Epidermal Growth Factor Inhibits Large Granulosa Cell Apoptosis by ...

0 downloads 0 Views 885KB Size Report
tained from Charles River Laboratory (Wilmington, MA) and housed under ..... Lederer K, Luciano A, Pappalardo A, Peluso J. Proliferative and steroidogenic ca-.
BIOLOGY OF REPRODUCTION 51, 646-654 (1994)

Epidermal Growth Factor Inhibits Large Granulosa Cell Apoptosis by Stimulating Progesterone Synthesis and Regulating the Distribution of Intracellular Free Calcium' A.M. LUCIAN0,

3

A. PAPPALARDO, C. RAY, and J.J. PELUSO 2

Department of Obstetrics and Gynecology, University of Connecticut Health Center Farmington, Connecticut 06030 ABSTRACT The initial study was designed to determine whether all granulosa cells (GCs) undergo apoptosis in vitro. GCs were isolated from immature rat ovaries and separated on a 15-45% Percoll gradient. Twelve fractions were collected, and GCs were pooled according to size: small GCs (- 50 tZ; fractions 2-5) and large GCs ( 75 2; fractions 6-8). GCs were cultured in serumfree medium for 24 h. After 24 h of culture, fragmented DNA, detected by in situ end labeling of the 3'OH ends of DNA fragments, was observed within 70-80% of large GCs. Similarly, in situ DNA staining demonstrated that at least 50% of large GCs possessed apoptotic nuclei. These degenerative changes in DNA were observed within < 5% of small GCs. These studies demonstrate that in serum-free medium, most large GCs die via an apoptotic mechanism within 24 h. Subsequent studies focused on the mechanism by which epidermal growth factor (EGF) inhibits large GC apoptosis. EGF reduced the percentage of large GCs with apoptotic nuclei from 47 1% for controls to 18 + 2% (p < 0.05). EGF also increased progesterone (P4 ) secretion from large GCs (6.3 0.7 for controls vs. 18.7 + 1.0 ng/ml for EGF treatment; p < 0.05). The effect of EGF on apoptosis was mimicked by P4 and attenuated by the P4 antagonist, RU 486, and aminoglutethimide (AG), an inhibitor of P4 synthesis. The effect of AG was overridden by P4. Therefore, EGF reduces large GC apoptosis by stimulating P4 synthesis, with P4 mediating its action through its receptor. Intracellular free calcium ([Ca2+]i), as assessed by fluo-3 fluorescence, was localized to cytoplasmic foci prior to culture and in the presence of P4. In the absence of P4 or in the presence of RU 486, calcium became dispersed throughout the cytoplasm and ultimately resulted in an apparent increase in [Ca2+] within 28% of the GCs. This suggests that P4 acts through its receptor to prevent a redistribution and increase in [Ca2+]i that may subsequently result in GC apoptosis.

INTRODUCTION

intrafollicular estrogen are associated with GC degeneration [1]. These studies support the concept that gonadotropins maintain estrogen biosynthesis within the follicle and that estrogen acts to prevent GC apoptosis. GCs also undergo DNA fragmentation in vitro, and this fragmentation is prevented by either epidermal growth factor (EGF) or basic fibroblast growth factor [9]. It is possible that estrogen, either acting directly on GCs or indirectly by stimulating the production of intraovarian growth factors, controls [Ca2+]i levels and thus apoptotic DNA fragmentation. However, in rat GCs isolated from eCG-primed immature rats, estrogen neither stimulates nor suppresses basal [Ca2+ ]i levels [10]. This observation supports the concept that estrogen prevents apoptosis indirectly by stimulating the synthesis of an intraovarian factor(s). While this concept of how GCs die is emerging, it is important to appreciate that there is considerable heterogeneity of GCs. Both small and large GCs are present within the immature rat ovary [11-13]. The large GCs secrete estrogen [11, 13] and progesterone [12], whereas the small GCs have a very limited steroidogenic capacity [11-13]. Further, ultrastructural analysis revealed that GCs degenerate by two morphological pathways. In one path, the nucleus becomes lobular and the chromatin condenses prior to alterations within the cytoplasmic organelles [14]. These nuclear changes are characteristic of cells undergoing apoptosis [15]. This pathway is induced by anti-estrogen, supporting the concept that estrogen prevents the GCs from undergoing apoptosis [7]. The alternate pathway is associ-

