Indian Journal of Experimental Biology Vol. 38, December 2000, pp. 1183-1191

Human granulosa cells in vitro: Characteristics of growth, morphology and influence of some cytokines on steroidogenesis*

A F G Stevenson

Institute for Toxicology, University of Kiel , Brunswiker Str. 10, D-24105 Kie1, Germany.

Granulosa ce1l.s (GCs) were characterised morphologically by light and electron microscopy. The steroidogenic capa­bility of GCs in vitro was estimated by radioimmunoassay (RIA): oestradiol (E:z), progesterone (P) and androstenedione (A) secreted into the culture medium were measured. The influence of several culture media and anchorage of the cell either to plastic vessels (monolayer) or to collagen fibrils (in gel) were studied. As the various culture media were assayed with re­gard to their suitability for IVF, it was found that Ham's FlO is quite satisfactory (in agreement with other observations on embryo cultures). A chemically defined medium BM 86 was found to be inadequate. In addition to the two cell types which are known, a third cell type which can perform efficient aromatisation (E:z production) in vitro is characterised here. The in~ fluence of cytokineslgrowth factors (GF) like insulin-like GF (lOF-I), epidermal GF (EGF), platelet-derived GF (POOF) and fibroblast GF (FGF) on steroidogenesis was tested either alone or with human chorionic gonadotrophin (hCG). Except for oestradiol (E:z) from early GCs, hCG generally stimulated progesterone (P) and E:z secretion. EGF by itself enhanced the secretion of P but not of E:z. EGF did not affect hCG stimulation of P, but reduced that of E:z. In contrast, in pre-ovulatory GCs IGF-I reduced the stimulatory effect of hCG on both E:z and P. In early GCs IGF-I potentiated hCG stimulation of P. In early GCs, neither hCG nor IGF-I nor a combination of lOF-1 with hCG had any effect on E:z production.

The development of ovarian follicles (folliculo­genesis) is a highly complex physiological process which involves not only several hormones of the endocrine system but also locally acting autocrine/paracrine growth factors which are known as cytokines. The influences of some of these on the growth and steroid producing behaviour of granulosa cells (GCs) of a few animal species have variously been reported. With the practice of in vitro fertilisation (IVF), human GCs have become available for experimental studies and current interests also focus on the role of cytokines as modulators of ovarian function. Despite a large number of reports, a clearcut picture has yet to emerge on the regulatory mechanisms involved in the triggering of the primordial oocyte to folliculogenesis, selection of the dominant follicle from the cohort of antral follicles and its development until ovulation, followed by luteinisation. GCs utilised in most experimental work have been obtained from Graafian Follicles and in the case of human GCs they are generally pre-ovulatory GCs. Thus, findings from such experimental studies represent temporal specificities in a much broader spectrum of events.

*Dedicated to Gertraud Teuchert-Noodt, PhD, MD, professor of Neuroanatomy and Human Biology, University of Bielefeld, Germany, on the occasion of her 60th Birthday.

Steroidogenesis is regulated by gonadothropins. Morphologically GCs are peculiar because they demonstrate histological ambivalence. Until luteinisation the GCs are true epithelial cells delimited by a basement membrane. On luteinisation, the basement membrane is destroyed. Vascularisation of the granulosa follows and the tissue trans­differentiates. The resulting granulosa lutein cells acquire an altered phenotype and a new steroidogenic profile.

GCs can grow in culture without any need of special growth factors. Pre-luteal GCs in culture have been reported to show a limited proliferative potential of about one population doubling (PD)l. Pre-luteal GCs may, thus, represent another peculiarity . since somatic euploid cells, excepting cardiomyocytes, generally have a large proliferative potential (50-70 PDs) in culture. The proliferative activity to which the granulosa cells are subjected during their growth from a primordial to a pre-ovulatory Graafian follicle could plausibly be the cause of proliferative exhaustion.

In vitro studies usually employ monolayers, which is the growth of cells on a two-dimensional growth surface. Under monolayer conditions, two populations of distinct morphological types have been described2

-4.

