Regulation of Steroidogenesis in NCI-H295 Cells: A Cellular Model of the Human Fetal Adrenal
Bart Staels*, Dean W. Hum*, and Walter L. Miller
Department of Pediatrics Metabolic Research Unit University of California San Francisco, California 94143-0978
NCI-H295 is a recently described human adrenocor- tical carcinoma cell line that makes a variety of steroid hormones. We sought to determine if steroi- dogenesis in these cells employs the same enzymes as those used in normal adrenal steroidogenesis, and if the genes encoding those enzymes exhibit characteristic responsiveness to activators of the protein kinase-A and -C pathways of intracellular second messengers. Northern blots show that NCI- H295 cells contain abundant mRNAs for three key steroidogenic enzymes, cytochrome P450scc, cyto- chrome P450c17, and cytochrome P450c21. These mRNAs accumulated in a time- and dose-dependent fashion in response to 8-bromo-cAMP (8Br-cAMP), forskolin, cholera toxin, and 3-isobutyl-1-methylxan- thine, all activators of the protein kinase-A pathway. Nuclear run-on assays and actinomycin-D transcrip- tional inhibition experiments show that cAMP regu- lates the expression of all three genes primarily at the transcriptional level. Inhibition of protein synthe- sis with cycloheximide did not prevent the cAMP- induced accumulation of P450scc or P450c17 mRNAs, but did inhibit accumulation of P450c21 mRNA, suggesting that cAMP is acting through a mechanism dependent on protein synthesis to pro- mote accumulation of P450c21 mRNA. Stimulation of the protein kinase-C pathway with phorbol ester decreased P450scc and P450c17 mRNAs, but stim- ulated the accumulation of P450c21 mRNA. RNase protection experiments, Northern blot hybridiza- tions, and reverse transcription-polymerase chain reaction show that NCI-H295 cells express both the 118-hydroxylase (P450c118) encoded by the P450c11B1 gene and the aldosterone synthetase (P450c11AS) encoded by the P450c11B2 gene. 8Br- CAMP increased the abundance of both of these mRNAs with similar kinetics, with maximal accumu- lation of both after about 24 h. NCI-H295 cells also contain the mRNAs for aromatase and insulin-like growth factor-II. 8Br-cAMP increased the abundance of aromatase mRNA and decreased the abundance
0888-8809/93/0423-0433$03.00/0 Molecular Endocrinology Copyright 1993 by The Endocrine Society
of IGF-Il mRNA. These studies show that NCI-H295 cells express most of the enzymes needed for hu- man adrenal steroidogenesis, and that the genes encoding these enzymes respond to stimulation of second messenger pathways in a manner similar to that of human adrenals. NCI-H295 cells appear to be a good model for studying the molecular regulation of human adrenal steroidogenesis. (Molecular En- docrinology 7: 423-433, 1993)
INTRODUCTION
In adrenal steroidogenesis, cholesterol is converted to mineralocorticoids, glucocorticoids, and adrenal andro- gens by a series of well characterized steroidogenic enzymes (for review, see Ref. 1). The first and rate- limiting step in steroidogenesis is the conversion of cholesterol to pregnenolone by cytochrome P450scc, which is thus the quantitatively regulating enzyme. P450scc is a single mitochondrial enzyme encoded by a single gene on human chromosome 15q23-q24 (2) that mediates three reactions, cholesterol 20@-hydrox- ylation, 22-hydroxylation, and C20,22 bond scission, all on a single active site, to yield pregnenolone. The next enzyme, cytochrome P450c17, determines the type of steroid produced. P450c17 is a single microsomal en- zyme encoded by a single gene on human chromosome 10q24-q25 (2) that mediates both 17a-hydroxylase and 17,20-lyase activities. If pregnenolone is not metabo- lized by P450c17 (e.g. in the zona glomerulosa) 17- deoxy C-21 precursors of mineralocorticoids are pro- duced. If pregnenolone is 17a-hydroxylated (e.g. in the zona fasciculata) 17-hydroxy C-21 precursors of glu- cocorticoids are made. If such 17a-hydroxylated ste- roids undergo the 17,20-lyase activity (e.g. in the ad- renal zona reticularis or in testicular Leydig cells), 17- hydroxy C-19 androgens are produced. In the adrenal cortex, both 17-deoxy C-21 steroids in the zona glo- merulosa and 17-hydroxy C-21 steroids in the zona fasciculata are 21-hydroxylated by the same cyto- chrome P450c21. P450c21 is a single enzyme encoded by the P450c21B gene on chromosome 6p21 that is
expressed solely in the adrenal cortex. Final metabolism of deoxycorticosterone to aldosterone requires 11- and 18-hydroxylations and 18-oxidation, all mediated by a single mitochondrial enzyme expressed solely in the zona glomerulosa and variously termed P450c11AS, aldosterone synthetase, P450c11aldo, and P450c11B2 (3-7). The 11-hydroxylation of 11-deoxycortisol to cor- tisol is catalyzed by a closely related enzyme expressed in the zonae fasciculata and reticularis and variously termed P450c118 and P450c11B1 (3-7). These two mitochondrial isozymes are encoded by the tandemly duplicated P450c11B1 and P450c11B2 genes on chro- mosome 8q22 that have 93% sequence identity (8).
