IN VITRO Volume14, No. 7, 1978 All rightsreserved 9

ISOLATION OF FUNCTIONAL HUMAN TROPHOBLAST AND THEIR PARTIAL CHARACTERIZATION IN PRIMARY CELL CULTURE

CELLS

KURT STROMBERG, J. C. AZIZKHAN, ANDK. V. SPEEG, JR. Viral Pathology Section, Laboratory of Viral Carcinogenesis, Division of Cancer Cause and Prevention, National Cancer Institute, Bethesda, Maryland 20014

SUMMARY Human trophoblast isolation and cell culture procedures were examined to identify variables that enhance secretion of chorionic gonadotropin (HCG) in primary culture. Brief exposure of unminced first-trimester placental specimens to a solution of trypsin-EDTADNAse, and isolation of the dispersed cells after Ficoll-hypaque centrifugation yielded primary cultures that were high in HCG secretion and content of epithelial-like cells. The gradual decline in HCG level with time in monolayer culture in these presumptive trophoblast ceils was retarded by treatment with theophylline and cyclic adenosine monophosphate. Exposure to methotrexate {MTX) did not increase HCG secretion in normal trophoblast cells, in contrast to a 5-fold stimulation by M T X in the JAR line of choriocarcinoma cells. Clusters of polygonal cells in primary culture progressively lost their capacity to secrete HCG and their epithelial-like morphology. However, they could be maintained as cell strains through approximately 15 passages over a period of 13 to 16 weeks. K e y words: human trophoblast; monolayer culture; chorionic gonadotropin.

INTRODUCTION Efforts to grow normal human trophoblast epithelium in vitro that retains prolonged capacity to secrete human chorionic gonadotropin {HCG) have met with limited success. Chung et al. (1) used a complex system that included human placental extract, a precisely controlled culture environment, and a rocker-platform method to maintain HCG-producing placental explants in organ culture for 4 months. A principal limitation of organ culture, in comparison to use of dispersed cells, has been that replicate samples cannot be readily obtained. Monolayer growth of normal human trophoblast in cell culture has been even less successful, with measurable HCG secretion continuing for only several days I2). If has not been possible to subculture normal trophoblast and maintain cell lines derived from it. Consequently, it has been difficult to study the modulation of HCG expression in normal primary trophoblast cells. In an effort to compare the factors that influence HCG expression in normal trophoblast with those shown to be important in the BeWo line of

choriocarcinoma cells (3), a simple, reliable procedure was developed to obtain functional human trophoblast cells for primary cell culture from first-trimester placental specimens. This report documents the features of the technique. Steps are described to limit yeast contamination of placental cell cultures. Use of unminced placental tissue and inclusion of DNAse in the trypsin-EDTA solution are used to obtain a high yield of trophoblast cells. In addition, the duration of trypsinization and the choice of media are assessed with respect to the amount of HCG released into the culture fluid. The types and morphology of the cells of the primary cultures are categorized by phasecontrast microscopy. The functional responsiveness of the primary placental cells is indicated by their enhanced HCG production following the addition of cyclic adenosine monophosphate (cAMP) and theophylline. However, treatment with methotrexate IMTX) failed to increase HCG secretion in normal primary placental cells, whereas M T X treatment of the JAR line (4) of choriocarcinoma cells caused a 5-fold stimulation of HCG secretion. 631

