0013.7227/92/1316-2503$03.00/0 Endocrinology Copyright C 1992 by The Endocrine

Tumor Necrosis Theta-Interstitial ROBERT Department

J. ZACHOW, of

Physiology,

Vol. 131, No. 6 Prrnted in U.S.A.

Society

JOSEPH University

Factor-a! Induces Cells in Vitro* S. TASH, of

AND

PAUL

Kansas Medical

Clustering

in Ovarian

F. TERRANOVA

Center, Kansas City, Kansas 66160-7401

ABSTRACT Tumor necrosis factor-a (TNF) has been implicated in the regulation of steroidogenesis in theta-interstitial cells (TIC). The purpose of this study was to evaluate any change in TIC morphology during the time course of TNF-induced inhibition of LH-stimulated androstenedione production. Ovaries from immature hypophysectomized rats were enzymatically digested and highly purified TIC were obtained by density gradient centrifugation. TIC treated with TNF (0.1-10 rig/ml) demonstrated distinct clustering in the presence and absence of LH (50 rig/ml). The number of clusters and the mean area per cluster were greatest after 4 days as a result of treatment with 1 or 10 ng TNF/ml. In addition, a dose-dependent inhibition of LH-supported androstene-

dione production was induced by TNF. TNF also inhibited LH-induced androstenedione in TIC after 2, 4, or 6 days of continuous LH treatment, and TIC clustering still occurred. TIC clustering was impeded by the protein kinase inhibitor H7 at 10 FM; however, the protein kinase inhibitor, HA 1004 (5 PM), did not inhibit TNF-induced clustering in TIC. Since H7 blocked TNF induced clustering, but did not block TNF inhibition of LH stimulated androstenedione synthesis, it is suggested that alternate signal transduction pathways for TNF induced inhibition of LH-stimulated androstenedione and stimulation of clustering of TIC may exist. The results also indicate that the TNFinduced TIC clustering may be independent of the TNF-induced inhibition of LH-stimulated androstenedione production and states of LHinduced differentiation of TIC. (Endocrinology 131: 2503-2513, 1992)

T

stimulated aromatase in cultured granulosa cells from rats, but TNF had no effect on aromatase activity already present in granulosa cells. Whether TNF inhibits further induction of androgen in TIC is not known. LH-stimulated androgen production in TIC is also abrogated by epidermal growth factor (EGF), in a manner similar to that exerted by TNF (15). Interestingly, whereas EGF inhibited LH-stimulated androgen it also induced morphological changes in TIC in vitro, whereby TIC aggregated into multilayered clusters (16). To date, we could find no evidence of studies showing the effects of TNF on morphology of steroid-producing cells of the ovary. Conversely, there are numerous examples of TNF-induced morphological changes in other cell types and epithelial ovarian cancer cell lines (17-20). Using 14 cell lines, Rosen et al. (19) showed that TNF induced motility of 12 of the 14 cell types in vitro. Two human ovarian cell lines of epithelial origin were also induced by TNF to migrate in vitro (19). The purpose of the present experiments was to determine any changes in TIC morphology during the period of TNF inhibition of LH stimulated androstenedione synthesis and attempt to determine whether any TNF-induced change in morphology correlated with inhibition of androstenedione production. It was also of interest to determine if TNF either prevented further increases of already established LH-stimulated androstenedione in a time dependent manner, or completely inhibited LH-stimulated androstenedione production, independent of the duration of prior exposure to LH.

UMOR necrosis factor-a (TNF) is a 17.3 kilodalton peptide which, in its traditional setting, is secreted by macrophages as a result of lipopolysaccharide/endotoxin stimulation (1). Classically, TNF has cytotoxic and growth inhibitory effects, as well as growth stimulatory effects on numerous cell types (2). Furthermore, TNF has been localized to the ovaries of rat, cow (3), human (4, 5), and rabbit (6, 7), and to specific ovarian cell types, such as antral granulosa cells, ovarian macrophages (3), and thecal cords (3). TNF has distinct effects on theta and granulosa cells within the ovary (8-12). Previous studies have shown inhibitory effects of TNF on LH/humanCG (hCG)-stimulated androsterone production in theta-interstitial cells (TIC) from immature, hypophysectomized rats (9), and porcine Leydig cells (13) in vitro. The inhibitory actions of TNF have been proposed to occur at locations proximal to CAMP generation (presumably at the level of adenylate cyclase) in TIC (9), Leydig (14), and granulosa cells (8). However, experimental evidence also supports the possibility for TNF actions at signal transduction loci distal to, or independent of, CAMP generation in Leydig cells (13) and granulosa cells (11). In the granulosa cells of rats, TNF has been shown to inhibit FSH-(8) and LH-(11) stimulated progesterone and estrogen biosynthesis, as well as FSH-stimulated aromatase activity (10). Time-course studies from Emoto and Baird (10) have shown that TNF inhibited the further induction of FSHReceived May 21, 1991. Address correspondence and requests for reprints to: Dr. Paul F. Terranova, Department of Physiology, University of Kansas Medical Center, 39th and Rainbow, Kansas City, Kansas 66103. *This work was supported by Grant CA-50616 from the National Cancer Institute (to P.F.T.), Grant GM-29496 from NIGMS, and a grant from the Andrew W. Mellon Foundation (to J.S.T.) This work was also supported by Grant HD-02528 from the NICHHD to the Kansas Center for Mental Retardation and Developmental Disabilities.

Materials

and Methods

Reagents and supplies Murine recombinant TNF (4.03 X lo7 U/m@ was supplied by Genentech, Inc. (South San Francisco, CA), and ovine LH (LH S-25; 2.3 U/

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 20:10 For personal use only. No other uses without permission. . All rights reserved.