The molecular biology of steroidogenesis has been studied in primary cultures of steroidogenic cells and in three transformed cell lines: mouse adrenocortical Y-1 cells, mouse Leydig MA-10 cells, and human JEG-3 choriocarcinoma cells. Studies with primary cultures of human adrenal cells are hampered by poor availability, rapidly diminishing responsiveness to ACTH and cAMP (9), and irreversible senescence (10). Transformed mouse Leydig MA-10 cells (11) make progesterone, but do not express P450c17 (12). JEG-3 choriocarcinoma cells (13) also make progesterone, closely resembling human cytotrophoblast cells (14), and are a good model for studying placental-specific steroidogenesis (15, 16), but this system is also confined to P450scc and its cofactors. A transformed rat granulosa cell line was described recently (17), but as expected for granulosa cells (26), these cells do not express steroidogenic enzymes distal to the synthesis of progesterone (18). Until recently, the only available adrenal steroidogenic cell line has been mouse Y-1 adrenocortical carcinoma cells (19). These have been widely used in transient transfection assays to study the transcription of genes for various steroidogenic enzymes (for reviews, see Refs. 20 and 22), but they do not express P450c21 or P450c17, and being mouse cells, they are sub-optimal for studying the expression of human genes.
Gazdar et al. (23) recently reported the establishment of NCI-H295 cells derived from a human adrenocortical carcinoma. These cells have been maintained in contin- uous culture for nearly 10 yr and make large amounts of the 45-steroids (pregnenolone, 17-hydroxypregnen- olone, and dehydroepiandrosterone), as is typical of human adrenal carcinomas, but also make small amounts of aldosterone and 11-deoxycortisol and trace amounts of cortisol and androstenedione (23), sug- gesting that these cells express all of the enzymes associated with adrenal steroidogenesis. However, it is well established that enzymes other than those nor- mally functioning in adrenal steroidogenesis can have activities similar to those of the “authentic” adrenal enzymes (24, 25). Therefore, the studies of NCI-H295 cell steroidogenesis did not establish the identities of the responsible enzymes. To evaluate NCI-H295 cells as a model of the human adrenal, we examined their expression of the genes encoding several steroidogenic enzymes. NCI-H295 cells express the genes for P450scc, P450c17, and P450c21; furthermore, these genes respond to activators of the protein kinase-A
(PKA) and protein kinase-C (PKC) pathways similarly to normal adrenals. These cells contain abundant mRNA for P450c118, P450c11AS, and insulin-like growth fac- tor-II (IGF-II), suggesting that they are less differentiated than the cells of the various zones of the adult human adrenal.
RESULTS
Expression of mRNAs for Steroidogenic Enzymes
Northern blots show that NCI-H295 cells contain the mRNAs for P450scc, P450c17, and P450c21, and that the abundances of these mRNAs increase when the cells are treated with 8-bromo-cAMP (8Br-cAMP) or other activators of the PKA pathway for 48 h (Fig. 1). The mRNAs for P450scc and P450c17 were more responsive to PKA agonists than the mRNA for P450c21, while the glyceraldehyde phosphate dehy- drogenase (GAPDH) control was unresponsive. Stimu- lation with 3-isobutyl-1-methylxanthine (MIX), which in- hibits phosphodiesterase, was less effective than the other PKA agonists, but elicited an additive effect when given in combination with cholera toxin, a stimulator of adenyl cyclase; however, the cAMP analog 8Br-cAMP always elicited the greatest response. The effect of 8Br-CAMP was maximal between 12-24 h of treatment, then remained essentially the same thereafter up to 48 h (Fig. 2). The mRNAs for P450scc, P450c17, and P450c21 responded to 8Br-cAMP in a dose-dependent fashion, showing maximal response in the range of 300-1000 µM (Fig. 3). Thus, both the kinetics and the dose-dependency of the responses of the steroidogenic
CONTROL
8Br-CAMP FORSKOLIN CT
CONTROL
MIX + CT
MIX
CONTROL
8Br-CAMP FORSKOLIN
CONTROL
MIX + CT
CT
MIX
A
B
SCC -
C17 -
C
D
C21 -
GAPDH -
NCI-H295 cells were treated for 48 h with 100 uM 8Br- CAMP, 50 KM forskolin, 10 ng/ml cholera toxin (CT), 0.5 mm MIX, or 0.5 mm MIX and 10 ng/ml CT (MIX+CT). The control incubations were performed with medium alone or medium containing 0.1% ethanol (the solvent for forskolin). A single Northern blot containing 20 µg total RNA in each lane was hybridized sequentially with human cDNA probes for P450scc (A; SCC), P450c17 (B; C17), P450c21 (C; C21), and GAPDH (D).
400
0 1 3 6 12 24 48 HOURS
- SCC
RELATIVE UNITS
300
- C17
200
SCC
C17
C21
GAPDH
- C21
100
- GAPDH
0
0
6
12
24
48
TIME (HOURS)
500
SCC
C17
400
C21
GAPDH
RELATIVE UNITS
300
200
100
0
0
10
30
100
300
1000
DOSE (LM)
enzymes in NCI-H295 cells are similar to those previ- ously found in primary cultures of both human fetal adrenal cells and luteinized human granulosa cells (9, 26, 27).
Effects of Phorbol Ester and Calcium lonophore
To examine the influence of other intracellular second messenger pathways, we used the phorbol ester phor-
PMA
A23187
0 1 3 6 12 24 48
0 1 3 6 12 24 48 HOURS
- SCC
- C17
- C21
- GAPDH
Cells were incubated with 100 ng/ml PMA (left) or 0.5 AM A23187 (right) for the times shown, and single Northern blots containing 20 µg RNA/lane were successively probed with CDNAs for P450scc (SCC), P450c17 (C17), P450c21 (C21), and GAPDH.
bol 12-myristate 13-acetate (PMA; also known as TPA) to activate PKC, and the calcium ionophore A23187 to mobilize intracellular calcium ion. These two pathways normally converge to mediate the response of the ad- renal zona glomerulosa to angiotensin-Il (for review, see Ref. 28). Incubating NCI-H295 cells with 100 ng/ml PMA diminished the abundance of P450scc mRNA within 3 h; this reached a nadir at 6 h (Fig. 4). However, by 12 h, a slight recovery was apparent, and by 24-48
| Time (h) | P450scc | P450c17 | P450c21 |
|---|---|---|---|
| ACTH | |||
| 0 | 100 | 100 | 100 |
| 12 | 104 | 98 | 107 |
| 24 | 129 | 108 | 119 |
| 48 | 88 | 115 | 115 |
| Angiotensin-II | |||
| 0 | 100 | 100 | 100 |
| 12 | 109 | 140 | 148 |
| 24 | 118 | 96 | 117 |
| 48 | 83 | 94 | 99 |
NCI-H295 cells were treated for the indicated periods of time with ACTH (0.1 AM) or angiotensin-Il (1 4M). RNA was ex- tracted, subjected to Northern blot analysis, and quantitatively analyzed on a Phosphorimager. Results are normalized to GAPDH mRNA and expressed in dimensionless units relative to the untreated control which was set equal to 100.
