method of Chen (8) on cells subcultured to one-chamber slides tLab-Tek Products). Chromosome analysis routinely employed G-banding by trypsin and Giemsa, as well as Q-banding by quinacrine dihydrochloride stain and fluorescence microscopy. Radioimmunoassay. Amounts of hCG were determined in the culture medium with the exception of one series of experiments that utilized cell homogenates lsee Fig. 3). Initially, a nonspecific radioimmunoassay was used which measures hCG and luteinizing hormone equally well due to the presence of a common a-subunit and homologies in the amino-acid sequence of the/3-subunits of these hormones t9). In later experiments, a specific radioimmunoassay for determination of hCG was used which employs antisera raised against the/~-subunit of hCG (purchased from the Institute of Bioendocrinology, Montreal, Quebec, Canada). This assay measures intact hCG and the
30
o 20
Nuclei/Field
l
& % M
~
15 IO 5 O
I
I
i
5
mIU/ml
I
I
I
4
g
12 8 4
~
DAYS
FI6. 1. The numbers of nuclei per unit area, the percent of nuclei in multinucleated cells, and the production of hCG with time in subculture. Cells for study are subcultured from AF mass strain PR-683 (chromosomally normal, two doublings after first subculture, from amniotic fluid, 16-week pregnancy of a 36-year-old). Cells are inoculated onto replicate one-chamber slides (Lab-Tek Products, Division Miles Laboratories, Inc., Naperville, IlL ) in 4 ml D&V medium, 15% fetal bovine serum. Each day, for 5 days after inoculation, medium is decanted and stored at -15 ~ C prior to a nonspecific radioimmunoassay for hCG (see text). Cell monolayers are rinsed twice with phosphate-buffered saline, chambers are removed, and slides are fixed in 95% ethanol for 15 min and stained by Papanicolaou's method. Analysis of cell nuclei is performed at x430 magnification using a 10-mm square grid in the ocular. The height of each bar ltop panel) represents mean number of nuclei in 50 fields examined; the line at the top of each bar represents _+ one standard deviation. /3-subunit of hCG with cross-reactivity of 100% (10). We determined that the antiserum has crossreactivities of less than 2% with follicle-stimulating hormone, less than 5~ with luteinizing hormone, less than 0 . ] % with thyroid-stimulating hormone, and less than 0.1% with growth hormone. The antiserum does not cross-react with
144
PRIEST ET AL.
I AF" 20
FiDermal broblastI I0~ 106 ~
DAYS FIG. 2. The number of cells per flask, semilog scale ~O--O) and production of hCG (e--Q) with time in subculture of three cell types. Cells for study are subcultltred from: (a) AF mass strain PR-713 (chromosomally normal, two doublings after first subculture, from amniotic fluid, 15.5-week pregnancy of a 38-year-oldl; (bl F cloned strain from Hoelger Hoehn, University of Washington, Seattle qchromosomallynormal, 55 doublings from single cell, from amniotic fluid, 21-week pregnancy of a mother with previous Down syndrome child; and (c) mass fetal dermal fibroblast strain PR-135 (chromosomally normal, 17 doublings after first subculture, 16-week therapeutic abortusl. Six replicative 75cm~ culttrre flasks for each culture type are inoculated with about 5 • 10~ceils per flask in 10 ml D&V medium, 15% fetal bovine serum. On days 2, 3, 4, 6, 8 and 13, medium is removed and stored at -15 ~ C prior to specific hCG/~-subunit assay (see text). Cell layers are rinsed with phosphate-buffered saline, suspended with 0.25% trypsin at 37~ C for 30 rain, diluted further with phosphate-buffered saline, and inoculated into a hemacytometer for cell counts. free c~-subunits of hCG and is capable of measuring 0.8 ng hCG per ml of culture medium. The antiserum was calibrated against complete hCG by radioimmunoassay using the second International Standard of hCG (from the World Health Organization). A highly purified hCG Ibatch, CR-119, Pituitary Hormone Distribution Program, National Institutes of Health) with known concentrations of hCG also was assayed at several dilutions concurrently. In addition, a homologous hCG assay employing antiserum to intact hCG (11) was used in order to determine if the activity measured in the hCG f~-subunit assay was the intact hormone or free/3-subunits. This assay has incomplete crossreactivity with luteinizing hormone and with the a- and f~-subunits ofhCG ~12). In all of these radioimmunoassays, we compensated for effect of protein concentration in the reaction mixture by first lyophilizing the specimens and then bringing them up to initial volume with human male serum. Polyethylene glycol was used as a precipitant for ['2SI]labeled antigen-
antibody complex. All assays were done in triplicate, and the results were calculated from linear regression curves of logit B/B0 versus log concentration (see Fig. 3). Unless otherwise specified, materials for radioimmunoassays were purchased from Serono Laboratories, Inc., Boston, Massachusetts, or from Pharmacia Fine Chemicals, Piscataway, New Jersey.
