Molecular and Cellular Endocrinology, 84 (1992) 243-251 0 1992 Elsevier Scientific Publishers Ireland. Ltd. 0303-7207/92/$05.00

MOLCEL

243

02732

Inhibition of adrenocortical steroidogenesis by a,-macroglobulin by associated transforming growth factor /3 Michelle Keramidas, Edmond M. Chambaz and Jean-Jacques

is caused

Feige

Unite’ INSERM 244, Laboratoire de Biochimie des Rigulations Cellulaires Endocrines, Dipartement de Biologie Molthdaire et Structurale, Centre d’Etude.7 Nuclkaires, 85X, F-38041 Grenoble, France (Received

Key words: Transforming

growth

factor

28 October

1991; accepted

p; cu,-Macroglobulin;

20 December

Steroidogenesis;

(Bovine

1991)

adrenal

cortex)

Summary

a,-Macroglobulin (cz,M) is the major protein secreted by bovine adrenocortical cells in primary culture and its synthesis-is stimulated by transforming growth factor /3 (TGFP). We investigated here the effects of a,M on adrenocortical steroidogenesis. We observed that commercial preparations of bovine plasma LY~Mwere able to mimic the inhibitory action of TGFP on adrenocortical cortisol production, with the same specificity of action directed at the steroid l’la-hydroxylation step. This inhibition was time-dependent and dose-dependent (50% inhibition observed with 2 mg/ml a,M). Acid/ethanol extracts of (Y*M appeared to retain the full inhibitory activity of (YAM.Anti-TGFP antibodies could reverse the inhibition caused by the acid/ethanol extract but not that caused by native CX~M.Taken together, these results indicate that the inhibition of adrenocortical steroidogenesis induced by a,M is caused by associated TGFP. We estimated that 2 mg of a,M contained approximately 0.1 ng of TGFP, corresponding to a molar ratio of l/700,000 between TGFP and a,M. These results also clearly indicate that the a,M-TGFP complexes are biologically active on adrenocortical cells, suggesting that these cells possess the enzymatic equipment that can activate the latent a,M-TGFP complexes.

Introduction

a,-Macroglobulin (a,M) is a major plasma protein (Barrett, 1981; Sottrup-Jensen, 1989). Its concentration in the plasma of healthy human

Correspondence to: Jean-Jacques Feige, Unite INSERM 244, Laboratoire de Biochimie des Regulations Cellulaires Endocrines, Departement de Biologie Moltkulaire et Structurale, Centre d’Etudes Nucleaires, 85X, F-38041 Grenoble Cedex, France. Tel. (33) 76.88.47.73; Fax (33) 76.88.51.55. Abbreviations: azM, cy,-macroglobulin; TGFP, transforming growth factor-p; ACTH, adrenocorticotropin; PBS, phosphate-buffered saline.

individuals varies between 2 and 4 g/l depending on age and sex. Relatively high concentrations of a,M are also found in a variety of extravascular fluids including lymph and colostrum. In rats, its blood concentration is much lower than in humans but it is dramatically increased during acute and chronic inflammation occurring in response to tissue damage or infections (Gehring et al., 1989). In this species, a2M represents the major acute-phase protein and the induction of its synthesis appears to be mediated by interleukin-6 (Gauldie et al., 1989). a,M is synthesized and secreted mainly by hepatocytes (Barrett, 1981; Bernuau et al., 1989) but recently its synthesis has

