Expression of 17\g=a\-hydroxylaseand 3\g=b\-hydroxysteroid dehydrogenase in fetal human adrenocortical cells transfected with SV40 T antigen C. Y.

Cheng,

M. V. Flasch

and P.

J. Hornsby

Departments of *Biochemistry and Molecular Biology and tPhysiology College of Georgia, Augusta, Georgia 30912, U.S.A. (Requests for offprints should be addressed to P. J. Hornsby) received

and

Endocrinology, Medical

18 November 1991

ABSTRACT

by cyclic AMP, as evidenced by Northern blotting and by the conversion of progesterone or 25-hydroxy-[ 1,2- 3H]cholesterol, this induction being blocked by low concentrations of 12-0\x=req-\ tetradecanoylphorbol-13-acetate (TPA). Cholesterol side-chain cleavage enzyme was strongly induced by cyclic AMP, and clones also showed low activities of 21-hydroxylase and 11\g=b\-hydroxylase. Under all circumstances levels of 3\g=b\-hydroxysteroiddehydrogenase (3\g=b\-HSD),as assessed by Northern blotting or by conversion of 25-hydroxycholesterol, were very low. 3\g=b\-HSD was not induced by cyclic AMP or TPA alone, but was induced by the combination of the two agents. The regulation of 17\g=a\-hydroxylase and 3\g=b\-HSDresembles that previously described in primary cultures of human fetal adrenocortical cells. Thus, transfection with SV40 T antigen resulted in the production of clones which preserve the unique

Primary fetal human adrenocortical cells of definitive zone origin were transfected by electroporation with pSV3neo, a plasmid coding for SV40 T antigen and neo, which confers resistance to the antibiotic G418. The clones obtained proliferated for 30 to 40 population doublings after isolation when grown under standard medium conditions, and then entered 'crisis'. When early-passage clones were incubated with cyclic AMP (1:1 N6-monobutyryl and 8-bromo analogues), cell rounding was observed, as in primary cultures of human adrenocortical cells. As previously shown in bovine adrenocortical cells, rounding was inhibited with a monoclonal antibody against urokinase plasminogen activator but not with a monoclonal antibody against tissue plasminogen activator. The regulation of the steroidogenic pathway in clones was investigated. The effects of cyclic AMP and activation of protein kinase C were examined in cells maintained in defined medium or in the presence of serum. 17\g=a\-Hydroxylase was strongly

Journal of Molecular Endocrinology (1992) 9,

INTRODUCTION

bond

A long-term cell culture system of human fetal adrenocortical cells would facilitate investigation of the unique aspects of the molecular biology of the steroid 17a-hydroxylase and 3ß-hydroxysteroid dehydrogenase A -isomerase (3ß-HSD) genes in the human adrenal cortex. The high rate of synthesis of the adrenal androgen dehydroepiandrosterone (DHEA), characteristic of adrenocortical steroido¬ genesis in humans and other primates, is attributable to the level of expression of 17a-hydroxylase, a single cytochrome P-450 enzyme responsible both for 17a-hydroxylation and for cleavage of the C-17/C-20 '

induced

characteristics of the human adrenal

cortex.

7-17

(Zuber et al. 1986), relative to that of 3ß-HSD (Hornsby, 1985; 1989). Primary cultures of fetal human adrenocortical cells have played an important role in elucidating the regulation of the enzymes involved in steroid biosynthesis in the human adrenal cortex. However, their usefulness is curtailed by several factors: a lim¬ ited supply of tissue, a low total replicative capacity in culture and a rapid loss of the ability to induce a complete steroidogenic pathway when cells are subcultured beyond the primary culture. Although fetal human adrenocortical cells proliferate in culture, they senesce after about 40 population doublings, a replicative life span shorter than that of bovine

adrenocortical cells but sufficient to allow primary clones of human adrenocortical cells to be isolated (McAllister & Hornsby, 1987). Clones had a short remaining life span and showed very low or negli¬ gible levels of steroidogenesis under conditions which induce steroidogenesis in primary cells (C. Y. Cheng & P. J. Hornsby, unpublished observations). Based on our work in bovine adrenocortical cells, the loss of steroid hydroxylase induction in human adrenocortical cells in culture may result from senescence processes that alter the expression of these genes (Hornsby, 1991). The introduction of SV40 T antigen into bovine adrenocortical cells pro¬ duced extended life span clones (Cheng et al. 1989; Hornsby, 1991). Moreover, these cells had an elevated expression of 17a-hydroxylase compared with non-transfected clones. Thus, SV40 T antigen is not only compatible with the expression of differ¬ entiated functions but may retard the process that otherwise causes loss of 17a-hydroxylase induction. These data indicated that it would be of interest to examine the effects of T antigen on human adrenocortical cell steroidogenic gene expression. We show that T antigen-transfected clones can be derived from the fetal human adrenal cortex, and also that regulation of 3ß-HSD and 17a-hydroxylase in these clones by cyclic AMP and by protein kinase C resembles that in primary cultures.