12 HOURS
24 HOURS
CONTROL
8Br-CAMP
ActD + 8Br-CAMP CONTROL
ActD
8Br-CAMP
ActD + 8Br-CAMP
ActD
- SCC
- C17
- C21
- GAPDH
Cells were treated with 100 AM 8Br-CAMP and/or 5 µg/ml actinomycin-D (ActD; added 90 min before 8Br-cAMP) for 12 or 24 h. Ten micrograms of RNA were loaded in each lane, and the blot was probed sequentially with human cDNAs for P450scc (SCC), P450c17 (C17), P450c21 (C21), and GAPDH.
h, the abundance of P450scc mRNA had returned to basal levels. In sharp contrast, reprobing of the same Northern blot showed very different patterns of P450c17 and P450c21 mRNAs. There was a slow steady decrease in the abundance of P450c17 mRNA for 24 h, with minimal recovery by 48 h, while the mRNA for P450c21 accumulated slowly and steadily, reaching
A
400
CONTROL
8Br-cAMP
RELATIVE UNITS
300
PMA
200
100
0
SCC
C17
C21
LAMIN
B
- C21
C
- YACTIN
- LAMIN
- C17
- SCC
BSK
CON 8Br-CAMP PMA
8Br-CAMP
PMA
a maximum after 48 h. The mRNA for GAPDH was slightly inhibited 48 h after PMA treatment.
Mobilization of intracellular calcium had little apparent effect on NCI-H295 cells. Incubation with 0.5 AM A23187 promoted a minimal accumulation of the mRNAs for P450scc, P450c17, and P450c21 after 6- 12 h. However, by 48 h, the mRNAs for all three steroidogenic enzymes and the GAPDH control were substantially diminished, suggesting general cellular toxicity (Fig. 4).
Effects of ACTH and Angiotensin-Il
To determine if NCI-H295 cells responded to extracel- lular stimulators of adrenal steroidogenesis, we incu- bated the cells with 10-7 M ACTH or 10-6 M angiotensin- II. As shown in Table 1, neither of these agents elicited changes in the mRNAs for P450scc, P450c17, or P450c21. Thus, NCI-H295 cells do not respond to ACTH or angiotensin-II, but the level of the defect is not known.
1000
RELATIVE UNITS
SCC
C17
C21
100
10
1
0
6
12
24
48 0
6
12
24
48
0
6
12
24
48
TIME (HOURS)
12 HOURS
24 HOURS
CONTROL
8Br-CAMP
CHX + 8Br-CAMP
CHX
CONTROL
8Br-CAMP CHX
CHX + 8Br-CAMP
- SCC
- C17
- C21
- GAPDH
Cells were treated with 100 AM 8Br-CAMP and/or 10 µg/ml cycloheximide (CHX). Ten micrograms of RNA were loaded in each lane, and the blot was probed with human cDNAs for P450scc (SCC), P450c17 (C17), P450c21 (C21), and GAPDH.
Transcription of Genes for Steroidogenic Enzymes in NCI-H295 Cells
Transcription is the principal mechanism for the accu- mulation of the mRNAs for most human steroidogenic enzymes (for review, see Refs. 20-22). To determine the relative contributions of transcriptional and post- transcriptional processes to the increase in the abun- dance of the mRNAs for P450scc, P450c17, and
P450c21, we measured rates of RNA synthesis and degradation. When nuclear transcription was inter- rupted with actinomycin-D (5 mg/ml) 90 min before treatment with 8Br-cAMP, the cAMP-mediated in- creases in P450scc, P450c17, and P450c21 mRNAs were abolished both 12 and 24 h after the addition of 8Br-cAMP (Fig. 5), suggesting that cAMP directly stim- ulates the transcription of these genes in NCI-H295 cells. Similarly, actinomycin-D reduced the abundance of GAPDH mRNA, consistent with its general effect to inhibit transcription.
To verify that 8Br-cAMP acts directly at the transcrip- tional level, we performed a nuclear run-on assay by hybridizing nuclear RNA labeled with [32PJUTP to the CDNAs for P450scc, P450c17, P450c21, y-actin, nu- clear lamin, and control pBluescript vector. The filters were analyzed by phosphorimaging (Fig. 6A) and au- toradiography (Fig. 6B). The phosphorimaging data are normalized to the transcriptional activity of y-actin, as its mRNA was unaffected by 8Br-cAMP or PMA in NCI- H295 cells (Fig. 6C). These assays showed that tran- scription of the P450scc, P450c17, and P450c21 genes increased about 4-, 3-, and 2-fold, respectively, after incubation with 100 AM 8Br-CAMP for 24 h. Treatment with 100 ng/ml PMA for 24 h stimulated P450c21 gene transcription about 2-fold, but decreased transcription of the P450c17 gene consistent with the mRNA re- sponses shown in Fig. 4, while the effect on P450scc was minimal.