Over the course of the reproductive life span, only a select few follicles mature and ovulate, with the majority of follicles ultimately undergoing atresia [1]. Although the mechanism responsible for follicular atresia has not been clearly defined, atresia seems to be initiated with the death of the granulosa cells (GCs) [1]. Recent studies have shown that GC death is characterized by DNA fragmentation, suggesting that GCs die through an active process referred to as programmed cell death or apoptosis [2,3]. In several nonovarian cell types, DNA fragmentation is triggered by an increase in intracellular free calcium ([Ca 2+ ]), which activates an endonuclease that eventually results in the cleavage of the DNA into multimers of about 180-bp nucleosomal units with free 3'OH ends [4, 5]. Since GCs possess a calcium/magnesium-dependent endonuclease that is activated by the addition of calcium [6], it has been proposed that an increase in [Ca2+]i initiates GC apoptosis. In vivo studies have shown that hypophysectomy induces and estrogen prevents rat GC DNA fragmentation [7]. In addition, decreased levels of aromatase mRNA [8] and Accepted May 17, 1994. Received March 10, 1994. 1 This work was supported in part by NIH Grant -R55-HD27578-01A2. Carolyn Ray was supported by a Summer Fellowship Grant provided by NIH (Grant Number DK 07661). 'Correspondence. FAX: (203) 679-1436. 3Current address: Instituto di Anatomia degli Animali Domesticic, Facolta di Medicina Veterinaria, Universita degli Studi di Milano, Via Celoria, 10, 20133 Milano, Italy.

646

PROGESTERONE AND GRANULOSA CELL APOPTOSIS ated with degenerative cytoplasmic changes, such as the dilation of the endoplasmic reticulum and mitochondria, in the absence of nuclear changes [14]. Because of this heterogeneity, it is important to assess apoptosis within individual GCs. Since DNA within cells undergoing apoptosis is cleaved, thereby generating numerous DNA fragments with free 3'OH ends, in situ methods have been developed to label these 3'OH ends with digoxigenin- or biotin-dUTP. The labeled nucleotide is subsequently detected by immunofluorescent or immunocytochemical procedures [16]. GCs with fragmented DNA, detected by immunocytochemical procedures, are observed only under conditions that result in GC DNA fragmentation, as revealed by electrophoretic assessment of DNA ladders [7]. Therefore, the initial study was designed to determine whether both small and large GCs undergo apoptosis in vitro. In order to achieve this goal, in situ 3'OH end-labeling and DNA staining procedures were used so that simultaneous determination of cell size and apoptosis could be made. Subsequent investigations were focused on the cellular basis by which EGF inhibits GC apoptosis. MATERIALS AND METHODS Animals Immature female Wistar rats (22 days of age) were obtained from Charles River Laboratory (Wilmington, MA) and housed under controlled conditions of temperature, humidity, and photoperiod (12L:12D; lights-on at 0700 h). Rats were cervically dislocated between 0930 and 1000 h when they were between 25 and 29 days of age. This protocol was approved by the Animal Care Committee of the University of Connecticut Health Center. Preparationof Culture Medium RPMI-1640 medium (GIBCO Laboratories, Grand Island, NY) was used in all culture experiments. It was supplemented with penicillin (0.14 g/L), streptomycin (0.27 g/L), HEPES (4.76 g/L), BSA fraction V (2 g/L), sodium selenite (5 ng/ml), transferrin (5 ,ug/ml), and sodium bicarbonate (2.2 g/L). The pH was adjusted to 7.4, and the medium was filtered through a 0.2-,am filter. EGF (8 nmol/L) and progesterone (320 or 640 nmol/L, depending on the experimental design) were added to the cultures (Sigma Chemical Co., St. Louis, MO). Progesterone (P4 ) was dissolved in ethanol and then diluted in RPMI-1640 to the desired final concentration. RU 486 was provided by Rousel-UCLAF (Romainville, France) and used at a concentration of 64 imol/ L. Aminoglutethimide (AG) was purchased from Sigma Chemical and used at a concentration of 0.5 mmol/L. GC Isolation and Culture GCs were isolated according to the procedure of Rao et al. [11] with slight modifications. Briefly, ovaries were placed

