Effects of Fibroblast and Epidermal Growth Factors on Ovarian Cell Proliferation in Vitro. I. Characterization of the Response of Granulosa Cells to FGF and EGF DENIS GOSPODAROWICZ,* CHARLES R. ILL, AND CHARLES R. BIRDWELL The Salk Institute for Biological Studies, P.O. Box 1809, San Diego, California 92112 final density reached was similar to that observed in 1% serum with EGF and FGF. Addition of EGF or FGF to 10% serum resulted in a final density 3 to 4-fold higher than that observed with 10% serum alone. The ultrastructure of the granulosa cells grown in the presence of EGF or FGF was similar to that of cells maintained in the absence of added mitogens. The only marked difference was that cells grown in the presence of FGF or EGF had a high lipid granule content while cells grown in their absence had a low lipid granule content. The effect of various concentrations of FGF and EGF on the proliferation of granulosa cells has been analyzed. The minimal effective dose of EGF was 3 x 10~uM and saturation was observed at 3 x 10~nM, with a half-maximal response at 6 x 10~13M. With FGF the minimal dose stimulating proliferation was 1.5 x 10~ I2 M and saturation was achieved at 1.5 x 10"'°M, with a half-maximal response at 3 x 10""M. Our results show that EGF and FGF are the most potent mitogens ever observed and are mitogenic for granulosa cells at 300 to 3000-fold lower concentrations than for other cell types which have been studied, such as fibroblasts or lens epithelial cells. (Endocrinology 100: 1108, 1977)
ABSTRACT. Despite numerous studies on the effects of gonadotropins on ovarian cells in tissue culture, the factors controlling the proliferation of granulosa cells in vitro remain unknown. We have examined the effect of fibroblast growth factor (FGF) and epidermal growth factor (EGF) on granulosa cell proliferation in vitro in an attempt to clarify their possible roles in the control of ovarian development. FGF and EGF both stimulate DNA synthesis in resting populations of granulosa cells. The halfmaximal response for this effect with FGF was observedat4 x lO^'MandwithEGFat 1.5 x 10-'3M. Autoradiography demonstrated that the whole cell population initiated DNA synthesis in the presence of either EGF or FGF, thus precluding an additive effect of the two mitogens. When cells were maintained at low density (100 cells/cm2) in the presence of low serum (1%) they divided with a doubling time of 48-72 h, but addition of either EGF or FGF accelerated their proliferation. The doubling time observed in the presence of FGF was 16 h versus 20 h with EGF and the final cell density reached in the presence of EGF or FGF was 20 times that of cells maintained in the presence of 1% calf serum alone. In the presence of 10% serum, granulosa cells had a doubling time of 24 h and the
G
RANULOSA cells maintained in tissue culture have been used extensively to study the agents and mechanisms involved in luteinization (1). However, little attention has been given to agents controlling the proliferation of these cells. While gonadotropins (LH and FSH) are believed to influence granulosa cell proliferation in vivo, they have been shown to have little or no effect on the growth of these cells
in vitro (1-3). In the few instances where granulosa cells have been observed to proliferate after the Received August 18, 1976. Supported by grants from the National Institutes of Health (No. 08180 and 11082), The American Cancer Society (VC 194), and the Rockefeller Foundation. * Present address: Cancer Research Institute, University of California School of Medicine, San Francisco, Ca. 96163.