The reported cell types are essentially a stellate fibroblastoid type and a polygonal epithelioid type. In collagen gels which permit growth in a

1184 INDIAN J EXP BIOL, DECEMBER 2000

three-dimensional matrix - the GCs express a completely different phenotype. One of the features of GCs is a conspicious hypertrophy of the cells discernible in monolayers in culture. Whether hypertrophy also occurs in collagen gel cultures, is' at present unresolved. Acquisition of an altered phenotype in collagen gel is accompanied by altered steroidogenic activity and reactions to stimulatory factors5

.

Of the generally known cytokines, the insulin-like growth factors (lGFs) probably constitute the major part of the studies. The effects of IGFs have been studied alone6-lO, in conjunction with somatotrophin ll

.13

or with gonadotrophins14. GCs have been reported to secrete IGF15 and IGF binding proteins l6 which are thought to in tum inhibit IGF action17

• GCs have receptors for IGF and IGF-1 has been shown to stimulate GC proliferation18, 19.0n the one hand, IGF-1 has been noted to interact with follicular stimulating hormone (FSH) in stimulating luteinising hormone (LH) receptor exrression8, and in stimulating steroid secretion by GCs .On the other hand, FSH and LH.are reported to promote the binding of IGFs to GCs7

. In contrast, epidermal growth factor (EGF) was reported to inhibit FSH induced development of LH receptors 20. EGF also inhibited hCG induced steroid secretion and ovulation in the rabbit21 . Basic fibroblast growth factor (FGF) was suggested to induce intraovarian oocyte maturation22 and to regulate FSH stimulation23. Platelet derived growth factor (PDGF) was found to enhance GC proliferation in combination with EGF and IGF-l 13

•

This paper deals with the steroidogenic activity of human GCs from pre-Iuteinised follicles of patients who opted for IVF. Comparison of cultures grown either as monolayers or in collagen gels are presented. Some results on the · influence of culture media, gonadotrophins (hCG) and effects of IGF-1, EGF, FGF and PDGF on the biosynthesis of steroids in pre­ovulatory human GCs cultured as monolayers are presented. The fine morphology of these cells under different culture conditions is discussed in connection with their steroidogenic behaviour.

Materials and Methods Granulosa cells (GCs)

Pre-ovulatory granulosa cells were obtained from follicular aspirates for oocyte retrieval from about 40 patients who opted for assisted reproduction at the Gynaecological Clinics of the University. The patients were hyperstimulated with human menopausal

gonadotrophin (hMG). Ovulation was induced by injection of 10000 I.u. of hCG. GCs from Graafian follicles from natural cycles were obtained from 13 patients with regular cycles who underwent laparoscopy because of benign gynaecological ailments between day 7 to 13 of cycle. These patients gave informed consent to follicular aspiration.

Separation and culture of GCs Follicular aspirates were centrifuged 10 min at

175 g. The cell pellets were resuspended in cold phosphate buffered saline (PBS), pooled and layered over 45% Percoll (Pharmacia, Sweden) in PBS and recentrifuged 20 min at about 700 g. When present, red blood corpuscles (RBC) pelleted at the tip of the conical tubes while the intact GCs settled at the interphase of the supernatant and Percoll. The GCs were picked with pasteur pipettes, washed in PBS, spun down and counted in Neubauer Cytometers after resuspension in a small volume of PBS and staining with trypan blue (TB). Approximately 5x104/0,5 ml viable GCs (TB exclusion was 70 to 95%) in culture medium were plated into multi well (growth surface 1,5 cm2) plastic culture dishes (Nunc, Denmark).

Culture media Several commercially available culture media were

tested for their adequacy in supporting GC cultures. With the exception of one chemically defined medium BM 86 (Boehringer Mannheim, Germany) which was obtained ready for use, all other media were made by dissolving the respective powders in distilled water of injection quality. The maximal storage time under refrigeration was 4 weeks. The only supplements to the media (excepting BM 86) were 10% foetal calf serum (FCS) and antibiotics (100 IU/ml penicillin + 100 J!g/ml streptomycin). The following media: MEM with Earle's salts, Ham's FlO, RPMI 1640, Medium 199, McCoy's 5a and BM 86 (all from Boehringer Mannheim, Germany) were bicarbonate-buffered and were, therefore, kept in an atmosphere of 5% CO2 in air. Two other media CMRL 1415 and Leibowitz 15 (SeromedlBiochrom, Berlin,Germany) were amino acid buffered and were, therefore, kept in air. All cultures were incubated at 37° C with saturated humidity.