632

STROMBERG, AZIZKHAN, AND SPEEG MATERIALS AND METHODS

Primary culture. All samples consisted of firsttrimester (8 to 12 weeks gestation} human placental specimens obtained after elective abortion by vacuum aspiration. Samples were selected for tissue integrity, small size (measuring approximately 3 by 3 by 4 cm in fluid suspensiont, and absence of clinical monilial vaginitis in the patient. The placental tissue was immediately immersed in a solution of sterile ice-cold Gey's balanced salt solution (BSS) in a screw-topped 8-oz specimen jar. The jar was shaken vigorously, and the tissue was drained and then transferred to a second j a r also containing Gey's BSS and supplemented with penicillin (100 U per m|L and streptomycin (100 #g per ml}, and gentamicin (50 #g per mD. This transfer and rinsing procedure was repeated three times. For transport to the laboratory the placental sample was placed in a jar containing ice-cold medium 199 with 20% fetal bovine serum (North American Biological Co., Miami, Fla.} and penicillin, streptomycin and gentamicin (as listed above), as well as amphotericin B t0.25/~g per ml} and mycostatin il00 U per ml~ to reduce the incidence of subsequent contamination with Candida albicans, an organism commonly present in the vaginal tract. The excess fibrous membrane tissue was excised under aseptic conditions in a finger bowl containing 0.2% trypsin (200 U per mg; Worthington Biochemical Co.} and 0.02% E D T A in Dulbecco's PBS without calcium or magnesium. The remaining stalk of the placental specimen was attached to a No. 10 snap swival (Mid-Atlantic Distributors, Columbia, Md. } on the end of a string. The placental sample then was suspended in a 150-ml Erlenmeyer flask, with a side-arm spout, containing 75 ml trypsin-EDTA solution and 1.5 ml bovine pancreatic DNAse I ~2200 Kurtz units per ml; Sigma Chemical Co., St. Louis, MoA in Dulbecco's PBS (5} with a Tefloncoated magnetic stirring bar. Generally three to five placental specimens were prepared as described and placed on a multimagnestir I Lab-line, Melrose Park, Ill.}. The dial of the stirrer was set on maximum so that when activated, the magnetic stirring bar kicked about the flask to agitate and massage the suspended placental specimen. Successive periods t0 to 5 rain, 5 min to 10 min, and 10 min to 20 rain} of exposure to the trypsinE D T A - D N A s e solution at 37 ~ C were carried out. Following each time period, the fluid from each flask was poured in equal amounts into two

50-ml plastic tubes (Falcon Plastics) containing 10 ml medium 199 with 20% fetal bovine serum. The fragments of tissue that rapidly settled to the bottom of the tube were removed by aspiration with a 9-inch pasteur pipette. The cells released into the fluid following exposure to the trypsinE D T A - D N A s e solution were collected by centrifugation ~400 x g for 6 min at room temperature}. The cells from resuspended pellets from each placental sample then were counted by a hemacytometer and assessed for viability by trypan blue exclusion. Specimens of placental cells that were free of gross yeast contamination by microscopic examination (over 95%) and contained over 2 x 106 trypan blue-negative, non-RBC cells were combined. Routinely, the concentration of dispersed cells was adjusted to 1.0 x 106 per ml and 4 ml was layered over 3 ml Ficoll-hypaque solution of 1.077 g per ml density (6) l"Lymphoprep," Nyegaard and Co., Oslo, Norway~ in a 15- by 150-mm screw-topped tube (Falcon No. 3033}. Following centrifugation at room temperature ~400 x g for 20 min}, the cell layer at the interface was withdrawn, diluted 5-fold with culture medium, and recentrifuged (300 x g for 5 min~ to a pellet. After one more rinsing step, aliquots (5 x 105 cells per ml} of these cells were plated into 35-mm Petri dishes to obtain replicate samples for subsequent evaluation. Unless otherwise stated, the medium used for cell culture was Waymouth's 752/1 and Gey's BSS in a 55/45 ratio with 20% fetal bovine serum. The usual duration of time between acquiring the placental specimen and plating the dispersed primary placental cells was 3 to 4 hr. After an overnight attachment period (approximately 18 hrL the dishes were gently rinsed twice and refed with medium. Subsequently, medium was changed every 24 hr. After cell debris was removed by eentrifugation, the used medium was frozen at - 7 0 ~ C for subsequent radioimmunoprecipitation assay for HCG t3L Each data point represents the average HCG value of two 35-mm dishes assayed in duplicate. Each separate experiment was carried out two or three times with comparable results. The tissue fragments remaining after the various periods of trypsinization were routinely fixed, paraffin-embedded, sectioned and stained with hematoxylin and eosin. They were microscopically examined to assess the ability of various trypsin concentrations and exposure times to denude the trophoblastic epithelium from the underlying basement membrane of the placental villi.