647

in Medium 199 containing 0.2% BSA, 9.1 mM EGTA at pH 7.4. The follicles within the ovary were punctured with 20gauge needles and then incubated for 5 min at 37°C in 5% CO 2:95% air. The ovaries were transferred to Medium 199 containing 0.2% BSA, 2.1 mM EGTA, and 0.5 M sucrose at pH 7.4 and then incubated at 370C in 5% CO2:95% air for 10 min. The ovaries were washed, resuspended in fresh Medium 199 containing 0.2% BSA, and then pressed to release the GCs. The cells were centrifuged for 30 sec at 13 000 rpm, resuspended in 1 ml of media, and then loaded onto the top of a 15-45% Percoll gradient (8 ml total volume). The gradient was centrifuged at 200 g for 5 min; and, starting from the top, 12 fractions of 666 Iil were collected. The cells in each fraction were washed and resuspended in RPMI1640. Small GCs were typically collected in fractions 2-5 and were pooled, while larger GCs were pooled from fractions 6-9. This procedure results in enriched populations of small and large GCs with the small- and large-GC pools being approximately 80% and 50% pure, respectively. Cells from the small- and large-GC pools were plated in 35-mm dishes or on glass cover slips and cultured for up to 24 h in a 5% C0 2 /air atmosphere. In Situ Assessment of Apoptosis Cells undergoing apoptosis were identified by two procedures. In the first procedure, cells with DNA fragments were revealed by extending the DNA fragments with digoxigenin-labeled dUTPs and immunocytochemically detecting these nucleotides. The second procedure involved staining the cells with specific DNA stains and visualizing apoptotic nuclei. For these studies, GCs were cultured for either 2 or 24 h, and the cells were washed 3 times in PBS and then stained to detect either free 3'OH ends of DNA fragments or apoptotic nuclei. In situ end-labeling of DNA fragments. DNA fragments were labeled and detected by use of the reagents and procedures provided in the ApopTag in situ apoptosis detection kit (Oncor, Gaithersdurg, MD). Briefly, GCs were fixed in 10% buffered formalin and washed twice in PBS. The cells were then incubated in a humidified chamber at 37°C for 1 h in the presence of terminal deoxynucleotidyl transferase (TdT) and digoxigenin-11 dUTP and dATP. The cells were washed with buffer and incubated with anti-digoxigenin-fluorescein antibody for 30 min at room temperature. The cells were then washed with buffer and observed under epifluorescence and brightfield optics. Visualization of apoptotic nuclei. The nuclear structures of individual cells were stained with either hydroethidine or propidium iodide. For hydroethidine staining, cells were incubated with hydroethidine (14 Ig/ml of PBS) for 15 min at room temperature in the dark [17]. In some experiments, cells were fixed in methanol, incubated with RNase (100 U/ml for 30 min at room temperature), and stained with propidium iodide according to the instructions provided by Becton Dickinson Immunocytometry Sys-

648

LUCIANO ET AL.

tems (San Jose, CA) as outlined by Dietch and associates [18]. After staining, the cells were observed with a fluorescent microscope with an excitation filter of 535 nm and a barrier filter of 585 nm. At these settings, the chromatin fluoresced bright red. Nuclei that possessed dispersed heterochromatin were considered normal, while apoptotic nuclei were characterized by nuclear fragmentation and/or condensation (see Fig. 2) [15]. Because both DNA stains yielded identical results, the data from these studies were pooled and indicated in the figure legends. Between 100 and 200 cells were observed for each treatment and used to calculate the percentage of cells with apoptotic nuclei. Only in situ DNA staining was used to generate the data shown in Figures 3 and 5. P4 Assay The P4 concentration in the culture medium was determined by a direct P4 RIA (Immunochem Co., Carson, CA). All samples were run in a single assay with an intraassay coefficient of variation of 11% and a sensitivity of 0.5 ng/ ml. The antiserum to P4 cross-reacted 100% with P4 , 4% with 20 alpha-hydroxyprogesterone and corticosterone, and less than 0.04% with several other steroids. Localization of [Ca2+]i [Ca 2+ ]i was localized by use of the calcium-sensitive dye, fluo-3. Fluo-3 was used for these experiments because it is easily loaded into GCs, is nonfluorescent unless bound to calcium, and undergoes an 80-fold enhancement of emission intensity upon calcium binding. In addition, the emission intensity is proportional to the amount of calcium bound [191. GCs were loaded with fluo-3 AM as previously described [20]. For this procedure, a 1-mM solution of fluo-3 AM was prepared by dissolving 50 pLg of fluo-3 AM in 44 pLl of dry dimethyl sulfoxide. Ten microliters of 1 mM fluo-3 AM was added directly to GCs that had previously been plated onto glass cover slips in 1 ml of RPMI-1640 within 35-mm culture dishes. Depending on the experimental design, the medium was supplemented with either 0, 320, or 640 nmol of P4 . The GCs were loaded with fluo-3 AM by placing the culture dishes in the dark on a rotary shaker at 60 rpm at room temperature for 45 min. The cover slip was removed from the 35-mm culture dish, drained of the fluo-3 AM-containing medium, and then exposed to various treatments. This was accomplished by constructing a well of approximately 1 cm in diameter on a 2.54- by 7.62-cm glass microscope slide by use of vaspar, a 20:1 mixture of vaseline and paraffin. This vaspar well was filled with approximately 500 Il of either control medium, or medium supplemented with P4 or P4 and RU 486. The cover slip was slowly lowered onto the vaspar until a complete air-tight seal was formed. The slide was then placed on a heated microscope stage that was maintained at 36°C. The GCs were

