0013.7227/92/1311-0471$03.00/0 Endocrinology Copyright CL’ 1992 by The Endocrine

Vol. 131, No. 1 Prrnted in U.S.A.

Sowsty

Glucocorticoid Regulation of Transforming Growth Factor-P Activation in Urogenital Sinus Mesenchymal Cells* DAVID

R. ROWLEY

Department

of

Cell Biology, Baylor College of Medicine,

One Baylor Plaza, Houston,

Texas 77030

ABSTRACT The transforming growth factors-p (TGF-0s) regulate many aspects of cell proliferation and differentiation. The TGF-8 isoforms are produced by most cell types in the biologically inactive, latent form. The physiological relevance of latent TGF-p and the regulation of activation to the biologically active form are not well understood. Although expression of TGF-8 messenger RNA is regulated by the steroid hormone family, the mechanisms of hormonal regulation of TGF-@ activation have not been well studied. Fetal rat urogenital sinus organ cultures and derived mesenchymal cell lines (U4F, U4Fl) have been established in order to analyze the expression of growth and differentiation regulatory factors which may function in mesenchymal induction of epithelial differentiation. The U4Fl cell line in chemically defined medium upon supplementation with dexamethasone (10 nM1.4 KM), produced an activity which was growth inhibitory to PC-3 prostatic carcinoma epithelial cells. Analysis of phsicochemical prop-

erties and purification of activity demonstrated a 25.kDa protein was responsible for activity. The activity cross-reacted to antisera specific for TGF-@1,@2 and for TGF-02 exclusively, but not with antisera to rat interferon (n-P) or rat interleukin 6. Acid treatment of control (unsupplemented) conditioned medium and cultures supplemented with other steroid hormones produced identical levels of activated TGF-0 as nonacid-treated conditioned medium from dexamethasone supplemented cultures which did not increase levels of activity upon acid activation. Activity from the acid-treated control conditioned medium was neutralized by TGF-fi antibodies. These data suggest latent TGF8 is produced constitutively by U4Fl mesenchymal cultures in steroid unsupplemented medium and these cultures are induced by dexamethasone to activate identical levels of TGF-0. These observations may be relevant to understanding diverse aspects of glucocorticoid regulation of tissue function and suggests that TGF-b may be relevant to paracrine and autocrine growth regulation in the developing urogenital sinus. (Endocrinology 131: 471-478, 1992)

T

becomes an important question. Mechanisms involving physiological activation of latent TGF-8 complexes remain unknown. TGF-Ds are expressed as large pre-pro forms. The pro domain is believed to be important for proper folding of mature TGF-/3 and the pro domain homodimer serves as a binding protein (@l-LAP) for mature, latent TGF-0 (17). Mature TGF-/3 monomers are cleaved from the carborxyl terminus of the pre-pro precursor with high conservation of sequence homology (70-80%) between the @l-83 isoforms (l-3). The pro domain is less conserved (less than 50%) and differences in the association and size of pro domain binding proteins varies with the specific isoform (18). These data support the concept that differential composition of latency associated proteins is relevant to differential regulation of activation of TGF-P isoforms. Latent TGF-P may be activated by either acidic (pH 3) or basic (pH 9) conditions, or by heat, and treatment with various proteases (17, l-3). Although several regions of the pro domain have been identified as important for TGF-Pl processing, secretion, and activation (19), the specific mechanisms of TGF-fi activation and hormonal control of activation have not been well studied. Previous studies by our laboratory have addressed the role of paracrine factors in mesenchymal-epithelial interactions in the development of the prostate gland from fetal urogenital sinus. Since TGF-@. are expressed by mesenchymal cells (20) and regulate epithelial cell growth and differentiation, these growth factors may play a role in mesenchymal induction of epithelial phenotype in the prostate gland. Mesen-

RANSFORMING growth factor-@ (TGF-0) regulate various parameters of cell growth and differentiation. Three TGF-P isoforms (fll, 82, and 03) have been characterized in mammalian species (1). These isoforms belong to an extended gene family which includes mullerian inhibitory substance, the inhibins-activins, Xenopus Vgl, and the decapentaplegic gene in Drosophilia (2, 3). The diverse cellular activities regulated by TGF-P include wound healing (4), formation of cartilage and bone in development and in fracture repair (5), the synthesis of extracellular matrix components (6), regulation of immune function (7), and the growth and differentiation of ovarian granulosa (8) and theta cells (9). Equally diverse is the hormonal regulation of TGF0 expression. Expression of TGF-Pl and 02 is differentially regulated by retinoic acid and epidermal growth factor (EGF) in NRK-49F and A549 cells (10). Conversely, dexamethasone acted to decrease TGF-P expression (lo), and down-regulate TGF-/3 type I and II receptor complexes (11). Androgen and estrogen increases TGF-0 expression in osteosarcoma cells (12, 13). Tamoxifin regulated TGF-P2 expression in prostatic carcinoma cell line (14). Since TGF-/3 is produced primarily as the latent, nonbiologically active form by most cell types (15) and TGF-P receptors are universally expressed (16), the regulation of TGF-0 activation to the biologically active form Received December 16, 1991. Address all correspondence and requests for reprints to: David R. Rowley, Department of Cell Biology, Baylor College of Medicine, One Baylor Plaza, Houston, Texas 77030. * Supported by National Science Foundation Grant DCB-9018138.

