MICROSCOPY RESEARCH AND TECHNIQUE 27:97-107 (1994)

The Role of Granulosa and Theca Cell Interactions in Ovarian Structure and Function
FUMIKAZU KOTSUJI AND TOSHIRO TOMINAGA
Department of Obstetrics and Gynecology, Fukui Medical School, Fukui-Ken 910-11, Japan

KEY WORDS

Cellular interaction, Granulosa cells, Theca cells, Morphology, Proliferation, Ovarian function

ABSTRACT The term paracrine control has recently become very fashionable in the field of endocrinology: however, no study has yet clarified directly the role of paracrine activity in the functional and morphological development of endocrine organs. We developed a culture system in which two types of cells were allowed to attach to opposite sides of a collagen membrane in order to observe cellular communication directly and used it to explore the role of granulosa and theca cell interactions in ovarian structure and function. In the first series of the study, we explored the interaction between theca and granulosa cells by investigating the morphology, proliferation, and steroidogenesis of the cells. Granulosa cells cultured alone were flattened and polygonal in shape and formed a monolayer sheet. Granulosa cells cocultured with theca cells formed multilayer sheets. The apical surface of each cell appeared convex. Numerous filopodia spread over the cellular surfaces to connect individual cells. Theca cells cultured alone were thin, flat, and spindle-shaped. Theca cells cocultured with granulosa cells were also spindle-shaped; however, the apical surface appeared convex. The cell numbers of both granulosa and theca cells in the cocultured group increased approximately twofold compared to control cells cultured alone. The progesterone producing activity of granulosa cells in the cocultured group was reduced to 40% of that of cells in the control group. In contrast, the androstenedione producing activity of theca cells in the cocultured group increased approximately threefold compared to that of control group cells. In the second series of experiments, effects of cellular communication on cellular response to gonadotropins were investigated. When granulosa cells were cultured with theca cells, FSH treatment (1 Fg/ml) promoted cellular growth: however, LH treatment (1Fg/ml) suppressed cellular growth and augmented their progesterone production. In contrast, such effects of gonadotropins were not detected when granulosa cells were cultured alone. LH induced estradiol production by granulosa cells both cultured alone and cocultured with theca cells: however, FSH-induced estradiol production was not detected in this experimental condition. LH treatment increased androstenedione production of theca cells cocultured with granulosa cells, but no increase was observed for theca cells cultured alone. These results demonstrate that communication between these two types of follicular cells results in reciprocal modulation of their morphology, structure, growth, and function and that the actions of gonadotropins on target cells in the follicular wall can also be modified by the communication between these cells. In contrast to granulosa and theca cells cultured alone, cells in coculture seemed to possess morphological and functional characteristics more similar t o those of cells in the growing follicular wall in vivo. Therefore, we speculate that interactions between these two types of follicular cells is essential in the maintenance of original structure and function of the bovine follicular wall. 0 1994 Wiley-Liss, Inc. INTRODUCTION
Many parenchymal organs possess a fundamental structure in which epithelial and mesenchymal cells face each other across a basement membrane and in which mesenchymal cells have been considered to be essential for epithelial proliferation, morphogenesis, true in Organs and differentiation This is with a rapidly renewing epithelium, such as intestine, and in organs that have cycles O f functional activity, such as those of the female reproductive system (reviewed by Donjacour and Cuna, 1991). However, the mechanisms by which mesenchyme modulates epithe-

Received November 14, 1991; accepted in revised form September 2, 1992. Address reprint requests to Fumikazu Kotsuji, Department of OBIGYN,Fukui Medical School, Matsuoka-Cho, Yoshida-Gun, Fukui-Ken 910.11, Japan.

0 1994 WILEY-LISS, INC.