2504

TNF

AND

THECA-INTERSTITIAL 4

1. Effect of TNF doses on LH-stimulated androstenedione production. TIC (4 x loj cells) were incubated in 1 ml M5A and exposed to either no LH (control), LH (50 rig/ml) alone, or LH plus increasing doses of TNF (0.1-10 rig/ml). Androstenedione levels were measured in media aspirated from cell cultures at 2-day intervals; each data point represents the mean from three different experiments. Two-way ANOVA demonstrated significant differences (P < 0.001) between both time and treatment effects. Significance among treatment groups either on day 2 or day 4 was determined using Fisher’s test after oneway ANOVA. Significance (P < 0.05) among groups is indicated by different letters on a day of culture.

-I

FIG.

c

CELL

a

tJ

Control

T

q

LI I (SO ngiml)

-t-

3

d

1992 No 6

LH + TNF (0.1 rig/ml)

E > ” ; 2 g 5 2 z

Endo. Voll31.

CLUSTERING

q

Lt-I + I-NE (1 .O ng/mI)

a

2

1

0

c Day of Culture

mg) was obtained from the National Hormone and Pituitary Program. l-(5isoquinolinesulfonyl)-2-Methylpiperazine dihydrochloride (H7) and N-(2-guanidinoethyl)-5-isoquinolinesulfonamide (HA 1004) were obtained from Seikagaku America (Rockville, MD). McCoy’s 5A media (modified, without serum; M5A) and Medium 199 (M199) (with Hank’s salts, IO0 mg/L L-glutamine and 25 rnM HEPES) were purchased from GIBCO (Grand Island, NY). Collagenase (type II, 530 U/m@, deoxyribonuclease (DNase type IV, 1800 Kunitz U/mg protein), BSA, Percoll (specific gravity = 1.130 g/ml) and phorbol 12.myristate 13.acetate (PMA) were obtained from Sigma (St. Louis, MO). Penicillin G/streptomycin were from JRH Scientific (Lenexa, KS). Twenty four and six well culture plates were from Costar, Inc. (Cambridge, MA) and Falcon, Inc. (Lincoln Park, NJ), respectively. Safranin 0 was obtained from Fisher (St. Louis, MO). Epon polybed 812 was obtained from Poly Science (Warrington, PA).

Animals Twenty one-day-old virgin hypophysectomized female rats were obtained from Harlan, Inc. (Madison, WI), and maintained in 22-24 C rooms on a 14-h light, 10 h-dark schedule. Food and water were provided ad libidum. Six to 7 days after hypophysectomy, animals were decapitated under ether anesthesia. Ovaries were isolated and placed in sterile ice-cold M199, containing 1 mg/ml BSA. The animal procedures followed the guidelines of the Institutional Animal Care and Use Committee of the University of Kansas Medical Center.

Cell dispersal Ovaries from lo-12 rats were cleaned of extraneous fat, bursa, oviduct, and uterine tissues with the aid of a dissecting microscope and ultrafine microforceps. The cleaned ovaries were finely minced to yield 6-8 fragments per ovary, after which, the pieces were washed 2 times in 10 ml fresh Ml99 containing 0.1% BSA. The ovarian fragments were placed in a 50.ml polystyrene conical centrifuge tube containing 2 ml collagenase-DNase solution (21) and 2 ml Ml99 (final volume equaling 4 ml) and incubated at 37 C in a humidified atmosphere of 5% CO1 and 95% air for 90 min. After 30 min of incubation, the ovarian pieces were flushed through a Pasteur pipette approximately 20 times, and at 60 and 90 min the dispersate was flushed as above using a 21-gauge needle after flushing through a Pasteur pipette. Upon termination of incubation, the cells were washed with M5A, pelleted, and resuspended in 2 ml M5A. Fifty microliters of the cell suspension were combined with 5 kl 0.5% trypan blue (5 mg trypan blue in 1 ml PBS, pH = 7.0), and cell

number viability

and viability was regularly

were determined in excess of 88%).

using

a hemacytometer

(cell

TIC purification TIC were isolated from the whole ovary dispersate as previously described by Magoffin and Erickson (22). Briefly, 1 ml 44% Percoll in Ml99 containing 0.1% BSA was pipetted into 12 x 75-mm polystyrene centrifuge tubes, Two milliliters of Percoll with the specific gravity adjusted to 1.055 g/ml were pipetted on top of the 44% Percoll, and 1 ml cell suspension (3-7 X lo6 cells) was carefully layered on top of the 1.055 g/ml Percoll. The tubes were capped and centrifuged at 400 x g at 4 C for 20 min. After centrifugation, the TIC were restricted to the 1.055 g/ml Percoll layer, with the majority of TIC forming the interface between the 1.055 g/ml and 44% Percoll layers. The TIC were carefully aspirated, pelleted, and resuspended in 2 ml M5A. TIC number and viability were evaluated as described above, and the cell density was adjusted to 1 X lo6 TIC/ml.

Cell culture TIC were plated at a density of either 3-4 x lo5 or 1 x lo6 cells/ml M5A (supplemented with 100 U/ml penicillin G and 100 pg/ml streptomycin), using 24 well or 6 well plastic culture dishes, respectively. Cells were allowed to attach for 18 h (at 37 C in a humidified atmosphere of 5% CO, in air) in M5A before the medium was removed and fresh M5A with or without treatments were added. Treatments were administered in duplicate and consisted of either LH (50 rig/ml), TNF (0.1, 1.0 or 10.0 rig/ml), H7 (10 PM), HA1004 (5 FM), PMA (50 nM), or combinations of the aforementioned. Doses of TNF were chosen based on the Kd for the binding to the TNF receptor (23). Controls received vehicle or no treatment. Medium was collected and fresh treatments were added on days 2, 4, and 6 of culture, and the experiments were terminated on days 6 or 8. Media were frozen at -20 C until RIA for androstenedione was performed. Each experiment was independently performed at three different times.

