The effect of insulin on progesterone production and cellular growth in long-term cultures of human granulosa lutein cells

FERTILITY AND STERILITY Copyright" 1987 The American Fertility Society

Vol. 48, No.5, November 1987 Printed in U.S.A.

The effect of insulin on progesterone production and cellular growth in long-term cultures of human granulosa lutein cells

Arye Hurwitz, M.D.*t Neri Laufer, M.D.* Ilana Keshet, Ph.D.§ Rony Rabinowitz, M.D.*

Aby Lewin, M.D.* Zvi Palti, M.D.* Joseph G. Schenker, M.D.*

Hadassah Hebrew University Sclwol of Medicine, Jerusalem, Israel

The direct action of insulin on human granulosa lutein cells (GLCs) in long-term cultures obtained from in vitro fertilization (IVF) cycles was investigated. Progesterone (P) secretion by GLC increased progressively in both basal and human chorionic gonadotropin (hCG; 100 mIU Iml) stimulated conditions up to 4 days in culture, and plateaued thereafter. Insulin (0.0025 mU/ml to 2500 mU/ml) had no effect on either basal or hCG stimulation during the culture period. GLC in culture formed a monolayer and multiplied at a rate of approximately once every 3 days. Neither morphology nor cell division was affected by insulin in supraphysiologic levels (25 mU/ml). These results suggest that GLC obtained from preovulatory follicles in an IVF program are already stimulated maximally by in vivo exposure to high doses of human menopausal gonadotropin (hMG)/hCG administered to the women. Contrary to its stimulatory effect on early preovulatory granulosa cells, insulin does not affect P production, cellular morphology, or growth rate of luteinized granulosa cells. Fertil Steril 48:791, 1987

Over the last few years, clinical and experimental data has accumulated to suggest a direct effect of insulin on ovarian steroidogenesis. Clinically, it was demonstrated that ovarian function is disrupted in women with diabetes mellitus. 1 In addition, insulin has been implicated as a possible cause for polycystic ovarian disease (PCOD) because of the association of hyperinsulinemia with elevated luteinizing hormone (LH) and testosterone (T) levels. 2,3 Several recent reports examined experimentally the effect of insulin on interstitial cells in the ovary. Barbieri et a1. 4 found that insulin Received March 9,1987; revised and accepted June 25, 1987.

* Department of Obstetrics and Gynecology, Hadassah Mt. Scopus. t Reprint requests: A. Hurwit~, M.D., Department of Obstetrics and Gynecology, Hadassah University Hospital, Mount Scopus, P.O. Box 24035, Jerusalem 91240, Israel. :j: Department of Obstetrics and Gynecology, Ein Kerem. § Department of Cellular Biology. Vol. 48, No.5, November 1987

augmented the stimulatory effect of LH on progesterone (P) production in porcine theca cells, and stimulated androgen accumulation in ovarian stroma obtained from women with hyperandrogenism.5 Similarly, insulin has been shown to promote the growth of granulosa cells in culture, enhance their P production, and induce LH/human chorionic gonadotropin (hCG) receptors. 6 - 8 Because no information exists concerning the role of insulin in the human corpus luteum, this study was undertaken to investigate the effect of insulin on P production, growth, and morphologic features of granulosa lutein cells (GLCs) in long-term cultures as a model of early corpus luteum function. MATERIALS AND METHODS

Cell Isolation

GLCs were retrieved from follicles obtained from normal ovulatory women undergoing ovulation inHurwitz et aI. Effect of insulin on granulosa cells

791

duction for in vitro fertilization (IVF). The follicles were stimulated by human menopausal gonadotropin (hMG) hCG, as described previously.9 All accessible follicles were harvested at the time of ultrasonography transvesically and oocytes removed thereafter from follicular aspirates. The follicular fluid was centrifuged at 300 X g for 10 minutes and the supernatant was decanted. Cells from all follicles aspirated from each woman were combined. The cells were layered onto 5-ml, 50% percoll columns (Pharmacia, Vppsala, Sweden) and centrifuged at 200 X g for 30 minutes to pellet red cells. Thereafter, these cells were aspir~ ated from the interface, resuspended, and counted in a hemocytometer. Viability tested by trypan blue was above 90% in all cases.