471

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 22 November 2015. at 06:26 For personal use only. No other uses without permission. . All rights reserved.

472

GLUCOCORTICOID

INDUCED

chymal-epithelial interactions are central to the development and differentiation of many tissues. In several tissues including mammary gland and urogenital sinus, the effects of steroid hormones on epithelium are mediated through the mesenchymal-stroma compartment (21). The molecular mechanisms of these steroid regulated mesenchymal-epithelial interactions are not understood. To address the actions of possible paracrine-acting factors from stroma, we have reported previously on organ cultures of fetal rat urogenital sinus (22, 23) and derived mesenchymal cell lines (24). These studies have characterized the expression of a growth inhibitory activity which acted to stimulate secretory protein synthesis and alter epithelial phenotypic morphology (22-24). The physicochemical, biological, and immunological properties of this activity (provisionally termed urogenital sinusderived growth inhibitory factor, UGIF) were found to be unrelated to other negative growth factors and cytokines including the TGF-Ps (24). The biological properties of UGIF activity suggested it functions as a differentiation-regulatory factor. Recent studies in our laboratory have reported the spontaneous development of a variant mesenchymal cell strain (U4Fl) with an altered phenotype and UGIF secretory properties (24). To identify growth factors expressed by this cell strain, cultures have been adapted to serum-free media and the secretion of growth regulatory activity into conditioned medium analyzed. We report here the constitutive secretion of latent TGF-0 by the U4Fl cell strain in serumfree, chemically defined medium and the regulation of TGF/3 activation by dexamethasone at physiological concentrations. The U4Fl mesenchymal cell line may serve as a useful model to address the specific mechanisms and regulation of TGF-P activation in chemically defined conditions. Moreover, this cell line may prove useful in studies regarding the mechanisms of stromal-epithelial interactions in the developing prostate gland.

Materials

and Methods

Materials The following materials were purchased: Dulbecco’s modified Eagles medium (DMEM) and Hanks balanced salt solution (calcium and magnesium free; cmf) from Grand Island Biological Company (GIBCO, Grand Island, NY); Nu-Serum and EGF from Collaborative Research (Lexington, MA); fetal bovine serum and calf serum from Hazelton Research Products (Denver, CO), Flow laboratories (McLean, VA), and HyClone (Logan UT); testosterone from Steraloids (Wilton, NH); MCDB110 basal media, insulin, trypsin type II, crystal violet dye, trypan blue dye, dihydrotestosterone, dexamethasone, estradiol-17/3, progesterone, cycloheximide, and sodium lauryl sulfate from Sigma (St. Louis, MO); penicillin and streptomycin from GIBCO and Sigma; methanol, trichloroacetic acid, and disodium EDTA from Fisher (Fair Lawn, NJ); glacial acetic acid from Chempure (Curtin Matheson, Houston, TX); Tris (ultra pure), [methyl-3H]thymidine, and [?S] methionine from ICN Radiochemicals (Irvine, CA); aqueous counting scintillant (BCS, No. NBCS104) from Amersham (Arlington Heights, IL); Bio-Gel P-200 (100-200 mesh) gel filtration resin, glycine, and acrylamide from Bio-Rad (Richmond, CA); diethylaminoethyl cellulose (DE-52) ion exchange resin from Whatman Biosystems (Maidstone, England); 96. and 24.well culture plates (Nos. 76-003-05 and 76-033-05) f rom Flow (Linbro) (McLean, VA); 25 cm2 tissue culture flasks (No. 25100) from Corning (Corning, NY); 75 cm’ and 175 cm’ tissue culture flasks from Falcon (No. 3024) (Becton Dickinson, Oxnard, CA); dialysis tubing No. 3 (m.w. cutoff = 3,500)

TGF-B ACTIVATION

Endo. Voll31.

1992 No 1

from Spectrapore (Los Angeles, CA); TGF-@l (porcine and human) and TGF-/31-@2 and TGF-62 specific neutralizing polyclonal antibodies from R & D Systems (Minneapolis, MN); rat fibroblast-derived (ol-fl) interferon (INF) and neutralizing antisera from Lee Biomolecular (San Diego, CA); and neutralizing antisera to rat interleukin 6 (IL-6) was kindly provided by Dr. Jack Gauldie (Dept. Pathology, McMaster University, Hamilton, Ontario, Canada).