98

F. KOTSUJI AND T. TOMINAGA

lial structure and function are not yet well understood. Development of a n in vitro model which allows us to directly observe epithelial-mesenchymal interactions could provide useful information about the physiological role for such communication. The ovarian follicle is one of the most rapidly growing normal tissues. The proliferation of granulosa and theca cells, the two somatic cell types most prominent in the follicle, is considered to be responsible for the majority of follicle expansion during the maturation process. However, the mechanism by which the growth and differentiation of granulosa and theca cells are controlled is poorly characterized. The ovarian follicle possesses a fundamental structure in which granulosa (epithelial-like) and theca (mesenchymal-like) cells face each other across a basement membrane. Some products of theca and granulosa cells are known to affect the function andlor proliferation of the cells on the opposite side of the basement membrane. For example, androgens synthesized by theca cells serve as a substrate for aromatization, act as a hormone to influence FSH-induced steroidogenesis (Armstrong and Dorrington, 1976, Goff et al., 19791, induce mitochondria1 changes and the acquisition of smooth endoplasmic reticulum, and lead to early atresia of granulosa cells in the rat (Anderson, 1989). Granulosa steroids are also known to stimulate theca cell production of androstenedione by supplying them with a progestin precursor (Fortune, 1986). It has recently become clear that a number of growth factors are not only synthesized by ovarian cells but also have paracrine and autocrine actions at the level of granulosa and theca cells. The effects of several growth factors, including epidermal growth factor (EGF), fibroblast growth factor (FGF), insulin-like growth factor I (IGF-I) and transforming growth factors-a and -p (TGF-a and -p) on granulosa cells and theca interna cells have recently been reported by numerous investigators. For example, bovine theca cells express the gene for TGF-P, and granulosa cells are shown to contain receptors for TGF-P (Skinner and Coffey, 1988). Theca cells also produce a mitogenic factor with a molecular weight greater than 10,000 (Bendell et al., 1988). Recently, the proteins inhibin and activin, which are produced by granulosa cells and control follicle-stimulating hormone (FSH) secretion from the anterior pituitary gland, have been purified from follicular fluid of several species. These proteins are also known to exert paracrine effects. Both proteins regulate theca cell androgen production (Bicsak and Hsueh, 1988). These studies suggest the importance of granulosa-theca interaction in the mechanism of controlling ovarian function. Recently, we have developed an in vitro culture system in which different types of cells are attached to the opposite sides of a collagen membrane (Kotsuji et al., 1990). This system has made it possible to explore the effects of intercellular communication on cellular morphology and function. In the present review, we will summarize recent work from our laboratory, in which the role of communication between two types of follicular cells in the structure and function of the ovary were explored.

Collagen membrane

II I

Apical medium

Plastic dish

Supporting apparatus

Fig. 1. Schematic diagram of the floating collagen membrane culture system used for the coculture of granulosa and theca cells. The membrane was made of Type-1 collagen and has a n area of 8 cm2. Apical and basal chambers contain 1.5 and 8.0 ml of culture medium, respectively.

DEVELOPMENT OF COLLAGEN MEMBRANE CULTURE SYSTEM Figure 1 represents the collagen membrane cell culture system. A supporting apparatus to which a Type I collagen membrane was attached is plated in a plastic dish. Apical and basal chambers were separated and factors less than 12.5 kd were permeable through a collagen membrane. Bovine ovaries were collected from heifers less than 20 min after slaughter at a local abattoir. Granulosa and theca cells were prepared a s described in Figure 2. Under aseptic conditions, granulosa cells were harvested by needle aspiration of small size follicles (3-5 mm in diameter) and washed three times in culture medium consisting of Waymouths MB 751/1, Hanks solution, fetal bovine serum (6:3:1), supplemented with 100 Uiml penicillin, and 100 pg/ml streptomycin. This is subsequently referred to as culture medium in this manuscript. Cell viability was 30-42%, based upon trypan blue dye exclusion. Theca cells were prepared according to the following procedure. Follicles 6-10 mm in diameter were collected and cut into hemispheres. The theca interna layer was then removed from the follicular sections with fine forceps. Granulosa cells together with a part of the theca cell layer were removed by scraping with a knife under a stereomicroscope. The thin theca interna layer was then rinsed thoroughly with culture medium. The theca cell layer thus obtained was minced and then treated with 0.4% collagenase (Type 1)and 0.03% DNase in HanksHEPES buffer containing 0.4% BSA and 0.2% glucose (pH 7.4). Digestion was continued for 30 min a t 37C with continuous stirring at 80 rpm. The theca cells were then further digested for 8 min with 0.25%pancreatin in Hanks-HEPES buffer. Dispersed cells were washed three times with culture medium. Cell viability was 93-96%. For further purification, dispersed cells were plated onto plastic dishes, the floating cells

GRANULOSA-THECA INTERACTION

99
T Cells

G Cells
Aspirated from 4-6mm follicles Washed 3 times by culture medium

J.

Theca interna layer fr m 10-15 mm follicles Scraped with knife to remove G cells Treated with 0.4 % Collagenase 0.03 % DNase 0.25 YOPancreatine Washed 3 times

3.

Plated on collagen membrane

Removed by trypsin treatment

J. .T

J.