sequentially photographed under both brightfield and fluorescent optics over a 2-h culture period. Depending on experimental design, photographs were taken at 10- or 30min intervals. Under these culture conditions, insulin stimulates small GCs to undergo mitosis (unpublished observation). The photographic images were used to assess the cellular localization of the [Ca2+]i. Free intracellular calcium was either localized to cytoplasmic foci or dispersed throughout the GC. In those GCs with dispersed intracellular calcium, the relative fluorescent intensity was classified as moderate or intense. The percentage of cells that showed an increase in fluorescent intensity during the 2-h culture was calculated. StatisticalAnalysis All experiments were replicated at least three times, regardless of the end-point used to assess apoptosis. When appropriate, the data were analyzed by a one-way analysis of variance followed by the Student-Newman-Keuls multiple range test. To assess the effect of RU 486 and/or P4 on the pattern of calcium distribution, the data from all the experiments were pooled. The percentage of cells with a specific pattern of intracellular calcium distribution and/or intensity was analyzed by Chi-square analysis. Regardless of the statistical test, onlyp values < 0.05 were considered to be significant. RESULTS DNA Fragmentation,Apoptotic Nuclei, and GC Size Cells, stained in the absence of TdT or anti-digoxigenin (i.e., negative controls), did not fluoresce, demonstrating that this procedure yields very little background staining. Therefore, few cells were incorrectly classified as being apoptotic. When freshly-isolated small and large GCs were stained in the presence of both TdT and anti-digoxigenin, most (2 90%) of the cells did not fluoresce. In contrast, 70-80% of large GCs fluoresced intensely after 24 h of culture in serumfree medium. Small GCs either did not fluoresce or fluoresced at a relatively low intensity (Fig. 1). Studies using both DNA stains confirmed these observations. Specifically, DNA staining revealed that prior to culture, the DNA within nuclei of freshly isolated large GCs was distributed in a distinct heterochromatin pattern. Nearly 50% of these large GCs possessed nuclei that were either fragmented or condensed after culture (Figs. 2 and 3). These nuclear changes are characteristic of cells that have undergone apoptosis [15]. Small GCs possessed nuclei with uniformly dispersed DNA that did not fragment into apoptotic bodies after 24 h of culture (Fig. 3). EGF and P4 as Regulators of Large GC Apoptosis EGF reduced apoptosis (p - 0.05; Fig. 3) and stimulated a threefold increase in the amount of P4 secreted by large

PROGESTERONE AND GRANULOSA CELL APOPTOSIS

649

FIG. 1. GCs observed under brightfield (A, C,E) and fluorescent optics (B, D, F) to detect the presence of free 3'OH ends of DNA fragments. (A, B) Small GCs collected after 24 h of culture. Note that these cells do not fluoresce even after culture, indicating that they have not undergone apoptosis. Large GCs are shown before (C, D) and after (E,F) culture. Notice that these cells are considerably larger than small GCs shown in A. Before culture, large GCs generally have intact DNA, as indicated by the absence of fluorescence (D). After culture, most of the large GCs fluoresce, indicating that they possess fragmented DNA and are undergoing apoptosis. x400.

