A Comparison of the Binding of Epidermal Growth Factor to Cultured Granulosa and Luteal Cells*

Epidermal growth factor (EGF) binds in a specific and saturable manner to both bovine granulosa (K,, = 2.4 x 10-l” M, n = 2.3 x lo4 EGF molecules/cell) and luteal cells (K,> = 7.0 x 10-l” M, n = 1.05 X lo” EGF molecules/cell). In luteal cells, however, as opposed to granulosa cells, the binding of EGF does not promote a mitogenic response. The spontaneous luteinization of cultured granulosa cells derived from large, preovulatory follicles is associated with an increased number of EGF receptor sites per cell and with a loss of responsiveness to EGF. Both EGF-responsive granulosa cells and unresponsive luteinizing cells showed a similar degree of internalization and lysosomal degradation of the bound EGF molecules as well as a specific loss of EGF receptor sites after exposure to EGF (“down regulation”). Preincubation with nonsaturating concentrations of EGF produced a degree of receptor loss which exceeded the percentage of receptors occupied at these concentrations. Nonetheless, even after an extended exposure to oversaturating concentrations of EGF (100 nglml), 15 to 25% of the receptor sites still remained available for EGF binding. This degree of receptor occupancy is sufficient to induce a maxima1 mitogenic response in granulosa cells. Exposure to either EGF or low density lipoprotein specifically affected the appropriate surface receptor sites, but neither molecule had any effect on the heterologous receptor sites. Our results on the binding of EGF to granulosa and luteal cells, as well as to early and late passages of granulosa cells, indicate a lack of correlation between EGF-binding capacity and mitogenic activity. The loss of mitogenic response in luteinizing cells is not due to a defect in the internalization and degradation of the cell-bound EGF, or to a defect in the EGF-induced receptor modulation and recovery, or both.

In luteal cells, however, as opposed to granulosa cells, the binding of EGF does not promote a mitogenic response.
The spontaneous luteinization of cultured granulosa cells derived from large, preovulatory follicles is associated with an increased number of EGF receptor sites per cell and with a loss of responsiveness to EGF. Both EGF-responsive granulosa cells and unresponsive luteinizing cells showed a similar degree of internalization and lysosomal degradation of the bound EGF molecules as well as a specific loss of EGF receptor sites after exposure to EGF ("down regulation").
Preincubation with nonsaturating concentrations of EGF produced a degree of receptor loss which exceeded the percentage of receptors occupied at these concentrations. Nonetheless, even after an extended exposure to oversaturating concentrations of EGF (100 nglml), 15 to 25% of the receptor sites still remained available for EGF binding. This degree of receptor occupancy is sufficient to induce a maxima1 mitogenic response in granulosa cells. Exposure to either EGF or low density lipoprotein specifically affected the appropriate surface receptor sites, but neither molecule had any effect on the heterologous receptor sites. Our results on the binding of EGF to granulosa and luteal cells, as well as to early and late passages of granulosa cells, indicate a lack of correlation between EGF-binding capacity and mitogenic activity. The loss of mitogenic response in luteinizing cells is not due to a defect in the internalization and degradation of the cell-bound EGF, or to a defect in the EGF-induced receptor modulation and recovery, or both.
In previous studies (1) it has been shown that cultured bovine granulosa cells, derived from medium-sized ovarian follicles, are highly responsive to the mitogenic stimulus * This work was supported by Grants HD 11082 from the National Institutes of Health and VC-194 from the American Cancer Society. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ Recipient of the Chaim Weizmann Research Training Fellowship. provided by both the epidermal and fibroblast growth factors. In contrast, bovine luteal cells, which are derived from granulosa cells through the process of cytomorphosis, are no longer sensitive to EGF,' although they do respond quite well to FGF (2).
In the present study we have looked at the quantitative and kinetic aspects of the binding of EGF to granulosa and luteal cells. Our results demonstrate that granulosa cells, as well as luteal cells, are capable of binding EGF in a highly specific manner. Thus, the lack of response of luteal cells to EGF cannot be explained on the basis of a loss of EGF receptor sites. Since the cytomorphosis of granulosa cells to luteal cells is associated with a loss of sensitivity to EGF but not of EGF receptor sites, the in vitro proliferative response of these cell types to EGF and FGF provides a direct way to study the relationship between mitogenicity, binding capacity, and the associated decreased concentration ("down regulation") (3-5) of surface receptor sites for EGF.