4-8

Cultured in plastic dished

culture medium

Fig. 2. Preparation of granulosa and theca cells. G, granulosa; T, theca.

were washed out 20 min later, and the remaining cells were cultured for 24 h. Cells were then removed from the dishes by treatment for 30 min with 0.1% trypsin (Sigma) and were used for culture on the collagen membrane culture. Figure 3 represents the cell culture procedures. Viable theca cells (5 x lo5) were plated onto Type 1 collagen membranes. On the second day of culture, each collagen membrane was turned over, and freshly prepared granulosa cells (1 x lo6) were plated onto the opposite surface. As controls, theca or granulosa cells were cultured alone on one side of a collagen membrane. Culture was continued until day 8, and medium was changed on days 2,5, and 7 of the culture. All cells were cultured a t 37C in a humidified atmosphere of 5%C 0 2 and 95%air to maintain the medium at pH 7.4. Granulosa and theca cells cultured alone became confluent on day 6 and 5, respectively. Those cells in the cocultured group became confluent earlier. The advantages of this system are 1) it more closely resembles the in vivo environment in which both types of the cells are separated by extracellular matrix; 2) changes in the granulosa and theca cells are able to be observed simultaneously; and 3) cells cultured on a floating collagen express their original function more closely than do cells cultured on a collagen-coated surface (Emerman and Pitelka, 1977; Emerman et al., 1977).

Collagen Membrane

I T Cells

Fig. 3. Culture of theca (TIand granulosa (G) cells. Viable theca cells (5 x lo5)were plated onto collagen membranes and cultured for 8 days. On the second day of the culture, each membrane was turned over, and freshly prepared granulosa cells (1 x lo6) were plated on the opposite surface of the membrane and cultured until day 8. As a control, theca or granulosa cells were cultured alone on one side of the collagen membrane.

EFFECT OF GRANULOSA-THECA CELL INTERACTIONS ON CELLULAR MORPHOLOGY, PROLIFERATION, AND BASAL STEROID HORMONE SYNTHESIS At the end of culture, cellular morphology, proliferation, and steroidogenesis were compared between the control and experimental groups. Figures 4 and 5 are light and scanning electron micrographs of granulosa cells, respectively. Granulosa cells cultured alone formed a monolayer sheet exhibiting a cobblestone-like appearance (Fig. 4A). Cells cultured with theca cells became smaller and highly condensed (Fig. 4B). When cells were observed using scanning electron microscopy, granulosa cells cultured alone had a thin, flat, and polygonal shape and were

densely packed. Microvilli, which were uniform in length, were apparent on the apical surface of the cells (Fig. 5A). Granulosa cells cultured with theca cells also had a polygonal shape. However, the apical surface of the cells appeared more convex than cells cultured alone. Some cells bore small cup-like ruffles. Scattered secretory products were also found on the cellular surface. Numerous filopodia spread over the cellular surfaces, and thereby across inter-connecting cells (Fig. 5B). The morphology of theca cells was also modified by the existence of granulosa cells. In the control, highly condensed fibroblastic cells formed characteristic parallel arrays and whorls (Fig. 6A). However, parallelism and the cellular whorls seemed to be disorganized in the presence of granulosa cells (Fig. 6B). When cells were observed using scanning electron microscopy, theca cells cultured alone were thin, flat, and spindleshaped (Fig. 7A). Theca cells cocultured with granulosa cells also had a spindle shape; however, the apical surface of the cells appeared more convex. The cells were more densely packed as compared with theca cells cultured alone. Scattered secretory products were also found on the cellular surface (Fig. 7B). In order to study the effect of cellular communication on cellular structure, the semi-thin sections of the collagen membranes from each group were prepared (Fig.

100

F. KOTSUJI AND T. TOMINAGA

Fig. 4. Phase contrast light microscopic observations of granulosa cells on day 8 of the culture. A Cells cultured alone. B: Cells cocultured with theca cells. Arrowheads represent areas suggesting multiple layering of the cells. x 200.

8). Granulosa cells cultured alone formed a monolayer sheet: however, granulosa cells cultured with theca cells formed double, and sometimes triple, layered sheets as well as single layers. Theca cells of both the control and cocultured groups formed multilayer sheets, although the cellular layers of the cocultured group were thicker than those of the control group. These observations indicate the existence of signals between theca and granulosa cells to modulate cellular morphology, structure, and activity. Scanning electron microscopic observations of the granulosa cells cultured with theca cells, such as cellular swelling and the appearance of small cup-like ruffles and secretory products on the apical surface, suggest that the levels of cellular activity are increased. Numerous filopodia connecting juxtaposed cells suggest the existence of intercellular communication among granulosa cells. Swelling of the cellular body, close cellular contact, and the appearance of secretory products on the surface observed in theca cells cultured with granulosa cells suggest a promotion of cellular activity. In order to detect the effects of heterologous cellular communication on granulosa and theca cell growth,