Cell culture Stock U4Fl mesenchymal cells were cultured in medium Bfs (DMEM 90%; calf serum 5%, Nu-Serum 5%, insulin 5 Kg/ml, testosterone 0.5 pg/ml, penicillin 100 U/ml, and streptomycin 100 fig/ml) according to procedures we have reported previously (24). A serum-free, chemically defined medium was developed empirically, based on parameters of cell viability and relative growth rates, Chemically defined media consisted of MCDB-110 basal media (Sigma) supplemented with insulin (5 pg/ ml), transferrin (5 pg/ml), selenium (5 rig/ml), and EGF (0.1 pg/ml) hereafter referred to as medium M. Steroid hormone supplementation was according to conditions discussed in the text. Conditioned medium was collected each 48 h, clarified by centrifugation, and was either used immediately, or frozen and stored at -20 C until subsequent use. The PC-3 prostatic carcinoma (epithelial) cell line (ATCC CRL 1435) was received from American Type Culture Collection (Rockville, MD) and was cultured in DMEM 93%, fetal bovine serum 7%, penicillin (25 U/ml), and streptomycin (25 pg/ml). Medium was replaced each 2-3 days, Monolayer cultures were passaged routinely at confluence by brief exposure to trypsin-EDTA (0.25% trypsin, 0.025% EDTA in cmf Hanks). Cell viability was established by the trypan blue dye exclusion test with an improved Neubauer type hemocytometer. All cultures were routinely tested for mycoplasma contamination (MycoTect Kit, GIBCO).

p’H]Thymidine

incorporation

assay

A miniassay procedure was used following steps we have reported previously (24). PC-3 cells were seeded at 6 X lo3 cells/l48 Fl/well in 96.well plates and allowed to attach for 24 h. Wells received a O-52 ~1 aliquot of sample (200 pi/well final vol) and allowed to incubate for an additional 24 h. Cultures were pulsed with [3H]thymidine (2 pCi/ml) during the final 3 h of incubation. Cell monolayers were fixed in situ with methanol:acetic acid (3:l) (200 pi/well) for 10 min at 22 C., washed with 100% methanol (5 min,. 22 C) followed by a wash with 5% trichloroacetic acid (5 min, 22 C), and a 3~ wash with methanol (22 C, 200 Fl/well). Plates were allowed to air dry at 37 C for 20-30 min. Plates could either be stored indefinitely at 22 C or processed immediately. The monolayers were hydrolyzed by the addition of 1 N NaOH (200 pi/well, 5 min, 22 C). Sample aliquots (180 ~1) were added to scintillation vials containing 180 ~1 of 1 N HCl to neutralize pH. Radioactivity was determined by scintillation counting. For analysis of activity, average incorporation of [3H]thymidine produced by each assayed sample was calculated as a percent change of the average [3H]thymidine incorporation produced by volume-matched fresh control media (identical hormonal supplementation) according to parameters we have reported previously (22-24). All results were analyzed for significance by the Student’s t test. In experiments using antibodies to various cytokines and in the testing of activation of latent TGF-@, identical concentrations of subsaturable activity were used and the incubation conditions for each antibody confirmed for neutralization of the natural antigen as we reported previously (24).

Ion exchange chromatography Biological activity from U4Fl cell conditioned medium was initially fractionated using ion exchange chromatography. Conditioned media (170 ml) was diluted (1:l) with 40 rnM ammonium carbonate buffer, pH 9.0, and applied to a diethylaminoethyl cellulose (DE-52) column (1.5 X 9.5 cm) equilibrated in 20 rnM ammonium carbonate, pH 9.0, with a hydrostatic pressure of 30 cm. The column was washed with S-bed vol 20 rnt.4 ammonium carbonate and step eluted with 150 ml of 20 mM ammonium acetate-acetic acid, pH 4.0, followed by 150 ml of 1 M acetic acid, pH 2.5. The pH and absorbance at 280 nm (AZ80) were determined

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 22 November 2015. at 06:26 For personal use only. No other uses without permission. . All rights reserved.

GLUCOCORTICOID

INDUCED

for each fraction. Aliquots (25 ~1) of each fraction were vacuum dried, resolubilized in medium Bfs (65 PI), and assayed for growth regulatory activity using the [3H]thymidine incorporation assay as described.

Gel filtration

chromatography

Samples were prepared by dialysis (Spectrapore 3 tubing) against 1 M acetic acid (33 vol, changed at 24 h) for 48 h at 2 C. The samples were dried (Savant Speed Vat) and used immediately or stored at -20 C until analyzed. Samples were solubilized in 1 M acetic acid (1 ml) and applied to a Bio-Gel P 200 column (1.4 x 70 cm) equilibrated in 1 M acetic acid at room temperature. Aliquots (100 ~1) from each fraction were vacuum dried and assayed for growth inhibitory activity using the [3H]thymidine incorporation miniassay. Protein size standards and relative molecular weights were: BSA (66 kDa); carbonic anhydrase (29 kDa); soybean trypsin inhibitor (20.1 kDa); and lysozyme (14.3 kDa).