Materials
-Fibroblast growth factor (FGF) was purified from bovine pituitary glands (6) and brains (7) as previously described. Both pituitary and brain FGF yielded single bands on polyacrylamide gel electrophoresis at pH 4.5. No bands were observed at pH 8.5. EGF was purified as described by Savage and Cohen (8,9)  Similar results were obtained with the two kinds of preparations.
ConA was labeled with [3H]acetic anhydride, and the specific binding of 13HlconA to cells in monolayers was measured as described (11). Low density lipoprotein (LDL) was a generous gift from Drs. P. and C. Fielding (University of California, San Francisco). It was obtained from human plasma by differential ultracentrifugal flotation (12, 13), radioiodinated, and tested for specific binding to the cell surface as described (13,14). Establishment of Primary

Granulosa
Cell Cultures-Ovaries obtained from l-to 2-year-old nonpregnant cows were collected at a local slaughterhouse and transported on ice to the laboratory, where they were immediately dissected free from adventitial tissues. Small-sized follicles (4 to 6 mm diameter), medium-sized follicles (10 mm diameter), and large-sized follicles (15 mm diameter1 were slit with a scalpel as described by Channing et al. (15). The follicular fluid was removed with a Pasteur pipette, and the granulosa cells were removed by gently rinsing the inner wall of the follicle with 0.1 ml of medium.
The rinsings from one follicle were pooled in a sterile centrifuge tube containing 5 ml of F12 medium supplemented with 10% calf serum and 50 pg/ml of gentamicin.
The cell suspension medium was then transferred to a 6-cm plastic tissue culture dish and incubated in a humidified CO, incubator at 37°C. After 18 h the medium was renewed to remove unattached cells, and the dish was left at 37°C for 3 more days before being subcultured.
Establishment of Primary Luteal Cell Cultures -Corpora lutea from dairy cows of mixed breeds in the first half of pregnancy were used. Corpora lutea from pregnancy were 2 to 3 months old as determined from the crown-rump length of the fetuses. In some instances corpora lutea from midluteal phase ovaries of nonpregnant cows were used. Ovaries were collected as soon as possible following slaughter and transported in ice-cold phosphate-buffered saline to the laboratory, where they were immediately dissected free from adventitial tissues. The dissociation of the tissue was carried out as already described (2,(15)(16)(17). The dissociated luteal cells were suspended in Ham's F-12 medium containing 10% calf serum and 50 PgI ml of gentamicin.
The cell suspension was collected in a polyethylene tube and spun in a clinical centrifuge at 150 x g for 4 min at 4°C. The supernatant was removed, and the pellet was resuspended in F-12 medium with a plastic pipette. This procedure was repeated three times, and the final pellet was resuspended in the incubation medium and filtered through a Japanese silk screen (12 mesh) to remove cell aggregates which formed during the centrifugation step. Finally, the concentration of luteal cells was adjusted to 1 x 10" cells/ml.
One milliliter of the cell suspension was then transferred to a lo-cm plastic tissue culture dish, and 15 ml of F-12 with 10% calf serum were added. When the binding of EGF to granulosa and luteal cell cultures was compared to its mitogenic effect on these cultures, it was observed, as previously reported (1,21, that although EGF did bind to luteal cells, it had no mitogenic eRect upon them, as indicated by the absence of a significant increase in cell number (2). By contrast, EGF was a strong mitogen for granulosa cell cultures and gave a maximal increase (lo-to 20-fold) in cell number and thymidine incorporation at concentrations as low as 0.2 rig/ml (3.3 X lo-" M) and a half-maximal effect at 15 pglml (2.5 x lo-'* M) (1). These results indicate that the presence of EGF receptor sites is, by itself, not sufficient for the induction of cell proliferation by EGF. cells obtained from small follicles fail to luteinize spontaneously in culture (18).
Using granulosa cells maintained in tissue culture, one can study the dependency or loss of dependency on growth factors which could be developed during the process of cytomorphosis, i.e. when granulosa cells luteinize to give a population of luteal cells.
The following experiments were undertaken to determine whether a longterm, luteinizing, culture of granulosa cells becomes unresponsive to EGF and whether changes in responsiveness are associated with changes in the EGF-binding capacity, internalization, and release and/or in the EGFinduced loss and recovery of surface receptors for EGF.