To determine whether mRNA stability was altered by treatment with cAMP, RNA synthesis was blocked with actinomycin-D, and the abundances of the mRNAs for P450scc, P450c17, and P450c21 were examined in cells untreated or pretreated with 8Br-CAMP for 24 h (Fig. 7). The rates of disappearance of P450scc, P450c17, and P450c21 mRNAs were very similar in NCI-H295 cells treated with or without 8Br-cAMP, in- dicating that cAMP does not have a major influence on the stability of these mRNAs in NCI-H295 cells.
The role of protein synthesis in mediating basal and cAMP-induced transcription of genes for various ster- oidogenic enzymes differs among various cell types (12, 15, 20, 29). The addition of 10 µg/ml cycloheximide
A
CONTROL
8Br-CAMP
FORSKOLIN
CONTROL
MIX + CT
B
MARKERS
HEP G2
HELA
SW-13
NCI-H295
CONTROL
F9
CT
MIX
1353
1070
872
603
-C11
310
281
271
- GAPDH
234
194
118
NCI-H295 RNA
HUMAN FETAL
ADRENAL RNA
P450 CIIB2
GENOMIC DNA
C
MARKERS
MARKERS
8Br-CAMP
+
+
Bgl I
+
+
+
+
2027
1353
1070
872
=C11 AS
603
310
-C11B
281
-C11AS
271
234
-GAPDH
194
for 12 h had little effect on P450scc or P450c17 mRNAs, whereas after 24 h of incubation, there was a 50% decrease in these two mRNAs, matched by a similar decrease in GAPDH mRNA, suggesting nonspe- cific cellular toxicity (Fig. 8). By contrast, P450c21 mRNA abundance decreased within 12 h and was less than 20% of the control value by 24 h. Cycloheximide in combination with 8Br-cAMP had no measurable ef- fect on the cAMP-induced accumulation of P450scc mRNA, but it blunted the accumulation of P450c17
mRNA and completely prevented the cAMP-induced accumulation of P450c21 mRNA. Thus, transcription of the P450c21 gene appears to require ongoing protein synthesis in NCI-H295 cells, whereas transcription of the P450scc gene does not.
Expression of mRNAs for P450c118 and P450c11AS
The aldosterone synthase, P450c11AS, is expressed only in the zona glomerulosa, whereas the 118-hydrox-
A
0 3 6 12 24 48
HOURS
B
MARKERS
PROBE
489-
404 - 364
9 0 3 6 12 24 48
HOURS
404-
364-
- C11B
242 -
- C11AS
-C11B
190 -
242-
- GAPDH
190-
-C11AS
157-
118 -
118-
110 -
C
67
500
57
-C11AS
RELATIVE UNITS
400
300
200
C11
100
C11 AS
0
036
12
24
48
Time (HOURS)
A, NCI-H295 cells were treated with 0.1 uM 8Br-CAMP for the times shown, and 50 ng total RNA were reverse transcribed and PCR amplified for 20 cycles, as described in Fig. 9B. The PCR amplification included 106 cpm end-labeled antisense P450c11 oligonucleotide and 106 cpm end-labeled sense GAPDH oligonucleotide as an internal control. The PCR products were then digested with Bg/l (to distinguish P450c118 from P450c11AS as in Fig. 9C) before electrophoresis through 2% agarose gel and autoradiography. B, RNase protection assay of total RNA extracted from NCI-H295 cells treated with 100 MM 8Br-CAMP for the times shown. Ten micrograms of each RNA sample were hybridized to a 386-base cRNA probe corresponding to exon 3 of P450c118 and digested with RNase-A. tRNA, Transfer RNA. C, Phosphorimager data of the experiment shown in B.
0 1 3 6 12 24 48 HOURS
800
RELATIVE UNITS
600
M
Arom
400
GAPDH
- AROM
200
- GAPDH
0
0
3 6
12
24
48
TIME (HOURS)
Cells were treated with 100 KM 8Br-CAMP for the times shown, and the relative amounts of mRNAs for aromatase (Arom) and GAPDH were determined by Northern blotting (right) and phosphorimaging (left).
+
+
-
-
IGF-II
GAPDH
ylase, P450c118, appears to be expressed only in the zonae fasciculata and reticularis (6, 7). To examine the expression of these two mRNAs (which bear 93% sequence identity) in NCI-H295 cells, we reprobed the Northern blot shown in Fig. 1 with a polymerase chain reaction (PCR)-amplified probe corresponding to exon 3 of the P450c11B1 gene. This probe detected the same pattern of bands seen in human fetal adrenals (3) and showed that all four RNA species accumulate in response to 8Br-cAMP or other activators of the PKA pathway (Fig. 9A). To confirm that these cells contain P450c11 mRNA, we used reverse transcription of mRNA, followed by PCR amplification (RT-PCR). Using PCR primers in exons 1 and 2 that do not distinguish between P450c118 and P450c11AS, Fig. 9B shows that NCI-H295 cells contain P450c11 mRNA, whereas non-adrenal human cell lines (HepG2, F9, and HeLa) and non-steroidogenic human adrenal SW-13 cells (30)
do not contain P450c11 mRNA detectable by this very sensitive procedure. To distinguish P450c118 from P450c11AS mRNA, we cleaved the products of the RT-PCR reaction with Bg/l, which cleaves at the se- quence encoding amino acid 29 of P450c11AS, but not in P450c118 (3, 8). Figure 9C shows that the NCI-H295 cells encode an RT PCR product that is cleaved to a 307-basepair (bp) band with Bg/l, as predicted for P450c11AS, and also encode an RT-PCR product of 392 bp that is not cleaved by Bg/l, as predicted for P450c118. The completeness of the Bg/l digestion was demonstrated by an internal control using GAPDH, which contains a single Bg/l site that converts the 239- bp GAPDH RT-PCR fragment to 226 bp. Furthermore, PCR amplification of a clone of P450c11B2 genomic DNA containing exons 1 and 2 and the intervening intron yielded the expected 774-bp P450c11B2 frag- ment, which was cleaved to 689 bp by Bg/l. This ability to distinguish P450c118 from P450c11AS shows that NCI-H295 cells contain both mRNAs, whereas no known human adrenal cell type contains both of these mRNAs.