Transfection and

growth of transfected cells

Primary cultures of fetal human adrenocortical cells transfected by electroporation as described by Cheng et al. (1989), using plasmid preparations pre¬ pared by pZ523 column (5 Prime—>3 Prime Inc., Boulder, CO, U.S.A.) and linearized with EcoRI. Cells were maintained in normal complete medium for 24 h before transfer into medium containing G418 (200mg/l; GIBCO/BRL, Gaithersburg, MD, U.S.A.). After 10 days incubation with G418, sur¬ viving cells were replated at low density. Subculturing was performed by incubation with Pronase in serum-containing medium. Clones were allowed to grow for 10 days, isolated with cloning rings, and grown up to an approximate area of 50 cm (equiva¬ lent to 25 population doublings). Clones were stud¬ ied within the next six passages (1:5 split ratio) beyond this point. For investigation of cell rounding, cultures were changed to defined, serum-free medium comprising DMEM/F-12 (1:1), bovine serum albumin (50 mg/1), recombinant human insulin-like growth fac¬ tor-I (10 nmol/1; Imcera, Terre Haute, IN, U.S.A.), ascorbic acid (2 mmol/1), selenite (20 nmol/1) and a-tocopherol (1 pmol/1), with optional addition of cyclic AMP analogues (N -monobutyryl cyclic AMP and 8-bromo cyclic AMP, Sigma), and monoclonal antibodies against human urokinase and human tissue plasminogen activator (Chemicon Inter¬ national Inc., El Segundo, CA, U.S.A.). were

MATERIALS AND METHODS

of primary human adrenocortical cell cultures

Preparation

Human adrenocortical cell cultures were prepared by collagenase/DNase digestion from the definitive zone of the fetal human adrenal gland ( 16 to 20 weeks of gestation), as described previously (Hornsby, 1980; Hornsby et al. 1983; Hornsby & Aldern, 1984; McAllister & Hornsby, 1987; Hornsby & McAllister, 1991). Primary cells were stored frozen in 5% dimethyl sulphoxide until required for experi¬ ments. Frozen cells were thawed and plated in cul¬ ture dishes coated with fibronectin (Hornsby & McAllister, 1991 ). Cells were grown in a 1: 1 mixture of Dulbecco's Modified Eagle's medium (DMEM) and Ham's F-12 medium (F-12) with 10% (v/v) fetal bovine serum (FBS; Irvine Scientific, Irvine, CA, U.S.A.), 10% (v/v) horse serum (HS; Sigma Chemical Co., St Louis, MO, U.S.A.), recombinant fibroblast growth factor (1 pg/1) and 2% (v/v) UltroSer G (IBF Inc., Savage, MD, U.S.A.). The gas phase was 5% 02, 90% N2 and 5% C02.

Induction and

analysis

of mRNA

Cultures were grown to confluence (10 cells) in 10 cm plates. The medium of the culture was changed to defined medium, described above, with optional addition of cyclic AMP analogues, 12-0tetradecanoylphorbol-13-acetate (TPA; Sigma), forskolin (Calbiochem Inc., La Jolla, CA, U.S.A.) or serum (10% (v/v) HS and 2% (v/v) FBS). The incu¬ bation was continued for 36 h prior to harvest of RNA. RNA was isolated using RNAzol (proprietary mixture of phenol, guanidinium isothiocyanate and ß-mercaptoethanol; Biotecx Laboratories, Houston, TX, U.S.A.) according to the manufacturer's proto¬ col. Northern blotting and hybridization methods were as previously described (Hornsby et al. 1987; Naseeruddin & Hornsby, 1990). Full-length cDNA clones for cholesterol side-chain cleavage enzyme (SCC; Chung et al. 1986) and 17a-hydroxy1ase (Chung et al. 1987) were generous gifts of W. Miller and S. Townsend, University of California, San Francisco, CA, U.S.A.; human placental 3ß-HSD cDNA (The et al. 1989) was from F. Labrie, Centre

de l'Université Laval, Quebec, Canada Superoxide dismutase cDNA (Sherman et al. 1983) was from Y. Groner, Weizmann Insti¬ tute, Tel Aviv, Israel. Rehybridization was per¬ formed after removal of probe by washing in water at 85 °C. Adult human adrenal cortex RNA was pre¬ pared using RNAzol from glands obtained through the Cooperative Human Tissue Network funded by the National Cancer Institute.

Hopitalier

and human

High performance liquid chromatography (HPLC) of steroid metabolites Cultures were grown to confluence (5 x 10' cells) in 35 mm plates. Cells were incubated under the con¬

ditions described for the induction of mRNA for 72 h. At that time, steroidogenic substrates were added in defined medium: progesterone (10pmol/l) for 2 h or 25-hydroxycholesterol (10pmol/l) with 25-hydroxy-[l ,2- HJcholesterol (1 mCi/1; Amer¬ sham Corp., Arlington Heights, IL, U.S.A.) for 18 h. The medium was extracted with dichloromethane and steroids were separated on a linear methanol : water gradient (20-100% methanol) or a step gradient, as previously described (Hornsby & Aldern, 1984). When 25-hydroxy-[l ,2-3H] cholesterol was used as substrate, total products (A' plus A ) were detected by flow-through scintillation counter (IN/US Corp., Fairfield, NJ, U.S.A.). In all HPLC separations, A steroid products were detected by absorbance at 240 nm. The identity of peaks was established by comparison with the reten¬ tion times of authentic standards. A major A" product was identified as 20,22,25-trihydroxycholesterol by gas chromatography—mass spectroscopy (Alsema et al. 1980). The SCC inhibitor

(20R)20-phenyl-5-pregnene-3ß,20-diol

erous

gift of

Irvine, CA,

L. Vickery, U.S.A.

was

a

gen¬

University of California,

RESULTS

Primary fetal human adrenocortical cells of defini¬ zone origin were transfected with plasmid pSV3neo containing the early region of the SV40 genome, coding for the oncogenic T antigen, and a gene for G418 resistance (Southern & Berg, 1982). The experiments reported here were performed on early passages of 12 different clones derived from primary cells in a transfection experiment using electroporation. Clones were also derived by transfecting the same plasmid using liposomes and by infecting cells with a defective retrovirus (Hornsby & Sal¬ mons, 1992). The plasmid pSV3neo and the retrovirus pZIPneoSV40Xl/T (Jat et al. 1986) both have tive