to the cells for 10 minutes and was decanted. The cells were dissolved in 2 ml NaOH 0.2 N. H3 concentration was measured with a liquid scintillation counter and was expressed as counts/min/24 hours. Cell Count

After cell attachement to the well at 2, 4, 6, and 8 days, respectively, the medium was decanted and the well was washed with saline to remove residual serum. Trypsine versene 1% was added for 1 minute, then removed, and thereafter 1 ml medium was added to the well and cell number was assessed with the hemocytometer. Autoradiography

Cells were cultured in triplicates (20,000 to 50,000 cells/well) in 1 ml Ham's F-10 medium (Gibco, Grand Island, NY) containing 10% inactivated human serum in plastic tissue culture wells (Falcon 3047, Falcon Plastics, Oxford, CA). The cells were incubated at 37°C in an atmosphere of 95% air, 5% CO 2 for 144 hours. Medium was changed every 48 hours with freshly prepared hCG and insulin solutions. After collection, medium was stored at -20°C until steroid assays were performed simultaneously. All experiments were repeated with a minimum of three separate cell preparations (different women).

The cells were incubated with thymidine H3 (1 JLCi/ml) for 24 hours. Fixation of the cells was performed by washing with phosphate buffered saline (PBS), one third fixer:two thirds PBS (10 minutes), one half fixer:one half PBS (10 minutes), fixer (10 minutes) (fixer = three parts methanol, one part acetic acid). After drying, the cells were washed with cold TCA 5% for 5 minutes, thereafter 5 minutes with ethanol (twice), and then ethanol 10% for 1 minute. A liquid emulsion (Kodak, Rochester, NY) was layered on the fixed cells in a dark room and the cells were left for exposure for 1 week. The developing was done with the developer and fixer of Kodak. Cells then were stained with Giemsa to delineate cellular borders.

Hormones

Statistical Analysis

Insulin Actrapid (100 IV/ml) was obtained from Novo Industrias (Copenhagen, Denmark), and hCG 5000 IV/ampule from Organon (Oss, Holland).

Differences between group means were calculated using the Student's t-test. Results are presented as mean ± standard error of the mean (SEM).

Cell Culture

Assays

P radioimmunoassay (RIA) was performed by kit code: PROGK (PROGK-M International-CIS, Cedex, France). DNA Labeling

One JLCi/ml thymidine H3 (Code 29B Nuclear Research Centre, Negev Beer Sheva, Israel) was added to the cells at 48, 96, and 120 hours of culture in three separate experiments for 24 hours of incubation. The cells then were washed twice with cold Hank's solution, which contained cold thymidine in excess. Trichloroacetic acid (TCA) 5% was added 792

Hurwitz et al. Effect of insulin on granulosa cells .

RESULTS Effect of Insulin on Basal and Human Chorionic Gonadotropin-Stimulated Progesterone Production

P production by unstimulated and stimulated (hCG 100 mIV/ml) GLCs increased 3- to 5-fold at 96 hours when compared with the first 48 hours of incubation. At 144 hours, P production plateaued in both conditions but was significantly higher (P < 0.05) in the stimulated compared with the unstimulated (basal) condition (Fig. 1). Insulin (0.0025 to 2500 mV/ml) had no effect on P production in Fertility and Sterility

%

::: 100

"

4000

L

co ..., 80

...... ~

""ii

.,~

"'"

~

u '''0 60

c

40

<:>

. - . hCG

0

...'"

3000

20

(100

mlu/mll

. - - - . Basal

c

o u ~:z:

48

96

2000

144 hr

Figure 1 Basal and stimulated (hCG 100 mIU/ml) P production by human GLC during 144 hours of incubation. Mean ± SEM (n = 5 - 9). *P < 0.05 compared with the basal state at 144.

1000

T 96

48

either basal or hCG-stimulated conditions at all incubation periods (Fig. 2). Growth Rate of Granulosa Lutein Cells in Long-Term Cultures

Radioactive count of GLCs after incorporation with labeled thymidine for 24 hours in two separate experiments is presented in Figure 3. In both experiments, there was a significant increase (P < 0.005) in thymidine incorporation at 96 and 120 hours of incubation, compared with 48 hours of

.