Morphology

and histology

TIC morphology was evaluated on days 0, 2, 4, and 6 via light microscopy using a Nikon diaphot-TMD inverted microscope (internal magnification = 2.5~) using either 4X, 10X, or 20X phase-contrast objectives. After termination of the incubation period, the cultures were fixed in 10% formaldehyde in phosphate buffer (pH = 7.0) and stained with 5% safranin 0 in distilled H20. Additional cultures were fixed in

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 20:10 For personal use only. No other uses without permission. . All rights reserved.

TNF AND THECA-INTERSTITIAL

CELL

CLUSTERING

2505

FIG. 2. Effect of TNF doses on TIC morphology. TIC (4 x lo5 cells/ml M5A) were cultured in the presence of LH (50 rig/ml) and increasing doses of TNF (0.1-10 rig/ml) for 6 days. Media were changed and treatments were replenished at P-day intervals. As observed using phase contrast microscopy, LHtreated TIC (a) remained in a confluent monolayer throughout the culture duration. Loose aggregates of TIC were present after treatment with 0.1 ng TNF/ml (b). TIC exposed to 1.0 (c) and 10 ng TNF/ml (d) formed the most well defined clusters.

2% glutaraldehyde in phosphate buffer and postfixed in 1% osmium tetroxide. Postfixed TIC were embedded in Epon polybed 812, sectioned at 1 pm, stained with toluidine blue and evaluated for both gross morphology, as well as changes in cluster pattern of cells.

per cluster were averaged for each analyzed by Systat statistical package

Quantifying

Concentrations media as previously

TIC clusters

Fixed and stained theta-interstitial cell clusters were quantitated by digital image analysis. Tissue culture plates were imaged using Optimas BioScan. Plate wells were imaged using a Nikon diaphot-TMD inverted microscope as above, using a 2X bright field objective. The final image magnification had a resolution of 6.41 pm/pixel. The background threshold gray level was set based on staining intensity of control plates (no TNF) in areas where only single cells were located. Objects above background larger than 2000 pm2 were counted as clusters. Identical areas of 70.8 mm* were counted in control and TNF-treated wells. The area occupied by each cluster was also determined. Three separate wells were imaged for each treatment group, and the 3 median values for area

well. Size distribution (Evanston, IL).

data

were

RIA of androstenedione of androstenedione described (24).

were

determined

in unextracted

Results Effects of TNF on LH-stimulated

androstenedione

production

To evaluate whether or not TNF would inhibit LH-induced androstenedione in a dose-dependent manner, TIC were treated with LH (50 q/ml) and either 0.0, 0.1 ng, 1.0 ng, or 10 ng TNF/ml. Figure 1 shows that TIC which received only

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 20:10 For personal use only. No other uses without permission. . All rights reserved.

2506

TNF

AND

THECA-INTERSTITIAL 200

CELL

a

Endo. 1992 Vol 131. No 6

CLUSTERING

n control

q

LH+TNF(l.Ong/ml)

q LH+TNF(lOng/ml) 100

b a

0

FIG. 3. The effect of TNF doses on TIC clustering: number of clusters and area per cluster. TIC (4 x lo5 cells/ml M5A) were cultured for 4 days and either untreated (control), exposed to LH (50 ng/ ml) alone, or treated with LH and increasing doses of TNF (0.1-10 rig/ml). Mean number of clusters (per 70.8 mm*) (a) and mean area (pm’) per cluster (b) for each treatment group were quantified as described in Materials and Methods. TIC aggregates with areas of less than 2000 pm* were determined not to be clusters and were not included in analyses. In separate experiments (c), TIC were treated with LH, TNF, H7, and/or HA 1004 as described in the text. Values for a, b, and c represent the mean from three separate experiments, and significance of P < 0.05 (as determined by one-way ANOVA followed by Fisher’s test) is denoted by different letters.

n ConId 8000-r

b

b

0 LH(50nghnl) LH

+ TNF

22 2 6000'

LH

+ TNF (1 .O rig/ml)

u V F: 4000B P,

LH

+TNF(lOng/ml)

2 2

(0.1 nghl)

l-

2000-

C

n H q 0 cl

COlllrOl

w

LIi

+ TNF (1.0 q/ml)

LH (50 nglml)

q

LIl

+ TNF + IIA

LI I + TNF (0.1 &ml) LII+TNF(l.Ong/ml) LH+TNF(IOng/ml)

+ II7 (10 uM)

1004 (5 uM)

C

T

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 20:10 For personal use only. No other uses without permission. . All rights reserved.

TNF AND

THECA-INTERSTITIAL

FIG. 4. Concentric layering and apparent cell-cell contacts induced in TIC as a result of TfiF treatment: TIC were cultured at a density of 4 x lo5 cells/ml M5A. TIC were treated with TNF (10 ne/mll for 4 davs. Media were changed and fresh TNF was added at !&da~ntervals. After 4 days, cells were fixed with 10% formaldehyde in phosphate buffer and stained using 5% safranin 0. When examined via phase contrast microscopy, TIC were organized into spheroid-shaped clusters consisting of concentric layers of cells (b; large arrow). TIC surrounding and within clusters appeared in contact with each other (a, b, small arrows). TIC not associated with well defined clusters were substantially elongated (a; large arrow).

