Regulation of apolipoprotein E synthesis in rat ovarian granulosa cells.

Apoprotein E (apo-E) is a surface component of several classes of plasma lipoproteins. It functions as a ligand for receptor-mediated uptake of lipoproteins. Granulosa cells from ovaries of diethylstilbestrol-stimulated hypophysectomized immature rats cultured in serum-free medium with [35S]methionine secretes a 34-kDa protein which reacts with a monospecific anti-rat apo-E antibody and represents 0.2% of total secreted protein. Protease mapping confirms that this protein is apoprotein E. The secreted apoprotein E may be complexed with lipid since it floats in the ultracentrifuge at density less than 1.21 micrograms/ml. Freshly isolated granulosa cells contain receptors for follicle stimulating hormone (FSH) but not for human chorionic gonadotropin (hCG) or prolactin. Apoprotein E secretion is stimulated 2-fold by FSH, but hCG and prolactin have no effect. When granulosa cells develop hCG and prolactin receptors after 48 h of culture with FSH, apoprotein E secretion is not stimulated by addition of FSH, hCG, or prolactin although steroidogenesis is induced. The addition of 10(-7) M androgen plus FSH stimulates a marked increase in progestin synthesis over FSH alone, but androgen has little added effect on apoprotein E secretion. Cholera toxin (1.25 micrograms/ml) and dibutyryl cAMP (5 mg/ml), both of which increase intracellular cAMP, stimulate apo-E secretion 9-fold and 12-fold, respectively. The dibutyryl cAMP effect is dependent on both dose (greater than or equal to 0.5 mg/ml required) and time (onset at 24 h, maximum at 48 h, and back to near baseline at 96 h). Isobutylmethylxanthine, a phosphodiesterase inhibitor, augments FSH-stimulated apoprotein E synthesis 2.5-fold, supporting a role for cAMP in mediating the FSH effect. This is the first demonstration of the hormonal regulation of apoprotein E synthesis in an extrahepatic tissue.


Regulation of Apolipoprotein E Synthesis in Rat Ovarian
Apoprotein E (apo-E) is a surface component of several classes of plasma lipoproteins. It functions as a ligand for receptor-mediated uptake of lipoproteins. Granulosa cells from ovaries of diethylstilbestrol-stimulated hypophysectomized immature rats cultured in serum-free medium with [S6S]methionine secretes a 34-kDa protein which reacts with a monospecific anti-rat apo-E antibody and represents 0.2% of total secreted protein. Protease mapping confirms that this protein is apoprotein E. The secreted apoprotein E may be complexed with lipid since it floats in the ultracentrifuge at density <1.21 pg/ml. Freshly isolated granulosa cells contain receptors for follicle stimulating hormone (FSH) but not for human chorionic gonadotropin (hCG) or prolactin. Apoprotein E secretion is stimulated 2fold by FSH, but hCG and prolactin have no effect. When granulosa cells develop hCG and prolactin receptors after 48 h of culture with FSH, apoprotein E secretion is not stimulated by addition of FSH, hCG, or prolactin although steroidogenesis is induced. The addition of 10" M androgen plus FSH stimulates a marked increase in progestin synthesis over FSH alone, but androgen has little added effect on apoprotein E secretion. Cholera toxin (1.25 pg/ml) and dibutyryl cAMP (5 mg/ml), both of which increase intracellular CAMP, stimulate apo-E secretion 9-fold and 12-fold, respectively. The dibutyryl cAMP effect is dependent on both dose (20.5 mg/ml required) and time (onset at 24 h, maximum at 48 h, and back to near baseline at 96 h). Isobutylmethylxanthine, a phosphodiesterase inhibitor, augments FSH-stimulated apoprotein E synthesis 2.5-fold, supporting a role for cAMP in mediating the FSH effect. This is the first demonstration of the hormonal regulation of apoprotein E synthesis in an extrahepatic tissue.
Apolipoprotein E is a 35,000-Da glycoprotein found in several classes of mammalian lipoproteins including VLDL,' * This work was supported by National Institutes of Health Grants HL 15062 and HD 17753. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "oduertisernent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 11 To whom correspondence and requests for reprints should be addressed.