SV40 T antigen and neo expressed under the control of separate SV40 promoters. Clones derived by these three methods all had similar growth and functional properties. After isolation of clones, continuous pro¬ liferation was observed for 30 to 40 population doub¬ lings before the culture entered 'crisis', as observed for SV40 T antigen-transfected fibroblasts (Stein,

1985). Figure

1 illustrates a clone of T antigentransfected human adrenocortical cells after incuba¬ tion with cyclic AMP, showing the characteristic response of rounding or retraction (Neville & O'Hare, 1982; Hornsby et a!. 1983; Rainey et al. 1983; McAllister & Hornsby, 1987; Hornsby et al. 1989). For comparison, human adrenal gland fibro¬ blasts transfected with pSV3neo were observed under the same conditions; no morphological response to cyclic AMP was evident (not shown). Rounding was inhibited by a monoclonal antibody against urokinase plasminogen activator but not by a monoclonal antibody against tissue plasminogen activator. Cholera toxin (1 nmol/1) and forskolin (10pmol/l) also caused rounding. A few clones showed a rounding response to adrenocorticotrophic hormone (ACTH; 1 pmol/1), but most did not. In the following experiments, cells were treated with cyclic AMP analogues or forskolin rather than ACTH. 17a-Hydroxylase activity in clones was assayed by induction of the enzyme with cyclic AMP analogues followed by incubation with progesterone and separ¬ ation of the products by HPLC. Products of progest¬ erone in one clone are illustrated in Fig. 2a; all others gave similar results. In 2-h incubations, progester¬ one was not detectably metabolized in cultures not previously incubated with cyclic AMP. After cyclic AMP, progesterone was converted to 17a-hydroxyprogesterone and 11-deoxycortisol and, in longer incubations, cortisol. The clones studied all had 17a-hydroxylase activity, but the cyclic AMPinduced levels varied, as shown in Fig. 2b. The induction of cytochrome P-450 mRNAs in clones was studied. Bands of the expected sizes for the mRNAs of human SCC and 17a-hydroxylase were detected. All clones showed cyclic AMPinducible expression of 17a-hydroxylase and SCC; data for one clone are shown in Fig. 3. Superoxide dismutase mRNA levels were used as an internal control; cyclic AMP has no apparent effect on this mRNA. The regulation of 17cc-hydroxylase and 3ß-HSD by cyclic AMP and activation of protein kinase C was investigated by conversion of 25-hydroxycholesterol and by Northern blotting. Results were similar in two different clones.

1. The phase-contrast appearance of fetal human adrenocortical cells transfected plasmid encoding SV40 T antigen, showing that cyclic AMP-induced morphol¬ ogical changes in cells are urokinase plasminogen activator dependent. (A) Cells were incubated in defined medium for 24 h without further additions; (B) cells were incubated with N -monobutyryl cyclic AMP (1 mmol/1) and 8-bromo cyclic AMP (1 mmol/1) for 24 h; (C) cells were incubated with the same cyclic AMP analogues together with a mono¬ clonal antibody (10mg/I) against human urokinase; (D) cells were incubated with the same cyclic AMP analogues together with a monoclonal antibody (10mg/l) against human tissue plasminogen activator. FIGURE

with

a

25-Hydroxycholesterol labelled with H at C-l and C-2 was used to examine the spectrum of ster¬ oids made by the cells, since such sterols are rapidly metabolized by human adrenocortical cell SCC independent of intracellular cholesterol-transport

mechanisms (Mason et al. 1978; Carr & Mason, 1988). Under control conditions, i.e. incubation in defined medium only, 25-hydroxycholesterol was converted into several A5 products (detected by scin¬ tillation counting) but no A products (by u.v.

Ha) 17aP4

17aP4

cA, 8 h Time

figure 2. Effects of cyclic AMP on the conversion of progesterone in an SV40 T antigentransfected human adrenocortical cell clone, (a) Cells were incubated for 72 h in defined medium alone (con) or with N -monobutyryl cyclic AMP (1 mmol/1) and 8-bromo cyclic AMP (1 mmol/1) (cA). Cells were then incubated with progesterone (10pmol/l) for 2 or 8 h. A4 products were separated by high-performance liquid chromatography. P progesterone; 17aP 17a-hydroxyprogesterone; S= 11-deoxycortisol; F cortisol. The peak indicated by the broken line is a u.v.-absorbing non-steroid component of the medium, (b) The conversion of progesterone to 17a-hydroxyprogesterone plus 11-deoxycortisol (total 17a-hydroxylase activity) is shown for five individual clones. Cultures were treated as described for (a), followed by a 2-h incubation with progesterone. Points are values for activity from replicate cultures; bars are mean values. Open bars indicate activity in control cells and stippled bars indicate activity in cells incubated with cyclic AMP analogues. Numbers (2 to 6) indicate different clones. =