GCs over the first 24 h of culture (p - 0.05; Fig. 4). Further, AG suppressed both basal and EGF-induced P4 secretion to nondetectable levels (p - 0.05; Fig. 4). On the basis of the relationship between EGF and P4 secretion, experiments were designed to assess the role of P4 in regulating large GC apoptosis. Compared to serumfree controls, treatment with either RU 486 or AG increased the percentage of apoptotic nuclei present after 24 h of culture (p - 0.05; Fig. 5). Both EGF and P4' reduced GC apoptosis, compared to control values (p - 0.05). However, EGF treatment was not sufficient to inhibit large GC apoptosis

in the presence of either RU 486 or AG. Similarly, P4 did not impede apoptosis in the presence of RU 486 but was effective in maintaining large GC viability in the presence of AG (Fig. 5). Identical results were obtained for large GCs isolated 48 h after eCG treatment (data not shown; n = 2). and Changes in [Ca2 + ]i Distribution After being cultured for 2 h in serum-free medium and then loaded with fluo-3 for 45 min, the large GCs showed three distinct patterns of free calcium distribution. In a few GCs, free calcium was localized to cytoplasmic foci (Fig.

P4

650

LUCIANO ET AL. 50* Fraclions 2-5 |] Fraclions 6-9

O

0)

40-

o

30-

20-

0

10-

.==

a-

Time Oh

r_=

-"

Time

24h

Time

24h

EGF FIG. 3. Effect of EGF (8 nmol) on percentage of small (fractions 2-5) and large (fractions 6-9) GCs undergoing apoptosis after 24 h of culture in serum-free media, as judged by in situ DNA staining. Values represent the mean + standard error of the mean (SEM) of three separate experiments. Two experiments were conducted with hydroethidine, one with propidium iodide. In each experiment, 100-200 cells were observed for each treatment and used to calculate percentage of cells with apoptotic nuclei.

evaluated by plating and loading large GCs with fluo-3 in the presence of P4. When loaded with fluo-3 in the presence of 320 nmol of P4 , free calcium was localized to cytoplasmic foci in 46% of the GCs (n = 85). Compared to the results with the 320-nmol concentration of P4, more GCs that were loaded with fluo-3 in the presence of 640 nmol P4 had free calcium localized to cytoplasmic foci (78%; n = 176, p < 0.05). The remaining cells possessed free cal20 h

FIG. 2. In situ DNA distribution patterns of large GCs stained with hydroethidine before (A) and 24 h after (B) culture in serum-free media. Cells stained before culture (A) show distinct heterochromatin staining pattern; DNA of cells after culture was condensed and/or fragmented (B). x750). Identical results were obtained with propidium iodide staining.

E U e

15 -

L u

10-

coV

6A). Most of the GCs had a diffuse distribution of calcium with moderately intense fluo-3 fluorescence (Fig. 6A) However, several GCs showed a very intense fluo-3 fluorescence. The intense fluorescence was primarily localized within the nucleus (Fig. 6B). Those cells with intense nuclear fluorescence may have been in the initial stages of degeneration, as judged by condensation of nuclear material and cytoplasmic blebbing (Fig. 6C). Since P4 regulated the rate of large GC apoptosis, timedependent changes in distribution of free calcium were

C

0

5-

0) L. M.

0-

Control

EGF

AG

EGF/AG

FIG. 4. Effect of EGF and AG on P4 secretion during a 24-h culture period. Values represent mean + SEM of one representative experiment (n = 4). This experiment was repeated twice.

PROGESTERONE AND GRANULOSA CELL APOPTOSIS

651

FIG. 5. Effect of EGF, P4 , RU 486, and AG on percentage of large GCs undergoing apoptosis after 24 h of culture in serum-free media, as judged by in situ DNA staining. Entire experiment was repeated three times. In each experiment, 100-200 cells were observed for each treatment and used to calculated percentage of cells with apoptotic nuclei. Values represent mean ± SEM of these experiments.