For collagen gel cultures, only Ham's FlO medium was employed, after it was established that this medium performed better than any other. Rat tail collagen solution 2 mg/ml in 0,1% acetic acid was obtained from Serva, Heidelberg, Germany. The cultures were set by mixing at ice-bath temperature 7

STEVENSON: HUMAN GRANULOSA CELLS IN CULTURE 1185

parts of collagen solu~ion, 1 part of lOx concentrated Ham's FlO and 2 parts NaHC03 (11.76 mglml). GCs suspended in FCS were incorporated by replacing 1 part of the mixture with FCS containing the cells. Thus, the final gel cultures contained 10% FCS.

Stimulation of GC cultures The influence of hCG on steroidogenesis was

tested. A lOOx concentrated hCG (Pregnesin, Serono, Germany) stock solution in PBS was made and kept at -20°C. After dilution in culture medium hCG was used at a concentration of 10 IV/mi.

Radioimmunoassay (RIA) The steroidogenic activity of the GC cultures was

moni'tored through daily estimations of oestradiol (Ez), progesterone (P) and androstenedione (A) secreted into the culture medium. The concentrations of these honnones in spent medium was measured by radioimmunoassay (RIA) using commercially available RIA kits (Biermann, Bad Nauheim, Gennany). 125I-activity was measured in a well-type crystal scintillation counter (LKB, Sweden).

Microscopy Routine inspections of GC cultures were done with

an inverted microscope equipped with phase contrast optics and a camera for microphotography. For transmission electron microscopy (TEM), cultures were fixed with 3% glutaraldehyde in cacodylate buffer at pH 7,4 containing 3% sucrose. Post-fixation in 2% osmic acid was followed by dehydration in ethyl alcohol. Specimens were embeded in araldite. Thin sections cut with a Reichert ultramicrotome using either glass knives or a diamond knife, were contrasted with uranyl acetate and lead citrate and examined in a Zeiss EM 9 electron microscope at 60 kV.

Results

General features of GCs in monolayers After plating GCs attached within 24 hr. The GCs

acquire stellate to elongated fibroblastoid shapes depending on density. Isolated single cells tend to be stellate while those in low density cukures displaying cell to cell contact appear elongated. Early cultures showed predominantly one or the other cell fonn, depending on dispersion. Fig. IA shows such an early culture of elongated cells in loose contact. The inset demonstrates stellate cells. The predominantly stellate cells seen in Fig. 2A, despite cell to cell contact, are the result of hCG treatment.

Often (in about 85% of cases), beginning about the third to fourth day of culture (or later), isolated clusters of actively proliferating oval or polygonal epithelioid cells appeared (Fig. IB) thus representing a second population. Active proliferation could be documented through the rapid increase in numbers as well as the presence of mitotic cells. The origin and honnonal relevance of these cells is unclear and will be discussed. Only rarely, a third cell type could be identified. These cells were in morphology and growth pattern indistinguishable from diploid human fibroblasts (HDF) in culture (Fig. IC). Owing to possible follicular mural ruptures during flushing (for oocyte retrieval), the presence of fibroblasts in the aspirates may be expected. However, the steroidogenic activity of these cells excludes this possibility. . Longer culture duration led to hypertrophy and loss

of filopodia. Excepting the larger size, such GC cultures (Fig. ID) resembled a senescing phase ill HDF culture. When the second epitheloid cell type (Fig. lB) appeared, they overgrew the fibroblastoid type.

The fine structure of monolayer GCs seen under TEM usually showed cytopathological features. The presence of condensed mitochondria, vacuoles and areas of rarified cytoplasm were typical. The vacuoles probably contained lipids which got lost during processing for TEM (see Fig. 2B and C). Completely degenerated cells were frequently observed. The presence of smooth endoplasmic reticulum (SER) was difficult to detect in monolayer GCs. The nuclei were predominantly euchromatic with only scarce peripheral heterochromatin. The presence of stress fibres (Fig. 2D) i~ typical for monolayer GCs.