244

been observed in a number of other cell types including fibroblasts (Brissenden and Cox, 1982), ovarian follicles and corpora lutea (Gaddy-Kurten et al., 19891, monocytes (Hovi et al., 1977) and adrenocortical cells (Kirshner et al., 1989; Shi et al., 1990). Synthesis of a,M by adrenocortical cells appears to be highly stimulated by transforming growth factor p (TGFP) (Shi et al., 1990). N?M is a broad spectrum protease inhibitor which complexes with proteases via an irreversible ‘molecular trapping’ mechanism that has been characterized in great detail (Boisset et al., 1989; Sottrup-Jensen, 1989; Sottrup-Jensen et al., 1989). A second function recently ascribed to cvzM that may also be fundamental is its capacity to bind cytokines, such as interleukin-lp (Barth and Luger, 1989) and interleukin-6 (Matsudo et al., 1989; James, 1990), as well as several growth factors. Platelet-derived growth factor (Huang et al., 1984), nerve growth factor (Ronne et al., 1979), insulin (Chu et al., 1991), fibroblast growth factor (Dennis et al., 1989) and transforming growth factor p (O’Connor-McCourt and Wakefield, 1987; Huang et al., 1988) have thus been reported to form complexes with (Y?M. Different roles have been proposed for a,M-growth factor complexes. The (YeM-TGFP complexes have been particularly studied (O’Connor-McCourt and Wakefield, 1987; Huang et al., 1988; Danielpour and Sporn, 1990). In serum, N?M was shown to participate in the latency of this growth factor and was suggested to be responsible for its clearance from the extracellular medium and plasma (O’Connor-McCourt and Wakefield, 1987). However, recent observations by Philip and O’Connor-McCourt (1991) rather favored a role in TGFP delivery. We identified recently cuzM as the major protein secreted by primary cultures of bovine adrenocortical cells (Shi et al., 1990). This led us to examine its possible biological function in this cell system. The present study reports that commercial preparations of cu,M purified from bovine plasma inhibit adrenocortical cell steroid production. We characterized this effect and could demonstrate that it was caused by associated TGFP, suggesting that these a,M-TGFP complexes are biologically active on adrenocortical cells.

Material

and methods

Chemicals TGFP, and TGFPz purified from human platelets were purchased from R&D Systems (Minneapolis, MN, USA). Rabbit polyclonal antiTGFP antibodies from R&D Systems (lot No. 5865) recognize and neutralize the biological activity of both TGFP, and TGFP, but do not react with the latent forms of these factors. Synthetic p,_z4 ACTH (Synacthene) and angiotensin I1 (Hypertensin) were provided by Ciba-Geigy (Basel, Switzerland). [ ‘H]Cortisol (96 Ci/mmol) was from the Commissariat a 1’Energie Atomique (Saclay, France). Steroids were purchased from Sigma (St. Louis, MO, USA). Bovine plasma cy?macroglobulin, Ham’s F12 medium and sera were purchased from Boehringer (Meylan, France). Cell culture Bovine adrenocortical ceils were prepared by successive tryptic digestions of fresh adrenal glands and seeded in 6-well plates (9.6 cm’ per well) or 24-well plates (2 cm’ per well) (Falcon, Oxnard, CA, USA) as previously described (Duperray and Chambaz, 1980). Cells were grown in Ham’s F12 medium supplemented with 12.5% horse serum and 2.5% foetal calf serum under an air/CO, (95% : 5%) atmosphere. Cell number was determined following lysis of cells with citric acid, by counting crystal violetstained nuclei in a Neubauer hematimeter. In the present study, the cells were routinely used on day 4 of culture. Cortisol production Cortisol was measured in the culture medium by a specific radioimmunoassay, as previously described (Duperray and Chambaz, 1980). Steroidogenesis inhibition assay Bovine adrenocortical cells were grown in 24well plates under standard conditions. On day 4 of culture, they were preincubated in F12 medium supplemented with 6% foetal calf serum and cu,-macroglobulin. At the end of the preincubation time, the cells were rinsed and incubated for 2 h with 3 x lo-’ M angiotensin II or 3 x lo-’ M ACTH in foetal calf serum-supplemented cul-

245

ture medium. Cortisol secreted in the medium during this acute hormonal stimulation was measured by radioimmunoassay.

Acid extraction of a,M The procedure was adapted from that employed in the purification of TGFP from platelets (Frolik et al., 1983). 25 mg of bovine a,M were added to 3 ml of extraction buffer (30% ethanol, 0.7 N HCl). After incubation for 24 h at 4°C under agitation, the suspension was centrifuged for 10 min at 10,000 Xg. The supernatant was lyophilized and dissolved in 0.25 ml of Tris-10 mM HCl pH 8.0, 1% bovine serum albumin (BSA). Results

a2 M potentiates steroidogenesis

TGFP-induced

inhibition

the presence or absence of 3.2 mg/ml of plasma cu,M. They were subsequently challenged with an optimal dose of angiotensin II and cortisol released in the medium was measured by radioimmunoassay (Fig. 1). Since a,M is able to bind TGFP (O’Connor-McCourt and Wakefield, 1987; Huang et al., 1988; Danielpour and Sporn, 19901, it was expected that the presence of the macroglobulin would result in a decrease of the TGFP activity. To our surprise, cu,M did not neutralize the inhibitory action of TGFPl and TGFPz but rather enhanced it. In the absence of TGFP, cu,M appeared to partly inhibit adrenocortical cell cortisol production. We thus decided to examine in more detail the effects of commercial preparations of plasma a,-macroglobulin on the differentiated functions of adrenocortical cells.