633

TROPHOBLAST CELLS IN PRIMARY CULTURE Growth o ] J A R cells. JAR cells (4) were a generous gift of Dr. R. A. Pattillo, Medical College of Wisconsin, Milwaukee, Wisconsin. They were maintained as monolayer cultures in 50% Waymouth's 752/1, 40% Gey's BSS, and 10% fetal bovine serum in an atmosphere of 5% CO2-95% air. The experimental protocol used for this cell line is identical to that previously described in detail for the BeWo line {3). Chemicals. Methotrexate (MTX) as a sodium salt was purchased from Lederle Laboratories, Division of American Cyanamid Co., Pearl River, New York. Theophylline, adenosine 3',5'monophosphate (cAMP), and thymidine were from Sigma Chemical Co., St. Louis, Missouri. Media. All cell culture media were purchased from Grand Island Biological Co., Grand Island, New York, and included medium 199, Waymouth's 752/1, Gey's BSS, NCTC 109, R P M I 1640, C M R L 1066, McCoy's 5A, and Ham's F10. EXPERIMENTAL RESULTS The HCG secretion by primary placental cells was compared in seven nutrient rich media (medium 199, Waymouth's 752/1, NCTC 109, R P M I 1640, C M R L 1066, McCoy's 5A, Ham's F-10) with concentrations of fetal bovine serum from 2% to 20%. There were no significant differences in the secretion of HCG from cells in primary cell culture in these different media. Only reduction of the fetal bovine serum component below 2% resulted in an accelerated decline in HCG production ~data not shown}. Consequently, because of its use in prior studies of the BeWo line ~3, 7) and the JAR line of choriocarcinoma cells (4), Waymouth's 752/1 and Gey's BSS in a 55/45 ratio with 20% fetal bovine serum was routinely employed in this study. The use of centrifugation through a layer of Ficoll-hypaque to remove red blood cells, white blood cell granulocytes and yeast forms from the trophoblast cells did result in a slower decline in HCG values per day. This step also greatly reduced the frequency of yeast contamination in the primary cell cultures. The data in Fig. 1 demonstrate that longer periods of trypsinization reduce the level of HCG production. When the 0- to 5-rain period is compared with the 10- to 20-min period, the reduction per dish over the first 3 days is approximately 5fold. This fall in HCG secretion from primary placental cells is related to the successive effacement of the trophoblastic epithelium from the supporting stroma, as shown by microscopy of H

& E stained, paraffin-embedded sections (Fig. 2 A - D . After a 20-min exposure to the trypsinE D T A - D N A s e solution, the epithelial surface of the chorionic villi is denuded (compare Fig. 2,4 with 2D). The addition of DNAse to the trypsinE D T A solution is used to reduce the formation of a clear, viscous material which surrounds the suspended placental specimen and prevents the removal of dispersed cells after each trypsinizing period. The conventional method of preparing human primary placental cells for monolayer culture involves mincing the dissected chorionic villi and straining the minced tissue through gauze to remove tissue fragments {2, 8-11). A comparison between placental cells derived from the portions of minced chorionic villi and companion portions of unminced villi from the same specimens revealed that a nearly 2-fold higher set of HCG values between day 1 and day 4 was obtained in cells isolated from the unminced portion as described in the Materials and Methods section (data not shownL

80(]

700

60(]

500

4~0

300

200

100

0

I

1

I 2

I 4

I

3

I 5

7 6

DAYS

Fie. 1. Effect of duration of trypsinization on levels of HCG secretion. Placental cells were isolated and plated into three duplicate sets of 35-mm Falcon Petri dishes, and daily culture supernates assayed in duplicate for HCG content, as described in Materials and Methods. During isolation the length of time in the trypsinEDTA-DNAse solution varied from 0 to 5 rain (e); 5 to 10 min (O); or 10 to 20 min (IlL

634

STROMBERG, AZIZKHAN, AND SPEEG

FI6. 2. Histologic appearance of unminced first-trimester human placental villi after different lengths of exposure to trypsin-EDTA-DNAse solution (see Materials and Methods). Fragments of villi before treatment (A, x55) are compared with villi after 0 to 5 min (B, x140), 5 to 10 rain (C, x140), and at the end of 20 min of exposure (D, x55). Note the progressive loss of trophoblast epithelium.

635

TROPHOBLAST CELLS IN PRIMARY CULTURE A. 4

A.

JAR CHORIOCARCINOMA LINE 4

B

3

3

2

2

1

I

0

0

1.0

.0

8.

