Lymphocytes stimulate progesterone production by cultured human granulosa luteal cells Nobuyuki Emi, MD, Hideharu Kanzaki, MD, Masumi Yoshida, MD, Kenji Takakura, MD, Masatoshi Kariya, MD, Norihiko Okamoto, MD, Kimitoshi Imai, MD, and Takahide Mori, MD Kyoto, Japan After follicular rupture, massive invasion of blood vessels with neovascularization of the developing corpus luteum takes place, providing many chances for direct contact of luteal cells with resident and migrating immune cells. We studied the effects of peripheral blood lymphocytes on progesterone production by human granulosa luteal cells isolated from women undergoing in vitro fertilization. During 6 days of culture, progesterone production by granulosa luteal cells was significantly increased when they were cultured together with autologous or allogenic peripheral blood lymphocytes. This stimulatory effect was also observed on the addition of medium conditioned with peripheral blood lymphocytes and was synergistic with gonadotropin stimulation. The activity was present in the fraction retained by ultrafiltration with a 30,000 molecular weight cutoff filter and was preserved after heating at 56° C for 30 minutes but disappeared after heating at 70° C for 15 minutes. These findings suggest that lymphocytes infiltrating the corpus luteum during early luteinization can stimulate the function of human granulosa luteal cells through the action of some protein-like humoral factor(s) of higher molecular weight than that of previously identified Iymphokines and indicate a possible paracrinologic regulatory role for lymphocytes in ovarian function. (AM J OSSTET GVNECOL 1991 ;165:1469-74.)
Key words: Human granulosa luteal cell, progesterone production, lymphocyte, cytokine In recent years a growing body of evidence has accumulated to suggest that the immune system is an additional local regulator of ovarian function. It is well known that immunoincompetent or immunosuppressed animals show numerous reproductive disorders. 1-3 Thymic control of ovarian function has clearly been demonstrated by the experimental induction of oophoritis in neonatally thymectomized mice}' 5 Macrophages have been identified in the developing corpus luteum in several species, including man,"-s and have been suggested to be putative intraovarian regulators. One macrophage-derived cytokine, interleukin-l, has been observed to suppress the functional and morphologic luteinization of cultured murine and porcine granulosa cells. g , 10 Another cytokine, tumor necrosis factor-a, has been reported to inhibit the gonadotropin-dependent differentiation of murine granulosa cells and to cause complex dose-dependent alterations in the elaboration of progesterone and androstenedi-
one. 1l - 13 On the other hand, lymphocyte secretory products (lymphokines) have been shown to have an effect on steroidogenesis in cultured rat granulosa luteal cells. However, it is unclear whether the function of granulosa luteal cells can be modulated by other immune cells, especially lymphocytes, which should playa central role in the immune response, and whether the stimulatory effect can be observed in an autologous combination occurring inside ruptured follicles in vivo. Accordingly, we examined the effect of lymphocytes of autologous origin and the conditioned medium derived from lymphocytes on progesterone production during the culture of granulosa luteal cells obtained from patients at oocyte retrieval in our in vitro fertilization program. The ovarian stimulation was controlled completely by exogenous gonadotropins that mimic the normal ovulatory cycle. Thus granulosa luteal cell culture for 6 days after oocyte retrieval is considered to be an appropriate in vitro model for studying early luteinization.
From the Department of G.vnecology and Obstetrics, Faculty ofMedicine, Kyoto University. Supported in part by Grants-in-Aid for Scientific Research (Nos. 01570928 and 02222104) from the Ministry of Education, Science and Culture ofJapan. Received for publication August 27, 1990; revised April 9, 1991; accepted April 22, 1991. Reprint requests: Hideharu Kanzaki, MD, Department ofGynecology and Obstetrics, Faculty of Medicine, Kyoto Universitv 54, Shogoin Kawahara-cho, Sakyo-ku, Kyoto 606, Japan. 6/1/30438
Material and methods Granulosa luteal cells were obtained from patients undergoing superovulation in the in vitro fertilization program at Kyoto University Hospital. They had primary or secondary infertility with fallopian tube occlusion but no endocrinologic disorders. Follicular development was induced with 150 IU human menopausal gonadotropin (Pergonal, Teikoku-zoki, Tokyo) given 1469