RESULTS Initially, to document increase in muhinucleation with time in culture of AF cells and to determine if glycoprotein hormone was present in the culture medium, mass AF subcultures were inoculated into replicate slide culture chambers. These AF subcultures were identified by morphologic characteristics (4, 5) which are also summarized in the Introduction section. The number of nuclei per microscope field increases with time (Fig. 1, top panelL and the percent of nuclei found in multinucleated cells also increases ~Fig. 1, center panel). Hormone levels in culture medium rise with time in culture {Fig. 1, bottom panel) as determined by the nonspecific radioimmunoassay that measures hCG and luteinizing hormone equally well due to presence of identical asubunits and homologies in amino-acid sequence of the fJ-subunits in these hormones (9L The experiment shown in Fig. 1 was repeated once with similar results. In these experiments, the percent of nuclei found in multinucleated ceils cannot be determined accurately after day 3 because cells are too crowded on the slides. In other experiments when cells are inoculated at lower density initially, or grown on larger surface area, the percent of nuclei found in multinucleated ceils increases to a maximum of 30% and reaches a plateau by 5 to 7 days after subculture. Next we compared mass AF, cloned F, and mass dermal fibroblasts during time in subculture by measuring hCG with the specific radioimmunoassay employing antiserum against the f~subunit to control for cross-reactivities with other glycoprotein hormones, particularly luteinizing hormone and free a-subunits (10). Hormone levels increase with time in culture of AF cells but similar increases are not present for F cells or dermal fibroblasts (Fig. 2). Then we studied culture media from a number of different confluent cultures in 25- or 75-cm2 flasks using the hCG /Jsubunit assay just described. The purpose was to look for variability between different AF mass
CHORIONIC GONADOTROPIN PRODUCTION
145
TABLE 1 SINGLE SPECIMEN COLLECTIONS FOR HCG ASSAYS ON CULTURE MEDIA (MEANS AND STANDARD D EVIATIONS ARE N OTED W HEN APPROPRIATE} Cultures a
Mass AF 739 742 744 747 749 753 755
PrimaryAF 749 765 772 777 780 781 783 785 786
Mass Dermal Fibroblasts 135 (fetal} 748 ~child)
SpecifichCG/~-Subunit Assay
Intact hCG Assay
m l U / ml
mlU/mg protein b
m l U / cm ~c
m l U / ml
98 44 52 72 95 48 55 66•
294 157 197 335 335 160 212 241•
11.8 5.3 6.2 19.2 11.4 5.8 6.6 9.5•
90 52 60 70 -52 55 63•
190 135 120 65 105 110 108 140 155 125•
2262 ---------
50.7 36.0 32.0 17.3 28.0 29.3 28.8 37.3 41.3 33.4•
210 132 102 76 115 116 119 129 168 130•
5 5
30 27
1.3 1.3
7 7
F 452 (cloned} 9 42 2.4 8 743 ( mass} 9 35 2.4 7 a Different specimens are identified by culture number. All cultures have 46,XX or 46,XY chromosome composition. Mass refers to subcultures that are not cloned. All amniotic fluid specimens for culture were obtained for genetic indications--either advanced maternal age or family member with a chromosome abnormality. Amniocenteses were performed between 15 and 18 weeks of pregnancy after signed informed consent. b Lowry protein assay of cell monolayers. c Surface area available for cell growth.
cultures as well as between culture types. A F culture media, particularly before subculture, consistently show the highest levels of hormone (Table 1). H o r m o n e levels in media from F and dermal fibroblast cultures are near the the lower limit of sensitivity of the assay. Radioimmunoassays of media from these AF, F and dermal fibroblast cultures at confluency do not detect the presence of h u m a n follicle-stimulating hormone or thyroidstimulating hormone. The h C G assay employing antisera to intact h C G (11), with incomplete cross-reaction to luteinizing hormone and a- and /3-subunits of h C G {12), then was employed. H o r m o n e analysis by both intact h C G and specific h C G / 3 - s u b u n i t
radioimmunoassay yielded the results recorded in Table 1 and they agree quite well. In parallelism studies, a typical dilution curve of culture medium gives a line parallel to that of h C G standard and to highly purified h C G (Fig. 3) indicating production by A F cultures of a substance antigenically identical to intact h C G . A dilution curve of the A F cells homogenate, on the other hand, deviates from parallelism indicating the presence in the cell homogenate of cross-reacting substances not identical to intact hCG. Further studies of A F cell homogenates are in progress. In primary cultures of minced placental villi from an 18-week, morphologically and chromosomally normal therapeutic abortus, we have
146
PRIEST ET AL.