=

=

absorbance) (Fig. 4). In longer incubations, the large peak of 20,22,25-trihydroxycholesterol decreased in size with concomitant increases in the peaks of pregnenolone and 17a-hydroxypregnenolone (not shown). The synthesis of all radioactive products of 25-hydroxycholesterol was abolished by the inclu¬ sion in the incubation medium of (20R)20-phenyl5-pregnene-3ß,20-diol (0-5 pmol/1), a potent and specific inhibitor of SCC (Ascoli et al. 1983). When cells were incubated with cyclic AMP analogues, there was a much larger peak of 17ahydroxypregnenolone, a peak of DHEA and smaller peaks of 20,22,25-trihydroxycholesterol and preg¬ nenolone (Fig. 4). As in control cultures, no A prod-

ucts were detected. In other experiments, cyclic AMP analogues added at 25 pmol/1 produced a con¬ version of 25-hydroxycholesterol to 17a-hydroxypregnenolone and DHEA that was very similar to that shown for 1 mmol/1; forskolin (10 pmol/1) also had a similar effect. TPA by itself had no effect on the conversion of 25-hydroxycholesterol (not shown). When added with cyclic AMP analogues, however, TPA decreased the peaks of 17a-hydroxypregnenolone and DHEA to sizes smaller than those seen in cyclic AMP-treated or control cultures, and A products (very small peaks of 17a-hydroxyprogesterone and progesterone) were detected (Fig. 4). The size of the

u.v.-absorbing 17a-hydroxyprogesterone peak indi¬ cates that the corresponding peak of radioactivity would be too small to be observed, because of the lower sensitivity of the flow-through radioactivity detector. Maximal enhanceme'nt of the production of A steroids was biphasic with respect to TPA con¬ centration; the greatest effect was usually observed at 10 nmol/1 rather than at higher or lower concen¬ trations (not shown). The influence of serum during incubation with cyclic AMP and TPA was studied. Interest in the effects of serum was based on previous data showing that growth factors affect the induction of 3ß-HSD in primary human adrenocortical cells in culture (McAllister & Hornsby, 1988), and that unknown factors present in the (serum-containing) culture

environment

FIGURE 3. Induction of cytochrome P-450 mRNAs in an SV40 T antigen-transfected human adrenocortical cell clone. Cells were incubated with or without cyclic AMP analogues; RNA was harvested, blotted and hybridized as described in the Materials and Methods. C control; cA incubated for 36 h with N -monobutyryl cyclic AMP (1 mmol/1) and 8-bromo cyclic AMP (1 mmol/1) prior to harvesting of RNA. The membrane was probed successively with cDNAs for cholesterol side-chain =

=

cleavage enzyme (SCC), 17a-hydroxylase (17a-HYD) and Superoxide dismutase (SOD). Numbers indicate sizes in kb.

induce

bovine

adrenocortical

cell

21-hydroxylase in co-operation with protein kinase C (Chang et al. 1991). Serum had no effect in cells not exposed to cyclic AMP or TPA (Fig. 4). It blunted the cyclic AMP-induced increase in 17ahydroxypregnenolone and DHEA, and caused the appearance of a minute peak of 17a-hydroxyprogesterone. The combination of serum, cyclic AMP and TPA decreased the peak of 17a-hydroxypregnenolone to a level well below that of the con¬ trol, whereas peaks of both progesterone and 17a-hydroxyprogesterone were seen in the u.v.absorbing products. Tri-iodothyronine was shown to synergize with cyclic AMP in the induction of 3ß-HSD in primary fetal human adrenocortical cells (Simonian, 1986), but did not reproduce the effects of serum in these experiments (not shown). In longer incubations than that shown in Fig. 4, cultures incu¬ bated with serum and cyclic AMP (with or without TPA) synthesized small amounts of 11-deoxycortisol and cortisol from 25-hydroxycholesterol. On Northern blots, the level of 17a-hydroxylase mRNA induced by cyclic AMP analogues added at 1 mmol/1 approached that in the intact adult adrenal cortex (Fig. 5). Forskolin at 10 pmol/1 induced a similar level. The induction by cyclic AMP was unaffected or perhaps slightly increased by the addi¬ tion of serum. Cyclic AMP analogues at 25 pmol/1 induced

a

low

but

detectable

level

of

17a-hydroxylase mRNA. TPA by itself had no effect on mRNA for 17a-hydroxylase or 3ß-HSD (not shown). However, at concentrations as low as 3 nmol/1, it almost completely suppressed the induc¬ tion of 17a-hydroxyIase mRNA by cyclic AMP. TPA synergized with cyclic AMP in the induction of 3ß-HSD mRNA. In SV40 T antigen-transfected adrenocortical cells, the level of mRNA for 3ß-HSD

much lower than that in RNA from the intact adult adrenal cortex, and was at the lower level of detectability by Northern blotting. In the presence was

20,22,2SOHC 20,22,250HC

17aP

250HC

17aP',|

DHEA

20,22,25GHC P5

250HC cA

20,22,250HC S

250HC 17aP

cA + TPA

17aP4

Time

Time

4. Synthesis of steroid products from 25-hydroxycholesterol in an SV40 T antigen-transfected human adreno¬ cortical cell clone. Cells were incubated in defined medium for 72 h without additions (con) or with the addition of N6-monobutyryl cyclic AMP (1 mmol/1) and 8-bromo cyclic AMP (1 mmol/1) (cA), 12-0-tetradecanoylphorbol-13acetate (TPA; 10 nmol/1) or serum (10% (v/v) horse serum and 2% (v/v) fetal bovine serum). Cells were then incubated with 25-hydroxy-[l,2- HJcholesterol as described in the Materials and Methods. Total products (A5 and A ) were detected by flow-through scintillation counting (indicated as c.p.m.). A steroid products were also detected by absorbance at 240 nm (A240). Note that the attenuation of the u.v. absorbance traces is 20-fold lower than that in Fig. 2, thus the traces show an upward slope due to the u.v. absorbance of methanol. DHEA dehydroepiandrosterone; 250HC 25-hydroxycholesterol; 20,22,250HC 20,22,25-trihydroxycholesterol; P4 progesterone; 17aP4 17a-hydroxyprogesterone; P5 pregnenolone; 17aP5 17a-hydroxypregnenolone. The peaks indicated by the broken lines are u.v.-absorbing non-steroid components of the medium. FIGURE

=

=

=

=

=

=

=

androgen synthesis. Thus, human cells probably show stabilization of steroidogenic gene expression after transfection with SV40 T antigen, as noted in bovine adrenocortical cells (Cheng et al. 1989; Hornsby, 1991).

figure

and

5.