The abbreviations used are: VLDL, very low density lipoprotein; apo, apolipoprotein; BtSAMP, dibutyryl 3',5'-cyclic adenosine monophosphate; HDL, high density lipoprotein; IBMX, 3'-isobutyl-lmethylxanthine; LDL, low density lipoprotein; SDS, sodium dodecyl sulfate; FSH, follicle stimulating hormone; LH, luteinizing hormone; hCG, human chorionic gonadotropin. LDL, chylomicron remnants, and certain subclasses of HDL (1). Apo-E plays a critical role in cholesterol metabolism since it functions as a ligand for the receptor-mediated uptake of plasma lipoproteins by the B,E receptor of peripheral cells and by the E receptor on hepatocytes (2-4). Binding of a lipoprotein to these receptors results in the internalization and degradation of the lipoprotein with subsequent delivery of cholesterol for membrane synthesis in all cells and for steroid synthesis in steroidogenic cells (4).
Although lipoproteins are synthesized predominantly in the liver and intestine, only the liver appeared to be responsible for the synthesis of apo-E (5-7). Recently, however, apo-E synthesis has been demonstrated in a wide spectrum of extrahepatic tissues including steroidogenic organs such as the adrenal, testis, and ovary (8-10). The only cells proven to synthesize apo-E in vitro are the hepatocyte, the macrophage, and smooth muscle cell (7,10,11). The fact that macrophages and smooth muscle cells can synthesize apo-E in vitro suggested that these cells may contribute to the synthesis of apo-E in extrahepatic tissues in vivo. However, the magnitude of apo-E synthesis in peripheral tissues argues that other cell types may be involved. Although the phenomenon of extrahepatic synthesis of apo-E is well documented, little is known about the function or regulation of apo-E synthesis in these peripheral tissues.
We have previously shown that apo-E represents approximately 0.15% of the total protein synthesized in the rat ovary (10). In this study, we have used the ovarian granulosa cell as a model to study apo-E synthesis in a steroidogenic cell. Primary cultures of rat ovarian granulosa cells have been extensively used to study the hormonal regulation of steroidogenesis and cell differentiation (12). Freshly isolated granulosa cells have cell-surface receptors for FSH and retain their hormonal responsiveness in vitro when cultured in serum-free medium (13, 14). The addition of FSH to these cells in vitro stimulates the production of estrogens and progestins (13,15,16). FSH also stimulates cell differentiation and the development of receptors for two other gonadotropins, LH or hCG, and prolactin (17, 18).
It is now well established that gonadotropins stimulate adenylate cyclase activity in hormonally responsive cells and that cAMP functions as the second messenger for gonadotropin action (19,20). The effects of FSH on steroidogenesis and cell differentiation in granulosa cells can be reproduced in vitro by the addition of exogenous cAMP (as the nonpolar analog Bt2cAMP) (21)(22)(23). Other agents which raise the intracellular levels of cAMP such as cholera toxin, phosphodiesterase inhibitors, and forskolin, also stimulate steroidogenesis in granulosa cells (22,24,25).
In this study we demonstrate that rat ovarian granulosa cells synthesize and secrete apo-E in vitro which is the first demonstration of apo-E synthesis by an isolated steroidogenic cell. W e also present data on the regulation of apo-E synthesis in granulosa cells by gonadotropins and by CAMP.
Antibody Preparation-Rat apo-E was purified from plasma VLDL by Sepharose 6B column chromatography and preparative gel electrophoresis (26) followed by heparin Sepharose 4B affinity chromatography (27). Antibody to rat apo-E was prepared in a female goat and affinity-enriched by absorption of the immune plasma to a rat HDL-Sepharose 4B column (26). This antibody has been shown to be monospecific for apo-E by immunodecoration of electrophoretically separated apolipoproteins transferred to nitrocellulose paper (10, 28). Preimmune serum globulin was purified by ammonium sulfate fractionation and chromatography on a tandem DEAE-cellulose/carboxymethylcellulose column (29).