cultured cells were removed from the collagen membrane by a 1 h treatment with 1%trypsin and 0.2% EDTA, and cell numbers were calculated (Fig. 9). The numbers of both granulosa and theca cells of the cocultured group (9.23 2 0.40 x lo5 and 3.73 0.09 x lo6, respectively) were significantly larger than those of the control group (4.14 0.26 x lo5and 2.03 _t 0.08 x lo6, respectively). These observations indicate the existence of signals between theca and granulosa cells to modulate cellular proliferation. To detect the effects of cellular communication on granulosa and theca cell function, steroid hormone contents in culture media were measured. At the end of the culture period, 24 h culture medium between days 7 and 8 was collected from the apical and basal chambers of each dish, and estradiol, progesterone, androstenedione, and testosterone were measured by double antibody radioimmunoassay. In medium of control granulosa cells, progesterone, but not androstenedione, was detected. In medium of control theca cells, androstenedione, but not progesterone, was detected. Medium of cocultured groups contained both progesterone and androstenedione. Estradiol was not detected in cul-

GRANULOSA-THECAINTERACTION

101

Fig. 5. Scanning electron microscopic observations of granulosa cells on day 8 of the culture. A Cells : cultured alone. B Cells cocultured with theca cells. Arrows, secretory product; arrowhead, a cup-like ruffle. x 1.500.

ture medium of either cocultured or control group in this experimental condition. Testosterone was detected in medium of control theca cells and cocultured cells; however, the content of testosterone was less than 10% of that of androstenedione. When progesterone production per 1 x lo5 of granulosa cells was compared (Fig. lOA), that of the cocultured group was significantly reduced as compared with that of the control group (17.33 r 0.40 ng vs. 43.61 2 2.42 ng; P < 0.01). In contrast, androstenedione production per 1 x lo5 of theca cells of the cocultured group increased significantly as compared with that of the control group (427 k 27 pg vs. 174 r 3 pg; P < 0.01) (Fig. 10B). One possible explanation for the phenomenon is conversion of progesterone to androgen by theca cells. However, the reduction of ng order amount of progesterone cannot be explained by the increase of pg order of androstenedione production. Therefore, it can be concluded that the progesterone producing activity of granulosa cells was reduced by the presence of the theca cells. In contrast, androgen producing activity of theca cells was augmented by the presence of granulosa cells. Taken together, it can be concluded that communication between the two types of follicular cells results in reciprocal modulation of morphologic structure, proliferation, and function. Granulosa cells, the growth and function of which are modified by signals from the theca cells, in t urn stimulate the proliferation and

modulate the function of the theca cells. Such interaction between these two types of follicular cells may play a n important role for the maintenance of follicular structure and function. What factorb) are involved in the intercellular communication between theca and granulosa cells? In all tissues, epithelial cells are separated from mesenchyma1 cells by a basement membrane which is composed of several extracellular molecules (ECM) such as collagens, fibronectin, laminin, and proteoglycans. It has been reported that alterations in the ECM(s) can modify function and morphologic structure in numerous cell types (Chrambard et al., 1983; Hadley et al., 1987). Therefore, ECMs produced by theca and granulosa cells, as well a s gonadal steroids and regulatory peptides produced by theca and granulosa cells, could play a n important role in the communication between cells. Further study is in progress to define what factor(s) are responsible for the observations in our laboratory.

EFFECTS OF CELLULAR COMMUNICATION BETWEEN TWO TYPES OF FOLLICULAR CELLS ON THE ACTION(S) OF PITUITARY GONADOTROPINS ON THEIR TARGET CELLS Gonadotropins have been shown to play the most important role in controlling ovarian function. In addition, gonadotropins induce various effects on follicular cells throughout the follicular maturation process and the ovulatory cycle. Therefore, it is critical to deter-

102

F. KOTSUJI AND T.TOMINAGA

Fig. 6. Phase contrast light microscopic observation of theca cells on day 8 of the culture. A Cells cultured alone. B: Cells cocultured with granulosa cells. x 200.

mine whether cellular communication between these two types of follicular cells modulates gonadotropin actions at their target cells or not. Granulosa and theca cells possess FSH and LH receptors, respectively, and FSH and/or estradiol act synergistically to induce LH receptor expression on the surface of granulosa cells (Kessel et al., 1985; Ireland, 1987). Therefore, we examined whether the response of granulosa cells to FSH and/or LH is modified by the coexistence of theca cells. How the response of theca cells to LH is modified by the coexistence of granulosa cells was also studied. In the first series of experiments, the response of granulosa cells to gonadotropins was investigated. Granulosa cells in the control and cocultured groups were treated with either 1)FSH (NIADDK-ovine FSH17, 1pg/ml; FSH-treated) or 2) LH (bovine LH-USDAB5, 1 p,g/ml; LH-treated) for 8 days (days 2-9 of culture), and/or 3) FSH for the first 7 days (days 2-29, followed by LH for 24 h (FSH-LH-treated). Untreated control cells were cultured with medium alone. Theca cells in the cocultured group were cultured without gonadotropin. During the final 24 h of culture, androstenedione was added into culture medium (final