Mitogenic
Activity and Binding Capacity -Freshly isolated granulosa cells and late passages of cultured granulosa cells derived from lo-mm follicles were tested for sensitivity to EGF and for EGF-binding capacity. The results indicate that cells from a late passage undergo luteinization in vitro and, as with luteal cells, EGF no longer had a significant effect on either cell proliferation (Fig. 4) or thymidine incorporation.' This change in sensitivity was associated with an EGF-binding capacity 6 times higher than the binding to cells freshly isolated (Fig. 5). Freshly isolated granulosa cells which are EGF-responsive show a saturation of EGF binding at 4 rig/ml, whereas cultures from late passages which no longer respond to EGF and show signs of terminal differentiation saturate at higher concentrations (10 rig/ml) (Fig. 5). In the following experiments EGF-responsive cells were obtained from small follicles (3 to 5 mm) and cultured in the presence of FGF for 1 to 3 weeks (two to four cell passages). The nonresponsive cells were derived from big follicles (10 to 15 mm) and cultured (2 to 3 weeks) in the presence of FGF and then in its absence for at least 4 weeks before use in the experiments. These cells stopped proliferating, became considerably enlarged, and showed the morphology of luteal cells.   (up to 24 h) to excess quantities of EGF (up to 100 rig/ml). It is of interest that with granulosa cells which do not luteinize in culture even less than 10% of the receptor sites for EGF have to be occupied in order to obtain a maximum mitogenic effect (1).
Recovery of EGF-binding Capacity-Cells that were first incubated with EGF to induce a 70 to 80% decrease in EGFbinding capacity were incubated for different time periods in the absence of EGF, and their ability to rebind fresh EGF was tested. The results (Fig. 8) indicate that both luteinizing and nonluteinizing granulosa cells are capable of gaining the initial EGF-binding capacity within 12 to 24 h of incubation in the presence, but not in the absence, of serum.
Effect of LDL and conA on EGF Receptors Granulosa cells show a specific binding capacity for concanavalin A (conA) and for low density lipoprotein (LDL) (at saturation, 2.8 x lo7 conA and 6 x lo4 LDL molecules are bound per ce11). 2 We have therefore studied whether the modulation of EGF surface receptors can affect the binding of conA or LDL molecules to their appropriate surface receptors and vice versa. The results (Table II) indicate that in cells that were preincubated (37°C 12 h) with EGF (10 rig/ml), a 70 to 80% decrease in the EGF-binding capacity was not associated with a detectable change in the binding of either conA or LDL molecules to the cell surface. Similarly, a 70% decrease in the binding of LDL was induced by preincubation (37"C, 24 h) with LDL (50 pg/ml), but this did not produce a change in the binding of EGF to these cells. The same experiment was not possible with conA, since it was found that this lectin blocks the binding of EGF. In cells that were preincubated (37°C 1 h) with 2.5 to 10 pg of conA/ml and washed free of unbound conA there was a 50 to 90% decrease in EGF-binding capacity. Since various conA-induced changes (cell agglutination, receptor redistribution, mitogenic stimulation) are temperature-dependent (11,24), cells were preincubated with con4 either at 4°C or 37°C and their capacities to bind EGF were compared. The results (Fig. 9) indicate that at 4"C, in contrast to 37"C, there was no effect of preincubation with 2.5 pg of conA/ml, and even at saturating concentrations of conA (10 to 100 pg/ml) there was only 20 to 50% inhibition of EGF binding,

DISCUSSION
We have shown in a previous study that the cytomorphosis of granulosa cells to luteal cells is associated with a loss of sensitivity to EGF but not to FGF (1, 2, 18). Our present results on the binding of EGF to granulosa and luteal cells as well as to early and late, luteinizing, passages of granulosa cells indicate a lack of correlation between EGFbinding capacity and mitogenic activity. Granulosa cells which proliferate in response to EGF bind 10 times more EGF than is needed to induce a maximal proliferative response, whereas luteal cells which do not respond at all can even bind 4 to 6 times more EGF molecules/cell than granulosa cells. These observations lead us to conclude that 1) binding studies should be correlated with biological activity as expressed by the increase in cell number, since it is often implied that binding of a given agent will induce a biological effect. The example of EGF binding to granulosa cells versus that to luteal cells demonstrates that this is not always the case; 2) the injection of a 12"1-labeled mitogen in vivo and its subsequent distribution in the body can lead to false conclusion regarding the locus of mitogenic activity. For example, with lZ"I-EGF the in viuo binding to granulosa cells will be barely detectable because of the high sensitivity of the cells, while luteal cells which do not respond to EGF will show, in comparison to granulosa cells, a detectable binding. This could lead to the wrong conclusion that luteal, and not granulosa, cells are the target cells for EGF.