Preparative

and analytical

gel electrophoresis

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDSPAGE), followed modifications of the procedures of Laemmli (25). Samples were vacuum dried and resolubilized in Laemmli sample buffer (nonreducing conditions), warmed (10 min, 70 C), and 100 ~1 aliquots loaded per gel lane. The running gel was made to 15% acrylamide. Samples were electrophoresed through the stacking gel at 40 mA and through the running gel at 70 mA continuous current. Molecular weight standard marker proteins were: lysozyme (14.3 kDa); soybean trypsin inhibitor (20.1 kDa); carbonic anhydrase (29 kDa); ovalbumin (43 kDa); and BSA (66 kDa). For preparative SDS-PAGE the position of marker proteins was visualized by incubating marker protein gel lanes with cold KC1 (0.25 M, 2 C) followed by cold water (2 C). Parallel lanes with sample proteins were cut in successive 3-mm wide transverse slices. Each gel slice was rinsed 3~ in double distilled water and diced with a methanol-washed razor blade to 1 mm3 cubes. The diced gel cubes were washed 3~ in 1 ml double distilled water for 5 min per wash with periodic vortex. The washed diced gel cubes were placed in a siliconized 10 X 75 glass tubes containing either 1 M acetic acid or growth medium Bfs and incubated 16 h, at 22 C on a rocker plate to extract proteins from gel cubes. Extracts were separated from gel cubes and extracts in Bfs media sterilized by membrane filtration through a 0.22-pm Millex filter and assayed for [ Hlthymidine incorporation in PC-3 cells to detect growth regulatory activity. Extracts in 1 M acetic acid were vacuum dried in sterile microfuge tubes and used both for the [3H]thymidine incorporation assay (vacuum dried sample resolubilized in sterile Bfs medium) and for analytical SDS-PAGE (dried sample resolubilized in Laemmli sample buffer). Analytical SDS-PAGE followed procedures of sample and gel preparation according to the procedures of Laemmli (25). Analytical SDS-PAGE gels were processed for visualization of proteins using silver staining procedures.

TGF-fi

ACTIVATION

ml), transferrin (5 pg/ml), selenium (5 rig/ml), and EGF (0.1 pg/ml) (medium M). Medium M formulation maintained growth rates and phenotypic morphology of U4Fl cells identical to serum-containing media reported for U4Fl cells previously (24). Conditioned medium (48 h) from each supplementation regime was assayed for expression of growth regulatory activity on target PC-3 prostatic epithelial cells based on [3H] thymidine incorporation assays following procedures we have reported previously (22-24). Growth regulatory activity was not detected from U4Fl cells in the unsupplemented MCDB-110 medium or in MCDB-110 plus any of the basal supplementations (insulin, transferrin, selenium, EGF) tested for optimal U4Fl cell proliferation. Unlike cultures in serumcontaining medium, significant levels of the previously described UGIF activity could not be identified in the crude conditioned medium under these conditions. Since mesenchymal-epithelial interactions leading to prostate phenotype are regulated by steroids, U4Fl culture medium M was supplemented with various concentrations (10 nM to 1.4 PM) of testosterone, dihydrotestosterone, estrogen, progesterone, and dexamethasone both separately and in combination and conditioned medium assayed for activity. A growth inhibitory activity was observed exclusively in conditioned medium from U4Fl cultures in medium M supplemented with dexamethasone (Fig. 1, MIED). Expression of this activity was enhanced to a minor extent by the addition of testosterone to media formulations containing dexamethasone (MIEDT). Dose-response studies with these samples indicated the percent inhibition of [3H]thymidine incorporation in PC-3 cells was a linear and saturable (max inhibition = 50-60% relative to volume and media composition matched controls) function of the conditioned medium concentration tested and was reversible (data not shown). The activity was stable to heat (80 C, 10 min), acid (dialysis

Results Expression conditioned

a negative growth medium

of

activity in serum-free M

A serum-free, chemically defined medium to support U4Fl mesenchymal cells was developed empirically based on cell viability (trypan blue dye exclusion test, crystal violet vital dye incorporation), growth rates (population doubling times), and long-term culture survival (greater than 8 weeks survival, multiple passages). MCDB-110 basal media, initially formulated to support human diploid fibroblasts was found most optimal for U4Fl mesenchymal cells. Progressive supplementation testing with hormones, carrier proteins, vitamins, nonessential amino acids, trace elements, and growth factors, formulated an optimal supplementation. Optimal minimal supplementation of MCDB-110 consisted of insulin (5 pg/

MI

MIE

hIlED

MIET

MIEDT

FIG. 1. Secretion of growth inhibitory activity from U4Fl mesenchyma1 cells in chemically defined medium. Replicate cultures in triplicate were exposed to basal medium MCDB-110 alone (M) and in combination with the following supplementations (I): ITS (5 fig/ml insulin, 5 pg/ml transferrin, 5 rig/ml selenium); (E): 0.1 pg/ml EGF; (D): 1.4 FM Dex; (T): 1.4 pM testosterone. Cultures were allowed to condition medium for 48 h, cells counted, and medium (52 ~1 aliquot) assayed for growth regulatory activity for PC-3 cells following the [“Hlthymidine incorporation assay as described in Materials and Methods. Shown is the average (n = 3) percent inhibition of [3H]thymidine incorporation (relative to control) and standard error (bar) of the mean for each regime. Columns with a to b labels represent statistically significant differences (P < 0.01). No significant differences existed in the number of U4Fl cells per dish between the treatment regimes (data not shown).