Granulosa Cell Cultures-Rat ovarian granulosa cells were isolated and cultured as previously described (13). Briefly, immature (23-25day-old) female rats were hypophysectomized by Hormone Assay Laboratories (Chicago, IL) with the concomitant subcutaneous placement of Silastic capsules filled with diethylstilbestrol (30). After 4-5 days, granulosa cells were isolated from the ovaries of these animals by puncturing the preantral follicles with an iris knife. Cells were washed and plated in Falcon tissue culture dishes (35 X 10 mm) at 4 X lo6 cells/ml in serum-free McCoy's medium containing 100 units/ ml of penicillin, 100 pg/ml streptomycin, and 2 mM L-glutamine. Cells were cultured for various lengths of time (as indicated in the figure legends) in the above medium or in the above medium supplemented with one of the following: 50 ng/ml FSH; M androstenedione; lo-' M dihydrotestosterone; 1 IU/ml of hCG; 1 pg/ml prolactin; 5 mg/ml Bt,cAMP; 1.25 pg/ml cholera toxin; 0.2 mM IBMX. After the indicated length of time, medium was removed from each plate, frozen, and analyzed for steroids by radioimmunoassay. The cells were then labeled with 10-30 pCi/ml of [=S]methionine for 8 or 16 h as noted in the figure legends. Incorporation of isotopic precursor into protein was linear with time from 4-16 h. Immunoprecipitation-After 8 or 16 h, the medium was removed from the [35Sjmethionine-labeled granulosa cells and centrifuged at 1000 X g for 5 min. The cells were rinsed with phosphate-buffered saline and lysed with phosphate-buffered saline containing 10 mM methionine, 1% Triton X-100, 1% Na+ deoxycholate and homogenized in a ground glass Dounce homogenizer. After determining the amount of trichloroacetic acid-precipitable radioactivity in the medium and in the cell extract, the total radiolabeled proteins were analyzed by SDS-polyacrylamide gel electrophoresis. Because apo-E represented a small percentage of the total proteins secreted by granulosa cells, relative changes in apo-E secretion were measured by a quantitative immunoprecipitation assay (10). Aliquots of medium and cell extract containing equal amounts of trichloroacetic acidprecipitable radioactivity were incubated for 1 h at room temperature with 20 pg of either preimmune globulin or with the monospecific goat anti-rat apo-E antibody.
Preliminary experiments established that the amount of antibody added to the medium was sufficient to effectively immunoprecipitate all the apo-E in the medium. When the proteins in the granulosa cell medium which were not immunoprecipitated by the anti-apo-E antibody were incubated with a second aliquot of the antibody, no radiolabeled apo-E was recovered, indicating that the amount of antiapo-E antibody added to the medium was not limiting. When tracer lZ5I-labeled plasma apo-E was added to granulosa cell medium and incubated with the anti-apo-E antibody, over 95% of the radioactivity was recovered in the immunoprecipitate. In addition, when samples of granulosa cell medium containing different amounts of apo-E were mixed together and then immunoprecipitated, the amount of apo-E recovered was additive.
After the granulosa cell medium was incubated with the anti-apo-E antibody, SDS was added to achieve a final concentration of 0.2% and the immunoprecipitated proteins were isolated by binding to Protein A-Sepharose as previously described (10). The Sepharose beads were washed and the bound proteins were eluted by boiling for 2 min in 62.5 mM Tris, pH 6.8,2% SDS, 5% B-mercaptoethanol. The samples were analyzed by electrophoresis on a SDS, 5-22% polyacrylamide gel (31). Rat plasma apo-E was iodinated using iodine monochloride (32), immunoprecipitated, and electrophoresed to identify plasma apo-E. The relative amount of radioactivity incorporated into apo-E in each lane was determined by quantitating radioactivity in gel slices by scintillation counting or by determining the absorbance of silver grains eluted from the autoradiogram as described below.
Fluorography-Gels were impregnated with 1 M Na+ salicylate (33) and exposed to Kodak X-Omat AR-5 film with a DuPont Lightening Plus Intensifying Screen. The relative amount of radiolabeled apo-E in each sample was determined by two methods. In the first method, bands were eluted from the dried gel in 30% H202 and the amount of radioactivity was determined by scintillation counting (34). In the second method, the bands on the autoradiogram itself were cut out of the film and incubated in 1 M NaOH for 2 h to elute the silver grains from the film. The absorbance of the eluted grains was determined spectrophotometrically at 500 nm (35). A standard curve was constructed for each technique by immunoprecipitating increasing amounts of 36S-labeled medium from unstimulated granulosa cells. Both techniques gave similar results.