addition of gonadotropins to the culture medium (2,3), the response may have been due to a potent mitogenic agent, distinct from known pituitary hormones, that is a common contaminant of gonadotropin preparations (4). Recently, several factors have been shown to stimulate the proliferation of a variety of cells in tissue culture. Two of these are epidermal growth factor (EGF) and fibroblast growth factor (FGF). EGF is found at a high concentration in the submaxillary gland of the adult male mouse (5). However, it is not produced solely in the submaxillary gland since extirpation of the gland, while lowering the level of EGF in plasma, does not make it disappear completely (6). First demonstrated as a mitogen for epidermal cells (7), EGF was later shown to stimulate fibroblast prolifera-
1108 The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
FGF AND EGF EFFECTS ON GRANULOSA CELLS tion as well (8,9) and to have a primary structure similar to that of urogastrone (10). FGF has been isolated from the brain and pituitary of mammals (11,12). First shown to be a mitogenic agent for fibroblasts (13), it was later found to promote the proliferation of a wide variety of mesoderm-derived cells (12). FGF has been shown to stimulate the division of vascular endothelial cells (14), cornea endothelial cells (15), chondrocytes (12,16), smooth muscle cells (17), myoblasts (18), glial cells (19), adrenal cortex cells (12,20), and amphibian regeneration cells (21). Since granulosa cells are also derived from the mesoderm and the mitogenic agents for these cells are unknown, we investigated the effect of FGF, as well as EGF, on these cells in vitro. We found that both EGF and FGF are potent mitogens for these cells. Materials and Methods Materials Fibroblast growth factor (FGF) was purified from bovine pituitary glands (11) and brains (12) as previously described. Both pituitary and brain FGF yield single bands on polyacrylamide gel electrophoresis atpH 4.5. No bands are observed at pH 8.5. EGF was purified as described by Savage and Cohen (22) From the submaxillary glands of adult male, Swiss Webster mice that had been given daily sc injections of testosterone propionate (1 mg per animal) for 8 days. The final preparation yielded a single band on polyacrylamide gel electrophoresis at pH 8.5 and, like the preparation of Savage and Cohen (22), three amino acids—lysine, alanineandphenylalanine—were absent from the final preparation. The biological activity of EGF, as measured by the stimulation of the initiation of DNA synthesis in human foreskin fibroblasts, was equal to that of a reference preparation from Dr. S. Cohen, Vanderbilt University. LH was purified from bovine pituitary glands by the method of Papkoffet al. (23), and further purified by diethylaminoethyl cellulose chromatography to remove contaminating thyroid stimulating hormone, as described by Pierce and Carsten (24). The biological activity as measured by the ovarian ascorbic acid depletion
1109
assay was 2.75 U/mg (95% confidence limits 2.1 to 3.7) compared with the standard NIH-LH-B9. The standard was a gift from the NIH Endocrine Study Section, Bethesda, Md. Partially purified preparations of FSH (NIHFSH-S9) were obtained from the NIH Endocrine Study Section, Bethesda, Md. Crystalline bovine serum albumin (BSA) was from Schwarz/Mann and insulin was from Sigma. [3H]Thymidine and Liquifluor solution were from New England Nuclear. Tissue culture media (F12, 199, Dulbecco's modified Eagle's medium [DME] were from GIBCO. Gentamicin was from Schering. Tissue culture dishes were from Falcon Plastics and Lab Tek tissue culture chambers (2 chambers per slide) were from Scientific Products. Methods Others (1-3) have reported the cultivation of granulosa cells at high densities in medium 199. However, since we were concerned with growth at low cell densities, we screened several media and found that media 199 and F10 did not support the survival of the granulosa cells at low density (100 cells/cm2). After a few days the cells became enlarged, vacuolated, and unresponsive to mitogens. In contrast, Ham's F12 and DME supported cell survival under the same conditions. Since F12 is routinely used in our laboratory for the culture of adrenal cortical cells, we chose to use that medium, supplemented with 10% calf serum and 50 /u-g/ml gentamicin, to culture granulosa cells. Establishment of primary cell cultures Ovaries obtained from 1- 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. Medium-sized follicles (1 cm diameter) were slit with a scalpel as described by Channing et al. (25). The follicular fluid was removed with a Pasteur pipette and the granulosa cells within were removed by gently rinsing the inner wall of the follicle with a Pasteur pipette containing 0.1 ml F12 medium. The rinsings from 5-10 follicles were pooled in a sterile centrifuge tube containing 5 ml F12 medium. Two milliliters of the eel suspension medium was then transferred to a 10 cm plastic tissue culture dish and 15 ml F12 with 10% calf serum
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
1110