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 22 November 2015. at 06:26 For personal use only. No other uses without permission. . All rights reserved.

474

GLUCOCORTICOID

INDUCED

TGF-B

IIS. 1 M acetic acid, 16 h), and unlike the previously described UGIF activity, was stable upon exposure to SDS and exhibited different levels of saturable activity with the PC-3 cell line. Activity was neither cytotoxic or cytostatic in mechanism of action. In contrast, activity was neutralized by reducing agents (dithiothreitol) and by exposure to trypsin (data not shown).

Endo. Vol131.

ACTIVATION

0.12 8.0

0.10

6.0

and identification

of

growth

inhibitory

activity

To assessmolecular weight and physicochemical properties, fractionation of activity using ion exchange chromatography, gel filtration chromatography, and preparative SDSPAGE gel electrophoresis was done. Conditioned medium (150 ml) was mixed (1:l) with 40 mM ammonium carbonate (pH 9) and applied to a DE-52 ion exchange column (1.5 x 9.5 cm) equilibrated in 20 mM ammonium carbonate (pH 9). The column was washed with 20 mM ammonium carbonate (pH 9) and step eluted with 20 mM ammonium acetate (pH 4.0) followed by 1 M acetic acid (pH 2.5). As shown in Fig. 2, the majority of protein (absorbanceat 280 nm, A2HO) eluted in the pH 4.0 step elution. A peak of growth inhibitory activity (l/cpm plotted to show inhibition of [3H]thymidine incorporation as a peak of activity) eluted in 1 M acetic acid (pH 2.5). Fractions from the peak region of activity were pooled (fractions underlined by bar, Fig. 2) and concentrated to 1 ml by vacuum drying. The activity was fractionated further by gel filtration chromatography (BioRad P-200) equilibrated in 1 M acetic acid asshown in Fig. 3. The majority of protein (AZHo)eluted in the void volume and a single peak of activity eluted consistent with a molecular size of approximately 25 kDa. Since the biological activity was stable to exposure to SDS, but sensitive to reducing agents, preparative SDS-PAGE in nonreducing conditions was used to assessdirectly the molecular weight of the growth inhibitory activity concentrated r

12,

10.5

T

1'0

3

2. 2

0.0 0

10

20

30

40

2

50

Fraction

Number

FIG. 2. Ion exchange chromatography of growth inhibitory activity from U4Fl cells. Conditioned medium from U4Fl cells cultured in medium MIEDT was prepared and chromatographed through a DE-52 ion exchange column as described in Materials and Methods. Aliquots (100 ~1) from each fraction were vacuum dried, resolubilized in sterile medium Bfs, and analyzed for incorporation of [“Hlthymidine into PC3 cultures as described. Total protein (A& and pH were determined from each fraction. Presented is the reciprocal of cpm (l/cpm) (incorporated [JH]thymidine) (closed circles) per fraction to show inhibition as a peak of activity. Open circles, total protein; dashed line, pH. Fractions of bioactivity were pooled (bar region).

0.08

z -

0.06

2 4”

0.04

4.0

Fractionation

1992 No 1

0.02

Fracl~on

Number

3. Gel filtration chromatography of growth inhibitory activity from U4Fl cells. Pooled fractions of bioactivity from the ion exchange step were concentrated to 1 ml by vacuum drying and applied to a P200 (Bio-Rad) gel filtration column equilibrated in 1 M acetic acid as described in Materials and Methods. Aliquots (100 ~1) from eluted fractions were analyzed for bioactivity as described in Fig. 2. Reciprocal of incorporated cpm (l/cpm) is shown for each fraction to show inhibition of [“Hlthymidine incorporation as a peak of activity (closed circles). Total protein (AZx,,) (open circles). Positions of molecular size calibration markers are shown at top of graph: BSA (66 kDa); chymotrypsinogen (27.5 kDa); soybean trypsin inhibitor (20.1 kDa); lysozyme (14.3 kDa). Bar, Fractions pooled. FIG.