cium that was distributed throughout the cytoplasm with either moderate or intense fluo-3 fluorescence. Over the course of the 2-h incubation with RU 486 in the presence of 320 nmol (n = 122) or 640 nmol (n = 69) of P4 , calcium became distributed throughout the cytoplasm in 28% of the GCs (Fig. 7, B and C). These GCs eventually showed an apparent increase in [Ca2+]i, as judged by an increase in the intensity of fluo-3 fluorescence (Fig. 7D). In contrast, continuous exposure to 320 or 640 nmol of P4 resulted in only 19% (n = 85) and 2% (n = 176) of the cells showing a

change in calcium distribution and subsequent elevation in the apparent [Ca2 +]i, respectively (p < 0.05 compared to the appropriate P4 +RU 486 treatment group). Observation of P4 + RU 486-treated cells at 10 min intervals over a 2 h period (n = 86) showed the apparent increase in [Ca 2+], to be transient, lasting 20-30 min. The time during the 2-h culture period when these changes in the distribution and [Ca2 +]joccurred varied from cell to cell. DISCUSSION

Ovarian follicles are composed of two sizes of GCs: small and large. These two sizes of GCs have different functional capacities, with large GCs being steroidogenic [11-13]. As the follicle matures, the percentage of large steroidogenic GCs increases, accounting for the increased ability of the follicle to secrete both estrogen and P4 [12, 13]. If these mature follicles do not ovulate, they become atretic [1]. Biochemical characterization of atretic follicles reveals that the GC DNA is fragmented, demonstrating that the GCs undergo apoptosis [3]. The present study shows that apoptosis occurs mainly in the large steroidogenic GCs. After culture, a few cells within the small-GC fractions possess fragmented DNA with free 3'OH ends. This could indicate that small GCs undergo apoptosis at a slower rate than do large GCs. Alternatively, the presence of GCs with fragmented DNA

FIG. 6. Intracellular calcium distribution pattern of large GCs cultured for 2 h in the absence of P4 and then loaded with fluo-3. Three patterns of calcium distribution are observed. Calcium is localized to cytoplasmic foci (arrows) or evenly distributed throughout the cytoplasm with moderate fluo3 fluorescence (A). In several large GCs, fluo-3 fluorescence is very intense (A). At higher magnification, the calcium within the intensely fluorescing GCs was concentrated within the nucleus (B). When observed under phase optics, these cells appeared to be degenerating C) (A x400; B/C x600).

could be due to the presence of large GCs that contaminate the small-GC fractions. Since cell size decreases during apoptosis, time-lapse studies are needed to determine whether the small GCs with fragmented DNA are derived from small or large GCs. If large-GC contamination accounts for the presence of apoptotic GCs in the small-GC fractions, then small GCs may not die via an apoptotic mechanism. This latter possibility correlates well with the ultrastructural findings that demonstrate that GCs die by two distinct pathways [14]. Some GCs die by apoptosis, as indicated by nuclear condensation and fragmentation. In other GCs, death is initiated by the deterioration of the cytoplasmic organelles such as the mitochondria and endoplasmic reticulum. Both types of GC death are observed within the same atretic follicle [14]. This could indicate that as the follicles undergo atresia, separate mechanisms are activated that result in the degeneration of both large and small GCs. Although virtually nothing is known about the factors that regulate viability of small GCs, estrogen inhibits GC apoptosis in vivo [7]. It is possible that estrogen prevents apoptosis by binding to its receptor, which then acts as a nuclear transcription factor to regulate estrogen-dependent gene

652

LUCIANO ET AL.

FIG. 7. Sequential phase (A, E) and fluorescent photomicrographs of large GCs, loaded with fluo-3 and cultured in the presence of P4 and then exposed to RU 486. Photomicrographs of GCs, shown in A and B, were taken after 1 h of RU 486 exposure. After 1.5 (C) and 2.0 h (D, E)of RU 486 exposure, some GCs showed an increase in fluo-3 fluorescence (x400). Photomicrographs were taken at a higher ASA setting so that the pictures would be underexposed, allowing only GCs with intense fluo-3 fluorescence to be imaged.