To examine the kinetics of the responses of P450c118 and P450c11AS to 8Br-cAMP, we per- formed RT-PCR reactions with a 32P-labeled P450c11 antisense primer, which will label both P450c118 and P450c11AS, and with 32P-labeled GAPDH sense primer in a single PCR reaction limited to 20 cycles. As shown in Fig. 10A, 8Br-cAMP increased the P450c118 signal within 3 h and was maximal by 24-48 h. Similarly, the P450c11AS signal was increased within 3-6 h and was also maximal by 24-48 h, although at a slightly lower level. Because this RT-PCR experiment is only semi- quantitative, we also examined the abundance of P450c118 and P450c11AS mRNAs by solution hybrid- ization/RNase protection, using a riboprobe generated from the P450c118 exon 3 cDNA probe used in Fig. 9A. RNA from NCI-H295 cells protected a 218-bp frag- ment corresponding to P450c118 and two internally cleaved fragments of 164 and 54 bp corresponding to P450c11AS (Fig. 10, B and C). Treatment with 8Br- CAMP increased the abundance of both P450c118 and P450c11AS. Both mRNAs increased between 6-12 h and reached maximal levels at 12-48 h, although P450c11AS rose only 3-fold, while P450c118 rose more than 5-fold.
Expression of Aromatase mRNA
Gazdar et al. (23) reported that NCI-H295 cells secrete small amounts of estrogen; therefore, we looked for aromatase mRNA in these cells. Northern blotting shows that NCI-H295 cells contain barely detectable amounts of aromatase mRNA, but that this RNA is readily induced by 8Br-cAMP (Fig. 11). Increases in aromatase mRNA were readily detectable after 12 h and were maximal at 24-48 h. By contrast, treatment with PMA elicited no detectable change in the very low basal levels of aromatase mRNA (not shown).
Expression of IGF-II mRNA
Human fetal adrenals express high levels of IGF-II mRNA, whereas no IGF-II mRNA is detected in adult adrenals (27, 31). Thus, in the adrenal, IGF-II mRNA may serve as a marker for fetal development. Northern blots show that IGF-Il mRNA is very abundant in NCI- H295 cells, being detectable within 1 h of autoradi- ographic exposure (Fig. 12). Treatment of NCI-H295 cells with 8Br-CAMP for 12 h decreased IGF-II mRNA (Fig. 12); by contrast, ACTH increased IGF-Il mRNA in primary cultures of human fetal adrenal cells (27, 31).
DISCUSSION
Studies of human adrenal steroidogenesis have been hampered by the lack of adequate cell culture systems; the establishment of the NCI-H295 cell system should change this dramatically. In their initial description of these cells, Gazdar et al. (23) provided steroidal evi- dence for the existence of all major pathways of adrenal steroidogenesis in these cells. Our data now show that these pathways are based on expression of the same enzymes found in the normal adrenal, and that the genes encoding these enzymes respond to intracellular second messengers in a physiologically meaningful fashion.
8Br-cAMP and other agonists of the PKA pathway stimulated P450scc, P450c17, and P450c21 mRNA accumulation similar to that in primary cultures of hu- man fetal adrenal cells (9, 27, 31). Furthermore, the endogenous human P450scc, P450c17, and P450c21 genes in these cells were transcriptionally induced by CAMP, similar to the behavior of the corresponding human promoter/reporter constructions when trans- fected into mouse Y-1 adrenocortical carcinoma cells (for review, see Refs. 21 and 22). As in various systems, the cAMP-induced accumulation of these mRNAs is primarily a reflection of transcriptional, rather than post- transcriptional, induction. The cAMP induction of P450scc and P450c17 did not require ongoing protein synthesis, but the cAMP induction of P450c21 did. The role of protein synthesis in basal and cAMP-induced transcription has not been studied previously in human adrenal cells. Although the transcription of bovine P450scc, P450c17, and P450c21 genes appears to require ongoing protein synthesis in primary cultures of bovine adrenal cells (for review, see Ref. 20), transcrip- tion of the human P450scc, P450c17, and P450c21 promoters appears to be direct and not require protein synthesis when these promoters are put into mouse adrenal Y-1 cells (32-34). The difference in the cyclo- heximide sensitivity of P450c21 gene expression vs. the insensitivity of P450scc and P450c17 gene expres- sion in NCI-H295 cells suggests that transcriptional activation of the human P450c21 gene is fundamentally different from activation of the P450scc and P450c17 genes, consistent with the view that some of the nuclear
transcription factors that activate these genes are the same, and some are different (35).
Stimulating the PKC pathway in NCI-H295 cells with the phorbol ester PMA decreased P450scc mRNA ac- cumulation after 3-12 h, but this recovered by 24 h (Fig. 4), a time at which no effect was seen on P450scc transcription (Fig. 6). Treatment with PMA for 6-12 h similarly suppresses transcription of human P450scc promoter/reporter constructions transiently transfected into mouse adrenal Y-1 cells (32). PMA also induced P450c21 gene transcription and mRNA accumulation, while decreasing P450c17 gene transcription and mRNA accumulation. Phorbol ester induces similar in- creases in P450c21 mRNA and decreases in P450c17 mRNA in primary cultures of human fetal adrenal cells (36), and PMA supresses transcription of the human P450c17 promoter/reporter constructions transiently transfected into Y-1 cells (33). By contrast, PMA in- creases P450c17 mRNA slightly in primary cultures of adult human adrenal cells (36), suggesting that NCI- H295 cells more closely resemble fetal than adult hu- man adrenal cells. This is consistent with their relatively high production of 45-steroids and relatively low pro- duction of 44-steroids (23) and the expression of abun- dant IGF-II mRNA (27, 31), which are typical features of human fetal adrenal cells. Finally, the expression of large amounts of both P450c118 (113-hydroxylase) and P450c11AS (aldosterone synthetase) mRNAs suggests that NCI-H295 cells represent an adrenocortical cell that has not fully differentiated into the phenotype of one of the three adrenocortical zones. This in combi- nation with the various fetal characteristics of these cells suggest that they are zonally undifferentiated hu- man fetal adrenal cells.