from gel filtration chromatography. The biologically active fractions from gel filtration chromatography were pooled (region underlined by bar, Fig. 3), divided in half, vacuum dried, resolubilized in Laemmli sample buffer, and electrophoresed through a 15% polyacrylamide gel according to procedures in Materials and Methods. One lane was stained according to the silver method and the companion lane was processed for extraction of proteins from gel slices. The preparative lane was sliced into consecutive 3-mm sections. Sections were diced, processed, and extracted overnight to elute constituent proteins according to procedures described in Materials and Methods. Figure 4 shows the stained protein profile and the corresponding biological activity from companion lane gel slices. Significant biological activity was associatedprimarily with a band of protein with an M, of 25 kDa, in agreement with molecular size of activity eluted from gel filtration chromatography. To confirm molecular weight, the extracted protein (extracted in 1 M acetic acid) was vacuum dried and reelectrophoresed using SDS-PAGE and visualized by silver staining. As illustrated in Fig. 5, the extract consisted of a single protein specieswith a corresponding M, of 25 kDa in nonreducing conditions. The physicochemical and biological properties of the activity including the secretion from mesenchymal cells were in partial or full agreement with those properties of transforming growth factor type betas (TGF-P), fibroblast derived rat interferons (rat-a and /3-INF), and rat interleukin 6 (IL-6). A seriesof neutralizing antibodies were tested for cross-reactivity to specifically identify the factor responsible for the activity. Figure 6 shows that antibodies neutralizing to rat INFs and rat IL-6 demonstrated no cross-reactivity. Conversely, an affinity purified antisera neutralizing to both the TGF-PI and fl2 isoforms (+abTGF) neutralized biological activity of both conditioned medium and the 25-kDa protein. To assessspecific isoform, an antisera specific for TGF-02 (+abB2) was tested and acted to neutralize activity to control

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 22 November 2015. at 06:26 For personal use only. No other uses without permission. . All rights reserved.

GLUCOCORTICOID Mr

I

INDUCED

TGF-B ACTIVATION

475

I

CM

+ablNF

+ablLG

+abTGF

+abB2

P25

+abTGF

6. Effects of cytokine specific antibodies on the growth inhibitory activity from U4Fl cells. Conditioned medium aliquots (CM) and purified 25 kDa protein (P25) from U4Fl cultures were incubated with antibodies neutralizing to rat fibroblast-derived (a! and fl) interferons, to rat interleukin 6, to transforming growth factors pl and p2, and neutralizing specific for transforming growth factor p2. Shown is the percent inhibition of [3H]thymidine in PC-3 cultures relative to control. Values equal mean of n = 3 separate determinations + SE (bars). Values with a to b labels indicate sianificant differences (P < 0.01). CM, U4Fl conditioned medium; + abINF, CM plus antisera to rat fibroblastderived interferon (12.5 U neutralizing activity/ml); + ab IL-6, CM ~1~s antisera to rat interleukin 6 (1:20 dilution): +ab TGF. CM incubated with neutralizing antisera to TGFPl,PL (40 fig/ml); +ab p2, CM incubated with neutralizing antisera to TGF-@2 (40 fig/ml); P25, 25 kDa nrotein extracted from nrenarative SDS-PAGE eels (Fies. 4 and 5); gab TGF, P25 incubated with antisera to transforming growth factor pl,p2 (40 fig/ml). FIG.

0

10

20

% Inhibition

30

40

[3H] Thymidine

50

60

70

incorporation

4. Preparative SDS-PAGE of growth inhibitory activity from U4Fl cells. The pooled fractions from the gel filtration chromatography step (Fig. 3) were vacuum dried, solubilized in Laemmli sample buffer, and electrophoresed through a 15% polyacrylamide gel (nonreducing), the gel processed and proteins extracted from gel slices as described in Materials and Methods. Extracts were assayed for bioactivity (percent inhibition of [3H]thymidine incorporation relative to control into PC3 cells) as described. Shown is a silver staining profile of proteins from a companion lane aligned with the plot of extract bioactivity from each gel slice. Values equal mean of n = 3 separate determinations + SE (bars). The position of molecular weight markers are shown on the left (M,). Lysozyme: (14.3 kDa); soybean trypsin inhibitor: (20.1 kDa); carbonic anhydrase: (29 kDa); ovalbumin (43 kDa); and BSA: (66 kDa). FIG.

levels. These data confirmed that the growth inhibitory activity produced from dexamethasone-treated U4Fl cells could be ascribed to TGF-82. Glucocorticoid-induced

:\ ! ::.,

::

,’

: : : .;

66,

; . . “, \ _i: ‘.,“\i

45.;:.

.:f::.jj i’

\:‘

J .:: f ?

2OW’

:

df,

FIG. 5. Analytical SDS-PAGE of bioactive proteins extracted from preparative SDS-PAGE gels. Extract (in 1 M acetic acid) from the bioactive gel slice shown in Fig. 4 was vacuum dried, resolubilized in Laemmli sample buffer, and electrophoresed through a 15% polyacrylamide gel (nonreducing) as described in Materials and Methods. Proteins were visualized by the silver staining method. A single protein band at the 25 kDa position was detected. Molecular weight markers are as indicated in Fig. 4 and relative positions are as indicated (M,); df, dye front.