Influence of culture media on steroidogenesis The synthesis of androstenedione (A), progesterone

(P) and oestradiol (E2) was detected through the presence of these honnones in the culture medium. The kinetics of A and P synthesis were closely correlated. Fig. 4 shows daily concentrations of P in the different culture media which were tested. The curves for A production are not shown since the kinetic pattern is identical to that for P. Both A and P production increased during the first three days of culture, reaching peak values about the third day, and declining to base-line values by the end of the first week. Noteworthy, was the difference between the perfonnances of two culture media: Ham's FlO

1186 . INDIAN J EXP BIOL, DECEMBER 2000

STEVENSON : HUMAN GRANULOSA CELLS IN CULTURE 1187

Fig. 3-GCs in collagen gel culture. (A) Phase contrast microphotograph showing a typical GC collagen culture. The inset shows a single cell with arborescent filopodia. (8) TEM of a gel GC. The relative absence of cytopathological features is conspicious. The arrows indi­cate intercellular contacts (x2500). (C) TEM of a GC in collagen gel showing smooth endoplasmatic reticulum (arrow heads) and tubular mitochondria. L = lipid droplet, N = nucleus and arrow points to filopodial contact (x8000). (D) TEM of a native GC obtained from the follicular aspirate. The condensed mitochondria is attributable to unavoidable time lags before cells from the aspirate can be processed for TEM (x8000).

[ pper panel] Fig. I-Phase contrast microphotographs of monolayer cell types grown from follicular aspirates. Arrow heads point to mitotic cells. (A) A typical early GC culture 24 to 48 hr after plating. The cells are in loose contact and represent the primary GC popula­tion. The inset shows isolated stellate cells in early culture. (8) Oval or polygonal cells which appear later. (C) HDF-like cells which also appear later. (D) Typical late cultures (about 2 weeks). The cell type here belongs to the primary population shown in A. The inset depicts isolated late cells which have hypertrophied. Microscope magnification was 20 X 10.

[Lower panel] Fig. 2-Primary GC type in monolayers. (A) Altered morphology upon hCG treatment: lamellipodia are retracted and replaced by filopodia. The inset shows the altered morphology after treatment with hCG either in combination with IGF-I or EGF. Mi­croscope magnification was 20 X 10 (8) TEM overview of monolayer GC (xI800). (C) TEM of monolayer GC showing cytoplasmic details. Mitochondria (M) are condensed; lipid droplets are typically lost leaving vacuoles (L) (x8000). (D) Late monolayer GCs hyper­trophy and often reveal extensive stress fibres (arrow head) (x1800)

showed high P and A levels with rather broad peaks while BM 86, which is a chemically defined medium (i.e. no serum supplementation) specially designed for the culture of mammalian cells (particularly hybridomas), performed exceedingly poorly owing to cell death. Cell growth and hormone biosynthesis could, however, be restored (not shown), by supplementing BM 86 with 10% FCS.

The synthesis of ~ dropped drastically within 48 hr of culture (see Fig. 5). The inverse levels of P and A imply the inability of the cells to synthesise aromatic

precursors of~, also shown by the fact that the initial ~ levels for the individual media corresponded approximately · to those of final P and A. The steroidogenic activity of monolayer GCs is easily stimulated by the addition of hCG (10 IU/ml) to the culture medium (see Fig. 6 for P and Fig. 7 for E2,). When on rare occasions the third cell type (vide supra), whose morphology was indistinguishable from typical human diploid fibroblasts (HOF) , appeared in cultures, a resurgence of E2 levels was measured. These cells were initially seen at about the seventh or

INDIAN J EXP BIOL, DECEMBER 2000

2

1200 (j) 1000

':2 800

I 800

~ R 400

200

0 0 2

2 3

2 3

3

. 3 4

4

day

4

day

5 6

day

5 6

day

6 7 8

c

5 8 7 8

b

Q

7 8 10 12 14

7 8 10 12 14

Fig. 4--Influence of different culture media on the secretion of progesterone in monolayer GC cultures. [(a) MEM, (b) HAM's FlO, (c) RPMI, (d) M 199, (e) McCoy, (0 CRML, (g) L IS, (h) BM86] Fig. S-Influence of diferent culture media on the secretion of oestradiol in monolayer GC cultures. [(a) MEM, (b) HAM's FlO, (c) RPMI, (d) M 199, (e) McCoy, (0 CRML, (g) L IS, (h) BM 86] Fig. 6--Comparison of progesterone secretion between mono­layer and collagen gel cultures treated either with or without hCG in vitro. [(a) Monolayer, (b) Monolayer + hCG, (c) Collagen gel, (d) Collagen gel + hCG] Fig. 7--Comparison of oestradiol secretion between monolayer and collagen gel cultures treated either with or without hCG in vitro. [(a) Monolayer, (b) Monolayer + hCG, (c) Collagen gel, (d) Collagen gel + hCG]