MATERIALS AND METHODS
Cell Culture
NCI-H295 cells were grown in RPMI-1640 supplemented with 2% fetal calf serum supplemented with selenium (5 ng/ml), insulin (5 µg/ml), transferrin (5 µg/ml), hydrocortisone (10-8 M), and 178-estradiol (10-8 M) at 37 C in 5% CO2-95% air. To eliminate the influence of steroid hormones in the medium, cells were switched to RPMI-1640 supplemented with 10% fetal calf serum, selenium (5 ng/ml), insulin (5 µg/ml), and transferrin (5 Mg/ml) 14 days before incubation with 8Br-CAMP at the indicated doses, 50 uM forskolin, 10 ng/ml cholera toxin, 0.5 mm MIX, 10 µg/ml cycloheximide, 100 ng/ml PMA, 0.5 AM calcium ionophore A23187, 0.1 AM ACTH, 1.0 uM angiotensin- Il (all from Sigma, St. Louis, MO), or 5 ug/ml actinomycin-D (Boehringer Mannheim, Indianapolis, IN).
RNA Analysis
RNA isolated by acid guanidinium thiocyanate/phenol-chloro- form extraction was electrophoresed through 1.2% agarose- formaldehyde gels and Northern-transferred to Hybond-N ny- lon membranes (Amersham, Arlington Heights, IL). Filters were successively hybridized to cDNAs for human P450scc (37), P450c17 (38), P450c21 (39), IGF-II (40), aromatase (41), and GAPDH (42), labeled with [32P]deoxy-CTP by random primers. The P450c21 probe was the 0.7-kb Kpnl/EcoRI fragment of
cDNA that does not extend into the overlapping XB gene (43). A clone of exon 3 of the P450c11B1 gene encoding P450c118 (8) was isolated from a human fetal adrenal cDNA library by PCR amplification and subcloned in pBluescript SK (sense primer, TGAATGGGCCTGAATGGCGC; antisense ·primer, AAGTTGCTGGCTTCTATGGT) and used as probe for north- ern blots. Filters were hybridized to 1.0 x 106 cpm/ml of each probe and washed in 500 ml 0.1 x SSC and 0.1% sodium dodecyl sulfate for 10 min at room temperature and twice for 30 min at 65 C. Filters were scanned on a PhosphorImager (Molecular Dynamics, Eugene, OR) and quantitatively analyzed using Imagequant software before autoradiography with Ko- dak X-Omat-AR film (Eastman Kodak, Rochester, NY). Blots were then stripped twice for 30 min at 95 C in 0.1% sodium dodecyl sulfate and 5 mm Tris-HCI (pH 7.5) and reautoradi- ographed between hybridization of each probe to ensure that all radioactivity had been removed. P450c118 and P450c11AS expressions were analyzed by RNase protection and PCR amplification, followed by P450c11AS-specific restriction en- zyme digestion. RNase protection assays were performed essentially as previously described (32), using a P450c116 exon 3 riboprobe synthesized from the cDNA clone. Hybridi- zation of 20 µg total RNA to 5 x 105 cpm probe was performed at 50 C, followed by digestion with 15 ug DNase-free RNase- A for 1 h at 37 C. For PCR amplification, 50 ng total RNA were reverse transcribed using random hexamer primers, then P450c11 exons 1 and 2 were PCR amplified (sense primer, ATGGCACTCAGGGCAAAGGCA; antisense primer, CAA- GAACACGCCACATTTGTGC), followed by P450c11B2-spe- cific Bg/l digestion and 2% agarose gel electrophoresis. GAPDH-specific primers were used as internal controls. Each experiment was performed at least twice.
Nuclear Run-On Assay
Nuclei were isolated and labeled with [a-32PJUTP (3000 Ci/ mmol) for RNA polymerase run-on assays, as previously de- scribed (44). Nuclear RNA was isolated, and dot blots of nuclear RNA were hybridized and washed, as described above. Equivalent counts of labeled nuclear RNA were hybrid- ized for 36 h at 42 C to 5 µg cloned cDNAs for human P450scc (37), P450c17 (38), P450c21 (39, 43), lamin-A (45), and y- actin (46) or pBluescript vector DNA (Stratagene, La Jolla, CA). Filter hybridizations were quantitated by phosphorimaging (Fig. 6A) and autoradiography (Fig. 6B), as described for RNA blots.
Acknowledgments
We thank Drs. Synthia Mellon and Dong Lin for productive discussions, H. K. Oie for providing the NCI-H295 cells, James Bristow for the GAPDH probe, Shiuan Chen for the aromatase probe, Carlos Fardella for the P450c11B2 genomic DNA clone, and Rudi Grosschedl for use of the phosphorimager.
Received October 20, 1992. Revision received December 10, 1992. Accepted December 17, 1992.
Address requests for reprints to: Dr. Walter L. Miller, Build- ing MR IV, Room 209, University of California, San Fran- cisco, California 94143-0978.
This work was supported by a fellowship from the D. Collen Research Foundation (to B.S.), a fellowship from the Fonds de la Recherche en Santé du Quebec (no. 910428-103; to D.W.H.), and grants from the NIH (DK-37922 and DK-42154) and the March of Dimes (6-0098; to W.L.M.).