concentration of mol/l), and culture media samples (24 h) were collected. Estradiol and progesterone contents were measured. Cell numbers were calculated. Table 1 shows granulosa cell numbers at the end of culture and the production of progesterone and estradiol by 1 x lo5 granulosa cells during the last 24 h of culture. Gonadotropin treatment did not affect progesterone production by granulosa cells cultured alone. In contrast, LH and FSH-LH treatments increased progesterone production by granulosa cells in coculture by 69 31 and 38 3 25%, respectively, although FSH did not affect progesterone production. Estradiol was not detected in the culture medium when granulosa cells were cultured without gonadotropins. However, when granulosa cells were treated with LH or FSH + LH, estradiol was detected in the medium of both single and combined cultures, and FSH + LH treatment was more effective than LH alone in both single and coculture groups. No significant difference in estradiol producing ability was detected between cells in single and combined cultures. Gonadotropin treatment did not affect the growing

GRANULOSA-THECA INTERACTION

103

Fig. 7. Scanning electron microscopic observation of theca cells on day 8 of the culture. A: Cells cultured alone. B: Cells cocultured with granulosa cells. Arrow, secretory product. x 1,500.

ability of granulosa cells culture alone. In contrast, when granulosa cells were cultured with theca cells, FSH treatment augmented granulosa cells number by 22 -+ 2%, and LH treatment reduced it to 69 2 3% of untreated control. In the second series of experiments, the effect of LH on both the production of androstenedione and the growth of theca cells in combined culture was compared with the effects on theca cells cultured alone. In this experiment, the collagen membranes were turned over again on day 3 to place the theca cells in the apical chamber. Theca cells were cultured in either medium alone or in medium containing LH (1 pg/ml) for 7 days (days 3 -9). Table 2 shows androstenedione production by 1 x lo5 theca cells during 24 h culture on the final day of culture and cell numbers at the end of culture period. LH treatment did not affect androstenedione production by theca cells cultured alone. In contrast, androstenedione production by LH-treated theca cells cultured with granulosa cells was increased to 135 2 3 of untreated control. LH did not affect the proliferation of either monocultured or cocultured theca cells. These observations indicate that communication between the two types of follicular cells also modulates the actions of gonadotropins on their target cells. Gonadotropins possess multiple actions, and their effects on follicular cells vary throughout folliculogenesis. Various follicular factors are known to modulate gonadotropin action. For example, estradiol, androgens,

progesterone, IGFs, and TGF-a all enhance various aspects of FSH action, including induction of FSH or LH receptors and progesterone or estradiol synthesis. In contrast, FGF, EGF, calcium inhibitors, TGF-P, and follicular regulatory protein all inhibit FSH action (reviewed by Ireland, 1987). These studies and our data suggest that granulosa-theca interactions may be one of the important factors determining which of the multiplicity of possible gonadotropic actions is elicited at the target cell.

ROLEOFGRANULOSAANDTHECACELL INTERACTIONS IN OVARIAN STRUCTURE AND FUNCTION Corpus luteum cells derived from bovine granulosa cells produce large amounts of progesterone, but LH stimulation does not affect the progesterone production by these cells (Koors and Hansel, 1981; Ursely and Leymarie, 1979). The changes in steroidogenesis by granulosa cells due to luteinization are reported to depend upon alterations in the levels of steroidogenic enzymes. After ovulation, there is a dramatic increase in the content of cholesterol side-chain cleavage cytochrome P-450 (P-450,,), l7a-hydroxylase cytochrome P-450 (P-450,,,), and low density lipoprotein receptor in the luteal cells derived from granulosa cells (Rodgers et al., 1986; Rogers et al., 1987). Granulosa cells cultured alone in vitro undergo luteinization, which is indicated by the production of progesterone

C
Fig. 8. Light microscopic observations of the semi-thin sections of the collagen membranes on day 8 of the culture. A: Granulosa cells (G) cultured alone, B: Granulosa (G) and theca cells (T)cultured cultured alone. x 1,000. together, C: Theca cells (T)

GRANULOSA-THECA INTERACTION

105

G Cell Number

T Cell Number

P<O.Ol

c 4.0 n

Fig. 9. Cell numbers of granulosa and theca cells of control and cocultured groups. The data are represented by mean f SEM (n = 5). G, granulosa; T, theca.