eight day (later than the ovaVpolygonal ones). The steep rise in ~ seen at the beginnig of the second week (see Fig. 8) could be explained on the basis of the exponential increase in cell number. P and A levels were less dramatically affected. This reactivation of aromatisation was observed only in medium 199, MEM and McCoy, not in the other media that were tested. The co-culture of GCs with a strain of HDF (IMR-90) did not mimic this effect.

GC morphology in collagen gels In collagen gels, GCs expressed a completely

different phenotype. As shown in Fig. 3A the cells showed ramified filopodia. The perinuclear cytoplasmic volume became greatly reduced presumably due to the extensive formation of fllopodia. No cellular hypertrophy was noticeable with increasing age of culture. Although collagen gel cultures always appeared scanty in cell numbers when compared to monolayers for the same initial plating, cell counts performed after enzymatic resuspension of the GCs revealed the same numbers for both gels and monolayers. Thus, the apparent density was the result of altered spatial distribution.

TEM fine structure (Fig. 3B & C) was free of cytopathologica] alterations, indicating healthy cells. The mitochondria were not condensed and lipid droplets which are typical for pre-Iuteinised GCs remained intact (cf. within Fig. 3 panels B & C with D which shows a native GC from an aspirate). Rough endoplasmic reticulum (RER) and SER were clearly identifiable, and SER predominated both here and in native GCs. Further, the mitochondria seen in Fig. 3C are of the tubular type characteristic of steroid producing cells. In short, it can be said that the fine structure of gel culture GCs was closely akin to that of native GCs. Frequent cell to cell contacts and the presence of gap junctions were observed.

GC steroidogenesis in collagen gels Figures 6 and 7 show the comparative hormone

production of P and ~ for GCs as monolayers or in . gel cultures, respectively. Ham's FlO medium was used for all gel cultures. As in the case of monolayers, the kinetics of A and P were identical and are, therefore, not shown.

The addition of hCG to monolayer cultures stimulated hormone production significantly. Although a similar trend was observed also for GCs in collagen gels, the increase was statistically not significant. The major difference between the steroidogenic activities in mono]ayers and coHagen

STEVENSON : HUMAN GRANULOSA CELLS IN CULTURE 1189

gels was the significantly higher base-line hormone production levels of the gel cultures.

Influence of some cytokines on steroidogenesis In monolayer cultures, at a concentration of 25

nglml culture medium, IGF-l alone did not affect the steroid biosynthetic activity of pre-ovulatory GCs. As shown in Figs. 9 and 10 the secretion of P and ~, respectively, were both unaffected by IGF-l alone. When IGF-l was applied together with hCG, the stimulatory effect of hCG was reduced. In gel cultures, neither hCG alone (data not shown) nor in combination with IGF-l (Figs. 9 and 10) affected hormone production significantly.

The temporal secretory pattern for both P and E2 from early GCs which were obtained from young follicles from natural cycles, was essentially the same as that of pre-ovulatory Gcs, as shown in Figs. 11 and 12. hCG could only slightly stimulate P secretion but not that of ~. Neither P nor E2 secretion was influenced by IGF-l alone. In combination with hCG, IGF-l provoked a significant elevation of P secretion (Fig. 11). This was accompanied by morphological change to a predominantly fibroblastoid phenotype with bipolar cells, similar to the change observed for hCG and EGF acting together as shown in the inset in panel A, Fig. 2. EGF stimulated P secretion (Fig 11). The level of EGF stimulation was lower than that of hCG stimulation and EGF did not interact with hCG. The level of EGF stimulation was significant from day 5 to 10. EGF alone or in combination with hCG inhibited E2 secretion (Fig: 12).