LH produced characteristically substantial amounts of androstenedione, as compared to TIC receiving no LH treatment. TIC which received LH and increasing dosesof TNF showed a dose-dependent inhibition of androstenedione production within 2 days of TNF treatment, and androstenedione production remained suppressed in TNF-treated TIC for the remainder of the duration of the culture (day 4). Effects of TNF doses on TIC morphology

Since TNF inhibited LH-induced androstenedione production in a dose-dependent fashion in TIC, it was of interest to determine if TNF, in a similar dose-dependent manner, would affect the morphology of TIC in vitro. TIC were treated with either LH alone or LH and increasing dosesof TNF (as

CELL

CLUSTERING

2507

described above). When TIC were exposed to LH (50 rig/ml) for up to 6 days, the cellsremained in a confluent, monolayer configuration (Fig. 2a); similar confluency was observed in the absenceof any treatment (figure not shown). However, when TNF was added to TIC, in the presence(Fig. 2, b, c, d) or absenceof LH (figure not shown), TIC exhibited distinct clustering within approximately 3 days of TNF treatment. Importantly, this clustering phenomenon was most apparent in TIC exposed to 1 or 10 ng TNF/ml (Fig. 2, c, d). Furthermore, the greatest number of clusters with the largest individual areasformed as a result of treatment with 1.0 and 10 ng TNF/ml (Fig. 3). Conversely, when TIC were exposed to 0.1 ng TNF/ml, some loose aggregatesof TIC were noticed within 4 days of TNF administration (Fig. 2b), however, these groupings of cellswere much lessdefined than clusters seen after treatment with either 1 ng or 10 ng TNF/ml. As observed in Fig. 3, exposure to 0.1 ng TNF/ml resulted in significantly lessnumbers of clusters formed after 4 days as compared to the number of clusters formed after exposure to 1.0 or 10 ng TNF/ml. Figure 4 shows that at 4 days, TIC which had been exposed to 10 ng TNF/ml appeared to contact each other (Fig. 4, a, b; small arrows), as well as demonstrate an elongated appearance (Fig. 4a, large arrow). TIC which had clustered showed a concentric layered configuration of closely arranged cells (Fig. 4b, large arrow). It is important to emphasize that the TIC isolation procedure, when carefully executed, results in an approximately 92% pure population of TIC, which confirms a study by Magoffin and Erickson (22). As a check of the purification procedure, we performed lz51-labeledhCG binding assaysof whole ovary dispersateas compared to isolated TIC, and our results (data not shown) closely corroborated those of Magoffin and Erickson (18) in that approximately a 90-92% pure population of TIC were obtained as a result of the density gradient centrifugation procedure; therefore, a B10% population of nonthecal cell types in our cultures is acknowledged. Thus, the presence, albeit slight, of other ovarian cell types (i.e. macrophages, granulosa, etc.) may have influenced TNF induced TIC clustering via paracrine or juxtacrine cellular interactions. Although we cannot define the exact percentage of nonthecal cell types which may be contaminating the TIC cultures, microscopic evaluations of TNF induced TIC clusters have shown that cells within the clusters were morphologically similar (Fig. 5a). These cells resembledsteroid producing (thecal) cellsdue to the presence of lipid droplets and abundant mitochondria with tubular cristae (Fig. 5b). Importantly, these (and other) organelles were located in all cells throughout the clusters. Time course of the effect of TNF on LH-stimulated androstenedione production in cultured TIC

In order to elucidate the effects of TNF on androstenedione production at different times after culture with LH, TNF (10 rig/ml) was added on either day 2, 4, or 6 after TIC received continuous LH treatment (50 rig/ml) from day 0 of culture. Production of androstenedione decreased precipitously by days 4, 6, and 8 after treatment with TNF on days 2, 4, and 6, respectively (Fig. 6). TNF was effective in inducing clus-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 20:10 For personal use only. No other uses without permission. . All rights reserved.

2508

TNF AND

THECA-INTERSTITIAL

CELL

CLUSTERING

Endo - 1992 Voll31. No 6

FIG. 5. Histologic evaluation of TNF-induced TIC clusters. TIC (4 x lo5 cells/ ml M5A) were treated with LH (50 ng/ ml) and TNF (10 rig/ml) for 4 days, after which cells were fixed in 2% glutaraldehyde in phosphate buffer and postfixed in 1% osmium tetroxide. TIC were embedded in Epon polybed 812, sectioned at 1 pm and stained with toluidine blue (a). TIC clusters evaluated using electron microscopy were fixed as above and sectioned at 800 nm, then stained with lead citrate uranyl acetate. TNF treatment resulted in the formation of tightly packed clusters of TIC (a). Cells within the clusters contained numerous mitochondria with tubular cristae, lipid droplets, as well as rough endoplasmic reticulum (b). The scale bar for a equals 30 pm. The final magnification for b is 8000x.

6-

o

FIG. 6. Time course of the effect of TNF on LH-stimulated androstenedione production. TIC (4 x lo5 cells) were incubated in 1 ml M5A in the presence of LH (50 rig/ml). TNF (10 rig/ml) was added to cell cultures after either 2,4, or 6 days of continuous LH treatment. Androstenedione was measured in media collected from cultures at 2-day intervals. Unpaired t tests revealed significance between treatment groups (LH vs. LH + TNF) on days 4, 6, and 8. Significance of P < 0.05 is denoted by different letters on a day of culture.