Peptide Mapping-Iodinated rat plasma VLDL and medium from [?'3]methionine-labeled granulosa cells were immunoprecipitated with the monospecific goat anti-rat apo-E antibody and subject to limited proteolysis mapping using Staphylococcus aureus V8 protease as described by Cleveland et al. (36).
Rudioimmumussay-The amount of estrogen, 20a-dihydroprogeskrone, and progesterone in the unextracted tissue culture medium was determined by radioimmunoassay (13, 37). Because similar results were obtained for 20a-dihydroprogesterone and progesterone, only the results for estrogen and 20a-dihydroprogesterone are presented.

RESULTS
Apo-E Secretion by Granulosa Cells-To obtain primary cultures of rat ovarian granulosa cells, immature female rats were hypophysectomized to eliminate exposure to endogenous gonadotropins and treated with diethylstilbestrol to stimulate granulosa cell proliferation. Granulosa cells were isolated from the preantral follicles and incubated for 16 h in serumfree medium containing [35S]methionine to determine whether these cells could synthesize and secrete apo-E. T h e medium was isolated and aliquots of medium were incubated with either nonimmune globulin or a monospecific goat antirat apo-E antibody. The fluorogram in Fig. 1 displays the electrophoretic profile of the total proteins secreted into the medium by granulosa cells ( l a n e I). The anti-rat apo-E antibody ( l a n e 3 ) precipitated a triplet of proteins with a molecular weight similar t o rat plasma apo-E ( l a n e 4 ) . Rat plasma apo-E often appears as a broad band or even as a doublet or triplet on SDS-polyacrylamide gels but the basis for this heterogeneity is not known (28, 38). The proteins secreted from the granulosa cells were not precipitated by nonimmune globulin ( l a n e 2). The relative amount of apo-E secreted into the medium was determined by measuring the radioactivity in gel slices or by measuring the absorbance of silver grains eluted from the autoradiogram as described under "Experimental Procedures." Apo-E represented approximately 0.12% of the total proteins secreted by the granulosa cells. The medium containing the 35S-labeled proteins was floated in the ultracentrifuge and the floated proteins were analyzed by electrophoresis ( l a n e 5 ) . The apo-E in the medium floated at a density less than 1.21 g/ml which suggests that apo-E is secreted from the granulosa cell as part of a lipid-protein complex. Granulosa cell extracts were also analyzed for the presence of radiolabeled apo-E. Apo-E represented less than 0.02% of the newly synthesized intracellular protein in the unstimulated granulosa cell (data not shown).
Proteolysis Mapping of Apo-E-The apo-E-like proteins secreted from the granulosa cells appeared to be apo-E by Ovarian granulosa cells were isolated from hypophysectomized, diethylstilbestrol-treated immature female rats and cultured for 16 h in serum-free medium containing 10 pCi/ml of [35S]methionine as described under "Experimental Procedures." The medium was isolated and aliquots were precipitated with either nonimmune globulin or the monospecific goat anti-rat apo-E globulin as described under "Experimental Procedures." An aliquot of medium was also floated in the ultracentrifuge at density less than 1.21 g/ml. their electrophoretic mobility and their immunologic reactivity with a monospecific antibody. To confirm these observations, Staphylococcus aureus V8 protease was used for limited proteolysis mapping of lZ5I-labeled rat plasma apo-E and [35S] methionine-labeled apo-E from granulosa cell medium. The fluorogram in Fig. 2 demonstrates that identical cleavage patterns were generated from rat plasma apo-E and the apo-E secreted by the granulosa cells.