of

We examined the effects of cu,M on TGFP-induced inhibition of corticosteroid production. Adrenocortical cells were incubated for 20 h with various concentrations of TGFP, or TGFP?, in

(Y-,M inhibits ACTH- and angiotensin II-stimulated steroidogenesis Bovine adrenocortical cells were incubated for various periods of time in the presence of a dose of a,M (3.2 mg/ml) that is within its physiologi-

A

0

TGFBl

(nglml)

0.01

0.1

TGFD2

1

10

100

(nglml)

Fig. 1. Effect of cu,M on TGFfl,- and TGFP,-induced inhibition of bovine adrenocortical cell steroidogenesis. Bovine adrenocorticat cells (2x IO5 cells/2 cm’) were incubated for 20 h at 37°C in serum-supplemented F12 medium containing the indicated concentrations of TGFP, (A) or TGFPz (B) in the presence (closed circles) or absence (open circles) of 3.2 mg/ml of cu,M. At the end of the preincubation time, the cells were rinsed with F12 medium and incubated for 2 h in serum-free F12 medium containing 3X IO-’ M angiotensin II. Cortisol secreted in the medium during the acute hormonal stimulation was measured by radioimmunoassay as previously described (Duperray and Chambaz, 1980). The data presented are the mean*SD of duplicate determinations and are representative of results obtained in three independent experiments.

200 -

A

B

I

100 -

0 I 1

0

a2

2

-macroglobulin

3

4

(mUmU

0

10

20

time

30

(h)

Fig. 2. Inhibition of ACTH-stimulated adrenocortical cell steroidogenesis by c**M. (A) Dose response: Bovine adrenocortical cells (2~ 10’ cells/2 cm’) were preincubated in serum-supplemented F12 medium with various concentrations of cuzM for 20 h at 37°C. At the end of the preincubation time. the cells were rinsed with fresh F12 medium and incubated for 2 h in serum-free F12 medium containing 3 x lo-’ M ACTH. Cortisol produced during the acute hormonal stimulation was measured by radioimmunoassay as previously described (Duperray and Chambaz, 1980). The bar represents the basal production of cortisol during a 2 h incubation in the absence of ACTH. The data presented are the mean + SD of duplicate determinations and are representative of results obtained in three independent experiments. (B) Time course: Bovine adrenocortical cells (2x 10’ cells/2 cm’) were preincubated for the indicated times in serum-supplemented F12 medium containing 3.2 mg/ml of (YAM. At the end of the preincubation time, the cells were rinsed with fresh F12 medium and incubated for 2 h in serum-free F12 medium containing 3~ 10~’ M ACTH. Cortisol produced during the acute hormonal stimulation was measured by radioimmunoassay as previously described. The data presented are the mean + SD of duplicate determinations and are representative of results obtained in three independent experiments.

cal concentration range in human plasma (Sottrup-Jensen, 1989) (Fig. 2B). Cells were then washed, stimulated for 2 h with an optimal dose of ACTH and cortisol secretion was measured by radioimmunoassay. cu,M appeared to inhibit ACTH-induced cortisol production in a time-dependent manner. Inhibition was not detected before 10 h of treatment and reached 75% after 25 h of treatment. In a dose-response study, adrenocortical cells were treated for 20 h with various concentrations of (YAM. Half-maximal inhibition was observed in the presence of 2 mg/ml cu,M (Fig. 2A). Similar experiments were performed to check the effects of preincubation with a,M on angiotensin II-stimulated cortisol production (Fig. 3). (YAM caused an inhibition of angiotensin II-induced steroidogenesis that was very similar to that of ACTH-induced steroidogenesis. Similar results were obtained whether cells were incubated with a,M in foetal calf serum-supplemented or in serum-free medium,