99 --4 -.3 90
-,2
80 -.I
~ 4c m ~ 2c
-i
LO 3125
623
125
23
12 5
25
50
I00
075
~5
3
~n~CR
50 200
H9
~t CULTURE REPARIO0 ATIONS
-2 -3
4O0 mlU t~CG
-4
rig, mIU orjJl Anhgenper Tube
FIG. 3. Dose-response curves for hCG standard iO--O), highly purified hCG, Cr-ll9 (A--A), AF culture medium (I~--D), and AF cell homogenate (O--O) in the specific hCG/3-subunit assay (see text). B/Bo x 100 = counts bound in the presence of labeled and unlabeled hCG; B0 : counts bound in the absence of unlabeled hCG {maximum binding). The scale at the right of the graph is the logit transform of percent bound. demonstrated hormone by specific hCG/3-subunit assay of culture medium. The levels are from 57 to 200 m I U per ml culture fluid collected over periods ranging between 2 to 4 weeks from several different primary cultures of the same placenta. Extraembryonic tissues from more abortuses of different ages will need to be studied for hCG production and correlations made to different cell classes in culture. DISCUSSION Successful in vitro growth of human placenta with production of chorionic gonadotropin by bioassay was reported as early as 1943 by Jones, Gey, and Gey (13) and 1948 by Stewart, Sano and Montgomery (14L Later investigators confirmed these earlier studies t15, 16L The first classification of cells in primary culture of abortus placenta used the terms "stromal," "syncytial" and "Langhans" (13, 14). The "Langhans" cell had a characteristic halo around the nucleus and was predominant in primary cultures of abortus placenta that produced chorionic gonadotropin by bioassay. Correlation of cultured cell morphology with histologic tissue morphology continues to be a problem in spite of advances in techniques that permit organ culture, cloning, subculture and longer maintenance. After careful examination of the original photographs (13, 14), we conclude that the older term "stromal" probably refers to what we label in this paper as F cells of fihroblastic origin whereas the older terms "syncytial" and "Langhans" probably are both included in
our cell class designated AF which is reported in this paper to produce hCG. In a primary colony of AF cells, as well as in subculture, we see both "syncytial" lmultinucleated) and "Langhanslike" Icytotrophoblast) cells (5). If our interpretation is correct, AF primary and subcultures are derived from trophoblast and can show morphologic characteristics of both cyto- and syncytiotrophoblast. This finding is not surprising since Hertig and Tao ~17~, from study of abortus placental villi in organ culture, concluded that syncytiotrophoblast is a mature (differentiated) form of cytotrophoblast. Lower levels of hormone in mass AF media after subculture probably are not due to progressive overgrowth by F cells. In our experience change of morphology from AF to F, due to F-cell overgrowth with prolonged in vitro life, is easy to detect in the minority of subcultures (about 15% of amniotic-fluid specimensL These latter cultures were excluded from study. The typical AF mass cultures selected for hormone assay undergo two but not more than five doublings from primary culture before phase I I I occurs. For this reason cloning is difficult. Suggested but not proved is the hypothesis that AF cells after subculture lose the differentiated function of hCG production. We have ruled out hCG production in fibroblastic IF) cells from culture of amniotic fluid, but we have not studied the epithelial-like ceils in primary amniotic-fluid cultures (4). These cells are resistant to subculture by trypsin, show more than one kind of morphology, and do not usually predominate in primary culture. Study of epitheliallike cells from second-trimester amnlotic fluids is an important area for future investigation. Possible tissue sources include fetal urine and skin as well as differentiated amnion. There are reports of glycoprotein hormone in human cell cultures from sources other than placenta and fetal membrane I18-22L In these instances the tissue source is neoplastic, or an inducing agent such as sodium butyrate is added to the cultures. Neither situation is a prerequisite for hCG production by AF cultures of amniotic fluid reported here. REFERENCES 1. Cox, D. M., V. Niewczas-Late, M. I. Riffell, and J. L. Hamerton. 1974. Chromosomal mosaicism in diagnostic amniotic fluid cultures. Pediatr. Res. 8: 679-683. 2. Sutherland, G. R., S. M. Bowser-Riley, and A. D. Rain. 1975. Chromosomal mosaicism in amniotic
CHORIONIC GONADOTROPIN PRODUCTION fluid cell cultures. Clin. Genet. 7: 400-404. 3. Priest, R. E., J. H. Priest, J. F. Moinuddin, and A . J . Keyser. 1977. Differentiation in human amniotic fluid cell cultures. I. Collagen production. J. Med. Genet. 14: 157-162. 4. Hoehn, H., E. M. Bryant, L. E. Karp, and G. M. Martin. 1974. Cultivated cells from diagnostic amniocentesis in second trimester pregnancies. I. Clonal morphology and growth potential. Pediatr. Res. 8: 746-754. 5. Priest, R. E., K. M. Marimuthu, and J. H. Priest. 1978. Differentiation in human amniotic fluid cell cultures: Ukrastructural features. Lab. Invest. 39: 106-109. 6. Megaw, J . M . , J . H . Priest, R . E . Priest, and L. D. Johnson. 1977. Differentiation in human amniotic fluid cell cultures. II. Secretion of an epithelial basement membrane glycoprotein. J. Med. Genet. 14: 163-167. 7. Priest, J. H. 1977. Medical Cytogenetics and Cell Culture. Lea and Febiger, Philadelphia, pp. 121-128. 8. Chen, T. R. 1977. In situ detection of mycoplasma contamination in cell cultures by fluorescent Hoechst 33258 stain. Exp. Cell Res. 104: 255-262. 9. Midgley, A. R., Jr. 1965. Radioimmunoassay: A method for human chorionic gonadotropin and human luteinizing hormone. Endocrinology 79: 10-18. 10. Vaitukaitis, J. L., G . D . Braunstein, and G. T. Ross. 1972. A radioimmunoassay which specifically measures human chorionic gonadotropin in the presence of human luteinizing hormone. Am. J. Obstet. Gynecol. 113: 751-758. 11. Franchimont, P. 1970. A study of the cross-reaction between human chorionic and pituitary luteinizing hormones. Eur. J. Clin. Invest. 1: 65-73. 12. Dattatreyamurty, B., A. R. Sheth, L. R. Joshi, and S. S. Rao. 1975. Changes in the ratio between serum and "specific" levels of human chor-
13.