activation

of expressed TGF-/3

Since expression of TGF-P biological activity was observed in U4Fl cell conditioned medium from cultures exposed to pharmacological concentrations of glucocorticoid (Fig. l), experiments were conducted to address physiological concentrations, Figure 7 summarizes these results. Repilcate U4Fl cultures were exposed to steroid-free medium M (control) or medium M containing 10 nM concentrations of either dihydrotestosterone (DHT), dexamethasone (Dex), estradiol17p (E2), progesterone (P4), or combinations of these steroids. As shown in Fig. 7, significant TGF-/3 biological activity (inhibition of [3H]thymidine incorporation) was observed exclusively in conditioned media from cultures exposed to 10 nM Dex. The additional inclusion of androgen (DHT or testosterone) with Dex produced no statistically significant increasesin the secreted activity (data not shown). Since TGF-P is produced primarily as the latent form by most cells, we next determined whether Dex was inducing an increase in the secretion of activated TGF-0 protein over controls, or whether Dex was inducing an activation of constitutively secreted latent TGF-P. To determine this, conditioned medium from each steroid regime (10 nM) was exposed to acid conditions (pH 3.0, 10 min, 22 C, and neutralized to pH 7.0) to fully activate putative latent TGFP in samples.Volume matched aliquots were used at concentrations determined previously in linear dose-responsecurves

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 22 November 2015. at 06:26 For personal use only. No other uses without permission. . All rights reserved.

GLUCOCORTICOID

476

Dex

BIT

E2

P4

M

INDUCED

TGF-p

ACTIVATION

Dex

IM

Endo. Voll31.

.M

DW

+ aud

FIG. 7. Production of TGF-fl biological activity by U4Fl cells cultured in 10 nM concentrations of steroid hormones. Replicate U4Fl cultures were cultured in medium M plus 10 nM concentration of either Dex, DHT, estradiol-170 (E2), progesterone (P4), or unsupplemented control medium M (M). Medium conditioned for 48 h was collected and volume matched aliquots assayed for inhibition of [“Hlthymidine incorporation in PC-3 cells as described in Materials and Methods. To detect latent TGF-P, conditioned medium from unsupplemented control (M) was titrated with acid (10 N HCl) to pH = 3.0 for 10 min, room temperature and neutralized to pH 7.0 with base (10 N NaOH) to activate putative latent TGF-@ and identical volume aliquots assayed for growth inhibitory activity (M + acid). To determine if observed latent activity was ascribed to TGF-0, acid activated medium was incubated with the antisera to TGF-@1,@2 (40 pg/ml) (M + acid + abTGF). Values = average of n = 3 separate determination f SE (bars). The Dex and M + acid (as indicated) and Dex + acid (data not shown) levels of activity were statistically identical and significantly different from the other samples analyzed (DHT, E2, P4, M) (P < 0.01).

to be subsaturable in activity and were assayed for biological activity (inhibition of [3H]thymidine incorporation). As shown in Fig. 7, acid-activated conditioned medium from steroid-free control cultures (M, + acid bar) demonstrated activity at statistically identical (P < 0.01) levels to the nonacid-treated conditioned medium from Dex-treated cultures. Acid-activation of the Dex-treated conditioned medium produced no statistically significant increases in biological activity, indicating the TGF-P in conditioned medium from Dex-treated cultures was in the mature, fully-activated form. The acid-activated biological activity in hormone-free control conditioned medium (M, + acid bar) was neutralized to control levels with neutralizing antibodies to TGF-P (+abTGF, Fig. 7), indicating the acid-induced activity was due to TGF-/?. Similar experiments using acid-activated conditioned medium from cultures treated with 10 nM androgen (DHT), estradiol-170 (E2), or progesterone (P4) produced identical results as shown in Fig. 8. Upon acid-activation, conditioned medium from U4Fl cultures treated with each hormone regime produced statistically identical (P < 0.01) levels of biological activity as compared to conditioned medium from Dex-treated cultures (Dex, Fig. 8). As shown in Fig. 9, the secretion of latent/active TGF-/3 resulted from.& noua synthesis since the addition of cycloheximide decreased levels of TGF-0 in Dex-treated cultures. Moreover, when Dex was added to control-conditioned medium (M) (conditioned in the absence of Dex), no increase in bioactivity was observed (Fig. 9), indicating that Dex did not activate TGF-P through direct interaction with latent TGF-P. Preliminary studies showed an induction of bioactivity when conditioned medium from Dex + cycloheximide treated cultures (no

E2

1992 No 1

P4

+ acid

FIG. 8. Acid activation of conditioned medium from steroid treated U4Fl cultures. Conditioned medium from U4Fl cultures treated with the steroid regimes (10 nM) shown in Fig. 7 were activated by acid as described in Fig. 7 and identical volume aliquots (subsaturable activity concentrations) assayed for inhibition of [“Hlthymidine incorporation in PC-3 cultures as described and levels of activity compared to nonacid activated conditioned medium from Dex-treated U4Fl cultures. (Dex): dexamethasone, non-acid treated; (M): unsupplemented control medium, acid treated = + acid; (DHT): dihydrotestosterone, acid treated = + acid; (E2): estradiol-17fl, acid treated = + acid; (P4): progesterone, acid treated = + acid. Values equal mean of n = 3 separate determinations f SE (bars). No statistically significant differences (P = 0.1) were detected between any regime.