We have shown that during the in vitro luteinization of granulosa cells the number of surface receptors for EGF is increased and, as with EGF-responsive cells, can be down regulated by the extracellular concentration of EGF. Our experiments on the binding of EGF clearly indicate that the loss of a mitogenic response to EGF in cells that undergo luteinization is not due to a defect in the internalization and release of cell-bound EGF and/or in the EGF-induced loss and recovery of surface receptor sites for EGF.
We have further shown that in both luteinizing and nonluteinizing granulosa cells the bound molecules are degraded and released into the medium subsequent to EGF binding and that the cells become incapable of rebinding the same amount of additional EGF molecules. The magnitude of this effect was dependent both on the concentration of EGF and on the duration of the exposure. A 40 to 60% decrease in EGF receptor sites was induced by the occupancy of only a small percentage of the receptors (25% and 10% occupancy in EGF-responsive and nonresponsive cells, respectively). On the other hand, even with very high concentrations of EGF (up to 100 rig/ml) and after an extended preincubation time, 15 to 25% of the initial EGF receptor sites remain available for EGF binding. This fraction is still higher than the percentage of EGF surface receptors that have to be occupied for a maximal mitogenic response (1). It therefore seems that, as receptor sites are lost, the capacity of the cells to further reduce the concentration of receptors is impaired, so that cells with fewer receptor sites become resistant to the hormone-induced decrease in their own surface receptor sites (5). We have also demonstrated that the EGF-induced receptor modulation was limited to EGF receptors, since conA or LDL receptors on the same cells were totally unaffected. Conversely, while exposure to LDL did affect the LDL receptor sites, it had no effect on the EGF receptor sites. Similar results on the effect of hormones upon heterologous receptors were obtained in other systems (3,5).
There are various examples, both in vivo and in vitro, in which the exposure of a target cell to high levels of hormone results in a decreased sensitivity to the hormone, whereas the deprivation of a stimulating agent may result in an increased sensitivity to the introduction of this agent (3,5). In both granulosa and luteal cells, surface receptors for EGF are subjected to modulation by EGF, and this might provide an initial locus for regulating the cellular sensitivity to EGF. The present results indicate, however, that in the ovarian cell system this type of receptor regulation cannot be a major factor in determining the target cell's sensitivity, since it exists in both EGF-responsive and nonresponsive cells. It is therefore necessary to look fpr other steps in the pathway within the target cell which might alter the cell's responsiveness to EGF. One possibility is to study whether EGF-binding Binding of EGF to Granulosa and Luteal Cells proteins are present in the cell's cytosol or nucleus or whether EGF surface receptor sites are transferred into the cell when down regulation takes place. Changes in these possible sites of interaction with EGF, rather than changes in the cell surface receptors, could eventually lead to a loss of sensitivity to EGF without a noticeable change in the properties of the cell's surface receptor sites.
The induction of cell division is accompanied by various changes in cellular physiology and properties of the cell surface (22,23). It is, however, not clear which of the induced alterations are correlative and which are both necessary and sufficient for the G,-G, transition.
Our results indicate that the internalization and release of cell-bound EGF as well as the induced loss and recovery of EGF receptor sites are by themselves not sufficient to commit the cells to undergo cell division. This type of receptor modulation might, however, have a role in transmitting the mitogenic signal. In view of our results it is more likely that the inactivation of both the cell-bound EGF and its appropriate surface receptors simply serves as a control mechanism by which the cell can degrade and release the cell-bound EGF and regulate the concentration of its surface receptors, so that an excess of hormone can specifically decrease the number of its own functional receptor sites.
Note Added in Proof-EGF concentration as low as 15 pglml is enough to induce in granulosa cells a half-maximal mitogenie response (1). It is therefore possible that the loss of EGF responsiveness in luteinizing cells is due to a loss of a very small fraction (4%) of receptor sites that are specifically involved in mediating the mitogenic effect of EGF. This receptor loss will be hardly or not detectable since luteinization is associated with an overall 5-to lo-fold increase in the number of EGF receptor sites per cell.