suggesting that a,M present in foetal calf serum (z 3 mg/ml, according to our estimations) did not affect the baseline conditions. cr,M treatment induces a loss of the BAC cell steroid 17whydroxylation acticity The kinetics and the amplitude of these inhibitory effects were reminiscent of those previously observed in response to TGFP (Feige et al., 1986). Since in our hands TGF/3 was found to inhibit mainly the steroid 17a-hydroxylase enzyme (Feige et al., 1987; Perrin et al., 1991), we attempted to determine whether that same step in the biosynthetic pathway of cortisol was inhibited by (YAM. a2M-pretreated adrenocortical cells were incubated for 2 h with various steroid precursors and the production of cortisol was quantitated by radioimmunoassay. The results shown in Fig. 4 indicate that TGFP, and a,M seem to act at the same level: they inhibit the transformation of 25hydroxy-cholesterol, pregnenolone or pro-

247 _

200

200

B

=

A

: c”

E .z zl ‘0

100

100

z P

0

k

Basal

o---l

I

0

.

1

1

I

2

3

.

I

Fig. 3. Inhibition

of angiotensin

cells (2x

0

cm’)

adrenocortical

cuzM for 20 h at 37°C. At the end of the preincubation serum-free

F12 medium

and are representative

of results obtained

end of the preincubation containing

3 X lo-’

II.

Cortisol

produced

and Chambaz. experiments.

II. Cortisol

say ds previously described (Duperray

produced

and Chambaz,

are representative

Dose response: Bovine

the acute

hormonal

and incubated

19X0). The data presented

3.2 mg/ml

stimulation

was measured

are the mean i SD of duplicate

in three independent

was of

determinations cells (2~ 10’

of cu>M. At the

for 2 h in serum-free

stimulation

of

for 2 h in

the basal production

are the mean f SD of duplicate

F12 medium containing

during the acute hormonal

of results obtained

during

(B) Time course: Bovine adrenocortical

times in serum-supplemented

time, the cells were rinsed with fresh F12 medium

M angiotensin

(A)

with various concentrations

1980). The bar represents

II. The data presented

in three independent

for the indicated

F12 medium

(h)

time, the cells were rinsed with fresh F12 medium and incubated

in the absence of angiotensin

cm’) were preincubated

by a?-macroglobulin.

in serum-supplemented

as previously described (Duperray

cortisol during a 2 h incubation cells/2

cell steroidogenesis

3 X IOY’ M angiotensin

containing

measured by radioimmunoassay

time

were preincubated

3

20

10

(mg/ml)

II-stimulated

10’ cells/2

0

.

4

a2 -macroglobulin

adrenocortical

.Y r 0 0

F12 medium

by radioimmunoasdeterminations

and

experiments.

50

,“”

a Z -macroglobulin

TGFDl

40

30

20

10

0 0

OH-Chol.

preg.

steroid Fig. 4. Inhibition

by TGFpl

and

17OHpreg

17OHprog

0

OH-Chol.

precursor aYz-macroglobulin

preg.

steroid of the conversion

of exogenous

steroid

P’O9.

t 70Hpreg

170Hpro~

precursor precursors

into cortisol.

Bovine

for 20 h at 37°C in serum-supplemented F12 medium containing 2 ng/ml TGFP, (left panel) or 3.2 mg/ml cu,M (right panel). At the end of the preincubation time, the cells were rinsed with F12 medium and incubated for 2 h in serum-free F12 medium in the absence or the presence of 25 hydroxy-cholesterol (50 pg/ml) or of various steroid precursors (at a concentration of 5 pg/ml each). Cortisol secreted in the medium during the acute hormonal stimulation was measured by radioimmunoassay, as previously described. The data presented are the mean f SD of duplicate determinations and are representative of results obtained in three independent experiments.

adrenocortical

cells were incubated

prog.

gcsterone into cortisol whereas they do not modify that of 17~-hydroxy-progesterone or 17a-hydroxy-pregnenolone. This points to the 17a-hydroxylation step as the major target of cu,M as well as TGFP, actions. cyI M-induced inhibition of steroidogenesis is prerserzted by anti-TGFP antibodies Previous reports indicated that cu,M is a potent TGFP binding protein (O’Connor-McCourt and Wakefield, 1987; Huang et al., 1988) and that the latent form of TGF/? present in the serum is a complex between N?M and TGFP. This prompted us to examine whether the inhibitory action of TGFP on adrenocortical steroidogenesis could be due to associated TGFP. As TGFP was initially purified from acid-ethanol extracts of bovine platelets (Frolik et al., 19X3), we first

Fig.