14.
15. 16. 17. 18[ 19.
20.
21.
22.
147
ionic gonadotropin in different trimesters of pregnancy. Am. J. Obstet. Gynecol. 121: 300-305. Jones, G. E. S., G. O. Gey, and M. K. Gey. 1943. Hormone production by placental cells maintained in continuous culture. Bull. Johns Hopkills Hosp. 72: 26-38. Stewart, H . L . , M . E . Sano, and T . L . Montgomery. 1948. Hormone secretion by human placenta grown in tissue culture. J. Clin. Endocrinol. 8: 175-188. Thiede, H. A. 1960. Studies of the human trophoblast in tissue culture. Am. J. Obst. Gynec. 79: 636-647. Soma, H., R. L. Ehrmann, and A. T. Hertig. 1961. Human trophoblast in tissue culture. Obstet. Gynec. 18: 704-709. Hertig, A . T . 1968. Human Cytotrophoblast. Charles C. Thomas, Springfield, Ill., pp. 99-105. Ghosh, N. K., and R. P. Cox. 1976. Production of human chorionic gonadotropin in HeLa cell cultures. Nature 259: 416-417. Ghosh, N . K . , A. Rukenstein, and R. P. Cox. 1977. Induction of human choriogonadotropin in HeLa-cell cultures by aliphatic monocarboxylates and inhibitors of deoxyribonucleic acid synthesis. Biochem. J. 166: 265-274. Hussa, R. O. 1977. Studies on human chorionic gonadotropin secretion: Effects of EGTA, lanthanum, and the ionophore A23187. J. Clin. Endocrinol. Metab. 44: 520-529. Pattillo, R . A . , R . O . Httssa, M . T . Story, A. C. F. Ruckert, M. R. Shalaby, and R. F. Mattingly. 1977. Tumor antigen and human chorionic gonadotropin in CaSki cells: A new epidermal cervical cancer cell line. Science 196: 1456-1458. Chou, J. Y., J. C. Robinson, and S. Wang. 1977. Effects of sodium butyrate on synthesis of human chorionic gonadotropin in trophoblastic and nontrophoblastic tumours. Nature (London) 268: 543-544.
D I F F E R E N T I A T I O N IN H U M A N A M N I O T I C F L U I D C E L L C U L T U R E S : CHORIONIC GONADOTROPIN PRODUCTION 1 ROBERT E. PRIEST, ~JEAN H. PRIEST, JESSIE F. MOINUDDIN, ANDDEMETRIOS S. SGOUTAS
Department of Pathology and Laboratory Medicine and Department of Pediatrics, Division of Medical Genetics, Emory University, Atlanta, Georgia30322
SUMMARY Two of the distinguishable cell classes subcultured from human amniotic fluid were examined for their capability to produce human chorionic gonadotropin (hCG) as determined by radioimmunoassay. The class that predominates in most cultures used for prenatal genetic diagnosis, previously termed AF (for amniotic fluidl, secretes hCG into the culture medium. Dermal fibroblasts do not, nor does another type of cultured cell from amniotic fluid, previously termed F because of a resemblance to fibroblasts. Primary AF cultures produce more hCG than do subcultures. Evidence that this hormone is intact hCG is provided by its immunoreactivity with antisera raised against the/~-subunit and against the intact molecule of hCG. Furthermore, a dose-response curve for hormone in culture medium is parallel to that of highly purified intact hCG. It is postulated that AF cultures are derived from fetal membranes and retain properties of trophoblast.
Key words: human amniotic fluid; cell cultures; hCG (human chorionic gonadotropink radioimmunoassay; trophoblast. INTRODUCTION Interpretation of some prenatal genetic diagnoses is difficult because we lack information about origins and characteristics of the different cells in human amniotic-fluid cultures from second-trimester pregnancies. For instance, evaluation of chromosomal mosaicism in amniotic-fluid cultures is a problem when one does not know with certainty if the cells originate from the fetus, from extraembryonic fetal membranes, or from both (1, 2). Another problem of interpretation concerns studies of enzymes and proteins that are tissue-specific, for example, inherited disorders of collagen (3). Three distinguishable cell classes cultured from amniotic fluid were termed E for epithelial, AF for amniotic fluid, and F for fibroblast by Hoehn et al. (4). Only the AF and F cell classes can be subcuhured routinely by trypsin, and are the subject of this report. Since the AF class predomi~Research supported by Grand HD 11379 from the National Institutes of Health. 2To whom request for reprints should be addressed at the Department of Pathology and Laboratory Medicine, Emory University, Atlanta, Georgia 30322.