, Dex

Dex+chx,

acid dialysis

M

lJ-lL-5 +Dex

1 1 m,x with Dex+chx

FIG. 9. Effects of cycloheximide on TGF-P activity. U4Fl cells were cultured in medium M plus Dex (1.4 pM) either with or without cycloheximide (+chx) (1 pg/ml) for 48 h. Conditioned medium was dialyzed us. 1 M acetic acid to remove cyclohexamide and samples assayed for inhibition of [“Hlthymidine incorporation as described in Materials and Methods. The direct effects of Dex were tested by assaying previously conditioned unsupplemented medium M (M) either plus or minus the addition of Dex (+ Dex under bracket) (20 nM) prior to assay. For mixing experiments, medium conditioned in the presence of Dex (20 nM) and cyclohexamide (1 pg/ml) (Dex+chx under bracket) was added 1:l to unsupplemented conditioned medium M (M) (1:l mix under bracket). Values equal the mean of n = 3 determinations + SE (bars). Values of (Dex) us. (Dexfchx) and of(M) vs. (M, 1:l mix with Dex+chx) were significantly different (P < 0.01).

secreted TGF-/3) was mixed with conditioned medium (M) from control cultures (secretion of latent TGF-0) as shown in Fig. 9. These results are consistent with the possibility that glucocorticoid may act to release a stored protease which, in turn, activates latent TGF-@ present in the untreated controlconditioned medium. Additional preliminary studies tested a variety of protease inhibitors to examine whether the respective protease inhibitor could block the formation of activated TGF-/3 in the conditioned medium. Bestatin, pepstatin A, leupeptin, aprotinin, and phenylmethylsulfonyl-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 22 November 2015. at 06:26 For personal use only. No other uses without permission. . All rights reserved.

GLUCOCORTICOID

INDUCED

fluoride were added separately at subtoxic concentrations in addition with glucocorticoid to U4Fl cultures. None of these protease inhibitors affected apparent levels of activated TGF/3 under the conditions as tested, which may not have provided for optimal protease inhibitor activity (data not shown). The release of a protease is one of several possible mechanisms by which glucocorticoid may induce TGF-P activation. Determination of specific mechanisms awaits further study and clarification. Discussion Data presented in this manuscript suggests that glucorticoid regulates the mechanism of TGF-/I activation by U4Fl mesenchymal cells cultured in chemically defined media. U4Fl mesenchymal cells constitutively express and secrete TGF-/3 in the latent, nonbiologically active form in hormonefree culture conditions, or in cultures treated with either estrogen, progesterone, or androgen. Dex-treated U4Fl cultures exhibited identical levels of TGF-P biological activity; however, this activity was in the fully activated form. These data suggest that Dex acted to induce activation of latent TGF-/? rather than increase expression of TGF-/3 over constitutively expressed levels in controls. The identity of the growth inhibitory activity in Dex-treated U4Fl cell conditioned medium as TGF-/? was suggested by the physicochemical properties of the activity and the molecular weight of the substantially purified protein in nonreducing conditions. Identity was confirmed by cross-reactivity with a TGF-PI $2 neutralizing antisera. Additional experiments with antibodies exclusively specific for TGF-/32 suggests the activity described here is due to the TGF-02 isoform. Experiments described in this manuscript also outline a rapid three-step purification protocol for mature TGF-0 homodimer which should be useful for assessing TGF-P biological activity produced by a variety of cell cultures in serum-free conditions. Specific mechanisms for the physiological activation of latent TGF-P to the biologically active homodimer are not understood (l-3, 18, 19). TGF-Bs are expressed as a pre-pro precursor which is subsequently processed, secreted as a latent complex composed of the mature TGF-fi homodimer, a homodimer of the pro-domain (PI-LAP), and other binding proteins that are less well defined and vary with the specific TGF-6 isoform (17-19). The specific and characteristic differences of the latent TGF-/3 isoforms are not well-characterized, but are believed to be important for differential regulation of their biological activity (18, 19). Understanding the differential regulation of cellular activation of latent TGF-fl isoforms is necessary to understanding the overall regulation of TGF-/3 biological activity. Together with understanding hormonal regulation of message expression, the analysis of the hormonal regulation of TGF-/3 activation addresses another step in the overall regulation of actual TGF-P biological activity. The steroid hormone family is apparently the most widely reported regulator of TGF-P message expression (l3, 10-14). It is likely that an array of steroid hormones coordinately regulate through different mechanisms not only TGF-/3 gene expression, but the regulation of TGF-P synthesis and differential activation of latent TGF-/3 complexes as well.

TGF-P

ACTIVATION

477

Indeed, MCF-7 cells produced activated TGF-01 constitutively and antiestrogens acted to increase secretion of both the latent and activated forms of TGF-/3 but did not increase steady state levels of transcripts, suggesting that antiestrogen induction resulted from an increased rate of translation as well as an induction of activation of some but not all of the secreted TGF-fl (26). In comparison, primary cultures of keratinocytes produced high levels of activated TGF$2 in response to retinoic acid owing to a major increase in levels of P2 transcripts (due to posttranscriptional events) as well as increased activation (25% of TGF-/32 in active form) over untreated control (