6. Reversal

adrenocortical

acid

extract

(uglml)

0.2 120

” 0

0.4

cells were incubated of TGF/3,

of an acid-ethanol

(dilution:

or the presence

in serum-free

medium

secreted

stimulation previously

and incubated

containing

in the medium

was measured (Duperray

and

3~

of cu>M or O.Ih

10~’

IgGs (dilu-

time, cells were for 2 h at 37°C. M angiotensin

by radioimmunoassay Chamhar.

of results obtained

was

of non-immune

II.

during the acute hormonal

sented are the mean f SD of duplicate representative

either no added

or anti-TGFP

At the end of the preincubation

rinsed with fresh F12 medium Cortisol

for 20 h at 37°C in

or 3.2 mg/ml

l/IO())

of

antibodies.

extract of cuzM, The incubation

in the absence

rabbit serum l&s

extract

inhibition

by anti-TGF/J

F12 medium containing

factor or 0.5 ny/ml

tion: l/100).

I

acid

u,-macroglobulin-induced

Bovine adrenocortical

performed 0

of

cell steroidogrnesis

serum-supplemented fig/ml

a2M

TGFfll

control

IWJ).

as described

The

data

determinations

in two independent

prc-

and are experi-

ments.

0

cx 2

Fig. 5. Inhibition acid-ethanol globulin.

-macroglobulin

of adrenocortical

concentrations

acid-ethanol

preparations

of

F12 medium cu,M

(open

extract of cu,M prepared

by an

of a2-macro-

cells were incubated

at 37°C in serum-supplemented indicated

(mglml)

cell steroidogenesis

extract of commercial

Bovine adrenocortical

4

3

2

1

for 20 h

containing

circles)

as described

or of

the an

in Mate-

rial and methods (closed circles). At the end of the preincubation time, cells were rinsed with F12 medium for 7 h at 37°C in serum-free 3 x 10

’ M angiotensin

Chamhar, duplicate

as

1980). The

stimulation

previously and

was measured

described

data presented

determinations obtained

with

II. Cortisol secreted into the medium

during the acute hormonal dioimmunoassay

and incubated

Fl? medium supplemented

are

are the mean *SD

representative

in two independent

by TB-

(Duperray

experiments.

and of

of results

prepared an acid-ethanol extract of bovine plasma a,M and tested the effects of this extract on adrenocortical steroidogenesis (Fig. 5). Whereas the protein concentration in the extract was reduced by a factor 2 X lo”, as compared to the initial cuzM solution, the potency to inhibit cortisol production was not changed. As shown in Fig. 5, the dose-response curves of steroidogenesis inhibition were similar between ctZM and the ethanol-acid extract, except that the latter one was shifted to the left by a factor 2 X lo’, on a protein concentration basis. Since the inhibitory activity of cu,M was entirely recovered in the acid-ethanol extract, we attempted to neutralize this activity with anti-TGFP antibodies (Fig. 6). We used commercial antibodies that can neutralize the biological activity of both mature TGFP, and TGFP, but that do not recognize TGFPs engaged in latent complexes. At a l/l00 dilution,