nates in primary and early subcultures of secondtrimester amniotic fluids, these cells presumably also predominate in most cultures used for prenatal genetic diagnosis. AF subcultures show cellular pleomorphism and multinucleation, a loose network appearance at confluency, and growth potential less than half that of human diploid fibroblasts {4, 5). On the other hand, F subcultures resemble typical fibroblast cultures from dermis and other connective tissues in every respect so far studied. The cells have a spindle shape and form parallel arrays at confluency. There is no tendency to become multinucleated and the growth potential is like that of other human diploid fetal fibroblasts {4, 5). Other differences between these two cell classes reported recently from this laboratory include: (a) production of a glycoprotein specific for epithelial basement membrane by AF but not F cultures {6); (b) extracellular material with ultrastructural characteristics of Type I {connective tissue) collagen fibers in F but not AF cultures {5); and (c) biochemical evidence for Type I collagen in F cultures and for a basement membrane collagen in AF cultures (3). 142
143
CHORIONIC GONADOTROPIN PRODUCTION The information accumulating about F cultures strongly indicates their connective-tissue origin, whereas a source for AF cultures that would account for all characteristics described thus far is trophoblast. In these studies we provide further evidence for this extraembryonic fetal origin of the predominant cell type used for prenatal genetic diagnosis. We show that A F cells produce human chorionic gonadotropin lhCG) in standard culture conditions and without inducing agents. MATERIALS AND METHODS
Cell culture. The routine establishment and maintenance of amniotic-fluid cultures have been described previously (3, 7). Culture conditions and descriptions for individual experiments are included in t h e legends for Figs. 1 and 2, and Table 1. Dulbecco and Vogt ID&V) medium was prepared in this laboratory from a standard recipe, filter-sterilized with 0.22-tam pore size, and stored at 5 ~ C not longer than 3 months. Glutamine, fetal bovine serum, and antibiotics (penicillin and streptomycin) were stored at - 1 5 ~ C and added to complete medium just before use with cells. Primary cultures were grown in 25% and subcultures in 15% fetal bovine serum. Antibiotics were included routinely in medium of primary cultures but not in subcultures. Microbial growth was monitored by a change from the normal clear appearance of culture medium, unusual pH or cell growth, and increase in chromosomal breaks and rearrangements in a diploid culture. Mycoplasma assay was run, as indicated, by the method of Chen (8) on cells subcultured to one-chamber slides tLab-Tek Products). Chromosome analysis routinely employed G-banding by trypsin and Giemsa, as well as Q-banding by quinacrine dihydrochloride stain and fluorescence microscopy. Radioimmunoassay. Amounts of hCG were determined in the culture medium with the exception of one series of experiments that utilized cell homogenates lsee Fig. 3). Initially, a nonspecific radioimmunoassay was used which measures hCG and luteinizing hormone equally well due to the presence of a common a-subunit and homologies in the amino-acid sequence of the/3-subunits of these hormones t9). In later experiments, a specific radioimmunoassay for determination of hCG was used which employs antisera raised against the/~-subunit of hCG (purchased from the Institute of Bioendocrinology, Montreal, Quebec, Canada). This assay measures intact hCG and the
30
o 20
Nuclei/Field
l
& % M
~
15 IO 5 O
I
I
i
5
mIU/ml
I
I
I
4
g
12 8 4
~
DAYS
FI6. 1. The numbers of nuclei per unit area, the percent of nuclei in multinucleated cells, and the production of hCG with time in subculture. Cells for study are subcultured from AF mass strain PR-683 (chromosomally normal, two doublings after first subculture, from amniotic fluid, 16-week pregnancy of a 36-year-old). Cells are inoculated onto replicate one-chamber slides (Lab-Tek Products, Division Miles Laboratories, Inc., Naperville, IlL ) in 4 ml D&V medium, 15% fetal bovine serum. Each day, for 5 days after inoculation, medium is decanted and stored at -15 ~ C prior to a nonspecific radioimmunoassay for hCG (see text). Cell monolayers are rinsed twice with phosphate-buffered saline, chambers are removed, and slides are fixed in 95% ethanol for 15 min and stained by Papanicolaou's method. Analysis of cell nuclei is performed at x430 magnification using a 10-mm square grid in the ocular. The height of each bar ltop panel) represents mean number of nuclei in 50 fields examined; the line at the top of each bar represents _+ one standard deviation. /3-subunit of hCG with cross-reactivity of 100% (10). We determined that the antiserum has crossreactivities of less than 2% with follicle-stimulating hormone, less than 5~ with luteinizing hormone, less than 0 . ] % with thyroid-stimulating hormone, and less than 0.1% with growth hormone. The antiserum does not cross-react with
144
PRIEST ET AL.