B
P<O.Ol

. I

P<O.Ol
- T

Fig. 10. Progesterone production by 1 x lo5 granulosa cells and androstenedioneproduction by 1 x lo6 theca cells during 24 h culture. The data are represented by mean ? SEM (control group n = 5; cocultured group n = 10). A: Progesterone production. B: Androstenedione production.

accompanied by cellular hyperplasia (Gier and Marion, theca cells became small and convex (Figs. 4, 5) and 1961; Henderson and Moon, 1979; Skinner and Osteen, that they formed multilayer sheets (Fig. 8) also support 1988). Consistent with these observations are the this concept, since granulosa cells in the in vivo antral present findings that granulosa cells cultured alone follicle form multilayer sheets and are smaller than showed progesterone production which was not affected luteinized granulosa cells. Incomplete luteinization may explain why granulosa by gonadotropin treatment, therefore indicating spontaneous functional luteinization. However, progester- cells cultured with theca cells demonstrated rapid one production by cocultured granulosa cells was de- growth (Fig. 9) and why FSH augmented their prolifcreased to one-third of the level for monoculture (Fig. eration, while LH diminished it (Table 1). Granulosa 10) and LH treatment increased their progesterone cells of the developing follicle wall grow rapidly, and production (Table l), suggesting that the process is less FSH acts synergistically with estrogen to exert a miadvanced in the presence of theca cells. Morphologic togenic effect on these cells in vivo (Hsueh and Adashi, findings that granulosa cells cultured together with 1984). When these cells luteinize in response to LH

106

F. KOTSUJI AND T. TOMINAGA

TABLE 1. Gonadotropin effects on cell number and steroidogenic activity of granulosa cells cultured alone or with theca cells'
Control culture Coculture

FSHUntreated
Granulosa cell number ( x 105/disk) Progesterone production (ng/i05cells) Estradiol production (pg/i05 cells) treated

LHtreated
5.5 f 0.3 43
f5

FSH + LHtreated Untreated

FSHtreated
12.1 f 0.2*
19 f 6

LHtreated

FSH + LHtreated
11.1 f 0.5 22 184 4*

5.3 f 0.2 37
f4

5.1 f 0.3

5.2 c 0.4 41 c 4 156 t 22

9.9 f 0.3

6.8 t 0.3* 27 t 5* 65 t 19

42 f 4

16 f 2

N.D.

N.D.

54 f 11

N.D.

N.D.

* 26

'Granulosa cell numbers were determined on day 10 of culture and progesterone and estradiol production was assayed during the last 24 h culture. During the last 24 h of culture, androstenedione was added into culture medium (final concentration of lo-' molil). Control: Granulosa cells were cultured alone. Coculture: Granulosa cells were cultured with theca cells. FSH-treated: Granulosa cells were treated with FSH alone for 8 days from day 2 of culture. LH-treated: Granulosa cells were treated with LH alone for 8 days from day 2 of culture. FSH + LH-treated Granulosa cells were treated with FSH alone for 7 days between days 2 and 8 and then treated with LH during day 9. Data are represented as the mean ? SEM. N.D., not detectable. *Significantly different from untreated (P< 0.05).

TABLE 2 . Effects of LH on growth and androstenedione production by theca cells cultured alone or
with granulosa cells'
Control culture Untreated LH-treated Untreated 22.1 Coculture LH-treated

Theca cell number

( X 105/disk) Androstenedione production (ng/i05 cells)

11.6 t 0.4 32.1 t 4.1

11.6 ? 0.3 27.1 2 3.3

* 0.4

22.8 5 0.8 133.4 2 8.3*

71.2 f 6.3

'Theca cell numbers were determined on day 10 of culture and androstenedione production was assayed during the last 24 h culture. Control: Theca cells were cultured alone. Coculture: Theca cells were cultured with granulosa cells. LH-treated: Theca cells were treated with LH alone for 7 days from day 3 of culture. Data are represented as the mean 2 SEM. *Significantly different from untreated (P< 0.05).