Hormonal Regulation of Apo-E Synthesis by Gonudotropins-Freshly isolated granulosa cells have cell-surface receptors for FSH and retain their hormonal responsiveness in vitro when cultured in serum-free medium (13, 14). To study the hormonal regulation of apo-E synthesis in granulosa cells, freshly isolated granulosa cells were cultured for 48 h in serum-free medium or serum-free medium supplemented with additives such as 50 ng/ml FSH, M androstenedione, or M dihydrotestosterone. The medium was removed from each plate and analyzed by radioimmunoassay for steroid production. The cells were then labeled for 16 h with [35S] methionine in fresh McCoy's medium. Because apo-E represented a small percentage of the total secreted proteins, relative changes in apo-E secretion were measured by a quan- titative immunoprecipitation assay, Aliquots of medium containing equal amounts of trichloroacetic acid-precipitable radioactivity were immunoprecipitated with the anti-rat apo-E antibody and analyzed by electrophoresis. The relative amount of radiolabeled apo-E in each lane was determined by measuring radioactivity in gel slices by scintillation counting or by measuring the absorbance of silver grains eluted from the film as described under "Experimental Procedures." The results from these experiments are shown in Fig. 3. There was a 2-fold increase in the amount of radiolabeled apo-E secreted into the medium of cells stimulated with FSH compared to unstimulated cells. This increase in apo-E was not due to a nonspecific stimulation by FSH of protein synthesis or secretion since in FSH-stimulated cells, the total amount of radioactivity incorporated into newly synthesized proteins or into secreted proteins was only 15% higher than in unstimulated granulosa cells. In addition, the immunoprecipitations were done using aliquots of medium containing equal amounts of trichloroacetic acid-precipitable radioactivity which adjusted for this slight stimulation of general protein secretion by FSH.
The FSH-stimulated cells also synthesized a small amount of 20a"ihydroprogesterone but no detectable estrogens. In vivo, granulosa cells produce estrogens by the aromatization of androstenedione which is provided by neighboring thecal cells (39). When androstenedione and FSH were both added to granulosa cells in vitro, the production of estrogens and progestins was stimulated 15-fold. Although FSH required the addition of androstenedione to fully stimulate steroid production, FSH stimulated apo-E secretion equally well in the absence or presence of androstenedione or dihydrotestosterone. When dihydrotestosterone was substituted for androstenedione, the cells still synthesized progestins but did not produce estrogens since dihydrotestosterone is a nonaromatizable substrate. Neither androstenedione nor dihydrotestosterone had an effect on apo-E secretion or steroid production in the absence of FSH. Immunoprecipitation of granulosa cell extracts with the anti-apo-E antibody did not detect any changes in the intracellular levels of apo-E under any of these conditions (data not shown).
The experiments illustrated in Fig. 3 were done at an FSH concentration of 50 ng/ml. To determine the optimal FSH concentration for the stimulation of apo-E secretion, a doseresponse curve was generated by stimulating granulosa cells for 48 h with increasing concentrations of FSH. Although the production of 20a-dihydroprogesterone could be further stimulated by increasing the concentration of FSH up to 300 ng/ ml, both estrogen production and apo-E secretion were maximally stimulated at a FSH concentration of 50 ng/ml (data not shown).

Regulation of Apo-E Synthesis by Cyclic AMP-Studies in
other laboratories have established that many of the effects of FSH on granulosa cells, including the stimulation of steroidogenesis and the induction of LH and prolactin receptors, are mediated by CAMP (12). To determine whether the secretion of apo-E is responsive to CAMP, freshly isolated granulosa cells were cultured for 48 h in medium alone or in medium containing either 1.25 pg/ml cholera toxin or 5 mg/ml BtzcAMP. Androstenedione was added to all cells in this experiment to assess the effects of cholera toxin and BtzcAMP on estrogen production. After 48 h, the cells were labeled for 16 h with [35S]methionine and the medium was analyzed as described above. The fluorogram from this experiment is shown in Fig. 4, left panel, and the quantititative data on apo-E secretion and steroid production are shown in Fig. 4, right panel. As has been previously reported, both cholera toxin and BtZcAMP stimulated the production of estrogens and progestins by granulosa cells (22). Neither cholera toxin nor BtzcAMP stimulated general protein secretion in granulosa cells as determined by incorporation of radioactivity into trichloroacetic acid-precipitable proteins in the medium. However, the secretion of apo-E was stimulated 9-fold by cholera toxin and 12-fold by BtZcAMP (Fig. 423). Intracellular levels of apo-E were not altered by either compound (data not shown).