I AF" 20
FiDermal broblastI I0~ 106 ~
DAYS FIG. 2. The number of cells per flask, semilog scale ~O--O) and production of hCG (e--Q) with time in subculture of three cell types. Cells for study are subcultltred from: (a) AF mass strain PR-713 (chromosomally normal, two doublings after first subculture, from amniotic fluid, 15.5-week pregnancy of a 38-year-oldl; (bl F cloned strain from Hoelger Hoehn, University of Washington, Seattle qchromosomallynormal, 55 doublings from single cell, from amniotic fluid, 21-week pregnancy of a mother with previous Down syndrome child; and (c) mass fetal dermal fibroblast strain PR-135 (chromosomally normal, 17 doublings after first subculture, 16-week therapeutic abortusl. Six replicative 75cm~ culttrre flasks for each culture type are inoculated with about 5 • 10~ceils per flask in 10 ml D&V medium, 15% fetal bovine serum. On days 2, 3, 4, 6, 8 and 13, medium is removed and stored at -15 ~ C prior to specific hCG/~-subunit assay (see text). Cell layers are rinsed with phosphate-buffered saline, suspended with 0.25% trypsin at 37~ C for 30 rain, diluted further with phosphate-buffered saline, and inoculated into a hemacytometer for cell counts. free c~-subunits of hCG and is capable of measuring 0.8 ng hCG per ml of culture medium. The antiserum was calibrated against complete hCG by radioimmunoassay using the second International Standard of hCG (from the World Health Organization). A highly purified hCG Ibatch, CR-119, Pituitary Hormone Distribution Program, National Institutes of Health) with known concentrations of hCG also was assayed at several dilutions concurrently. In addition, a homologous hCG assay employing antiserum to intact hCG (11) was used in order to determine if the activity measured in the hCG f~-subunit assay was the intact hormone or free/3-subunits. This assay has incomplete crossreactivity with luteinizing hormone and with the a- and f~-subunits ofhCG ~12). In all of these radioimmunoassays, we compensated for effect of protein concentration in the reaction mixture by first lyophilizing the specimens and then bringing them up to initial volume with human male serum. Polyethylene glycol was used as a precipitant for ['2SI]labeled antigen-
antibody complex. All assays were done in triplicate, and the results were calculated from linear regression curves of logit B/B0 versus log concentration (see Fig. 3). Unless otherwise specified, materials for radioimmunoassays were purchased from Serono Laboratories, Inc., Boston, Massachusetts, or from Pharmacia Fine Chemicals, Piscataway, New Jersey.
RESULTS Initially, to document increase in muhinucleation with time in culture of AF cells and to determine if glycoprotein hormone was present in the culture medium, mass AF subcultures were inoculated into replicate slide culture chambers. These AF subcultures were identified by morphologic characteristics (4, 5) which are also summarized in the Introduction section. The number of nuclei per microscope field increases with time (Fig. 1, top panelL and the percent of nuclei found in multinucleated cells also increases ~Fig. 1, center panel). Hormone levels in culture medium rise with time in culture {Fig. 1, bottom panel) as determined by the nonspecific radioimmunoassay that measures hCG and luteinizing hormone equally well due to presence of identical asubunits and homologies in amino-acid sequence of the fJ-subunits in these hormones (9L The experiment shown in Fig. 1 was repeated once with similar results. In these experiments, the percent of nuclei found in multinucleated ceils cannot be determined accurately after day 3 because cells are too crowded on the slides. In other experiments when cells are inoculated at lower density initially, or grown on larger surface area, the percent of nuclei found in multinucleated ceils increases to a maximum of 30% and reaches a plateau by 5 to 7 days after subculture. Next we compared mass AF, cloned F, and mass dermal fibroblasts during time in subculture by measuring hCG with the specific radioimmunoassay employing antiserum against the f~subunit to control for cross-reactivities with other glycoprotein hormones, particularly luteinizing hormone and free a-subunits (10). Hormone levels increase with time in culture of AF cells but similar increases are not present for F cells or dermal fibroblasts (Fig. 2). Then we studied culture media from a number of different confluent cultures in 25- or 75-cm2 flasks using the hCG /Jsubunit assay just described. The purpose was to look for variability between different AF mass
CHORIONIC GONADOTROPIN PRODUCTION
145
TABLE 1 SINGLE SPECIMEN COLLECTIONS FOR HCG ASSAYS ON CULTURE MEDIA (MEANS AND STANDARD D EVIATIONS ARE N OTED W HEN APPROPRIATE} Cultures a
Mass AF 739 742 744 747 749 753 755
PrimaryAF 749 765 772 777 780 781 783 785 786
Mass Dermal Fibroblasts 135 (fetal} 748 ~child)
SpecifichCG/~-Subunit Assay
Intact hCG Assay
m l U / ml
mlU/mg protein b
m l U / cm ~c
m l U / ml
98 44 52 72 95 48 55 66•
294 157 197 335 335 160 212 241•
11.8 5.3 6.2 19.2 11.4 5.8 6.6 9.5•
90 52 60 70 -52 55 63•
190 135 120 65 105 110 108 140 155 125•
2262 ---------
50.7 36.0 32.0 17.3 28.0 29.3 28.8 37.3 41.3 33.4•
210 132 102 76 115 116 119 129 168 130•
5 5
30 27
1.3 1.3
7 7
F 452 (cloned} 9 42 2.4 8 743 ( mass} 9 35 2.4 7 a Different specimens are identified by culture number. All cultures have 46,XX or 46,XY chromosome composition. Mass refers to subcultures that are not cloned. All amniotic fluid specimens for culture were obtained for genetic indications--either advanced maternal age or family member with a chromosome abnormality. Amniocenteses were performed between 15 and 18 weeks of pregnancy after signed informed consent. b Lowry protein assay of cell monolayers. c Surface area available for cell growth.