=‘ & & Z x

5.

b LH (50 q/ml) d O-8

........*

LH d 04 + q?q d 2-4

---a--. -.-. a-*-

LHdOd+TNFd4-6

4-

LHdO-8+TNFd6-8

a

-5 i 2 4 2

0

2

4 Day

tering of TIC following 2 or 4 days of LH treatment (figure not shown). Importantly, the aggregations began forming after approximately 3 days of exposure to TNF, with well defined clusters forming 4 days after addition of TNF. Cultures were not observed beyond day 8 and thus, whether clustering occurred after TNF treatment on day 6 is not known. The effect of protein kinase inhibitors on TNF inhibition of LH-stimulated androstenedione production and morphology

H7 an inhibitor of protein kinases (Ki = 3..0 PM, 5.8 PM, and 6.0 PM for protein kinase A, G, and C, respectively) (25) was used to investigate the effects on TNF-induced inhibition of LH-stimulated androstenedione, and also on TNF-induced changesin morphology. TIC were treated with 10 PM H7 for 4 or 6 h, after which LH (50 rig/ml) and TNF (10 rig/ml) were added to the culture. TIC were pretreated with H7 or

of

6

8

Culture

HA1004 (seebelow) in order for these protein kinase inhibitors to have sufficient time to enter the cells and attenuate protein kinase activity before the addition of LH and/or TNF. As observed in Fig. 7, the H7 treatment did nothing to alleviate the repressive effect of TNF on LH-stimulated androstenedione production. Figure 8a shows that H7 attenuated the ability of TNF to cause clustering in TIC for approximately 6 days (compared to those receiving only LH + TNF, Fig. 8d). Statistical analysis (one-way analysis of variance (ANOVA) followed by Fisher’s test) revealed that the mean number of clusters formed as a result of H7 pretreatment was significantly reduced (Fig. 3c) as compared to the number of clusters formed in TNF treated cells not receiving H7; however, the number of clusters in H7 treated TIC did not significantly differ from the number of clusters in control or LH treated TIC (Fig. 3~). It is important to note that the cells which surrounded the aggregatesin the H7-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 20:10 For personal use only. No other uses without permission. . All rights reserved.

TNF

AND THECA-INTERSTITIAL

effect of protein kinase inhibitors on TNF suppression of LH-induced androstenedione production. TIC (4 x 105cells) were cultured in 1 ml M5A and treated with LH (50 rig/ml), LH + H7 (10 /.tM), LH + HA 1004 (5 /AM), LH + 10 ng TNF/ml, LH + TNF + H7, or LH + TNF + HA 1004. Media were collected from cultures at 2-day intervals and androstenedione levels were determined by RIA. Unpaired t tests revealed no significant differences in androstenedione production among groups treated with TNF/LH and H7 or HA 1004. Media from TIC treated with LH alone showed significantly greater amounts of androstenedione; H7 and HA 1004 did not significantly alter LH-stimulated androstenedione production by TIC. Significance of P < 0.02 is denoted by different letters within a particular day of culture.

41

a

2509

-I--.-

a

a

a

n

LH

q

LH+

H7 (10 uM)

q

LH+

HA

q

LH+

TNF (IO nghnl)

rig/ml)

(50

1004

(5 uM)

TNF+H7 TNF

b

+ HA

1004

b

T-I-

2

treated group had remained in a more confluent arrangement (Fig. 8a) in contrast to TIC treated with TNF + LH (Fig. 8d). HA1004, an inhibitor of protein kinase A (K, = 2.3 PM) and protein kinase G (K, = 1.3 PM), only mildly inhibits protein kinase C (PKC) (K, = 40 PM). HA1004 in low concentrations, was therefore a useful control for H7 because it would presumably block cyclic A and G-dependent protein kinasesat a concentration of 5 PM, but not PKC (26). When TIC were preincubated for 4 or 6 h with 5 PM HA1004 and were then given LH (50 rig/ml) and TNF (1 rig/ml or 10 ng/ ml), LH-stimulatable androstenedione production remained suppressed over 4 days in vitro, as shown in Fig. 7. TIC treated with 5 PM HA1004 (in the presence of LH), and exposed to TNF (1 rig/ml), exhibited cell clustering after 3 days of culture, and clustering was still evident at 6 days (Fig. 8b). Clustering in TIC which received HA1004 and TNF resembledclustering observed in TIC which were given TNF in the absenceof HA1004 (Fig. 8d). Importantly, TIC treated with only LH (50 rig/ml) remained in a confluent monolayer throughout the culture duration (Fig. 8~). Furthermore, statistical analysis (one-way ANOVA followed by Fisher’s test) showed no significant difference in mean number of clusters as a result of treatment with either 1.0 ng TNF/ml, or 1.0 ng TNF/ml in combination with HA1004 (Fig. 3~). ester on LH stimulated

CLUSTERING

C

FIG. 7. The

The effect of phorbol and TIC morphology

CELL

androstenedione

In order to further examine the involvement of Ca*+dependent PKC in attenuation of LH stimulated androstenedione synthesis and stimulation of TIC clustering, TIC were pretreated with the phorbol ester, PMA (50 nM), for 4 or 6 h, after which, LH (50 rig/ml) was added to the cultures. As observed in Fig. 9, TIC exposed to LH produced androstenedione (day 2 and 4) in amounts resembling those described above. Addition of PMA (without LH) did not result in a deviation from basal (control) androstenedione production (data not shown). However, when TIC were pretreated with

Day of culture

PMA for 4 or 6 h, then given LH, a noticeable inhibition of androstenedione production was observed by day 2 of culture, and androstenedione synthesis remained repressed throughout the culture duration. Interestingly, TIC began to form some aggregations by day 4; therefore, in separate experiments TIC were pretreated with PMA for 4 or 6 h, then given LH. TIC were cultured for 8 days, at which time light microscopic examination revealed TIC clusters (Fig. 9, inset, b) which resembled the clusters formed by TIC after treatment with TNF (1 rig/ml) for 4 days. Discussion

These studies revealed that TNF, in a dose-dependent manner, caused morphological alterations in TIC in vitro. Clustering of TIC was observed when using all dosesof TNF (0.1-10 rig/ml), whether in the presence or absence of LH; however, both the number of clusters and the median area per cluster was greatest when TIC were exposed to the higher doses of TNF (1 rig/ml and 10 rig/ml). These findings may be of physiological importance, as the dosesof TNF which were given to TIC were within the Kd for TNF observed in pig granulosa cells (23). The effect on TIC morphology by TNF could have significant physiological ramifications with respect to the processof early follicular development. TNF has been histochemically localized to the ovary (3), and to the ooplasm of rat primordial follicles (unpublished observations by J. Marcinkiewicz and I’. Terranova). Perhaps TNF, acting alone, or in concert with other growth factors, may function as a thecal (and/or granulosa cell) organizing factor that attracts these cells to the oocyte during follicular development. That TNF may be exerting its morphological effects via a PKC-mediated pathway is possible since 1) TNF has been shown to operate via PKC controlled mechanismsin regulating steroidogenesisof rat preovulatory follicles in vitro, with theta as a target (26), 2) H7, an inhibitor of PKC, greatly

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 20:10 For personal use only. No other uses without permission. . All rights reserved.