PRIMARY TROPHOBLAST CELLS

1.8

0.6

.6

0.4

.4

0.2

.2

1

2

3

4

5

6

DAYS

FIG. 3. Stimulation of HCG secretion by exposure to cyclic adenosine monophosphate (cAMP) and theophylline. Duplicate sets of 35-mm Petri dishes, containing either 1 x 105 ceils of the JAR line of ehoriocarcinoma (AL or 5 x l0 s primary placental cells isolated as described in Methods and Materials (B), were fed untreated media as a control (e ~or fed media containing 1 mM each cAMP and theophylline ( 9 during the first 24 hr (black bar). After rinsing each set once, the dishes were daily given 2 ml fresh medium, and the HCG content in the conditioned medium was determined by radioimmunoassay (3). Fig. 3 and 4 compare the effect of c A M P and theophylline, and M T X , respectively, on H C G secretion by primary placental cells and the J A R line of choriocarcinoma cells. Exposure to an optimal concentration t 1 mM~ of c A M P and theophylline results in a 3- to 4-fold elevation in H C G secretion in J A R line ceils, and similarly retards the decline of H C G production in primary placental cells (Fig. 3). No H C G response was noted in the primary trophoblasts with a treatment of 0.1 mM, and 10 mM was toxic to the ceils. T h e exposure of J A R cells to M T X results in a 5-fold increase in H C G secretion. Note that the H C G stimulation is blocked by simultaneous addition of M T X and excess thymidine (Fig. 4). In contrast, the addition of M T X over a broad concentration range to normal placental cells does not af-

m

S.

D

0.8

0

JAR CHORIOCARCINOMA LINE

0

PRIMARY TROPHOBLAST CELLS

I

2

3

4

5

6

DAYS

FIG. 4. Effect of methotrexate (MTX) treatment on normal and transformed placental cells. Duplicate sets of 35-mm Petri dishes, containing either 1 x l0 s cells of the JAR line of choriocarcinoma (A), or 5 x l0 s primary placental cells isolated as described in Materials and Methods (B), were fed during the first 24 hr (black bar} media that contained the following: MTX, 10-' M (ll); MTX, 10-6 M(A); MTX, 10.5 M(A); MTX, 10-6 M, and thymidine, 10-5 M(D); thymidine, 10.5 M(nl); and, as a control, no additional agents (S). Only Waymouth's 752/1 and Gey's BSS medium containing 2% fetal bovine serum was used throughout the course of this experiment. After the initial 24-hr exposure each dish was rinsed and given 2 ml fresh medium, and the HCG content in the conditioned medium was determined after every 24 hr by radioimmunoassay (3).

feet the decay curve of H C G activity with time in primary placental cell culture. During the first week of primary monolayer culture, three broad categories of cell morphology are apparent. First, the most abundant type of attached cell is a medium-sized polygonal cell with a large clear nucleus containing one to three nucleoli and numerous mitochondria (Fig. 5.4 ~. T h e second category of cells consists of larger rounded cells often containing 10 to 15 nuclei with smooth cell borders (12). Third, and infrequently, spindle-shaped cells with very few mitochondria are observed. T h e proportion of the second category declines over the first week in primary cul-

636

STROMBERG, AZIZKHAN, AND SPEEG ture, whereas the cells of the third category proliferate to loose clusters of spindle-shaped cells. By the end of the second week in primary culture, cells of the first category contain several nuclei with prominent nucleoli, reduced numbers of mitochondria, and have irregular cell borders with cytoplasmic projections (Fig. 5B). Sequential Polaroid phase-contrast photomicroscopy of the first category suggests that these polygonal epithelial-like cells tend to become multinucleated by endomitosis and not through fusion with adjacent cells. In agreement with prior morphological studies t2, 10, 13), these three categories may represent, respectively, cytotrophoblast, syncytiotrophoblast, and fibroblast cells. Repeated attempts with cloning cylinders to subculture islands of placental cells with the first morphologic category of polygonal cells yielded populations with predominantly spindle-shaped morphology after several passages. Ten such initially epithelial-like clusters were serially subcultured 1:2 in weekly passages for approximately 15 passages when senescence occurred in the form of increasingly high proportions of multinucleated cells that failed to proliferate (Fig. 5C). DISCUSSION This report describes a simple, rapid and reliable technique for isolation of primary placental cells from first-trimester human placental specimens. This tissue source represents perhaps the most readily available of all human epithelial cell sources with a well defined biochemical marker ~HCG). Because adjustments in parameters of the technique were compared so as to obtain increased HCG secretion, the procedure was designed to isolate maximal amounts of human trophoblast cells. That the procedure isolates functional trophoblast cells is indicated by the ability of exposure to cAMP and theophylline to retard the decline in HCG secretion. The daily reduction in HCG secretion into the media from placental cells in primary culture may

FIG. 5. Phase-contrast photomicrographs of cultured placental cells with initially epithelial-like morphology. Presumptive cytotrophoblast cells after 4 days in primary culture ~A) are compared to their appearance after 10 days in primary culture {B) when multinucleation and microtubular formation are evident. After 13 passages over 14 weeks, during which period the dominant morphology of the culture was fibroblastic, the morphologic pattern is one of senescent cells with few mitoehondria and multiple nuclei IC ). x118.