Regulation

of

17a-hydroxylase (17a-HYD)

3ß-hydroxysteroid dehydrogenase (3ß-HSD) mRNA

by cyclic AMP and 12-0-tetradecanoylphorbol-13acetate (TPA) in an SV40 T antigen-transfected human adrenocortical cell clone. RNA was from the following sources: lane 1, adult adrenal gland; lane 2, SV40 T antigen-transfected adrenocortical cells incubated in defined medium for 36 h; lane 3, cells incubated in defined medium with N -monobutyryl cyclic AMP (25 pmol/1) and 8-bromo cyclic AMP (25 pmol/1); lane 4,

cells incubated in defined medium with N6-monobutyryl cyclic AMP (1 mmol/1) and 8-bromo cyclic AMP (1 mmol/ 1); lane 5, as for lane 4 but in serum-containing medium (see Fig. 4) rather than defined medium; lane 6, cells incubated in defined medium with forskolin (10 pmol/1); lanes 7, 8 and 9, as for lane 4 but with the addition of TPA at 3, 10 and 30 nmol/1 respectively. RNA was harv¬ ested, blotted and hybridized as described in the Mat¬ erials and Methods. The membrane was probed successively with cDNAs for 3ß-HSD and 17a-hydroxylase. Note that exposure times were sixfold longer for 3ß-HSD than for 17a-hydroxylase. Numbers indicate sizes in kb.

of

cyclic AMP, TPA induced

a greater level of mRNA when used 10 at nmol/1 rather than 3ß-HSD at 3 or 30 nmol/1. Inclusion of serum with TPA and cyclic AMP gave results similar to those shown in Fig. 5; a quantitative comparison was not feasible because of the low mRNA levels.

DISCUSSION

The SV40 T antigen-transfected human adrenocor¬ tical cell clones isolated here have a replicative potential that is extended beyond that of nontransfected cells (McAllister & Hornsby, 1987). The transfected cells are not immortalized; immortaliza¬ tion in T antigen-transfected human fibroblasts is a rare event (Shay & Wright, 1989) and we anticipate that adrenocortical cells will show similar character¬ istics. These clones are useful in that they show that human adrenocortical cells can be grown for long periods in culture with maintenance of differentiated characteristics, particularly the regulation of adrenal

The features studied in the initial characterization of these clones were the induction of 17a-hydroxylase by cyclic AMP analogues and the characteristic rounding response to cyclic AMP. A synergistic 1:1 combination of N -monobutyryl cyclic AMP and 8-bromo cyclic AMP was used (Naseeruddin & Hornsby, 1990). All clones had high levels of inducible 17a-hydroxylase activity as measured by the production of 17a-hydroxyprogesterone from progesterone, and had high 17a-hydroxylase mRNA levels. Like bovine adrenocortical cells, human adrenocortical cells round in response to cyclic AMP by a process that appears to be mediated by cell surface urokinase (Hornsby et al. 1989). The major aspect of steroidogenesis investigated in these clones was the regulation of adrenal androgen synthesis. Other properties of the clones were studied in less detail. A rounding response to ACTH was observed in some clones, suggesting that some clones have ACTH-stimulated adenylate cyclase, but cyclic AMP analogues rather than ACTH were used in most experiments. Previously, we observed a clonal loss of responsiveness to ACTH in normal human and bovine adrenocortical cells (Hornsby et al. 1986; McAllister & Hornsby, 1987). SCC was strongly induced by cyclic AMP in T antigen-transfected human adrenocortical cell clones, as previously observed in primary cultures (Ohashi et al. 1983). Clones had 21-hydroxylase and 11 ß-hydroxylase activity, demonstrated by the production of 11deoxycortisol in short incubations and cortisol in longer incubations, but levels of these enzymes were low in comparison with 17a-hydroxylase. In bovine adrenocortical cells, additional factors are required for the expression of these genes in long-term culture (Cheng & Hornsby, 1992). The high rate of synthesis of the adrenal androgen DHEA, characteristic of the human adrenal cortex, was maintained in SV40 T antigen-transfected cells. DHEA is synthesized when 17a-hydroxylase activity is high and 3ß-HSD activity is limiting (Hornsby & Aldern, 1984; Hornsby, 1985). In primary cultures of fetal human adrenocortical cells, the only factor required for induction of 17a-hydroxylase in defined medium is cyclic AMP, or hormones that raise intra¬ cellular cyclic AMP such as ACTH (Hornsby & Aldern, 1984; McAllister & Hornsby, 1988). SV40 T antigen-transfected cells synthesize DHEA in defined medium with cyclic AMP without special adrenal androgen-stimulating factors, as in primary cultures of human adrenocortical cells (Hornsby &