_

90

96hr

70

50

120 hr

Figure 3 Increase in radioactive count incorporation of GLC with labeled thymidine for 24 hours after 48, 96, and 120 hours in culture (mean ± SEM). Experiment with 50,000 cells/well (--). Experiment with 30,000 cells/well (- - -). Both experiments were performed in basal conditions only (medium + 10% serum). *P < 0.005 compared with the respective count at 48 hours.

incubation, respectively. Moreover, there was a significant increase (P < 0.001) in cell count for every 2-day period of culture, with an approximate multiplication rate of once every 3 days (Fig. 4). The addition of hCG (100 mIU Iml) or insulin (25 mU/ml) to the growth medium did not affect the growth rate (Fig. 4). Additional proof of cell division after 5 days in culture was obtained by autoradiographic developing demonstrating active incorporation of thymidine H3 into deoxyribonucleic acid (DNA), indicating DNA synthesis prior to mitosis. In some cells, active mitosis could be seen clearly by the presence of two nuclei in one cell

30

....

~

10

CI

90

~~~~~--~~~~~-

..

2

1." hr

-

"i -:!-4 D

~

70

3

"! 50

'0 2

30

Z

~ ~

.me .me

hCG

sme

Ins.

0

10

-

~

tn,. mu/ml

~

o

~

0

d

:::c:i

~

~

1'0

g

on ...

Figure 2 The effect of insulin on basal (open bars) and stimulated (hCG 100 mIU /ml, crossed bars) P production per 48 hours of incubation. Upper panel, 48 hours; Middle panel, 96 hours; lower panel, 144 hours. Mean ± SEM (n = 5 - 9). *P < 0.05 compared with basal conditions at 144 hours. Vol. 48, No.5, November 1987

Days

in culture

Figure 4 Effect of hCG (100 mIU/ml and insulin [Ins] 25 mU/ml) upon cell count in long-term cultures compared with the basal condition (mean ± SEM). s, 10% serum; m, medium; c, cell (mean ± SEM). There is a significant increase in cell count between each group (P < 0.001).

Hurwitz et aI. Effect of insulin on granulosa cells

793

Figure 5 Photomicrograph of GLCs in long-term cultures (5 days) showing the cells with a dark nucleus with some cells in active mitosis (arrows; phase contrast X 400; Olympus inverted microscope model CK, Olympus, Tokyo, Japan).

(Fig. 5). The cells showing mitosis had a granulosa-like appearance with rounded contours and a high nucleus to cytoplasm ratio. DISCUSSION

This study demonstrates that GLCs retain their capacity to produce P and respond to hCG stimulation in long-term primary cultures. Basal P production was maximal after 96 hours of incubation, while hCG further stimulated P secretion slightly after 6 days of culture. These findings corroborate previous observations 10 and suggest that GLCs obtained at the time of IVF may serve as an appropriate in vitro model for early corpus luteum activity. Insulin was shown to exert a number of diverse effects at the cellular level,l1 including enhancement of steroidogenesis by porcine ovarian theca cells,4 human stroma,S and porcine granulosa cells.6 However, its effect on human corpus luteum, luteal tissue, or GLCs has not been investigated. In order to assess insulin's ability to potentiate P production, we used a submaximal stimulatory dose of hCG (100 mIU/ml) because it was found that the maximal stimulatory dose ofhCG for GLCs is 1000 mIU/mL12,13 This consideration evolved from the finding of Ladenheim et al.,14 who showed that no effect of insulin could be discerned when rat luteal cells were exposed to a maximal dose of hCG. Our results demonstrate that insulin at a wide range of concentrations (0.0025 mU/ml to 2500 mU/ml) had no stimulatory effect on P production by long-term culture of either nonstimulated or hCG-stimulated GLCs, despite the presence of insulin receptors on human granulosa cells. 15 This observation is in contrast to the findings of Channing et al.,6 May 794

Hurwitz et al. Effect of insulin on granulosa c,ells