cultures as well as between culture types. A F culture media, particularly before subculture, consistently show the highest levels of hormone (Table 1). H o r m o n e levels in media from F and dermal fibroblast cultures are near the the lower limit of sensitivity of the assay. Radioimmunoassays of media from these AF, F and dermal fibroblast cultures at confluency do not detect the presence of h u m a n follicle-stimulating hormone or thyroidstimulating hormone. The h C G assay employing antisera to intact h C G (11), with incomplete cross-reaction to luteinizing hormone and a- and /3-subunits of h C G {12), then was employed. H o r m o n e analysis by both intact h C G and specific h C G / 3 - s u b u n i t
radioimmunoassay yielded the results recorded in Table 1 and they agree quite well. In parallelism studies, a typical dilution curve of culture medium gives a line parallel to that of h C G standard and to highly purified h C G (Fig. 3) indicating production by A F cultures of a substance antigenically identical to intact h C G . A dilution curve of the A F cells homogenate, on the other hand, deviates from parallelism indicating the presence in the cell homogenate of cross-reacting substances not identical to intact hCG. Further studies of A F cell homogenates are in progress. In primary cultures of minced placental villi from an 18-week, morphologically and chromosomally normal therapeutic abortus, we have
146
PRIEST ET AL.
99 --4 -.3 90
-,2
80 -.I
~ 4c m ~ 2c
-i
LO 3125
623
125
23
12 5
25
50
I00
075
~5
3
~n~CR
50 200
H9
~t CULTURE REPARIO0 ATIONS
-2 -3
4O0 mlU t~CG
-4
rig, mIU orjJl Anhgenper Tube
FIG. 3. Dose-response curves for hCG standard iO--O), highly purified hCG, Cr-ll9 (A--A), AF culture medium (I~--D), and AF cell homogenate (O--O) in the specific hCG/3-subunit assay (see text). B/Bo x 100 = counts bound in the presence of labeled and unlabeled hCG; B0 : counts bound in the absence of unlabeled hCG {maximum binding). The scale at the right of the graph is the logit transform of percent bound. demonstrated hormone by specific hCG/3-subunit assay of culture medium. The levels are from 57 to 200 m I U per ml culture fluid collected over periods ranging between 2 to 4 weeks from several different primary cultures of the same placenta. Extraembryonic tissues from more abortuses of different ages will need to be studied for hCG production and correlations made to different cell classes in culture. DISCUSSION Successful in vitro growth of human placenta with production of chorionic gonadotropin by bioassay was reported as early as 1943 by Jones, Gey, and Gey (13) and 1948 by Stewart, Sano and Montgomery (14L Later investigators confirmed these earlier studies t15, 16L The first classification of cells in primary culture of abortus placenta used the terms "stromal," "syncytial" and "Langhans" (13, 14). The "Langhans" cell had a characteristic halo around the nucleus and was predominant in primary cultures of abortus placenta that produced chorionic gonadotropin by bioassay. Correlation of cultured cell morphology with histologic tissue morphology continues to be a problem in spite of advances in techniques that permit organ culture, cloning, subculture and longer maintenance. After careful examination of the original photographs (13, 14), we conclude that the older term "stromal" probably refers to what we label in this paper as F cells of fihroblastic origin whereas the older terms "syncytial" and "Langhans" probably are both included in
our cell class designated AF which is reported in this paper to produce hCG. In a primary colony of AF cells, as well as in subculture, we see both "syncytial" lmultinucleated) and "Langhanslike" Icytotrophoblast) cells (5). If our interpretation is correct, AF primary and subcultures are derived from trophoblast and can show morphologic characteristics of both cyto- and syncytiotrophoblast. This finding is not surprising since Hertig and Tao ~17~, from study of abortus placental villi in organ culture, concluded that syncytiotrophoblast is a mature (differentiated) form of cytotrophoblast. Lower levels of hormone in mass AF media after subculture probably are not due to progressive overgrowth by F cells. In our experience change of morphology from AF to F, due to F-cell overgrowth with prolonged in vitro life, is easy to detect in the minority of subcultures (about 15% of amniotic-fluid specimensL These latter cultures were excluded from study. The typical AF mass cultures selected for hormone assay undergo two but not more than five doublings from primary culture before phase I I I occurs. For this reason cloning is difficult. Suggested but not proved is the hypothesis that AF