Aldern, 1984; Mellon et al. 1991; Penhoat et al. 1991). The data from SV40 T antigen-transfected

human adrenocortical cells confirm data from pri¬ mary cultures that the activation of protein kinase C by phorbol esters synergizes with cyclic AMP to increase 3ß-HSD, whereas 17a-hydroxylase requires only cyclic AMP for induction and is decreased by protein kinase C activation (McAllister & Hornsby, 1988; Voutilainen et al. 1991). The pattern of products from 25-hydroxycholes¬ terol (A products without detectable A products) in defined medium with or without cyclic AMP indicates active SCC and 17a-hydroxylase and an absence of 3ß-HSD activity. 20,22,25-Trihydroxycholesterol was a prominent product of 25-hydroxy¬ cholesterol in control cells. This compound is produced by adrenocortical SCC from 25-hydroxy¬ cholesterol (Alsema et al. 1980), and is the equivalent of 20,22-dihydroxycholesterol, a normal intermedi¬ ate in the conversion of cholesterol to pregnenolone (Lambeth & Stevens, 1984). Whereas 20,22dihydroxycholesterol is not normally released from SCC, the trihydroxycholesterol presumably has sufficient polarity to escape from the active site of the enzyme. In time-course experiments in these cells, initial large peaks of 20,22,25-trihydroxycholesterol and pregnenolone were replaced over time by the large peaks of 17a-hydroxypregnenolone and DHEA, suggesting that the trihydroxycholesterol can be slowly converted by SCC to pregnenolone. Cyclic AMP produced a shift in A3 products con¬ sistent with the induction of 17a-hydroxylase/ C17,20-lyase activity, but the continued absence of A products indicates that 3ß-HSD activity is not induced by cyclic AMP alone. The lack of A prod¬ ucts from cholesterol does not imply a lack of the steroidogenic pathway beyond progesterone, since when cultures were incubated with progesterone rather than 25-hydroxycholesterol, cyclic AMPinducible production of 17a-hydroxyprogesterone, 11-deoxycortisol and cortisol was observed. The induction of 3ß-HSD by cyclic AMP plus TPA, shown by Northern blotting and by the syn¬ thesis of A steroids, is consistent with data on the effects of these agents in primary fetal human adrenocortical cells (McAllister & Hornsby, 1988; Voutilainen et al. 1991). As noted in primary fetal human adrenocortical cell cultures (Voutilainen et al. 1991), the level of mRNA for 3ß-HSD is at the lower level of detectability by Northern blotting, even with cyclic AMP plus TPA. As in primary cul¬ tures, a biphasic response to TPA for 3ß-HSD induction was observed. The maximal effect was at 10 nmol/1, a concentration about tenfold higher than that required for the same effect in primary adreno¬ cortical cells (McAllister & Hornsby, 1988), perhaps

indicating some differences in protein kinase C between the clones and their precursor cells. Vouti¬ lainen et al. (1991) suggested that TPA would act to down-regulate protein kinase C and that therefore 3ß-HSD induction is associated with lowered acti¬ vation of protein kinase C rather than elevated activation. However, the biphasic response of 3ß-HSD to protein kinase C activation by TPA sug¬ gests the alternative explanation that high, protein kinase C-down-regulating levels of TPA do not induce 3ß-HSD, and that sustained activation from lower concentrations is associated with induction. The variability in the concentration of TPA required for maximal effects may reflect the difficulty in achieving long-term activation of protein kinase C without down-regulation. There was a greater induction of 3ß-HSD and a greater suppression of cyclic AMP-induced 17a-hydroxylase by TPA in the presence of serum than by TPA alone. The components of serum responsible for these effects are unknown. The appearance of 17a-hydroxyprogesterone in excess of progesterone in the presence of serum, cyclic AMP and TPA presumably indicates that 17a-hydroxylase activity, although low under these conditions, is adequate for the complete conversion of the minute amount of progesterone synthesized. The biphasic response to TPA and synergism with other factors in the (serum-containing) culture environment resembles our previous findings on the regulation of 21-hydroxylase in early-passage bovine adreno¬ cortical cells (Chang et al. 1991). Like 3ß-HSD, 21-hydroxylase appears to be jointly regulated by cyclic AMP-dependent protein kinase and by other kinases including protein kinase C. All clones of T antigen-transfected cells had a lower level of 3ß-HSD activity than that observed in similarly treated primary cultures. In these clones, 3ß-HSD is clearly detectable (by Northern blotting or 25-hydroxycholesterol conversion) only after combined cyclic AMP and protein kinase C acti¬ vation, whereas in primary cultures there is syner¬ gism between cyclic AMP and protein kinase C, but 3ß-HSD can be induced by cyclic AMP or ACTH alone (Hornsby & Aldern, 1984; McAllister & Hornsby, 1988; Voutilainen et al. 1991). The differ¬ ence could indicate a change in behaviour resulting from an effect of SV40 T antigen, but such an effect was not seen in bovine adrenocortical cells. Bovine adrenocortical cells have a very high activity of 3ß-HSD in comparison with human cells (Hornsby & Aldern, 1984), and this was also a characteristic of SV40 T antigen-transfected bovine cell lines (Cheng et al. 1989; Hornsby, 1991). Indeed, such cell lines produce only A products from 25-hydroxy[l,2-3H]cholesterol (Cheng & Hornsby, 1992). One

explanation for the difference in the behaviour of 3ß-HSD between these lines and primary cultures of fetal human adrenocortical cells may be that primary

cells carry over factors from the in-vivo environment that act, like unidentified compounds in serum, as co-inducers of 3ß-HSD with cyclic AMP. Thus, pri¬ mary cultures may not yet be deficient in factors required for the induction of some enzymes. Another possibility is that, although clones were derived from definitive zone cells, they have acquired character¬ istics of the fetal zone of the adrenal cortex (or zona reticularis of the adult cortex) which has extremely low levels of 3ß-HSD (Hornsby, 1985). This may result from the effects of long-term growth or some feature of the culture environment. In summary, data from SV40 T antigen-trans¬ fected adrenocortical cell clones suggest that these cells will be useful in the investigation of the 17a-hydroxylase and 3ß-HSD genes, expression of which determines the balance between synthesis of glucocorticoids and synthesis of androgen precursor by the human adrenocortical cell. ACKNOWLEDGEMENTS