TROPHOBLAST CELLS IN PRIMARY CULTURE result from the absence of an essential growth factor, programmed senescence of trophoblast, or the necessity of syncytical integrity of the trophoblastic epithelium for prolonged HCG production in vitro. However, when comparing the initial 24 hr in cell culture ~day 0 to day 1), the amount of H C G produced by the normal trophoblasts on a per cell basis (14 x 10 -4 mIU HCG per cell) is equivalent to the J A R line of choriocarcinoma cells (12.5 x 10-4 mIU HCG per cell). The JAR cells have a generation time of approximately 28 hr and thus when plated at 1 x l0 s per dish after 24 hr contain 2 x l0 s cells that secrete 250 mIU of HCG (Figs. 3A, 4A). In comparison, the normal trophoblasts, plated at 5 x l0 s cells per dish, apparently fail to proliferate and secrete about 700 mIU HCG ~Figs. 1, 3B, 4B) over the first 24-hr period. This study did not attempt to extend HCG secretion by use of conditioned media, use of various lots of human placental cord serum, use of first-trimester placentas obtained by hysterotomy instead of vacuum aspiration, or plating isolated trophoblast cells at increased and perhaps more optimal concentrations. The usefulness of isolation of trophoblast cells over explants of chorionic villi is that ~a) replicate samples can he quantitatively obtained for studies of HCG modulation; ~b) isolated trophoblast cells are required for cell fusion in studies of the somatic cell genetics of normal trophoblast; and ~c) placental cell population enriched for trophoblast cell content opens the possibility of more precise isolation of the specific cell type or types responsible for HCG secretion by means of density centrifugation in Ficoll-hypaque or albumen gradients. While it is tempting to assign a histologic origin to the several categories of cells described in the primary placental cell cultures, cell morphology is not a reliable index in cell culture. Moreover, hesides the three major categories, placental cells of intermediate sizes and shapes were not infrequently seen. Consequently, though our observations are similar to previous reports (1, 2, 20, 12), we question the appropriateness of relating types of placental cells in culture to morphologic counterparts in intact tissue. We believe functional markers like HCG secretion define trophoblast cells. Modulation of H C G expression by metabolic factors that elevate levels of cAMP has been reported recently in both the JAR and the BeWo line of choriocarcinoma cells (14), and in organ cultures of normal human term placenta (15). In

637

addition to these examples, this report demonstrates that normal first-trimester placental cells in primary cell culture respond to exposure of theophylline and cAMP with a several-fold higher level of H C G synthesis when compared to the untreated control cells. The precise mechanism of this cAMP-induced increase in HCG synthesis is unknown but apparently requires R N A synthesis ~16). The 5-fold stimulation of H C G secretion following M T X exposure of the JAR line of choriocarcinoma cells contrasts with the absence of a M T X effect in normal placental cells in primary cell culture. A study of M T X stimulation of HCG synthesis in the BeWo line (3) and the JAR line (17) of choriocarcinoma cells related the M T X induced inhibition of D N A synthesis to enhanced H C G secretion. The lack of M T X stimulation in normal trophoblast cells in primary culture may simply reflect limited ability of normal trophoblast to proliferate in cell culture. In agreement with this interpretation, methotrexate has been reported to stimulate intercellular HCG in normal human trophoblast growing in the hamster cheek pouch (18). Alternatively, it is possible that there is a fundamental difference between normal and malignant trophoblasts in regard to the regulation of HCG synthesis. Enhanced stimulation of certain differentiated gene products may in fact be related to malignancy. For example, the level of phenylalanine hydroxylase enzyme after hydrocortisone treatment of cultured rat hepatoma cells was greater than that found in normal adult rat liver (19D. However, until growth factors (1) are identified that maintain growth of normal trophoblast cells with stable H C G secretion in monolayer culture, it will remain speculative as to which interpretation is appropriate. REFERENCES 1. Chung, H. K., W. F. McLimans, J. Horoszewicz, and M. M. Hreshchyshyn. 1969. In vitro studies of human trophoblast. Am. J. Obstet. Gynecol. 104: 945-952. 2. Taylor, P. V., and K. W. Hancock. 1973. Viability of human trophoblast in vitro. J. Obstet. Gynaecol. Br. Common. 80: 834-838. 3. Speeg, K. V., J. C. Azizkhan, and K. Stromberg. 1976. The stimulation by methotrexate of human chorionic gonadotropin and placental alkaline phosphatase in cultured choriocarcinoma cells. Cancer Res. 36: 4570-4576. 4. Pattillo, R. A., A. Ruchert, R. Hussa, R. Bernstein, and E. Dells. 1971. The JAR cell line--continuous human muhihormone production and controls. In Vitro 6: 398.