The effects of BtzcAMP on steroidogenesis and apo-E secretion in granulosa cells were both dose-and time-dependent. A dose-response curve for BtzcAMP is shown in Fig. 5. Freshly isolated granulosa cells were cultured for 48 h in medium containing M androstenedione and increasing concentrations of Bt2cAMP (ranging from 0.05 to 5 mg/ml). Concentrations greater than or equal to 0.5 mg/ml BtzcAMP were required to stimulate apo-E secretion or steroidogenesis in granulosa cells. Increasing concentrations of Bt2cAMP up to 5 mg/ml stimulated an increased production of apo-E and steroids.
A time course of BtzcAMP stimulation is shown in Fig. 6. Freshly isolated granulosa cells were incubated with or without 5 mg/ml BtzcAMP for various lengths of time and aliquots of medium were removed and analyzed for 20a-dihydroprogesterone. The cells were then labeled for 8 h with [35S] methionine and the amount of apo-E secreted into the medium during the 8-h incubation was determined as described above. Results are presented as the ratio of apo-E secreted by granulosa cells cultured in the presence of BtzcAMP to that secreted by cells cultured for the same length of time without BtzcAMP. As shown in Fig. 6A, Bt,cAMP had no effect on apo-E secretion during the first 8 h of culture. There was a 4fold increase in apo-E secretion at 24 h and a 14-fold increase at 48 and 72 h due to Bt,cAMP. After the first 16 h of culture, there was a linear increase in 20a-dihydroprogesterone production with time (data not shown). The effect of BtzcAMP on apo-E secretion at 48 and 96 h is shown in Fig. 6B. Following the media change at 48 h, Bt2cAMP was only 17% as effective in stimulating apo-E secretion at 96 h as it was at 48 h. Interestingly, 20a-dihydroprogesterone synthesis was 50% greater during the second 2 days of culture than it was in the first 2 days (data not shown).
Effect of IBMX on Apo-E Secretion-In addition to stimulating the adenylate cyclase, FSH also activates a phosphodiesterase activity in granulosa cells to prevent a prolonged response to gonadotropins (40). The effect of the phosphodiesterase inhibitor, IBMX, on apo-E secretion and steroidogenesis, was assessed as illustrated in Fig. 7. The culture of freshly isolated granulosa cells for 48 h in the presence of IBMX alone had no effect on either process, since these cells have very little endogenous cyclase activity. However, the addition of IBMX plus FSH to granulosa cells potentiated the effects of FSH on apo-E secretion and steroidogenesis. The secretion of apo-E was stimulated &fold in cells cultured with FSH plus IBMX compared to the 2-fold stimulation in cells incubated with FSH alone. The addition of IBMX plus FSH was still not as effective as 5 mg/ml Bt2cAMP in stimulating apo-E secretion. In contrast, the addition of IBMX to BtcAMP-stimulated cells did not further enhance the effects of Bt2cAMP on apo-E secretion or steroid production (data not shown).
Developmental Regulation of Apo-E Synthesis in Granulosa Cells-Granulosa cells isolated from the ovaries of hypophysectomized, diethyl stilbestrol-treated immature female rats contain FSH receptors but not receptors for LH (or hCG) or prolactin. The culture of these cells for 48 h in the presence of FSH or Bt2cAMP will induce cell differentiation, including the induction of receptors for LH and prolactin (12). The results presented in Fig. 6B suggested that the secretion of apo-E declined as Bt2cAMP-stimulated granulosa cells differentiate in culture. To examine this phenomenon in more detail, the regulation of apo-E synthesis was compared in granulosa cells cultured for various lengths of time as shown measured by radioimmunoassay. The apo-E results represent the mean of 4 samples (+/-S.D.). The steroid results are the mean of 2 representative samples.