* B.S. and D.W.H. should be considered equally as first author
REFERENCES
1. Miller WL 1988 Molecular biology of steroid hormone synthesis. Endocr Rev 9:295-317
2. Sparkes RS, Klisak I, Miller WL 1991 Regional mapping of genes encoding human steroidogenic enzymes: P450scc to 15q23-q24, adrenodoxin to 11q22; adreno- doxin reductase to 17q24-q25; and P450c 17 to 10q24- q25. DNA Cell Biol 10:359-365
3. Kawamoto T, Mitsuuchi Y, Ohnishi T, Ichikawa Y, Yokoy- ama Y, Somimoto H, Toda K, Miyahara K, Kuribayashi I, Nakao K, Hosoda K, Yamamoto Y, Imura H, Shizuta Y 1990 Cloning and expression of a cDNA for human cyto- chrome P-450aldo as related to primary aldosteronism. Biochem Biophys Res Commun 173:309-316
4. Malee MP, Mellon SH 1991 Zone-specific regulation of two distinct messenger RNAs for P450c 11 (11/18-hy- droxylase) in the adrenals of pregnant and non-pregnant rats. Proc Natl Acad Sci USA 88:4731-4735
5. Curnow KM, Tusie-Luna M, Pascoe L, Natarajan R, Gu J, Nadler JL, White PC 1991 The product of the CYP11 B 2 gene is required for aldosterone biosynthesis in the human adrenal cortex. Mol Endocrinol 5:1513-1522
6. Domalik LJ, Chaplin DD, Kirkman MS, Wu RC, Liu W, Howard TA, Seldin MF, Parker KL 1991 Different iso- zymes of mouse 118-hydroxylase produce mineralocorti- coids and glucocorticoids. Mol Endocrinol 5:1853-1861
7. Ogishima T, Suzuki H, Itata J, Mitani F, Ishimura Y 1992 Zone-specific expression of aldosterone synthase cyto- chrome P 450 and cytochrome P450-118 in rat adrenal cortex: histochemical basis for the functional zonation. Endocrinology 130:2971-2977
8. Mornet E, Dupont J, Vitek A, White PC 1989 Character- ization of two genes encoding human steroid 11-beta- hydroxylase (P-450-11beta). J Biol Chem 264:20961- 20967
9. DiBlasio AM, Voutilainen R, Jaffe RB, Miller WL 1987 Hormonal regulation of messenger ribonucleic acids for P450scc (cholesterol side-chain cleavage enzyme) and P450c 17 (17a-hydroxylase/17,20 lyase in cultured hu- man fetal adrenal cells. J Clin Endocrinol Metab 65:170- 175
10. Hornsby PJ, Ryan RF, Cheng CY 1989 Replicative senes- cence and differentiated gene expression in cultured adre- nocortical cells. Exp Geronotol 24:539-558
11. Ascoli M 1981 Characterization of several clonal lines of cultured leydig tumor cells: gonadotropin receptors and steroidogenic responses. Endocrinology 108:88-95
12. Mellon SH, Vaisse C 1989 cAMP regulates P450scc gene expression by a cycloheximide-insensitive mechanism in cultured mouse Leydig MA- 10 cells. Proc Natl Acad Sci USA 86:7775-7779
13. Kohler PO, Bridson WE 1971 Isolation of hormone-pro- ducing clonal lines of human choriocarcinoma. J Clin Endocrinol Metab 65:122-126
14. Ringler GE, Strauss III JF 1990 In vitro systems for the study of human placental endocrine function. Endocr Rev 11:105-123
15. Picado-Leonard J, Voutilainen R, Kao L, Chung B, Strauss III JF, Miller WL 1988 Human adrenodoxin: cloning of three cDNAs and cycloheximide enhancement in JEG-3 cells. J Biol Chem 263:3240-3244 (corrected 11016)
16. Moore CCD, Hum DW, Miller WL 1992 Identification of positive and negative placenta-specific basal elements and a cyclic adenosine 3’,5’-monophosphate response element in the human gene for P450scc. Mol Endocrinol 6:2045-2058
17. Suh BS, Amsterdam A 1990 Establishment of highly steroidogenic granulosa cell lines by co-transfection with SV 40 and Ha-ras oncogenes: induction of steroidogene- sis by cAMP and its supression by phorbol ester. Endo- crinology 127:2489-2500
18. Hanukoglu I, Suh BS, Himmelhoch S, Amsterdam A 1990
Induction and mitochondrial localization of cytochrome P450scc system enzymes in normal and transformed ovarian granulosa cells. J Cell Biol 111:1373-1381
19. Yasumura Y, Buonassisi V, Sato GH 1966 Clonal analysis of differentiated function in animal cell cultures. I. Possible correlated maintainence of differentiated function and the diploid karyotype. Cancer Res 26:529-535
20. Waterman MR, Simpson ER 1989 Regulation of steroid hydroxylase gene expression is multifactorial in nature. Recent Prog Horm Res 45:533-566
21. Moore CCD, Miller WL 1991 The role of transcriptional regulation in steroid hormone biosynthesis. J Steroid Biochem Mol Biol 40:517-525
22. Hum DW, Miller WL 1993 Transcriptional regulation of human genes for steroidogenic enzymes. Clin Chem 39:333-340
23. Gazdar AF, Oie HK, Shackleton CH, Chen TR, Triche TJ, Myers CE, Chrousos GP, Brennan MF, Stein CA, La- Rocca RV 1990 Establishment and characterization of a human adrenocortical carcinoma cell line that expresses multiple pathways of steroid biosynthesis. Cancer Res 50:5488-5496
24. Casey ML, MacDonald PC 1982 Extra-adrenal formation of a mineralo-corticoid: deoxycorticosterone and deoxy- corticosterone sulfate biosynthesis and metabolism. En- docr Rev 3:396-403