stimulation, they lose their proliferative capacity (Henderson and Moon, 1979). We speculate that the luteinization of granulosa cells cultured with theca cells was promoted by LH, resulting in a reduction of their proliferative activity. During folliculogenesis, theca cells respond to LH primarily by activation of the side-chain cleavage enzymes and the 17-20-desmolase enzyme systems (Erickson et al., 1985), allowing these cells to synthesize androgens. When bovine theca cells are luteinized, both P-45OI7, and its messenger ribonucleic acid decline to undetectable levels (Rodgers et al., 1986; Rogers et al., 19871, resulting in a decrease in basal androstenedione production and the loss of responsiveness to LH stimulation. More recently, Roberts and Skinner (1990) also reported that androgen production by bovine theca cells cultured on a collagen-coated plastic surface declined to undetectable levels with prolongation of the culture period and that human chorionic gonadotropin had a negligible effect on androgen production by these cells. In the present study, androstenedione production by theca cells cultured alone was smaller than that of theca cells in coculture and was not affected by LH treatment. In contrast, basal androgen producing activity of theca cells was increased and LH stimulation significantly augmented androstenedione production when theca cells were cultured with granulosa cells (Fig. 10; Table 2). Thus, it is possible that theca cells cultured alone are also functionally luteinized but that the process is not complete

for theca cells in combined culture. Theca cells cultured together with granulosa cells became convex and formed thicker cellular layers than theca cells cultured alone (Figs. 7, 8). Theca cells in coculture grew more rapidly than theca cells in single culture (Fig. 9). These findings may also support this concept, since theca cells grow rapidly and form multilayer sheets in the in vivo ovarian antral follicle and lose their proliferative capacity after luteinization (Ireland, 1987). Taken together, it is suggested that cocultured granulosa and theca cells display more characteristics of in vivo antral follicle cells compared to cells cultured alone. It was reported that bovine non-luteinized follicular granulosa cells can aromatize exogenous androgen to estradiol (Lacroix et al., 1974). In this study, however, even FSH-treated granulosa cells in coculture did not produce estradiol. More recently, two groups investigated the effects of FSH on estradiol production by cultured bovine granulosa cells derived from antral follicles using a serum-free cell culture system. Skinner and Osteen (1988) reported that aromatase activity was highest on day 1of culture and was stimulated by FSH, while both aromatase activity and FSH-induced estradiol production were not detected on day 6 of culture. They also reported that no response of granulosa cells to FSH was detected when serum was added to the culture medium. Saumande (1991) investigated the effect of FSH on bovine granulosa cells cultured with or without fibronectin and reported that the cells main-

GRANULOSA-THE(:A INTERACTION

107

tained their aromatase activity for at least 4 days and that estradiol production peaked a t 2 ngiml of FSH, decreasing at higher concentrations. Our experiment was performed on day 9 of culture and the culture medium contained bovine calf serum. Therefore, it is possible that the responsiveness of granulosa cells to FSH was lost before day 9 of culture. Another possibility is that the culture medium used and/or the concentration of FSH in the present experiment was not suitable for studying the effect of FSH on estradiol production. Granulosa cells cultured alone or with theca cells produced estradiol when treated with LH, suggesting that LH may play an important role in the aromatizing activity of these cells. Pretreatment of granulosa cells with FSH seemed to enhance the effect of LH on estradiol production by these cells through some unknown mechanism since FSH + LH was more effective than LH alone.