This work was supported by grant AG-06108 from the National Institute on Aging. REFERENCES

Alsema, G. J., Degenhart, H. J. & Hoogerbrugge, J. (1980). Side chain cleavage of hydroxylated sterols by bovine adrenal

cortex

mitochondria: Observation of

25-hydroxylase

activity during incubation at pH 7-80. Journal of Steroid Biochemistry 13, 539-543. Ascoli, M., Freeman, D. A., Sheets, J. J. & Vickery, L. E. (1983). Inhibition of steroidogenesis in cultured Leydig tumor cells by 22-amino-23,24-bisnor-5-cholen-3ß-ol and (20R)20-phenyl-5-pregnene-3ß,20-diol. Endocrinology 113, 127-132.

Carr, B. R. & Mason, J. I. (1988). The effects of a-human atrial natriuretic polypeptide on steroidogenesis by fetal

zone cells of the human fetal adrenal gland. American Journal of Obstetrics & Gynecology 159, 1361-1365. Chang, C.-W., Naseeruddin, S. A. & Hornsby, P. J. (1991). The response of 21-hydroxylase messenger ribonucleic acid levels to adenosine 3',5'-monophosphate and 12-0tetradecanoylphorbol-13-acetate in bovine adrenocortical cells is dependent on culture conditions. Endocrinology 128,

604-610. C. Y. & Hornsby, P. J. (1992). Expression of llß-hydroxylase and 21-hydroxylase in long-term cultures of bovine adrenocortical cells requires extracellular matrix factors. Endocrinology. (In Press.) Cheng, C. Y., Ryan, R. F., Vo, T. P. & Hornsby, P. J. (1989). Cellular senescence involves stochastic processes causing loss of expression of differentiated function genes: Transfection with SV40 as a means for dissociating effects of senescence on growth and on differentiated function gene expression. Experimental Cell Research 180, 49-62.

Cheng,

K. J., Voutilainen, R., Mohandas, T. K. & Miller, W. L. (1986). Human cholesterol sidechain cleavage enzyme, P450scc: cDNA cloning, assignment of the gene to chromosome 15, and expression in the placenta. Proceedings of the National Academy of Sciences of the U.S.A. 83, 8962-8966. Chung, B. C, Picado-Leonard, J., Haniu, M., Bienkowski, M., Hall, P. F., Shively, J. E. & Miller, W. L. (1987). Cytochrome P450cl7 (steroid 17a-hydroxylase/17,20 lyase): Cloning of human adrenal and testis cDNAs indicates the same gene is expressed in both tissues. Proceedings of the National Academy of Sciences of the U.S.A. 84, 407-411. Hornsby, P. J. (1980). Regulation of cytochrome P-450supported 11 ß-hydroxylation of deoxycortisol by steroids, oxygen, and antioxidants in adrenocortical cell cultures. Journal of Biological Chemistry 255, 4020-4027. Hornsby, P. J. (1985). The regulation of adrenocortical function by control of growth and structure. In Adrenal Cortex, pp. 1—31. Eds D. C. Anderson & J. S. D. Winter. London: Butterworth. Hornsby, P. J. (1989). Regulation of steroid 17a-hydroxylase in adrenocortical cells in culture. Endocrine Research 15, 159-182. Hornsby, P. J. (1991). Regulation of steroid hydroxylases in normal and SV40 T antigen-transfected bovine adrenocortical cells in long-term culture. Endocrine Research 17, 109-134. Hornsby, P. J. & Aldern, K. A. (1984). Steroidogenic enzyme activities in cultured human definitive zone adrenocortical cells: Comparison with bovine adrenocortical cells and resultant differences in adrenal androgen synthesis. Journal of Clinical Endocrinology & Metabolism 58, 121—130. Hornsby, P. J., Aldern, K. A. & Harris, S. E. (1986). Clonal variation in response to adrenocorticotropin in cultured bovine adrenocortical cells: Relationship to senescence. Journal of Cellular Physiology 129, 395^102. Hornsby, P. J., Hancock, J. P., Vo, T. P., Nason, L. M., Ryan, R. F. & McAllister, J. M. (1987). Loss of expression of a differentiated function gene, steroid 17a-hydroxylase, as adrenocortical cells senesce in culture. Proceedings of the National Academy of Sciences of the U.S.A. 84, 1580-1584. Hornsby, P. J. & McAllister, J. M. (1991). Culturing steroidogenic cells. In Methods in Enzymology, vol. 206, pp. 371-380. Eds M. R. Waterman & E. F. Johnson. San Diego: Academic Press. Hornsby, P. J., Maghsoudlou, S. S., Cheng, V. & Cheng, C. Y. (1989). Cyclic AMP-mediated cytoskeletal effects in adrenal cells are modified by serum, insulin, insulin-like growth factor-I, and an antibody against urokinase plasminogen activator. Molecular & Cellular Endocrinology 67, 185-194. Hornsby, P. J. & Salmons, B. (1992). Oncogene-expressing cell lines from steroidogenic tissues. In Cell and Tissue Culture: Laboratory Procedures. Eds J. B. Griffiths, A. Doyle & D. G. Newell. Chichester: John Wiley & Sons Ltd. (In

Chung, B., Matteson,

Press.) Hornsby, P. J., Sturek, M., Harris, M. H. (1983). Serum and growth

S. E. & Simonian, factor requirements for proliferation of human adrenocortical cells in culture: Comparisons with bovine adrenocortical cells. In Vitro 19, 863-869. Jat, P. S„ Cepko, C. L., Mulligan, R. C. & Sharp, P. A. (1986). Recombinant retroviruses encoding simian virus 40 large T antigen and polyomavirus large and middle T antigens. Molecular & Cellular Biology 6, 1204-1217. Lambeth, J. D. & Stevens, V. L. (1984). Cytochrome P-450scc: Enzymology, and the regulation of intramitochondrial cholesterol delivery to the enzyme. Endocrine Research 10, 283-310.