25. Mellon SH, Miller WL 1989 Extra-adrenal steroid 21- hydroxylation is not mediated by P450c21. J Clin Invest 84:1497-1502
26. Voutilainen R, Tapanainen J, Chung B, Matteson KJ, Miller WL 1986 Hormonal regulation of P450scc (20,22-des- molase) and P450c 17 (17a-hydroxylase/17,20-lyase) in cultured human granulosa cells. J Clin Endocrinol Metab 63:202-207
27. Voutilainen R, Miller WL 1987 Coordinate tropic hormone regulation of m RNAs for insulin-like growth factor Il and the cholesterol side-chain cleavage enzyme, P450scc, in human steroidogenic tissues. Proc Natl Acad Sci USA 84:1590-1594
28. Barrett PQ, Bollag WB, Isales CM, McCarthy RT, Ras- mussen H 1989 Role of calcium in angiotensin II-mediated aldosterone secretion. Endocr Rev 10:496-518
29. Golos TG, Miller WL, Strauss III JF 1987 Human chorionic gonadotropin and 8-bromo cyclic AMP promote an acute increase in cytochrome P450scc and adrenodoxin m RNAs in cultured human granulosa cells by a cyclohex- imide-insensitive mechanism. J Clin Invest 83:6450-6454
30. Leibovits A, McCombs WB, Johnston D, McCoy CE, Stinson JC 1973 New human cancer cell culture lines. ISW-13 small-cell carcinoma of the adrenal cortex. J Natl Cancer Inst 51:691-697
31. Voutilainen R, Miller WL 1988 Developmental and hor- monal regulation of m RNAs for insulin-like growth factor and steroidogenic enzymes in human fetal adrenals and gonads. DNA 7:9-15
32. Moore CCD, Brentano ST, Miller WL 1990 Human P450scc gene transcription is induced by cyclic AMP and repressed by 12-O-tetradecanoylphorbol-13-acetate and A2 3187 through independent cis-elements. Mol Cell Biol 10:6013-6023
33. Brentano ST, Picado-Leonard J, Mellon SH, Moore CCD, Miller WL 1990 Tissue-specific, cyclic adenosine 3’,5’- monophosphate-induced, and phorbol ester-repressed transcription from the human P450c17 promoter in mouse cells. Mol Endocrinol 4:1972-1979
34. Kagawa N, Waterman MR 1990 cAMP-dependent tran- scription of the human CYP21B (P450c21) gene requires a cis-regulatory element distinct from the consensus CAMP-regulatory element. J Biol Chem 265:11299-11305
35. Lala DS, Rice DA, Parker KL 1992 Steroidogenic factor 1, a key regulator of steroidogenic enzyme expression, is the mouse homolog of fushi tarazu factor I. Mol Endocrinol 6:1249-1258
36. Ilvesmaki V, Voutilainen R 1991 Interaction of phorboles- ter and adrenocorticotropin in the regulation of steroido- genic P 450 genes in human fetal and adult adrenal cell cultures. Endocrinology 128:1450-1458
37. Chung B, Matteson KJ, Voutilainen R, Mohandas TK, Miller WL 1986 Human cholesterol side-chain cleavage enzyme, P450scc: cDNA cloning, assignment of the gene to chromosome 15, and expression in the placenta. Proc Natl Acad Sci USA 83:8962-8966
38. Chung B, Picado-Leonard J, Haniu M, Bienkowski M, Hall PF, Shively JE, Miller WL 1987 Cytochrome P450c17 (steroid 17a-hydroxylase/17,20 lyase): cloning of human adrenal and testis cDNAs indicates the same gene is expressed in both tissues. Proc Natl Acad Sci USA 84:407-411
39. Matteson KJ, Phillips ilI JA, Miller WL, Chung B-C, Orlando PJ, Frisch H, Ferrandez A, Burr IM 1987 P450 XXI (steroid 21-hydroxylase) gene deletions are not found in family studies of congenital adrenal hyperplasia. Proc Natl Acad Sci USA 84:5858-5862
40. Bell GI, Merryweather JP, Sanchez-Pescador R, Stempien MM, Priestlet L, Scott J, Rall LB 1984 Sequence of a cDNA clone encoding human preproinsulin-like growth factor Il. Nature 310:775-777
41. Chen S, Besman MJ, Sparkes RS, Zollman S, Klisak I, Mohandas TK, Hall PF, Shively JE 1988 Human aroma- tase: cDNA cloning, Southern blot analysis, and assign- ment of the gene to chromosome 15. DNA 7:27-38
42. Tokunaga K, Nakamura Y, Sakata K, Fujimori K, Ohkubo M, Sawada K, Ssakiyama S 1987 Enhanced expression of a glyceraldehyde-3-phosphate dehydrogenase gene in human lung cancers. Cancer Res 47:5616-5619
43. Morel Y, Bristow J, Gitelman SE, Miller WL 1989 Tran- script encoded on the opposite strand of the human steroid 21-hydroxylase/complement component/C4 gene locus. Proc Natl Acad Sci USA 86A:6582-6586
44. Nevins JR 1987 Isolation and analysis of nuclear RNA. Methods Enzymol 152:234-241
45. Fisher DZ, Chaudhary N, Blobel G 1986 cDNA sequencing of nuclear lamins A and C reveals primary and secondary homology to intermediate filament proteins. Proc Natl Acad Sci USA 83:6450-6454
46. Gunning P, Ponte P, Okayama H, Engel J, Blau H, Viedes L 1983 Isolation and characterization of full-length cDNA clones for human a-, B-, and y-actin m RNAs: skeletal but not cytoplasmic actions have an amino-terminal cysteine that is subsequently removed. Mol Cell Biol 3:787-795