PROGESTERONE AND GRANULOSA CELL APOPTOSIS transcription. As part of this proposed mechanism, estrogen could stimulate the synthesis of EGF and/or a related growth factor, transforming growth factor alpha (TGFt) [21], since both are produced in the ovary [22, 23]. Both EGF and TGFa stimulate P4 secretion within the first 24 h of culture (present study, [12]). It appears that enhanced P4 synthesis is an essential part of the mechanism through which EGF and presumably TGFax maintain large GC viability. This conclusion is based on the observations that the anti-apoptotic action of EGF is inhibited by both RU 486, a progesterone receptor antagonist, and AG, which reduces basal and EGFinduced P4 secretion. Further, P4 alone is sufficient to extend large GC survival in vitro, and its action is attenuated only by RU 486 and not by AG. P4 also prevents apoptosis in other systems. For example, administration of P4 prevents and RU 486 triggers apoptosis within the uterine epithelium [24]. In addition, a decline in P4 secretion precedes apoptotic changes within luteal cells [25]. Taken together, these studies indicate that P4 acts as an autocrine/ paracrine factor to maintain cell viability within the ovary and uterus and that factors inhibiting P4 secretion ultimately induce apoptotic changes in these cells. While previous work has shown that a decline in serum P4 precedes apoptosis within the corpus luteum and uterine epithelium, serum P4 levels increase as the GCs of eCGinduced preovulatory follicles undergo DNA fragmentation [2]. The reason for this contradiction is mostly likely that the enzyme responsible for the conversion of pregnenolone to P4 , 3-hydrosteroid dehydrogenase (3p-HSD), is expressed in the GCs of healthy preovulatory follicles as well as the associated thecal and interstitial cells [26]. As these follicles undergo atresia, GCs, but not the thecal and interstitial cells, lose their ability to express 3-HSD [26]. This selective loss of 313-HSD within the GCs of atretic follicles could result in a decrease in P4 in the GC component of the follicle. Since P4 maintains GC viability in vitro, the putative decline in P4 within the GC layers could result in a corresponding increase in GC DNA fragmentation, even in the presence of elevated serum P4 levels. These experiments clearly indicate that P4 plays a pivotal role in regulating large GC survival. Previous work with luteal cells has shown that prostaglandin F2, induces an increase in [Ca 2 +]i that lasts for 20-30 min prior to decreasing P4 synthesis [27]. A similar transient increase in [Ca2+]i is observed as the large GCs undergo apoptosis (present study). Not only does the apparent [Ca 2+]i increase, but the calcium distribution changes prior to the large GCs' becoming apoptotic. In the presence of P4, large GC viability is maintained, and calcium is localized to cytoplasmic foci. This pattern of calcium distribution is also observed in GCs that are isolated and immediately loaded with fluo-3 (unpublished observation). When apoptosis is initiated by inhibiting the action of P4 with RU 486, calcium becomes uniformly distributed throughout the cytoplasm. At this stage, the apparent [Ca2+ ]i is relatively low, as judged by intensity

653

of the fluo-3 fluorescence. As the apoptotic process continues, the apparent [Ca2+]i increases, and calcium becomes concentrated within the nucleus. The increase in the apparent [Ca2 +]i may eventually lead to an increase in calcium concentration within the nucleus that could activate nuclear endonucleases and thereby result in the cleavage of the GC DNA.

ACKNOWLEDGMENTS The authors thank Madame M. Garnier of Rousel-UCLAF for the gift of RU 486.