FGF and PDGF were tested at concentrations of I and 5 nglml culture medium (data not shown). Both of these growth factors did not affect P and ~ secretion, respectively, in pre-ovulatory GCs. Morphological effects were also absent.

Discussion The presence of two cell types in cultured GCs

from pre-Ieuteinised human follicles is known from the literature2. 3. We have, in addition, observed a third cell type whose unique E2 biosynthetic potential was maximally express~d in Medium 199. Whether follicular GCs in vivo consists of a homogenous or a mixed population of cells is not definitely known. The phenotype of the major population of these cells in culture (stellate and elongate cells) is not typical of epithelial cells. No attempt has been made to demonstrate the secretion of a basal lamina or laminin in GCs cultured from preluteinised follicles, although

5 6 7 8 10 12 14

day

5 6 8 10 12 14

day

1200 @ 1000

b

J 800 a

600 c

i 400 e

200

0 2 3 4 5 6 8 10 12 14

day

Fig. 8-Oestradiol secretion of HDF-Iike GCs in culture grown as monolayers. A rebound in ~ production is observed after the first week. Medium 199 but not Ham's FlO supported the restoration of ~ production. [(a) Medium 199, (b) HAM's FlO]

Fig. 9-Comparison of the influences of IGF-l and IGF-I in combination with hCG on the secretion of progesterone in pre­ovulatory GCs in monolayer and gel cultures, respectively. [(a) HAM's FlO, (b) HAM's FlO + hCG (10 IU/mI), (c) HAM's FlO + IGF (25 nglml), (d) HAM's F 10 + hCG (10 IU/mJ) + IGF (25 nglmJ), (e) Collagen gel, (f) Collagen gel + hCG + IGF (i5 nglmJ)]

Fig. lO-Comparison of the influences of IGF-l and IGF-l in combination with hCG on the secretion of oestradiol in pre­ovulatory GCs in monolayer and gel cultures, respectively. [(a) Monolayer, (b) Monolayer + hCG, (c) Collagen gel, (d) Collagqn gel + hCG + IGF (25 nglml), (e) HAM's F 10 + IGF(25 nglml) , (f) HAM's F 10 + hCG (10 IU/mJ) + IGF (25 ng/mJ)]

Schmidt et af demonstrated in 1984 the presence of Call-Exner Bodies. If the features described here for the major cell po;mlation is representative of leuteinised cells24

, 2, then the oval or polygonal subpopulation may be regarded as the non-Iuteinised fraction or, in other words, the true epithelial GC

1190 INDIAN J EXP BIOL, DECEMBER 2000

2()()

~ 150 I!

i ~ 1()() Q.

:E 01 c: fil 50 ><

0

20

15 :2 ! ~ 10

i 5

0

®

2 3

L

2

4 5

3

day

4

day

d I

6 II 10 12 14

5 6 7 6

Fig. II-Comparison of the influences of EGF and IGF-l either alone or in combination with hCG on the secretion of progester­one in early GCs in monolayer cultures. [(a) HAM's F 10, (b) HAM's F 10 + hCG (10 IU/ml), (c) HAM's F 10 + EGF (2 nglml), (d) HAM,s F 10 + hCG (10 IU/mI) + EGF (2 nglml), (e) HAM's FlO + IGF (25 nglml), (f) HAM's 10 + hCG (10 IU/mI) + IGF (25 ng/ml)]

Fig. 12-Comparison of the influences of EGF and IGF-I either alone or in combination with hCG on the secretion of oestradiol in early GCs in monolayers. [(a) HAM's F 10, (b) HAM's F 10 + hCG (10 IU/mI), (c) HAM's F 10 + IGF (25 nglml), (d) HAM's F 10 + hCG (10 IU/mI) + IGF (25 ng/ml), (e) HAM's 10 + EGF (2 nglml), (f) HAM's FlO + hCG (10 IU/mI) + EGF (2 ng/ml)]

component observed in culture. As luteinisation progresses, this fraction vanishes. Despite active steroidogenesis, the fine structure of the oval cells does not demonstrate the features regarded as typical for luteinisation i.e. the accumulation of lipids and the abundance of SER. Basing on ultrastructural features, these cells would not be considered to be steroid producing cells.