McAllister, J. M. & Hornsby, P. J. (1987). Improved clonal and non-clonal growth of human, rat, and bovine adrenocortical cells in culture. In Vitro Cellular & Developmental Biology 23, 677-685. McAllister, J. M. & Hornsby, P. J. (1988). Dual regulation of 3ß-hydroxysteroid dehydrogenase, 17a-hydroxylase, and dehydroepiandrosterone sulfotransferase by adenosine 3',5'monophosphate and activators of protein kinase C in cultured human adrenocortical cells. Endocrinology 122, 2012-2018. Mason, J. I., Arthur, J. R. & Boyd, G. S. (1978). Control of sterol metabolism in rat adrenal mitochondria. Biochemical Journal 174, 1045-1051. Mellon, S. H., Shively, J. E. & Miller, W. L. (1991). Human

proopiomelanocortin-(79-96),

a

proposed androgen

does not affect steroidogenesis in cultured human fetal adrenal cells. Journal of Clinical Endocrinology & Metabolism 72, 19—22. Naseeruddin, S. A. & Hornsby, P. J. (1990). Regulation of llß- and 17a-hydroxyIases in cultured bovine adrenocortical cells: Cyclic adenosine 3',5'-monophosphate, insulin-like growth factor-I, and activators of protein kinase C. Endocrinology 127, 1673-1681. Neville, A. M. & O'Hare, M. J. (1982). The Human Adrenal Cortex. Pathology and Biology An Integrated Approach. Berlin: Springer-Verlag. Ohashi, M., Simpson, E. R., Mason, J. I. & Waterman, M. R. (1983). Biosynthesis of cholesterol side-chain cleavage cytochrome P-450 in human fetal adrenal cells in culture. Endocrinology 112, 2039-2045.

stimulatory hormone,

-

Penhoat, A., Sanchez, P., Jaillard, C, Langlois, D., Begeot, M. & Saez, J. M. (1991). Human proopiomelanocortin(79—96), a proposed cortical androgen-stimulating hormone, does not affect steroidogenesis in cultured human adult adrenal cells. Journal of Clinical Endocrinology & Metabolism 72, 23-26. Rainey, W. E., Hornsby, P. J. & Shay, J. W. (1983).

Morphological correlates of adrenocorticotropin-stimulated steroidogenesis in cultured adrenocortical cells: Differences between bovine and human cells. Endocrinology 113,

48-54.

Shay, J. W. frequency

& Wright, W. E. (1989). Quantitation of the of immortalization of normal human diploid fibroblasts by SV40 large T-antigen. Experimental Cell Research 184, 109-118. Sherman, L., Dafni, N., Lieman-Hurwitz, J. & Groner, Y. (1983). Nucleotide sequence and expression of human chromosome 21-encoded Superoxide dismutase mRNA. Proceedings of the National Acad' ..y of Sciences of the U.S.A. 80, 5465-5469. Simonian, M. H. (1986). ACTH and thyroid hormone regulation of 3ß-hydroxysteroid dehydrogenase activity in human fetal adrenocortical cells. Journal of Steroid Biochemistry 25, 1001-1006. Southern, P. J. & Berg, P. (1982). Transformation of mammalian cells to antibiotic resistance with a bacterial gene under control of the SV40 early region promoter. Journal of Molecular & Applied Genetics 1, 327-341. Stein, G. H. (1985). SV40-transformed human fibroblasts: Evidence for cellular aging in precrisis cells. Journal of Cellular Physiology 125, 36-44. The, V. L., Lachance, Y., Labrie, C, Leblanc, G., Thomas, J. L., Strickler, R. C. & Labrie, F. (1989). Full length cDNA structure and deduced amino acid sequence of human 3ß-hydroxy-5-ene steroid dehydrogenase. Molecular Endocrinology 3, 1310-1312. Voutilainen, R., Ilvesmaki, V. & Miettinen, P. J. (1991). Low expression of 3ß-hydroxy-5-ene steroid dehydrogenase gene in human fetal adrenals in vivo; adrenocorticotropin and protein kinase C-dependent regulation in adrenocortical cultures. Journal of Clinical Endocrinology & Metabolism 72, 761-767. Zuber, M. X., John, M. E., Okamura, T., Simpson, E. R. & Waterman, M. R. (1986). Bovine adrenocortical cytochrome P-45017u. Regulation of gene expression by ACTH and elucidation of primary sequence. Journal of Biological Chemistry 261, 2475-2482.

Expression of 17 alpha-hydroxylase and 3 beta-hydroxysteroid dehydrogenase in fetal human adrenocortical cells transfected with SV40 T antigen.

Primary fetal human adrenocortical cells of definitive zone origin were transfected by electroporation with pSV3neo, a plasmid coding for SV40 T antig...
5MB Sizes 0 Downloads 0 Views