1470 Emi et al.
from day 3 of the cycle and with 900 fLg gonadotropinreleasing hormone analog (buserelin acetate, Hoechst Japan, Tokyo), provided as a nasal spray, daily from day 1 of the cycle. Follicular maturation was assessed each day by measurement of serum estradiol levels and ultrasonographic scanning. When the leading follicle reached 18 mm in diameter, human menopausal gonadotropin was discontinued and 5000 lU human chorionic gonadotropin (hCG) (hCG Mochida, Mochida Pharmaceutical, Tokyo) was injected 52 hours after the last human menopausal gonadotropin administration. Oocytes were recovered 36 hours after the injection of hCG through the ultrasonographically guided transvaginal route. Granulosa luteal cells were obtained by using a heparinized buffer (Dulbecco's phosphate-buffered saline solution, Gibco, Grand Island, N. Y.) to wash the follicles so as to enhance the recovery of oocytes. After removal of the oocytes, the granulosa luteal cells were pooled, centrifuged at 200 g for 20 minutes, and resuspended in RPMI 1640 medium (Gibco). Separation of the granulosa luteal cells from red blood cells was achieved with a Ficoll gradient (Lymphocyte Separation Solution, Nakalai Tesque, Kyoto). The viability of the granulosa luteal cells was assessed by trypan blue staining and was usually >90%. The granulosa luteal cells were then centrifuged and resuspended in RPMI 1640 medium containing 10% fetal calf serum and 100 U I ml penicillin (complete medium). The cells were plated at a density of 1 x 105 viable cells per well in 24-multiwell culture plates (Corning Glass Works, Corning, N.Y.) containing 1 ml complete medium and precultured for 24 hours at 37° C under humidified 5% carbon dioxide in air. Peripheral blood mononuclear cells were isolated from male or female volunteers by means of centrifugation with a Ficoll gradient at 24 hours after oocyte retrieval. Peripheral blood mononuclear cells were resuspended in complete medium and then counted with a hemacytometer. They were then incubated for 1 hour at 37° C in plastic dishes that had been precoated with autologous serum for 15 minutes at 37° C. Nonadherent cells were collected from the dishes and were shown to contain 96% Leu M3-positive cells and were defined as monocytes. Granulosa luteal cells were cultured for 6 days with the medium being replaced every 2 days, and the basal
November 1991 Am J Obstet Gynecol
progesterone production was determined. To examine the effect of hCG on progesterone production, hCG was added to the granulosa luteal cell culture in concentrations from 5 to 50,000 mlU/ml. Autologous or allogenic peripheral blood mononuclear cells, peripheral blood lymphocytes, and monocytes prepared as described were added to the precultured granulosa luteal cells at a density of I x 10' to 2 x 106 viable cells per well. Each well contained 2 ml complete medium, and cells were added in the presence or absence of hCG (500 mllU Iml). All cultures were prepared in triplicate, and the number of granulosa luteal cells in each well was counted by the citric acid-crystal violet method at the completion of culture. Conditioned medium was prepared by culturing autologous or allogenic peripheral blood lymphocytes in complete medium for 6 days and added to granulosa luteal cell cultures. In addition, we examined the effect of conditioned medium after heat treatment (56° C for 30 minutes and 70° C for 15 minutes). Conditioned medium from peripheral blood lymphocytes was also prepared in RPMI 1640 containing 0.5% bovine serum albumin (BSA, Sigma, St. Louis) and subjected to ultrafiltration with a centrifugal microconcentrator (Centricon-30; 30,000 molecular weight cutoff; 2000 g for 30 minutes; Amicon Division, W.R. Grace, Danvers, Mass.). The retained or filtered fractions were added to the granulosa luteal cell cultures. Spent culture medium was frozen and assayed for progesterone by radioimmunoassay with a commercial kit (progesterone-iodine 125 kit; Sorin Biomedica, Saluggia, Italy). Progesterone production was expressed in micrograms per 10 5 cells per 48 hours or 6 days. All data are presented as the mean ± SD of triplicate cultures. Student's t test was used for statistical analysis. Results
Cultured granulosa luteal cells derived from preovulatory follicles after exogenous administration of hCG in vivo secreted large amounts of progesterone during the first 2 days of culture. In the subsequent culture period, from 2 to 6 days, basal progesterone production tended to decrease. Progesterone production by granulosa luteal cells was little enhanced even by the addition of hCG during the first 2 days of culture. However, from 2 to 6 days, the level of progesterone production of granulosa luteal cells could be boosted by doses of 50 to 50,000 mIU/ml hCG (Fig. 1).