Regulation of Apo-E Synthesis
FIG. 6. Time course of BtacAMP stimulation of apo-E secretion by granulosa cells. Freshly isolated granulosa cells were cultured in serum-free medium or in serum-free medium containing 5 mg/ml Bt,cAMP. A, cells were cultured for 8, 24, 48, or 72 h. Eight h prior to the indicated times, an aliquot of medium was removed for steroid analysis and the cells were then labeled during the last 8 h with 20 kCi/ml of ["Slmethionine and the amount of radiolabeled apo-E secreted into the medium was determined as described under "Experimental Procedures." B, cells were cultured for 48 or 96 h, with a medium change at 48 h in the latter group. Eight h prior to the indicated time, an aliquot of medium was removed for steroid analysis and the cells were then labeled with 20 pCi/ml of [%]methionine. One apo-E secretion unit = the amount of apo-E secreted by unstimulated granulosa cells in the 8 h of exposure to [35S]methionine. The results represent the mean of 3 samples (+/-S.D.).
in Fig. 8. In these particular experiments, hCG was used rather than LH. Androstenedione was included in all cultures in order to assess the effects of gonadotropins on estrogen production.
In the experiments depicted in Fig. 8, left panel, freshly isolated granulosa cells were cultured for 48 h in serum-free medium or serum-free medium containing FSH, hCG, or prolactin. After 48 h, an aliquot of medium was analyzed for steroids and the cells were then labeled with [35S]methionine for 16 h. As in the previous experiment, the addition of FSH stimulated apo-E secretion 2-fold and also induced steroidogenesis. As expected, the addition of hCG or prolactin to these cells had no stimulatory effect on either the secretion of apo-E or steroidogenesis since these cells cultured for only 2 days lack the appropriate receptors. In the experiment illustrated in Fig. 8, right panel  ferentiated cells, neither gonadotropin stimulated the secretion of apo-E above baseline. Interestingly, a second exposure to FSH also markedly stimulated steroid production but FSH no longer stimulated the secretion of apo-E.

DISCUSSION
Although apo-E synthesis has been detected in a number of steroidogenic organs, the finding that the rat ovarian granulosa cell synthesizes apo-E is the first demonstration of apo-E synthesis by an isolated steroidogenic cell. Based on its electrophoretic mobility, immunologic reactivity, and limited proteolysis mapping, the protein synthesized by the granulosa cells was identified as apo-E. The apo-E which was secreted from the granulosa cell floated in the ultracentrifuge at a density less than 1.21 g/ml which suggests that the apo-E is part of a lipid-protein complex. Studies in mouse peritoneal macrophages have shown that apo-E is secreted from these cells as part of a bilayer disc of phospholipid and protein which has a density of 1.075 g/ml (41). The nature of the product secreted from the granulosa cells is currently being investigated.
Freshly isolated granulosa cells have cell-surface receptors for FSH and the addition of FSH to these cells in culture stimulated the production of estrogen and progestins. Although FSH did not notably stimulate total protein synthesis or total protein secretion in the granulosa cells, the secretion of apo-E by these cells was specifically stimulated 2-fold by FSH. FSH also stimulated steroidogenesis but apo-E synthesis and steroid hormone production by granulosa cells were not always closely parallel. As shown in Fig. 3, the addition of M androgen plus FSH stimulated a marked increase in progestin synthesis over FSH alone, but added androgen had little effect on apo-E secretion.
Many of the effects of FSH on steroidogenesis and cell differentiation in granulosa cells are mediated by cAMP (21-23). The results presented here demonstrate that cAMP can also modulate apo-E secretion by granulosa cells. In addition to stimulating adenylate cyclase activity, FSH also induces a phosphodiesterase, increasing the cell's capacity to degrade cAMP which may prevent a prolonged response to gonadotropin stimulation (40). The addition of IBMX, a phosphodiesterase inhibitor, potentiated the effects of FSH on both ste-roidogenesis and apo-E secretion, suggesting that cAMP mediated the effects of FSH on apo-E synthesis and secretion. In support of this, two agents which elevate intracellular cAMP levels (BtKAMP and cholera toxin) markedly stimulated apo-E secretion. The effect of BbcAMP on apo-E secretion was both dose-and time-dependent.