Despite its spectacular ability to carry out aromatisation in medium 199, the third cell type described here is also devoid of lipid droplets and SER. Obviously, GCs are able to realise their steroidogenic potential through an alternative pathway. The mitochondria, instead of the SER, might serves as the site of synthesis26. It is also worthy of mention that mitochondrial pleomorphism was frequently observed in the GCs. It may be said that both the oval and IIDF-like GCs display an

abundance of structurally intact mitochondria. The origin of the IIDF-like cells may be disputable and it remains to be shown whether these cells are GCs or cells of thecal origin.

Despite interest in the sustained production of E2 in chemically defined culture medium27

, no other study has compared the influence of culture media on the steroidogenic activity of GCs. The interesting finding on the restoration of aromatisation, i.e. E2 production in the presence of a third population of cells, in certain culture media, should be critically appraised with regard to composition. A perusal of the components of medium 199, MEM and McCoy draws attention to the higher concentrations of Ca2

+ in these media. The importance of calcium in steroidogenesis has been discussed by others28. 29. Interestingly the relative levels of Ca++ correlates to E2 production. As molecular O2 is known to be essential for the process of aromatisation, the availability of O2 ought to be investigated in view of the high levels of P and A. Cells in monolayers are covered by a 2 to 3 mm layer of medium which certainly exceeds the mean diffusion path of O2 several fold. In collagen gels, the degree of hypoxia would depend on the distance of the cells from the culture surface. The active production of Ez by the IIDF-like cells in medium 199 etc. may depend on another mechanism of production, as already discussed.

In this context, it is certainly interesting to know how hCG re-activates E2 production in monolayers and why the same reaction fails to occur in collagen gel GCs. A peculiar correlation between the production of steroid by GC and changes in morpholo~y has been reported for monolayer cultures3

, 0, 31 . These observations were affirmed for monolayers but not for GCs in gels. The refractoriness to hCG observed for gel GCs could, therefore, be attributed to the inability of gel GCs to undergo certain reactions which become manifested as morphological changes observed in monolayers.

Although the gel GCs do not respond to hCG stimulation, a higher base-line level of Ez, P and A production respective was recorded. Despite the better correlation of ultrastructural features between gel and native GCs, failure to react to hCG stimulation makes it questionable which of the two culture systems (monolayer vs. collagen gels) reflects GC-behaviour in vivo. Elucidation of the mode of action of FSH and LH together with the intricate role of cellular, microenvironmental and growth factors would contribute to understanding the nature of GCs.

STEVENSON : HUMAN GRANULOSA CELLS IN CULTURE 1191

Follicular growth is the nett result of a concerted action of several concornittant factors in addition to gonadotrophins.

Of the four frequently studied cytokines that have been tested here, two of them viz. FGF and PDGF showed absolutely no effects with regard to steroidogenesis and morphology. Owing to the already mentioned complexity of granulosa development during folliculogenesis, a possible involvement of these two growth factors at some other developmental stage connot be ruled out, even if unlikely, because of th5 epithelial nature of the granulosa until luteinisation commences, which is at or about the time of ovulation. Thereupon, the granulosa transdifferentiates into a non-epithelial tissue and sensitivity to FGF and PDGF could develop.

Conversely, GC reaction to EGF is only to be expected. The nature of this reaction is, nonetheless, interesting. It is the only growth factor of the four tested which by itself stimulated steroidogenesis and also provoked a morphological change in the GCs. This could be interpreted as an indication of the still intact epithelial nature of the GCs shortly before ovulation. Steroidogenic stimulation was otherwise recorded only for hCG which could stimulate both P and E2 production. In the presence of EGF, hCG stimulation of E2 was inhibited.

Although the greater part of studies have concentrated on IGF which comprises of a group numbering from I to 5, the data here on IGF-l shows that IGF-1 by itself does not influence steroidogenesis. It also by itself did not provoke morphological change. This was the case for both pre­ovulatory as well as early GCs from developing follicles . However, IGF-l consistently interacted with hCG in that it altered hCG effects. The nature of this interaction depended on the age of the GCs. In mature GCs from the pre-ovulatory follicles IGF-1generally inhibited hormone production, but in young GCs, the reverse was the case. This enhancing effect was, however, selective and restricted only to P secretion.