cells after subculture lose the differentiated function of hCG production. We have ruled out hCG production in fibroblastic IF) cells from culture of amniotic fluid, but we have not studied the epithelial-like ceils in primary amniotic-fluid cultures (4). These cells are resistant to subculture by trypsin, show more than one kind of morphology, and do not usually predominate in primary culture. Study of epitheliallike cells from second-trimester amnlotic fluids is an important area for future investigation. Possible tissue sources include fetal urine and skin as well as differentiated amnion. There are reports of glycoprotein hormone in human cell cultures from sources other than placenta and fetal membrane I18-22L In these instances the tissue source is neoplastic, or an inducing agent such as sodium butyrate is added to the cultures. Neither situation is a prerequisite for hCG production by AF cultures of amniotic fluid reported here. REFERENCES 1. Cox, D. M., V. Niewczas-Late, M. I. Riffell, and J. L. Hamerton. 1974. Chromosomal mosaicism in diagnostic amniotic fluid cultures. Pediatr. Res. 8: 679-683. 2. Sutherland, G. R., S. M. Bowser-Riley, and A. D. Rain. 1975. Chromosomal mosaicism in amniotic
CHORIONIC GONADOTROPIN PRODUCTION fluid cell cultures. Clin. Genet. 7: 400-404. 3. Priest, R. E., J. H. Priest, J. F. Moinuddin, and A . J . Keyser. 1977. Differentiation in human amniotic fluid cell cultures. I. Collagen production. J. Med. Genet. 14: 157-162. 4. Hoehn, H., E. M. Bryant, L. E. Karp, and G. M. Martin. 1974. Cultivated cells from diagnostic amniocentesis in second trimester pregnancies. I. Clonal morphology and growth potential. Pediatr. Res. 8: 746-754. 5. Priest, R. E., K. M. Marimuthu, and J. H. Priest. 1978. Differentiation in human amniotic fluid cell cultures: Ukrastructural features. Lab. Invest. 39: 106-109. 6. Megaw, J . M . , J . H . Priest, R . E . Priest, and L. D. Johnson. 1977. Differentiation in human amniotic fluid cell cultures. II. Secretion of an epithelial basement membrane glycoprotein. J. Med. Genet. 14: 163-167. 7. Priest, J. H. 1977. Medical Cytogenetics and Cell Culture. Lea and Febiger, Philadelphia, pp. 121-128. 8. Chen, T. R. 1977. In situ detection of mycoplasma contamination in cell cultures by fluorescent Hoechst 33258 stain. Exp. Cell Res. 104: 255-262. 9. Midgley, A. R., Jr. 1965. Radioimmunoassay: A method for human chorionic gonadotropin and human luteinizing hormone. Endocrinology 79: 10-18. 10. Vaitukaitis, J. L., G . D . Braunstein, and G. T. Ross. 1972. A radioimmunoassay which specifically measures human chorionic gonadotropin in the presence of human luteinizing hormone. Am. J. Obstet. Gynecol. 113: 751-758. 11. Franchimont, P. 1970. A study of the cross-reaction between human chorionic and pituitary luteinizing hormones. Eur. J. Clin. Invest. 1: 65-73. 12. Dattatreyamurty, B., A. R. Sheth, L. R. Joshi, and S. S. Rao. 1975. Changes in the ratio between serum and "specific" levels of human chor-
13.
14.
15. 16. 17. 18[ 19.
20.
21.
22.
147
ionic gonadotropin in different trimesters of pregnancy. Am. J. Obstet. Gynecol. 121: 300-305. Jones, G. E. S., G. O. Gey, and M. K. Gey. 1943. Hormone production by placental cells maintained in continuous culture. Bull. Johns Hopkills Hosp. 72: 26-38. Stewart, H . L . , M . E . Sano, and T . L . Montgomery. 1948. Hormone secretion by human placenta grown in tissue culture. J. Clin. Endocrinol. 8: 175-188. Thiede, H. A. 1960. Studies of the human trophoblast in tissue culture. Am. J. Obst. Gynec. 79: 636-647. Soma, H., R. L. Ehrmann, and A. T. Hertig. 1961. Human trophoblast in tissue culture. Obstet. Gynec. 18: 704-709. Hertig, A . T . 1968. Human Cytotrophoblast. Charles C. Thomas, Springfield, Ill., pp. 99-105. Ghosh, N. K., and R. P. Cox. 1976. Production of human chorionic gonadotropin in HeLa cell cultures. Nature 259: 416-417. Ghosh, N . K . , A. Rukenstein, and R. P. Cox. 1977. Induction of human choriogonadotropin in HeLa-cell cultures by aliphatic monocarboxylates and inhibitors of deoxyribonucleic acid synthesis. Biochem. J. 166: 265-274. Hussa, R. O. 1977. Studies on human chorionic gonadotropin secretion: Effects of EGTA, lanthanum, and the ionophore A23187. J. Clin. Endocrinol. Metab. 44: 520-529. Pattillo, R . A . , R . O . Httssa, M . T . Story, A. C. F. Ruckert, M. R. Shalaby, and R. F. Mattingly. 1977. Tumor antigen and human chorionic gonadotropin in CaSki cells: A new epidermal cervical cancer cell line. Science 196: 1456-1458. Chou, J. Y., J. C. Robinson, and S. Wang. 1977. Effects of sodium butyrate on synthesis of human chorionic gonadotropin in trophoblastic and nontrophoblastic tumours. Nature (London) 268: 543-544.