When 1 x 105 granulosa luteal cells were cultured with various densities of autologous peripheral blood mononuclear cells, 1 x 10' or 1 X 10 5 peripheral blood mononuclear cells did not affect progesterone production by granulosa luteal cells in comparison with
Lymphocytes stimulate human luteal function
Volume lli5 :\umher 5. Pan I
,... .r=
IX)
Key words: Human granulosa luteal cell, progesterone production, lymphocyte, cytokine In recent years a growing body of evidence has accumulated to suggest that the immune system is an additional local regulator of ovarian function. It is well known that immunoincompetent or immunosuppressed animals show numerous reproductive disorders. 1-3 Thymic control of ovarian function has clearly been demonstrated by the experimental induction of oophoritis in neonatally thymectomized mice}' 5 Macrophages have been identified in the developing corpus luteum in several species, including man,"-s and have been suggested to be putative intraovarian regulators. One macrophage-derived cytokine, interleukin-l, has been observed to suppress the functional and morphologic luteinization of cultured murine and porcine granulosa cells. g , 10 Another cytokine, tumor necrosis factor-a, has been reported to inhibit the gonadotropin-dependent differentiation of murine granulosa cells and to cause complex dose-dependent alterations in the elaboration of progesterone and androstenedi-
one. 1l - 13 On the other hand, lymphocyte secretory products (lymphokines) have been shown to have an effect on steroidogenesis in cultured rat granulosa luteal cells. However, it is unclear whether the function of granulosa luteal cells can be modulated by other immune cells, especially lymphocytes, which should playa central role in the immune response, and whether the stimulatory effect can be observed in an autologous combination occurring inside ruptured follicles in vivo. Accordingly, we examined the effect of lymphocytes of autologous origin and the conditioned medium derived from lymphocytes on progesterone production during the culture of granulosa luteal cells obtained from patients at oocyte retrieval in our in vitro fertilization program. The ovarian stimulation was controlled completely by exogenous gonadotropins that mimic the normal ovulatory cycle. Thus granulosa luteal cell culture for 6 days after oocyte retrieval is considered to be an appropriate in vitro model for studying early luteinization.
From the Department of G.vnecology and Obstetrics, Faculty ofMedicine, Kyoto University. Supported in part by Grants-in-Aid for Scientific Research (Nos. 01570928 and 02222104) from the Ministry of Education, Science and Culture ofJapan. Received for publication August 27, 1990; revised April 9, 1991; accepted April 22, 1991. Reprint requests: Hideharu Kanzaki, MD, Department ofGynecology and Obstetrics, Faculty of Medicine, Kyoto Universitv 54, Shogoin Kawahara-cho, Sakyo-ku, Kyoto 606, Japan. 6/1/30438
Material and methods Granulosa luteal cells were obtained from patients undergoing superovulation in the in vitro fertilization program at Kyoto University Hospital. They had primary or secondary infertility with fallopian tube occlusion but no endocrinologic disorders. Follicular development was induced with 150 IU human menopausal gonadotropin (Pergonal, Teikoku-zoki, Tokyo) given 1469