249

this antibody was able to completely prevent the inhibition of adrenocortical cell cortisol production induced either by TGFP, or by the acidethanol extract of bovine plasma c-u,M (Fig. 6). At the same dilution, non-immune rabbit serum IgGs did not modify significantly the level of cortisol production in any of the conditions tested. Inhibition of steroidogenesis by untreated azM was not prevented by anti-TGFP antibodies, an observation which is in agreement with the specificity of the antibody for the mature forms of TGFP. Discussion Bovine adrenocortical cell steroidogenic functions are very efficiently inhibited by low doses of TGFp (Feige et al., 1986; Hotta and Baird, 1986). TGF/? inhibits several important steps in the biosynthetic pathway leading to cortisol. These include the low-density lipoprotein receptors (Hotta and Baird, 1987), the angiotensin II receptors (Feige et al., 1987) and the steroid 17a-hydroxylase enzyme (Feige et al., 1987; Perrin et al., 1991). Since adrenocortical cells possess TGFP receptors (Cachet et al., 1988) and synthesize a latent form of TGFP (Feige et al., 1991; Keramidas et al., 1991), we have proposed that this factor could represent an autocrine regulator of adrenocortical differentiated functions (Feige et al., 1991; Feige and Baird, 1991). The biochemical nature of the latent TGFP complex secreted by adrenocortical cells is not fully characterized. In other tissues, mature TGFP appears to remain associated with its precursor part (usually named ‘latency protein’), even after the proteolytic processing of the pro-factor (Wakefield et al., 1989; Lioubin et al., 1991). In platelets, an additional protein (usually named latency associated protein) is covalently bound to the latency protein (Miyazono et al., 1988). In serum, latency of TGFP is mostly due to the association of the mature factor with a,M and it has been proposed that this complex could represent a clearance form of the factor (O’Connor-McCourt and Wakefield, 1987). Adrenocortical cells secrete substantial quantities of a,M (Kirshner et al., 1989; Shi et al., 1990), but it is not known whether a,M participates in the latency of TGFP in this tissue. How-

ever, the ability of this protein to neutralize the activity of TGFP in serum prompted us to examine its effects on adrenocortical cell functions. The integrity of the commercial preparation of a,M that we used in this study was checked by Dr. E. Delain (Institut Gustave Roussy, Villejuif, France) by electron micrography after negative staining with uranyl acetate (Boisset et al., 1989). This analysis revealed that the preparation was totally under a native form (S form as defined by Barrett (1981)). To our surprise, we observed that cu,M did not block but rather potentiated the inhibition of steroidogenesis induced by TGFP on bovine adrenocortical cells at 37°C. Even in the absence of TGFP, a,M inhibited cortisol production in a dose-dependent and time-dependent manner. Half-maximal inhibition was observed for an a,M concentration of g 2 mg/ml, within the range of its physiological concentration in plasma (2-4 mg/ml). The kinetics of steroidogenesis inhibition induced by a,M (l/2 inhibition reached after 20 h of treatment) was slower than that induced by TGFP (l/2 inhibition reached after 6-7 h of treatment). This could suggest that the mechanisms of action of these two regulators are different. However, since azM has been demonstrated to bind TGFp (O’Connor-McCourt and Wakefield, 1987; Huang et al., 1988), we investigated whether a,M effects on adrenocortical cells could be mediated by associated TGFB. Several observations confirmed this hypothesis. First, under our experimental conditions, (YAM and TGFp both inhibited the steroid hormone biosynthetic pathway at the same step, i.e. the steroid 17a-hydroxylation. Second, the inhibitory activity of a,M could be totally recovered in an ethanol-acid extract, as observed during the purification of TGFP (Frolik et al., 1983). Third, the inhibitory activity of the ethanol/acid extract of a,M could be totally neutralized by anti-TGFP antibodies. Fourth, these anti-TGF/3 antibodies which do not recognize the latent forms of TGFP, did not block the inhibitory activity of au,M. Taken together, these observations suggest that inhibition of adrenocortical steroidogenesis by bovine plasma cr,M is caused by associated TGFP. The difference in the kinetics of action of TGFP and (Y*M and the existence of an = 10 h

250

lag period before cu,M starts to inhibit cortisol production suggest either that the ~u,M-TGFP complexes are internalized through a pathway independent of TGFP receptors or that some activation step is taking place that renders a,Massociated TGFP biologically active. Internalization through the ctlM receptors is a possibility. The presence of cu,M receptors at the surface of adrenocortical cells has not been reported yet but the recent identification of the cu,M receptor as a lipoprotein receptor (Herz et al., 19X8; Strickland et al., 1990) makes it very probable. However, since n,M receptors only bind LY~M under an F form (i.e., azM that has reacted with a protease) (Imber and Pizzo, 1981; Ashcom et al., 1990), then this hypothesis would imply that the substoichiometrical a,M-TGF/3 complexes are under an F form. By analogy with other cell systems, activation of the latent cu,M-TGF/3 complexes could be caused by secreted proteases such as plasmin (Lyons et al., 1990) or could require contact with the cell surface (Dennis and Rifkin, 1991). Whether the biological activity of (u,M-TGFP complexes is specific for adrenocortical cells or is also observed in other cell types is not precisely known at the moment. O’Connor and Wakefield (1987) have reported that cu,M at a concentration of 0.2 mg/ml is not able to stimulate the growth in soft agar of human lung carcinoma AS49 cells and of NRK fibroblasts, whereas picomolar concentrations of TGFP do. Also, Danielpour and Sporn (1990) have reported that cu,M at a concentration of 0.2 mg/ml did not inhibit the growth of CCL-64 mink lung epithelial cells whereas picomolar concentrations of TGFP, and TGFP, completely inhibited it. In adrenocortical cells, an ru,M concentration of 0.2 mg/ml is not sufficient to inhibit steroidogenesis. It would thus be necessary to check whether higher concentrations of N?M can mimic the biological effects of TGFP on AS49, NRK or CCL-64 cells as they do on adrenocortical cells, in order to know whether these azM-TGFP complexes are biologically active on diverse cell types. From the dose-response studies shown in Figs. 2 and 3 one can extrapolate that 1 mg of the commercial preparation of azM contains about 0.05 ng of TGFP. Given the different molecular weights of these two proteins, the stoichiometry