638

STROMBERG, AZIZKHAN, AND SPEEG

5. Salomon, D., and M. I. Sherman. 1975. The biosynthesis of progesterone by cultured mouse midgestation trophoblast cells. Dev. Biol. 47: 394-4O6. 6. Boyum, A. 1968. Separation of leucocytes from blood and bone marrow. Scand. J. Clin. Lab. Invest. Suppl. 21: 97. 7. Pattillo, R. A., and G. O. Gey. 1968. The establishment of a cell line of human hormone-synthesizing trophoblast cells in vitro. Cancer Res. 28: 1231-1237. 8. Theide, H. A., and J. H. Rudolph. 1961. A method for obtaining monolayer cultures of human letal cells from term placentas. Proc. Soc. Exp. Biol. Med. 107: 565-569. 9. Igarashi, A. 1964. On the identification of trypsinreleased human placental cells and their separation method. Tohoku J. Exp. Med. 83: 15-28. 10. Fox, H., and F. N. Khargongor. 1970. Morphology and enzyme histochemistry of cells derived from placental villi in tissue culture. J. Pathol. 101: 267-276. 11. Loke, Y. W., and R. Borland. 1970. Immunofluorescent localization of chorionic gonadotropin in monolayer cultures of human trophoblast cells. Nature 228: 561-562. 12. Foldes, J., T. Kehaty, and J. Schwartz. 1973. Interpretation of the muhinucleated giant cell in human trophoblast cultures. Gynecol. Invest. 4: 254-262. 13. Thielde, H. A. 1960. Studies of the human trophoblast in tissue culture. Am. J. Obstet. Gynecol. 79: 637-647.

14. Hussa, R.O., M . T . Story, R . A . Pattillo, and R. G. Kemp. 1977. Effect of cyclic 3':5'-AMP derivatives, prostaglandins, and related agents on human ehorionic gonadotropin secretion in human malignant trophoblast in culture. In Vitro 13: 443-449. 15. Handwerger, S., J. Barrett, L. Tyrey, and D. Schomberg. 1973. Differential effects of cyclic adenosine monophosphate on the secretion of human placental lactogen and human chorionic gonadotropin. J. Clin. Endoerinol. Metab. 36: 1268-1270. 16. Pattillo, R . A . , and R . O . Hussa. 1975. Early stimulation of human chorionie gonadotropin secretion by dibutyryl cyclic AMP and theophylline in human malignant trophoblast cells in vitro: Inhibition by actinomycin D, a-amanitin, and cordycepin. Gynecol. Invest. 6: 365-377. 17. Azizkhan, J. C., K. V. Speeg, K. Stromberg, and D. M. Goode. 1977. Stimulation of human thorionic gonadotropin synthesis in the JAR line of choriocarcinoma cells by inhibitors of DNA synthesis. Proc. Am. Assoc. Cancer Res. 18: 107. 18. Koide, Y., T. Aoki, and M. M. Hreshchyshyn. 1971. Effects of hormones, methotrexate, and dactinomycin on benign trophoblast. Am. J. Obstet. Gynecol. 109: 453-456. 19. Haggerty, D. F., G. Popjak, and P. L. Young. 1976. Properties of phenylalanine hydroxylase of cultured hepatoma cells. J. Biol. Chem. 251: 6901-6908.

We are grateful for the assistance of Denice Cora and Carlos Berry, and the support of Dr. J o h n W. Armstead, Clinical Director, and the staff of P R E TERM.