REFERENCES 1. Greenwald G, Terranova P. Follicular selection and its control. In: Knobil E, Ewing LL, Greenwald GS, Markert CL, Pfaff DW (eds.), The Physiology of Reproduction. New York: Raven Press; 1988: 387-446. 2. Hughes Fl, Gorospe WC. Biochemical identification of apoptosis (programmed cell death) in granulosa cells: evidence for a potential mechanism underlying follicular atresia. Endocrinology 1991; 129:2415-2422. 3. Tilly JL, Kowalski KI, Johnson AL, Hsueh A Involvement of apoptosis in ovarian follicular atresia and postovulatory regression. Endocrinology 1991; 129:27992801. 4. Schwartzman RA, Cidlowski JA. Apoptosis: the biochemistry and molecular biology of programmed cell death. Endocr Rev 1993; 14:133-151. 5. Hurwitz A, Adashi EY. Ovarian follicular atresia as an apoptotic process: a paradigm for programmed cell death in endocrine tissues. Mol Cell Endocrinol 1992; 84:C19-C23. 6. Zeleznik A, Ihrig L, Bassett S. Developmental expression of Ca+ +/Mg++-dependent endonuclease activity in rat granulosa and luteal cells. Endocrinology 1989; 125:2218-2220. 7. Billig H, Furuta I, Hsueh A Estrogens inhibit and androgens enhance ovarian granulosa cell apoptosis. Endocrinology 1993; 133:2204-2212. 8. Tilly JL, Kowalski KI, Schomberg DW, Hsueh A Apoptosis in atretic ovarian follicles is associated with selective decreases in messenger ribonucleic acid transcripts for gonadotropin receptors and cytochrome P450 aromatase. Endocrinology 1992; 131:1670-1676. 9. Tilly JL, Billig H, Kowalski KI, Hsueh A Epidermal growth factor and basic fibroblast growth factor suppress the spontaneous onset of apoptosis in cultured rat ovarian granulosa cells and follicles by a tyrosine kinase-dependent mechanism. Mol Endocrinol 1992; 6:1942-1950. 10. Morley P, Whitfield J, Vanderhyden B, Tsang B, Schwartz J. A new, nongenomic estrogen action: the rapid release of intracellular calcium. Endocrinology 1992; 131:1305-1312. 11. Rao IM, Mills TM, Anderson E, Mahesh VB. Heterogeneity in granulosa cells of developing rat follicles. Anat Rec 1991; 229:177-185. 12. Sanbuissho A, Lee GY, Anderson E. Functional and ultrastructural characteristics of two types of rat granulosa cell cultured in the presence of FSH or transforming growth factor alpha (TGF-alpha). J Reprod Fertil 1993; 98:367-376. 13. Lederer K, Luciano A, Pappalardo A, Peluso J. Proliferative and steroidogenic capabilities of different-sized rat granulosa cells. In: 44th SGI meeting; 1993; Toronto. 14. Peluso JJ, Charlesworth J, Egland-Charlesworth C. Role of estrogen and androgen in maintaining the preovulatory follicle. Cell Tissue Res 1981; 216:615-624. 15. Wyllie AH. Cell death. Int Rev Cytol 1987; 17:755-785. 16. Gavrieli Y, Sherman Y, Ben-Sasson S. Identification of programmed cell death in situ via specific labeling of nuclear DNA fragmentation. J Cell Biol 1992; 119:493501. 17. Bucana C, Saiki I, Nayar R. Uptake and accumulation of the vital dye hydroethidine in neoplastic cells. J Histochem Cytochem 1986; 34:1109-1115. 18. Dietch A,Law H, White R. A stable propidium iodide staining procedure for flow cytometry. J Histochem Cytochem 1982; 30:967-972. 19. Haugland R Handbook of Fluorescent Probes and Research Chemicals. 5th ed. Eugene, OR: Molecular Probes; 1992: 113-128.

654

LUCIANO ET AL.

20. Tsien R Fluorescence measurement and photochemical manipulation of cytosolic free calcium. Trends Neurochem Sci 1988; 11:419-432. 21. Lippman M, Dickson R. Mitogenic regulation of normal and malignant breast epithelium. Yale J Biol Med 1989; 62:459-480. 22. Maruo T, Ladines LC, Samoto T, Matsuo H, Manalo AS, Ito H, Mochizuki M. Expression of epidermal growth factor and its receptor in the human ovary during follicular growth and regression. Endocrinology 1993; 132:924-931. 23. Yeh J, Lee GY, Anderson E. Presence of transforming growth factor-alpha messenger ribonucleic acid (mRNA) and absence of epidermal growth factor mRNA in rat ovarian granulosa cells, and the effects of these factors on steroidogenesis in vitro. Biol Reprod 1993; 48:1071-1081.

24. Rotello R, Lieberman R, Lepoff R, Gerschenson L. Characterization of uterine epithelium apoptotic cell death kinetics and regulation by progesterone and RU 486. Am J Pathol 1992; 140:449-456. 25. JuengelJL, Garverick HA, Johnson AL, Youngquist RS, Smith MF. Apoptosis during luteal regression in cattle. Endocrinology 1993; 132:249-254. 26. Teerds KJ, Dorrington JH. Immunohistochemical localization of 33-hydroxysteroid dehydrogenase in the rat ovary during follicular development and atresia. Biol Reprod 1993; 49:989-996. 27. Wegner J, Martinez-Zaguilan R, Wise M, Gillies R, Hoyer P. Prostaglandin F2aphainduced calcium transient in ovine large luteal cells: I. alterations in cytosolicfree calcium levels and calcium flux. Endocrinology 1990; 127:3029-3037.