of TGFP bound to cu,M can be estimated at 1 molecule of TGFP/7 X 10’ molecules of a,M. This is of course extremely substoichiometrical but one must keep in mind that the physiological concentration of cu,M in serum is in the range of 3-6 PM and that the doses of TGFP that inhibit adrenocortical steroidogenesis are in the range of 2-4 pM. Extremely substoichiometrical (Y?MTGFP complexes like those observed in the preparations of bovine plasma cu,M that we used in this study appear to be sufficiently abundant to trigger a TGFP-like effect on adrenocortical cells. These new observations imply that TGFP may also act on the adrenal cortex as an endocrine factor, delivered to the tissue as a complex with a,M. This would represent an additional regulatory mechanism to the autocrine regulatory loop previously proposed (Feige et al., 1991; Fcige and Baird, 1991). Acknowledgements We are indebted to Dr. Etienne Delain for his helpful electron microscope analysis of the cu,M preparation. We are grateful to Claude BlancBrude and lsabelle Gaillard for their helpful contribution to the preparation of adrenocortical cell primary cultures and to Sonia Lidy for her editorial help. This work was supported by the Institut National de la Sante et de la Recherche Medicale (INSERM U 2441, the Commissariat a I’Energie Atomique (DSV/DBMS/BRCE) and the Liguc Nationale Frangaise contre le Cancer. References Ashcom, J.D.. Tiller. S.E., Dickerson, K., Cravens, J.L., Argraves. W.S. and Strickland, D.K. (1990) J. Cell Biol. I IO. 1041-104X. Barrett, A.J. (1981) Methods Enzymol. 80, 737-753. Bernuau, D., Leg&s. L.. Lamri, Y.. Giuily, N., Fey, G. and Feldmann. G. (19X9) J. Exp. Med. 170. 34%34. Boisset. N., Taveau, J.-C., Pochon, F.. Tardieu, A.. Barray, M., Lamy, J.N. and Delain. E. (1989) J. Biol. Chem. 2h4. I2046 12052. Borth. W. and Luger, T.A. (19X9) J. Biol. Chem. 264, 5XlXp 5825. Brissenden, J.E. and Cox. D.W. (lYX2) Somat. Cell Genet. X, 2x’)-305. Chu, C.T., Rubenstein. D.S., Enghild. J.J. and Pizza S.V. (1991) Biochemistry 30, lS5l-1560.