2510

TNF AND

THECA-INTERSTITIAL

CELL

CLUSTERING

Endo * 1992 Voll31. No 6

FIG. 8. Effect of the protein kinase inhibitors H7 and HA 1004 on TNF-induced TIC clustering. TIC (4 x lo5 cells/ ml M5A) were pretreated with either H7 (10 pM) or HA 1004 (5 @M) and LH (50 rig/ml) for 4 or 6 h, after which TNF (10 rig/ml) was added to the cultures. Treatments were applied for 6 days, media were changed and fresh treatments were administered at P-day intervals. Cells were fixed with 10% formaldehyde in phosphate buffer and stained using 5% safranin 0 as outlined in Materials and Methods. Phase contrast microscopy revealed that TIC treated with a combination of H7, LH, and TNF had formed reduced numbers of clusters, as well as less defined clusters (a), resembling TIC which were treated with LH alone (c). In contrast, TIC receiving a combination of HA 1004, LH, and TNF still showed clustering (b), as was also observed in TIC receiving TNF and LH in the absence of HA 1004 (d). The 50-pm scale bar corresponds to a, b, and c. The 100 lrn scale bar corresponds to d only.

reduced the ability of TNF to induce morphological changes in TIC in vitro and 3) PMA, a stimulator of PKC, caused similar morphological changes in TIC, and 4) PMA has been shown to inhibit LH stimulated androsterone production (27), as well as androstenedione synthesis by TIC. Importantly, an inhibitor of the cyclic nucleotide dependent protein kinases, HA1004, had no significant effect on reducing TIC clustering induced by TNF. Although the results of this study suggest a role for PKC and not the cyclic nucleotide-dependent protein kinases in mediating TNF-induced TIC clustering, the involvement of other cell signal transduction mechanisms cannot be disregarded at this time. Based on the Ki for HA1004, the concentration of HA1004 used in these studies (5 PM) should inhibit cyclic nucleotide-dependent protein kinases while not inhibiting PKC (26). However, we cannot be certain that HA1004 was an effective protein kinase

inhibitor in TIC in vitro because no effects were observed. Hence, it will be imperative to quantitate the relative activities of the protein kinases in question, as this may aid in further defining the cell signaling mechanisms used by TNF in TIC. TNF caused TIC to change from a cobblestone-like monolayer of cells to star-shaped clusters after 72 h of treatment. Although this is the first documented account of such an occurrence caused by TNF, it was noted in a review on ovarian TIC that similar morphologic changes are induced in TIC by epidermal growth factor (EGF) (16). In addition, when TIC are treated with TNF for greater than 4 days, the TIC masses formed well defined, spheroid-shaped clusters of cells. Within and surrounding the TIC masses, cells appeared in contact with one another, although whether gap junctions are present is not known. The possibility that TNF is operating via induction of an EGF regulated mechanism in caus-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 20:10 For personal use only. No other uses without permission. . All rights reserved.

TNF AND THECA-INTERSTITIAL

FIG. 9. The effect of PMA on LH-stimulated androstenedione production and TIC morphology. TIC (4 x lo5 cells/ml M5A) were given either LH (50 rig/ml), or pretreated with the phorhol ester PMA (50 nM) for 4 or 6 h, after which LH was added to the cultures. TIC were cultured for 4 days, with fresh media and treatments added on day 2. RIA for androstenedione content of TIC conditioned media (after day 4) showed a PMA-induced inhibition of LH stimulated androstenedione production by day 2 of culture, and this inhibition lasted throughout the culture duration. In separate experiments TIC were cultured using the above protocol (with fresh media and treatments added at 2-day intervals); however, culture duration was extended to 8 days in order to further examine the effects of PMA on TIC morphology. The inset shows that TIC treated with LH (a) remained in a confluent monolayer through 8 days of culture. Alternatively, TIC treated with PMA (b) formed clusters by day 8 in vitro. Scale bar (9a-b) = 100 Wm.

a

CELL CLUSTERING

2511

n LH (50 rig/ml) q PMA (50 nM) q

PMAtLH

c 2

ing TIC clustering is one attractive hypothesis which is under current investigation by our lab. Androstenedione is the main aromatizable androgen produced by TIC as a consequence of LH stimulation (16). However, the 5a-reduced androgens (androstanedione, androsterone, and androstanediol) constitute the majority of the LH-stimulatable C 19 androgens produced by TIC from immature rats (16). We chose to evaluate androstenedione concentrations in TIC conditioned media due to the physiological importance of this aromatizable thecal androgen. Therefore, the present study corroborates and extends beyond the findings of Andreani et al. (9) who observed TNFinduced inhibition of LH-induced androsterone production in TIC. Our data contrastswith the action of TNF on FSH-induced aromatase in rat granulosa cells. Emoto and Baird (10) observed that TNF inhibited further induction of FSH-induced

Day

of culture

b, c

4

aromatase, but TNF did not inhibit existing aromatase. Our time course studiesusing TIC showed that TNF was capable of attenuating androstenedione production after either 2, 4, or 6 days of continuous LH treatment, indicating that the more differentiated TIC still respond to the TNF, probably by a reduction of 17a-hydroxylase/l7,20-lyase activity (9). Interestingly, TNF was still able to induce clustering in TIC using these experimental conditions. Hence, the more differentiated thecal cells were still able to respond morphologically to TNF despite the inhibitory actions of TNF on LHsupported androstenedione. The effect of TNF on LH receptor content and LH binding in TIC have not been reported thus far. This is an extremely import area concerning the action of TNF and should be investigated. Importantly, Mauduit et al. (13) have shown that TNF decreasedLH/human CG binding to Leydig cells (the testicular homolog of ovarian TIC), and this is believed

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 20:10 For personal use only. No other uses without permission. . All rights reserved.