monal effects on intracellular and secreted casein in cultures of mouse mammary epithelial cells of floating collagen membrane. Proc. Natl. Acad. Sci. U S A . , 74:4466-4470. Emerman, J.T., and Pitelka D.R. (1977) Maintenance and induction of morphological differentiation in dissociated mammary epithelial function on floating collagen membranes. In Vitro, 13:316-328. Erickson, G.F., Magofin, D.A., Dyer, C.A., and Hofeditz, C. (1985) The ovarian androgen producing cells: A review of structureifunction relationships. Endocr. Rev., 6:371-399. Fortune, J.E. (1986) Bovine theca and granulosa cells interact to promote androgen production. Biol. Reprod., 35:292-299. Gier, H.T., and Marion, G.B. (1961) Formation of the bovine corpus luteum. J. Dairy Sci., 441187-1192. Goff, A.K., Leung, P.C.K., and Armstrong, D.T. (1979)Stimulatory action of FSH and androgens on the responsiveness of rat granulosa cells to gonadotropins in vitro. Endocrinology, 104:1124-1129. Hadley, M.A., Djakiew, D., Eyers, S.W., andDym, M. (1987) Polarized secretion of androgen-binding protein and transferrin by Sertoli cells grown in a bicameral culture system. Endocrinology, 120: 1097-1103. (1979) Luteinization of bovine granHenderson, K.M., and Moon, Y.S. ulosa cells and corpus luteum associated with loss of androgenaromatizing ability. J . Reprod. Fertil., 56:89-97. Hsueh, A.J.W., and Adashi, E.Y. (1984)Hormonal regulation of the CONCLUSIONS differentiation of cultured ovarian granulosa cells. Endocr. Rev., 5:76-127. We have developed a useful in vitro cell culture system for the investigation of communication between Ireland, J.J. (1987)Control of follicular growth and development. J. Reprod. Suppl., 34:39-54. two types of follicular cell. Using this system, the role Kessel, B.,Fertil.Y.X., Jia, X.C., and Hsueh, A.J.W. (1985) Liu, Autocrine of granulosa and theca cell interactions in ovarian role of estrogens in augmentation of luteinizing hormone receptor formation in cultured rat granulosa cells. Biol. Reprod., 32:1038structure and function was investigated. Our results 1050. suggested that communication between these two Koors, R.D., and Hansel, W. (1981) large and small cells of the The types of cells results in reciprocal modulation of their bovine corpus luteum: Ultrastructural and functional differences. growth, morphological structure, and function and that In: Dynamics of Ovarian Function. N.B. Schwartz, and M. Hunzicker-Dunn, eds. Raven Press, New York, pp. 197-203. the actions of gonadotropins at the target cell level are Bovine also modulated by cellular communication. The obser- Kotsuji, F., Kamitani, N., Goto, K., and Tominaga, T. (1990) theca and granulosa cell interactions modulate their growth, morvations further suggest that granulosa and theca cells phology, and function. Biol. Reprod., 43:726-732. cultured in combination more closely resemble in vivo Lacroix, E., Eechaute, W., and Leusen, I. (1974) biosynthesis of The estrogens by cow follicles. Steroids, 23:337-356. follicular cells than cells cultured alone. Therefore, we speculate that interaction of granulosa and theca cells Roberts, A.J., and Skinner, M.K. (1990)Hormonal regulation of thecal cell function during antral follicle development in bovine ovais essential for maintaining the original structure and ries. Endocrinology, 127:2907-2917. function of the bovine follicular wall. Rodgers, R.J., Waterman, M.R., and Simpson, E.R. (1986) Cytochromes P-450,,,, P 4 0 , , adrenodoxin, and reduced nicotina-5,, REFERENCES mide adenine dinucleotide phosphate-cytochrome P-450reductase in bovine follicles and corpora lutea. Changes in specific contents Anderson, E. (1989) vitro and in vivo effects of androgens on rat In during the ovarian cycle. Endocrinology, 118:1366-1374. ovarian granulosa cells. Am. J . Obstet. Gynecol., 160:782-788. Armstrong, D.T., and Dorrington, J.H. (1976)Androgens augment Rogers, R.J., Waterman, M.R., and Simpson, E.R. (1987)Levels of messenger ribonucleic acid encoding cholesterol side-chain cleavFSH-induced progesterone secretion by cultured rat granulosa age cytochrome P-450,17u-hydroxylase cytochrome P-450, adrenocells. Endocrinology, 9 91411-141 4. doxin, and low density lipoprotein receptor in bovine follicle and Bendell, J.J., Lobb, D.K., Chuma, A., Gysler, M., and Dorrington, J.H. corpora lutea throughout the ovarian cycle. Mol. Endocrinol., (1988) Bovine theca cells secrete factorb) that promote granulosa 1274-279. cell proliferation. Biol. Reprod., 38:790-797. Bicsak, T.A., and Hsueh, A.J.W. (1988)Cell to cell communication in Saumande, J. (1991)Culture of bovine granulosa cells in a chemically defined serum-free medium: The effect of insulin and fibronectin on the ovary. In: Cell to Cell Communication in Endocrinology. F. the response to FSH. J. Steroid. Biochem. Mol. Biol., 38:189-196. Piva, C.W. Bardin, F. Forti, and M. Motta, eds. Raven Press, New Skinner, M.K., and Coffey, R.J., J r . (1988)Regulation of ovarian cell York, pp. 167-180. growth through the local production of transforming growth facChrambard, M., Verrier, B., Gabrion, J., and Mauchamp, J . (1983) tor-p by theca cells. Endocrinology, 123:2632-2638. Polarization of thyroid cells in culture: Evidence for the basolateral Developmental and hormonal localization of the lodide pump and of the thyroid-stimulating Skinner, M.K., and Osteen, K.G. (1988) regulation of bovine granulosa cell function in the preovulatory hormone receptor-adenyl cyclase complex. J . Cell. Biol., 96:1173follicle. Endocrinology, 123:1668-1675. 1177. Donjacour, A.A., and Cunha, G.R. (1991) Stromal regulation of epi- Ursely, J., and Leymarie, P. (1979)Varying response to luteinizing hormone of two luteal cell types isolated from bovine corpus luteum. thelial function. Cancer Treat. Res., 53:335-364. Emerman, J.T., Enami, J., Pitelka D.R., and Nandi, S. (1977)HorJ . Endocrinol., 83:303-310.