251 Cachet, C., Feige, J.-J. and Chambaz, E.M. (1988) J. Biol. Chem. 263, 5707-5713. Danielpour, D. and Sporn. M.B. (1990) J. Biol. Chem. 265. 6973-6977. Dennis, P.A. and Ritkin, D.B. (1991) Proc. Natl. Acad. Sci. IJSA 88, 5X0-584. Dennis, P.A., Saksela. O., Harpel, P. and Rifkin, D.B. (19x9) J. Biol. Chem. 264, 7210-7216. Duperray, A. and Chambaz, E.M. (19X0) J. Steroid Biochem. 13, 135991364. Feige, J.-J. and Baird, A. (1991) Prog. Growth Factor Res. 3, 1033113. Feige, J.-J., Cachet. C. and Chambaz, E.M. (1986) Biochem. Biophys. Res. Commun. 139, 693-700. Feige. J.-J., Cachet, C., Rainey, W.E., Madani, C. and Chambaz, E.M. (1987) J. Biol. Chem. 262, 13491-13495. Feige, J.-J., Cachet, C., Savona, C., Shi, D.L., Keramidas, M.. Defaye, G. and Chambaz, E.M. (1991) Endocr. Res. 17, 267-279. Frolik, CA., Dart, L.L., Meyers, CA., Smith, D.M. and Sporn. M.B. (1983) Proc. Nat]. Acad. Sci. USA 80, 367636X2. Gaddy-Kurten, D., Hickey, G.J., Fey, G.H., Gauldie, J. and Richards, J.S. (1989) Endocrinology 125, 29X5-2995. Gauldie, J., Richards, C., Northemann, W., Fey, G. and Baumann, H. (1989) Ann. NY Acad. Sci. 557, 46-59. Gehring, M.R., Shiels, B.R., Northemann, W., de Bruijn, M.H.L., Kan, C-C., Chain, A.C., Noonan, D.J. and Fey, G.H. (1989) J. Biol. Chem. 262, 446-454. Herz, J., Hamann, U., Rogne, S., Myklebost, O., Gausepohl, H. and Stanley, K.K. (1988) EMBO J. 7, 4119-4127. Hotta. M. and Baird, A. (1986) Proc. Nat]. Acad. Sci. USA 83, 77Y5-7799. Hotta, M. and Baird, A. (1987) Endocrinology 121, 150-159. Hovi, T., Mosher, D.F. and Vaheri, A. (1977) J. Exp. Med. 145, 1580~158Y. Huang, J.S., Huang, S.S. and Deuel, T.F. (1984) Proc. Natl. Acad. Sci. USA 81, 342-346.

Huang, S.S., O’Grady, P. and Huang. J.S. (IYXX) J. Biol. Chem. 263, 153551541. Imber, M.J. and Pizzo. S.V. (1981) J. Biol. Chem. 256, X1348139. James, K. (1900) Immunol. Today 11, 163-166. Keramidas, M., Bourgarit, J.-J.. Tabone. E., Chambaz, E.M. and Feige, J.-J. (1YY1) Endocrinology 129, 517-526. Kirshner, N., Corcoran, J.J. and Erickson, H.P. (IYXY) Am. J. Physiol. 256, C79-C785. Lioubin, M.N., Madisen, L., Roth, R.A. and Purchio, A.F. (1991) Endocrinology 128, 2291-2296. Lyons, R.M.. Gentry, L.E., Purchio. A.F. and Moses, H.L. (1990) J. Cell Biol. 110, 1361-1367. Matsudo, T., Hirano, T., Nagasawa, S. and Kishimoto, T. (1989) J. Immunol. 142, 14X-152. Miyazono, K., Hellman, U., Wernstedt, C. and Heldin, C.-H. (1988) J. Biol. Chem. 263, 6407-6415. O’Connor-McCourt, M.D. and Wakefield, L.M. (1987) J. Biol. Chem. 262, 14090-14099. Perrin, A., Pascal, 0.. Defaye, G., Feige, J.-J. and Chambaz, E.M. (1991) Endocrinology 128, 3577362. Philip, A. and O’Connor McCourt, M.D. (1991) J. Biol. Chem. 266, 22290-22296. Ronne, H., Anundi, H., Rask, L. and Peterson, P.A. (1979) Biochem. Biophys. Res. Commun. X7, 330-336. Shi, D.L., Savona, C., Gagnon, J., Cachet, C., Chambaz, E.M. and Feige, J.-J. (1990) J. Biol. Chem. 265, 28X1-2887. Sottrup-Jensen, L. (1989) J. Biol. Chem. 264, 11539-11542. Sottrup-Jensen, L., Sand, O., Kristensen, L. and Fey, G.H. (1989) J. Biol. Chem. 264, 157X1-15789. Strickland, D.K., Ashcom, J.D., Williams, S., Burgess, W.H., Migliorini, M. and Argraves, W.S. (1990) J. Biol. Chem. 265, 17401-17404. Wakefield, L.M., Smith, D.M., Broz, S., Jackson, M., Levinson, A.D. and Sporn, M.B. (1989) Growth Factors 1, 203-218.