It is now well established that in eukaryotes, many of the effects of cAMP are mediated by CAMP-dependent protein kinases. The binding of cAMP to the regulatory subunit of the protein kinase activates the catalytic subunit of the kinase, which in turn phosphorylates a substrate protein (42). The mechanism by which cAMP stimulates apo-E secretion was not investigated in this study. This regulation could be a transcriptional, post-transcriptional, translational, or even post-translational event. Recently, it has become appreciated that cAMP can stimulate the synthesis of specific proteins by elevating the levels of the mRNAs which encode them. cAMP stimulates the transcription of the genes which code for discoidin I (43), prolactin (44), phosphoenolpyruvate carboxykinase (45), and lactate dehydrogenase subunit A (46). cAMP also stimulates the synthesis of a number of proteins by stabilizing their mRNAs (46, 47). The synthesis of tyrosine aminotransferase is stimulated by CAMP, which increases the rate of nascent chain elongation. This effect is mediated by the catalytic subunit of the CAMP-dependent protein kinase (48, 49). cAMP may also act at the post-translational level. The secretion of insulin from pancreatic acinar cells is stimulated by cAMP which promotes the lysis of secretory granules (50).
The results in Fig. 6 and 8 suggest that the synthesis of apo-E may be developmentally regulated in the rat granulosa cell. These findings are interesting in light of the observation that apo-E synthesis also appears to be developmentally regulated in bone marrow-derived macrophages (51). FSH and BbcAMP both stimulated steroidogenesis and apo-E secretion in freshly isolated granulosa cells. As the granulosa cells differentiated in uitro, they were still able to respond to gonadotropins and BtzcAMP as evidenced by the increased steroidogenesis induced by these agents. However, in these FSH-differentiated cells, the secretion of apo-E was no longer responsive to FSH or BtzcAMP (Figs. 6B and 8). In both these experiments, the media were changed at 48-h intervals in order to assure adequate cell nutrition (12). Viability of the cells at 96 h is not an issue since we have shown previously that these granulosa cells produce progestins in response to lipoproteins even after 144 h of culture, when the media are changed every 48 h (37). One possible explanation for the decline in apo-E secretion at 96 h is that the mediator of cAMP action on apo-E secretion is present in the freshly cultured cells but not in the granulosa cells that have been held in culture for 3-4 days. In this case, different mediators of cAMP action on apo-E secretion and steroidogenesis would be implied. There is evidence in Dictyostelium for the existence of CAMP-dependent protein kinases which are developmentally regulated (52). There could also be structural changes in the apo-E gene as granulosa cells differentiate in uitro. If cAMP acts at the transcriptional level, changes in the nucleosome structure or the methylation pattern of the apo-E gene during differentiation could render the gene unresponsive to CAMP. Demethylation of certain cytosine residues has been correlated with the expression of developmentally regulated genes (53).
The function of apo-E synthesized in extrahepatic tissues is not understood. Because lipoproteins containing apo-E are recognized and removed from the plasma by the liver, it has been suggested that apo-E may function to transport excess cholesterol from cholesterol-loaded macrophages to the liver for excretion (11). Blue et al. (8) have suggested that apo-E functions as a shuttle protein to transport cholesterol between tissues or cells. Because of its high affinity for phospholipid (54), apo-E could have a role in phospholipid synthesis, transport, or metabolism.
The apo-E secreted by granulosa cells may function in the uptake of lipoproteins. Although the rat ovarian granulosa cell has both LDL and HDL receptors, the granulosa cells of most species only have LDL receptors (55). However, LDL cannot penetrate the basal lamina which surrounds the avascular ovarian follicle prior to ovulation. Thus, the only lipoprotein that is available to the granulosa cell prior to ovulation is HDL, which is present in follicular fluid at near plasma concentrations (56,57). A granulosa cell with a need for cholesterol may secrete apo-E essentially generating its own receptor ligand. The secreted apo-E could bind to an apo-Epoor lipoprotein such as HDL, thus converting it to a lipoprotein which could be recognized by the cell's LDL receptor. Apo-E may even play a role in areas other than cholesterol transport. Because apo-E has been shown to inhibit lymphocyte proliferation in vitro (58, 59), apo-E may function as a more general messenger or chalone between cells. Apo-E secreted by granulosa cells in vivo may regulate the proliferation of neighboring thecal cells during the estrous cycle. Clearly, the function of apo-E synthesized by peripheral tissues needs to be investigated in the future.