and Schombers,s and Otani et al.,16 demonstrating a stimulatory effect of insulin on basal P production by porcine granulosa, theca cells,4 rat luteal cells,14 and induction of LH/hCG receptors in porcine ovarian cellsP The reason for using a wide range of concentration in this work was based on previous observations that cellular growth was enhanced at supraphysiologic levels (25 mU/mI), while cell metabolism was stimulated at low concentrations. 1S The normal level of insulin in the plasma and follicular fluid is in the range of 1 to 10 ~U /ml. The discrepancy between the latter observations and our own are best explained by the different cell populations used to test the effect of insulin. In most previous observations, the ovarian cells were obtained during a natural cycle either prior to or after a spontaneous LH surge, while GLCs in this study were derived from follicles sti~­ ulated by pharmacologic doses of both FSH/LH (20 ampules/cycle of Pergonal (Serono Laboratories, Randolph, MA) and hCG (10,000 IU/cycle). Whereas cells from natural cycles produce low basal levels of P and are extremely sensitive to hCG, GLCs are strongly stimulated in vivo, produce large amounts of P in culture, and are only mildly sensitive to hCG. Any additional effect of insulin is therefore amplified in cells from natural cycles and dampened in GLCs. For the first time, we were able to show in this cell model that GLCs in culture do actually divide. This evidence was obtained by increase in thymidine H3 incorporation into DNA, increase in cell count, and the demonstration of active mitosis by autoradiography. Based on DNA labeling and cell count, we estimated the rate of cell multiplication to be approximately once every 3 days. This rate, in luteinized GLCs, is similar to the division rate measured for porcine granulosa cells prior to an endogenous LH surge.s Our results contradict previous observations demonstrating that hCG inhibits DNA synthesis of luteinized bovine and rat granulosa cells,19 and that mitotic activity of luteal cells is arrested just after the LH surge. 20 It may be that the latter observed arrest in mitotic activity is only a temporary phenomenon or, alternatively, that our population of cells is heterogenous. It was previously demonstrated that the population of GLCs is composed of two cell types: one subpopulation of large luteinized cells rich in lipid droplets that also contain an abundance of LH/hCG receptors (about 70% of total cell population) and a subpopulation of small cells resembling nonluteinized granulosa cells that are devoid of lipid droplets and have only a few hCG receptor sites (about 30% of Fertility and Sterility

total cell population).21 Furthermore, Schmidt et al. 22 clearly demonstrated that there are two morphologically distinct populations: elongated luteal cells with granular cytoplasm and a round to polygonal granulosa-type cell with an increased nucleus-to-cytoplasm ratio. As we have shown (Fig. 5), it may well be that the population of cells that divide in culture are predominantly of the small, round, granulosa type. These cells probably retain their capacity to divide from the follicular phase and are less active with respect to P production compared with the fully luteinized large cells. These latter cells, which are steroidogenically active, may have truly lost their capacity to divide, as Bomsel-Helmreich et a1. 20 suggested. Such a speculation awaits further confirmation by the use of P monoclonal antibodies. There was no increase in cell count when insulin (25 mU/ml) was added to the culture medium, nor was there any difference in the morphologic appearance of the cells. In conclusion, this work demonstrates that P production increases during the first 96 hours of GLC culture, irrespective of the presence of hCG, but that, at 144 hours, hCG slightly augments P production compared with basal conditions. GLCs were shown to divide at a rate of approximately once every 3 days. Despite the stimulatory effect of insulin on granulosa cells prior to the LH surge, we were unable to demonstrate a similar effect of insulin on human granulosa lutein cells following hCG administration. Acknowledgment. This work is dedicated to the memory of Professor W. Z. Polishuk, a teacher whose concepts outlived his time.