1470 Emi et al.
from day 3 of the cycle and with 900 fLg gonadotropinreleasing hormone analog (buserelin acetate, Hoechst Japan, Tokyo), provided as a nasal spray, daily from day 1 of the cycle. Follicular maturation was assessed each day by measurement of serum estradiol levels and ultrasonographic scanning. When the leading follicle reached 18 mm in diameter, human menopausal gonadotropin was discontinued and 5000 lU human chorionic gonadotropin (hCG) (hCG Mochida, Mochida Pharmaceutical, Tokyo) was injected 52 hours after the last human menopausal gonadotropin administration. Oocytes were recovered 36 hours after the injection of hCG through the ultrasonographically guided transvaginal route. Granulosa luteal cells were obtained by using a heparinized buffer (Dulbecco's phosphate-buffered saline solution, Gibco, Grand Island, N. Y.) to wash the follicles so as to enhance the recovery of oocytes. After removal of the oocytes, the granulosa luteal cells were pooled, centrifuged at 200 g for 20 minutes, and resuspended in RPMI 1640 medium (Gibco). Separation of the granulosa luteal cells from red blood cells was achieved with a Ficoll gradient (Lymphocyte Separation Solution, Nakalai Tesque, Kyoto). The viability of the granulosa luteal cells was assessed by trypan blue staining and was usually >90%. The granulosa luteal cells were then centrifuged and resuspended in RPMI 1640 medium containing 10% fetal calf serum and 100 U I ml penicillin (complete medium). The cells were plated at a density of 1 x 105 viable cells per well in 24-multiwell culture plates (Corning Glass Works, Corning, N.Y.) containing 1 ml complete medium and precultured for 24 hours at 37° C under humidified 5% carbon dioxide in air. Peripheral blood mononuclear cells were isolated from male or female volunteers by means of centrifugation with a Ficoll gradient at 24 hours after oocyte retrieval. Peripheral blood mononuclear cells were resuspended in complete medium and then counted with a hemacytometer. They were then incubated for 1 hour at 37° C in plastic dishes that had been precoated with autologous serum for 15 minutes at 37° C. Nonadherent cells were collected from the dishes and were shown to contain 96% Leu M3-positive cells and were defined as monocytes. Granulosa luteal cells were cultured for 6 days with the medium being replaced every 2 days, and the basal
November 1991 Am J Obstet Gynecol
progesterone production was determined. To examine the effect of hCG on progesterone production, hCG was added to the granulosa luteal cell culture in concentrations from 5 to 50,000 mlU/ml. Autologous or allogenic peripheral blood mononuclear cells, peripheral blood lymphocytes, and monocytes prepared as described were added to the precultured granulosa luteal cells at a density of I x 10' to 2 x 106 viable cells per well. Each well contained 2 ml complete medium, and cells were added in the presence or absence of hCG (500 mllU Iml). All cultures were prepared in triplicate, and the number of granulosa luteal cells in each well was counted by the citric acid-crystal violet method at the completion of culture. Conditioned medium was prepared by culturing autologous or allogenic peripheral blood lymphocytes in complete medium for 6 days and added to granulosa luteal cell cultures. In addition, we examined the effect of conditioned medium after heat treatment (56° C for 30 minutes and 70° C for 15 minutes). Conditioned medium from peripheral blood lymphocytes was also prepared in RPMI 1640 containing 0.5% bovine serum albumin (BSA, Sigma, St. Louis) and subjected to ultrafiltration with a centrifugal microconcentrator (Centricon-30; 30,000 molecular weight cutoff; 2000 g for 30 minutes; Amicon Division, W.R. Grace, Danvers, Mass.). The retained or filtered fractions were added to the granulosa luteal cell cultures. Spent culture medium was frozen and assayed for progesterone by radioimmunoassay with a commercial kit (progesterone-iodine 125 kit; Sorin Biomedica, Saluggia, Italy). Progesterone production was expressed in micrograms per 10 5 cells per 48 hours or 6 days. All data are presented as the mean ± SD of triplicate cultures. Student's t test was used for statistical analysis. Results
Cultured granulosa luteal cells derived from preovulatory follicles after exogenous administration of hCG in vivo secreted large amounts of progesterone during the first 2 days of culture. In the subsequent culture period, from 2 to 6 days, basal progesterone production tended to decrease. Progesterone production by granulosa luteal cells was little enhanced even by the addition of hCG during the first 2 days of culture. However, from 2 to 6 days, the level of progesterone production of granulosa luteal cells could be boosted by doses of 50 to 50,000 mIU/ml hCG (Fig. 1).
When 1 x 105 granulosa luteal cells were cultured with various densities of autologous peripheral blood mononuclear cells, 1 x 10' or 1 X 10 5 peripheral blood mononuclear cells did not affect progesterone production by granulosa luteal cells in comparison with
Lymphocytes stimulate human luteal function
Volume lli5 :\umher 5. Pan I
,... .r=
IX)