2512

TNF AND

THECA-INTERSTITIAL

to be one of several mechanisms whereby TNF attenuated LH stimulated androgen production in this cell type. H7 and HA1004 did not alleviate the repressive effect of TNF on LH-induced androstenedione synthesis, although H7 did impede TIC clustering caused by TNF. Thus, it appears that H7 was effective in these cells, and it does not appear that clustering is required for TNF-induced inhibition of androstenedione production. This is further supported by the fact that the inhibition of LH-stimulated androstenedione was evident as early as 2 days of TNF treatment, whereas clustering was not apparent until 3-4 days of TNF treatment. The present results may also indicate the involvement of a different, or more than one, signal transduction mechanism by which TNF exerts its inhibitory effects on steroidogenesis, and its stimulatory effects on morphological changes in the TIC. Other growth factors interfere with LH-supported androstenedione production in TIC. EGF, in addition to its aforementioned morphologic effects in TIC, also inhibits LHinduced androgen production (15). Importantly, interactions between TNF and EGF have been previously demonstrated (20). Questions arise concerning the possible interactions between TNF and EGF within TIC. First, is EGF produced in response to TNF? Thus, the actions of TNF could be mediated through EGF. If this is the case, it would be expected that pharmacological inhibition of EGF expression or pretreatment of TIC with an antibody to EGF may interfere with the actions of TNF in TIC. Second, does TNF-increase thecal EGF receptor number, thus leading to enhanced response of TIC to endogenously produced EGF? Transforming growth factor-a (TGFa) possesses structural homology to that of EGF (28) and binds the same ligand-activated receptor which is a protein tyrosine kinase (28). Also, TGFa mRNA has been localized to ovarian theta in the rat (29). Thus, TGFa may mediate some of the TNF-induced responses in TIC. The alternative possibility exists whereby TNF may stimulate TGFa production by TIC. TNF-induced TGFa could act in an autocrine manner by binding the EGF receptor, and the resultant intracellular cascade could lead to observed morphological changes in TIC, as well as inhibition of LHsupported androstenedione production. Inhibition of LH-induced androstenedione production by ovarian TNF could have important consequences in vim. Inhibition of androgen precursors could alter the amount of estradiol-17/3 (EZ) produced by the granulosa cells of the preovulatory follicle. It would therefore be crucial to determine the in viva time course for TNF appearance within the ovary. If TNF is present early on during the course of folliculogenesis, it may be an important factor in repressing androgen in preantral follicles, when corresponding low levels of E2 are observed. If endogenous factors impede the inhibitory action of TNF on androgen production as follicles entered the preovulatory stage, TNF would not be detrimental to the normal ovulatory process. However, the sustained presence of TNF during the follicular cycle could result in long-term EZ repression, and inhibition of further expression of EZ by the follicle and result in the onset of follicular atresia. It is hoped that future research will provide insight into the

CELL

CLUSTERING

Endo. 1992 Vol 131. No 6

mechanism by which TNF inhibits LH-stimulated androstenedione production, as well as the mechanisms by which TNF induces clustering in TIC. Acknowledgments We thank Dr. R. DeLisle and Dr. G. Enders (Department of Anatomy and Cell Biology, University of Kansas Medical Center) for their assistance with the morphological evaluations and microscopy. The authors gratefully acknowledge the supply of murine recombinant tumor necrosis factor alpha from Genentech, Inc. (South San Francisco CA). The authors thank the NIDDK (Bethesda, MD), the CPR of the NICHHD, the ARS of the USDA (Beltsville, MD), and The University of Maryland School of Medicine (Baltimore, MD) for the gift of ovine LH.