REFERENCES 1. Djursing H, Nyholm HC, Carstensen L, Pedersen LM: Clinical and hormonal characteristics in women with anovulation and insulin treated diabetes mellitus. Am J Obstet Gynecol 143:878, 1982 2. Shoupe D, Kumar DD, Lobo RM: Insulin resistence in polycystic ovary syndrome. Am J Obstet GynecoI147:588, 1983 3. Burgham FA, Givens JR, Kitbachi AE: Correlation of hyperandrogenism with hyperinsulinism in polycystic ovarian disease. J Clin Endocrinol Metab 50:113, 1980 4. Barbieri RL, Makris A, Ryan KJ: Effects of insulin on steroidogenesis in cultured porcine ovarian theca. Fertil Steril 40:237, 1983 5. Barbieri RL, Makris A, Randall RW, Daniels G, Kistner RW, Ryan KJ: Insulin stimulates androgen accumulation in incubations of ovarian stroma obtained from women with hyperandrogenism. J Clin Endocrinol Metab 62:904, 1986 6. Channing CP, Tsai V, Sach D: Role of insulin thyroxine and cortisol in luteinization of porcine granulosa cells grown in clinically defined media. BioI Reprod 15:235, 1976

Vol. 48, No.5, November 1987

7. Veldhuis JD, Kolp LA, Toaff ME, Struass JF, Demers LM: Mechanisms subserving the trophic actions of insulin on ovarian cells. J Clin Invest 72:1046, 1983 8. May JV, Schombers DW: Granulosa cell differentiations in vitro: effect of insulin on growth and functional integrity. BioI Reprod 25:421, 1981 9. Laufer N, DeCherney AH, Haseltine FP, Polan ML, Mezer HC, Dlugi AM, Sweeney D, Nero F, Naftolin F: The use of high dose human menopausal gonadotropin in an in vitro program. Fertil Steril 40:734, 1983 10. Hillensjo T, Sjogren A, Strander B, Nilsson L, Wikland M, Hamberger L, Roos P: Effect of gonadotropins on progesterone secretion by cultured granulosa cells obtained from human preovulatory follicles. Acta Endocrinol (Copenh) 110:401, 1985 11. Goldfine ID: Interaction of polypeptide hormone with cell membranes specific receptors: studies with movlin and glucagon. Life Sci 23:2639, 1978 12. Polan ML, Seu D, Tarlatzis B: Human chorionic gonadotropin stimulation of estradiol production and androgen antagonisms of gonadotropin stimulated responses in cultured human granulos luteal cells. J Clin Endocrinol Metab 62:628, 1986 13. Laufer N, Barr I, Lewin A, Rabinowitz R, Elmaleh U, Kisselevitz R, Diamnt Z, Schenker JG: Aromatase activity in human granulosa lutein cells in long term cultures. Proceedings of the Fourth World Conference on in vitro fertilization. Melburn, November 1985, p 173 14. Ladenheim RG, Tesone M, Charreau EH: Insulin action and characterization of insulin receptors in rat luteal cells. Endocrinology 115:752, 1984 15. Poretsky L, Grigarescu F, Seibel M, Moses AC, Flier JS: Distribution and characterization of insulin and insulin like growth factor receptors in normal ovary. J Clin Endocrinol Metab 61:728, 1985 16. Otani T, Mauro T, Yukimura N, Mochizuki M: Effect of insulin on porcine granulosa cells: implication of possible receptor mediated action. Acta Endocrinol (Copenh) 108:104, 1985 17. Otani T, Mauro T, Ashitaka Y, Tojo S: Multiple inhibitory effects of luteinizing hormone releasing hormone agonist on hCG-dependent receptors in ovarian cells in vivo and in vitro. Endocrinol Jpn 29:597, 1982 18. Coppock DL, Covey LR, Srauss DS: Growth response to insulin in mouse melanoma cells and fibroblast x-melanoma hybrids. J Cell Physiol 105:81, 1980 19. Hsueh AJW, Adashi EY, Jones PBC, Welsh TH: Hormonal regulation of the differentiation of cultured ovarian granulosa cells. Endocr Rev 5:76, 1984 20. Bomsel-Helmreich 0, Gougeon A, Thebault A, Saltarelli D, Milgram E, Frydman R, Papiernik E: Healthy and atretic human follicles in the preovulatory phase: differences in evolution of follicular morphology and steroid content of follicular fluid. J Clin Endocrinol Metab 48:686, 1979 21. Laufer N, Luborsky J, Tarlatzis B, Merino M, Polan ML, DeCherney AH: Asynchrony between oocyte-corona-cumulus complex (Occc) maturation and mural granulosa lutein cell in hMG-hCG cycles for IVF (Abstr). Presented at the SGI 32nd annual meeting, Phoenix Arizona, March 22 to 23,1985 22. Schmidt CL, Kendall JZ, Dandekar PV, Quigley MM, Schmidt KL: Characterization of long term monolayer cultures of human granulosa cells from follicles of different size and exposed in vivo to clomiphene citrate and hCG. J Reprod Fert 71:279, 1984

Hurwitz et al. Effect of insulin on granulosa cells

795