Acknowledgement The support of the Gynaecological Clinics of the

University and technical aid provided by Regine Gaetje, graduate student, is acknowledged.

References 1 Orly J, Sato G. & Erickson GF, Cell, 20 (1980) 817.

2 Schmidt C L, Kendall J Z, Dandekar P V, Quingley M M & Schmidt K L, J Reprod Fert, 71 (1984) 279 .

3 Hurwitz A, Laufer N, Keshet I, Rabinowitz R, Lewin A Palti & Schenker J G, Fertil Steril, 48 (1987) 791

4 Gutierrez C G, Glazyrin A L, Robertson G W, Campbell B K, Gong J G, Bramley T A & Webb R, Mol Cell Endocrinol, 134 (1997) 51.

5 Sites C K, Kessel B & LaBarbera A R, Proc Soc Exp Biol Med, 212 (1996) 78 .

6 Adashi E Y, Resnick C E, D 'Ercole A J, Svoboda M E & van Wijk J J, Endocr Rev, 6 (1985) 400

7 Adashi E Y, Resnick C E, Fernandez E R, Hurwitz A, Rob­erts C T, Leroith D & Rosenfeld R, Ann NY Acad Sci , 626 (1991) 161.

8 Marno, T, Hayashi M, Matsuo H, Ueda y, Morikawa H & Mochizuki M, Acta Endocrinol, 117 (1988) 230.

9 Willis D S, Mason H D, Watson H & Franks S, J Clill Elldo­crillol Metab, 83 (1998) 1256.

IO Maestro M A, Planas J V, Moriyama S, Gutierrez J, Planas J & Swanson P, Gen Comp Endocrinology, 106 (1997) 189.

11 Bergh C, Carlsson B, Olsson J H, Billig H & Hillensjo T, Ann NY Acad Sci, 626 (1991) 169.

12 Foster C M, Hashimoto N, Randolph J F & Beitins I Z, Pedi­atr Res, 38 (1995) 763.

13 Gong J G, McBride D, Bramley T A & Webb R, Endocrinol­ogy, 143 (1994) 157.

14 Pawshe C H, Rao K B & Totey S M, Mol Reprod Dev, 49 (1998) 277.

15 Hsu C J & Hammond J M, Endocrinology, 120 (1990) 198. 16 Mondschein J S, Smith S A & Hammond J M, Endocrinol­

ogy, 127 (1990) 2298 . 17 Guidice L C, von Dessel H J, Cataldo N A, Chandrasekher

Y A, Yap 0 W & Fauser B C, Prog Growth Res, 6 (1995) 397.

18 May J V, Frost J P & Bridge A J, Endocrinology, 126 (1990) 2896.

19 Ovesen P, lngerslev H J, Orskov H & Ledet T, Endocrinol­ogy, 140 (1994) 313 .

20 Mondschein J S & Schomberg D W, Science, 211 (1981) 1179.

21 Endo K, Atlas S J, Rone J D, Zanagnolo V L, Kuo T C, Dharmarajan A M, & Wallsch E E, Endocrinology, 130 (1992) 186.

22 LaPolt P S, Yamamoto M, Veljkovic M, Sinciah C, Ny T, Tsafriri & Hsueh A J W, Endocrinology, 127 (1980) 2357.

23 Oury F & Darbon J-M , Biochem Biophys Res Comm, 156 (1988) 634

24 Crisp T M & Channing C P, Bioi Reprod, 7 ( 1972) 55. 25 Spaniel-Borowski K, & Sterzik K, Arch Gynecol, 240 (1987)

137. 26 Marsh J M, Bioi Reprod, 19 (1976) 30. 27 Gutierrez C G, Campbell B K & Webb R, Bioi Reprod, 56

(1997) 608. 28 Carnegie J A & Tsang B K, in Factors requlating ovarian

junction, edited by GS Greenwald & PF Terranova (Raven press, New York), 1983,375.

29 Carnegie J A & Tsang B K, Bioi Reprod, 36 (1987) 1007 30 Soto E A, Kliman H J, Strauss III J F & Pavola L G, Biol Re­

prod, 34 (1986) 559. 31 Carnegie J A, Byard R, Dardick I & Tsang B K, Bioi Reprod,

38 (1988) 881