References 1. Jaattela M 1991 Biologic activities and mechanisms of action of tumor necrosis factor-alpha/cachectin. Lab Invest 64:724-742 2. Baglioni C 1992 Mechanisms of cytotoxicity, cytolysis and growth stimulation by TNF. In: Beutler B (ed) Tumor necrosis factor: the molecules and their emerging role in medicine. Raven Press, NY, pp 425-438 3. Roby KF, Terranova PF 1989 Localization of tumor necrosis factor (TNF) in the rat and bovine ovary using immunocytochemistry and cell blot: Evidence for granulosal production. In: Hirshfield AN (ed) Growth Factors and the Ovary. Plenum Publishing Corporation, New York, pp 273-278 4. Roby KF, Weed J, Lyles R, Terranova PF 1990 Immunological evidence for a human ovarian tumor necrosis factor-a. J Clin Endocrinol Metab 71:1096-1102 5. Zolti M, Meirom R, Shemesh M, Wollach D, Mashiach S, Shore L, Ben Rafael Z 1990 Granulosa cells as a source and target organ for tumor necrosis factor-a. FEBS 261:253-255 6. Bagavandoss P, Wiggins RC, Kunkel SL, Remick DG, Keyes PL 1990 Tumor necrosis factor production and accumulation of inflammatory cells in the corpus luteum of pseudopregnancy and pregnancy in rabbits. Biol Repro 42:367-376 7. Bagavandoss P, Kunkel SL, Wiggins RC, Keyes PL 1988 Tumor necrosis factor cy (TNF-ol) production and localization of macrophages and T lymphocytes in the rabbit corpus luteum. Endocrinology 122:1185-1187 8. Adashi EY, Resnick CE, Croft CS, Payne DW 1989 Tumor necrosis factor (Y inhibits gonadotrophin hormonal action in nontransformed ovarian granulosa cells. J Biol Chem 264:11591-11597 9. Andreani CL, Payne DW, Packman JN, Resnick CE, Hurwitz A, Adashi EY 1991 Cytokine-mediated regulation of ovarian function, J Biol Chem 266:6761-6766 10. Emoto N, Baird A 1988 The effect of tumor necrosis factor/cachectin on follicle-stimulating hormone-induced aromatase activity in cultured rat granulosa cells. Biochem Biophys Res Commun 153:792-798 11. Darbon JM, Oury F, Laredo J, Bayard F 1989 Tumor necrosis factor-o inhibits follicle-stimulating hormone-induced differentiation in cultured rat granulosa cells. Biochem Biophys Res Commun 163:1038-1046 12. Roby KF, Terranova PF 1990 Effects of tumor necrosis factor-a in vitro on steroidogenesis of healthy and atretic follicles of the rat: theta as a target. Endocrinology 126:2711-2718 13. Mauduit C, Hartmann DJ, Chauvin MA, Revol A, Morera AM, Benahmed M 1991 Tumor necrosis factor alpha inhibits gonadotropin action in cultured porcine Leydig cells: site(s) of action. Endocrinology 129:2933-2940 14. Calkins LH, Guo H, Sigel MM, Lin T 1990 Tumor necrosis factor alpha enhances inhibitory effects of IL-I on leydig cell steriodogenesis. Biochem Biophys Res Commun 166:1313-1318 15. Erickson GF, Case E 1983 Epidermal growth factor antagonizes ovarian theta-interstitial cell cytodifferentiation. Mol Cell Endocrino1 31:71-76 16. Erickson GF, Magoffin DA, Dyer CA, Hofeditz C 1985 The

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 20:10 For personal use only. No other uses without permission. . All rights reserved.

TNF AND

17.

18.

19.

20.

21.

22

THECA-INTERSTITIAL

ovarian androgen producing cells: a review of structure/function relationships. Endocr Rev 6:371-399 Stolpen AH, Guinan EC, Fiers W, Pober JS 1987 recombinant tumor necrosis factor and immune interferon act singly and in combination to reorganize human vascular endothelial cell layers. Am J Path 123:16-24 Stolpen AM, Golan DE, Pober JS 1988 Tumor necrosis factor and immune interferon act in concert to slow the lateral diffusion of proteins and lipids in human endothelial cell membranes. J Cell Biol 107:781-789 Rosen E, Goldberg I, Liu D, Setter E, Donovan M, Bhargava M, Reiss M, Kacinski 1991 Tumor necrosis factor stimulates epithelial tumor cell motility. Cancer Res 51:5315-5321 Mawatari M, Kohno K, Mizoguchi H, Matsuda T, Asoh K, VanDamme J, Welgus HG, Kuwano M 1989 Effects of tumor necrosis factor and epidermal growth factor on cell morphology, cell surface receptors, and the production of tissue inhibitor of metalloproteinases and IL-6 in human microvascular endothelial cells. J lmmunol 143:1619-1627 Magoffin DA, Erickson GF 1982 Primary culture of differentiating ovarian-androgen producing cells in defined medium. J Biol Chem 257:4507-4513 Magoffin DA, Erickson GF 1988 Purification of ovarian thecainterstitial cells by density gradient centrifugation. Endocrinology

CELL

CLUSTERING

2513

122:2345-2347 23. Veldhuis JD, Garmey JC, Urban RJ, Demers LM, Aggarwal BB 1991 Ovarian actions of tumor necrosis factor-d (TNFa): Pleiotropic effects of TNFa on differentiated functions of untransformed swine granulosa cells. Endocrinology 129:641-648 24. Terranova PF, Garza F 1983 Relationship between the preovulatory luteinizing hormone (LH) surge and androstenedione synthesis of preantral follicles in the cyclic hamster: detection by in vitro responses to LH. Biol Reprod 29:630-636 25. Hidaka H, Inagaki M, Kawamoto S, Sasaki Y 1984 Isoquinolinesulfonamides, novel and potent inhibitors of cyclic nucleotide dependent protein kinase and protein kinase C. Biochem 23:50365041 26. Sancho-Tello M, Terranova PF 1991 Involvement of protein kinase C in regulating tumor necrosis a-stimulated progesterone production in rat preovulatory follicles in vitro. Endocrinology 128:1223-1228 27. Hofeditz C, Magoffin D, Erickson G 1988 Evidence for protein kinase C regulation of ovarian theta-interstitial cell androgen-biosynthesis. Biol Repro 39:873-881 H, Hunkapiller MW, Hood LE, Todaro GJ 1984 Rat 28. Marquardt transforming growth factor type I: structure and relation to epiderma1 growth factor. Science 223:1079-1081 29. Skinner M, Lobb D, Dorrington J 1987 Ovarian thecal/interstitial cells produce an epidermal growth factor-like substance. Endocrinology 121:1892-1899

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 20:10 For personal use only. No other uses without permission. . All rights reserved.

Tumor necrosis factor-alpha induces clustering in ovarian theca-interstitial cells in vitro.

Tumor necrosis factor-alpha (TNF) has been implicated in the regulation of steroidogenesis in theca-interstitial cells (TIC). The purpose of this stud...
2MB Sizes 0 Downloads 0 Views