GOSPODAROWICZ, ILL AND BIRDWELL
was added. The dish was then incubated in a humidified CO2 incubator at 37 C. After 18 h the medium was exchanged with fresh medium to remove unattached cells and the dish was left at 37 C for 3 more days. Measurement of the initiation ofDNA synthesis Stock plates containing granulosa cells were trypsinized with a thin film of 0.25% trypsin in phosphate-buffered saline. After the cells began to detach, they were harvested by suspending the cells in 5 ml F12 with 10% calf serum. The number of cells in the suspension was determined by Coulter counting. Aliquots containing 30,000 cells each were distributed in 3.5 cm dishes with 2 ml F12 with 10% calf serum added. The dishes were incubated for 16 h at 37 C after which the medium was removed, the cells were washed once with F12 and 2 ml F12 with 1% calf serum were added to each dish. The cells were further incubated for 72 h at 37 C. Fifty microliter aliquots of various concentrations of the different mitogens dissolved in 0.5% crystalline BSA in F12 (to reduce the non-specific adsorption of mitogens to vessel walls) were added to the dishes. Twelve hours later [3H]thymidine was added to the dishes as already described (26) at a final concentration of 1 /xCi/ml. After 16 h, the medium was removed, the cells were lysed by the addition of 2 ml of 0.5N NaOH per dish, and the dishes were further incubated for 8 h at 37 C. DNA was precipitated by the addition of cold trichloroacetic acid at a final concentration of 30%. The precipitates were collected on glass fiber filters, washed 3 times with 15% trichloroacetic acid, once with ethanol and then dried. To determine the radioactivity incorporated into DNA, the filters were counted in a liquid scintillation counter (Nuclear Chicago Mark I) with a 4% Liquifluor solution in toluene. The efficiency of [3H]thymidine counting was 48%. The assay was done in triplicate. Autoradiography For autoradiography, 100,000 cells were plated onto Lab Tek 2 chamber slides in 2 ml F12 with 10% calf serum. Sixteen hours later the medium was changed to 2 ml F12 with 1% calf serum and the cells were left at 37 C for 72 h. The different mitogens were then added as already described for measurement of the initia-
Endo • 1977 Vol 100 • No 4
tion of DNA synthesis. Twelve hours later, 1 fiCi of [3H]thymidine per ml was added. The cells were pulsed for 12 h at 37 C after which the slides were washed 3 times in phosphate-buffered saline, immersed in .10% formalin for 20 min, washed in distilled water, and allowed to dry at room temperature. The slides were then dipped in Ilford L4 nuclear emulsion, diluted 1:3 with distilled water, allowed to dry in a vertical position, and placed at 4 C in light-tight slide boxes containing desiccant. After 7 days, the autoradiographs were developed in Kodak D 19, fixed in Kodak Rapid-Fix, rinsed thoroughly, stained with hematoxylin, dehydrated and mounted. Morphology of the cells To study the morphology of the cells maintained in the presence of various concentrations of serum with and without FGF, the granulosa cells were plated at 3000 cells per 6 cm dish in 5 ml F12 with 10% calf serum. Sixteen hours later the medium was removed and the cells were washed once with F12. Then, F12 containing 10, 5, 2.5, 1, 0.5, or 0.1% calf serum was added to the cells with or without EGF or FGF at a concentration of 100 ng/ml. The cells were maintained at 37 C and the media were renewed every other day. When clones were visible to the naked eye, photographs were taken using a Nikon phase contrast microscope. The plates were fixed with 10% formalin and stained with 1% Giemsa for light photomicroscopy in order to count and measure the clones. Electron microscopy Cells were grown in tissue culture dishes as already described and fixed in situ with 2.5% glutaraldehyde-O.lM cacodylate (pH 7.2) for 20 min at room temperature. After washing with the same buffer, the cells were postfixed in 2% OsO4 in 0.1M cacodylate (pH 7.2) for 20. min at room temperature, stained with 2% uranyl acetate in 3% ethanol, dehydrated with ethanol and finally embedded in Epon 812 on the plate to give a layer 1-2 mm thick. When the Epon was cured, the plastic was peeled away from the Epon and areas of interest were cut from the Epon mold and glued to Epon blocks. Thin sections were cut parallel to the monolayer, stained with lead citrate and examined in an Hitachi HU-12 electron microscope.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
FGF AND EGF EFFECTS ON GRANULOSA CELLS
1111
FIG. 1. Initiation of DNA synthe213 sis in subcultures of granulosa a. it cells in response to increasing concentrations of EGF, FGF and serum. (A) Granulosa cells were plated at 30,000 cells per 3.5 cm dish in Ham's F12 me2 | ,.0001 .001 .01 .1 I 10 KX) KXX) | 0 -3 0-4 io-3 K)- O- 0.51 5 I 0 5 0 O 0 dium with 10% calf serum. PROTEINS CONCENTRATION(ng/irt) PROTEINS (ng/ml) Twenty-four hours later the medium was removed and replaced with F12 with 1% calf serum. ~ The cells were left for 3 days before sample additions. FGF ( • • ) , EGF ( • • ) , highly purified LH ( • • ) , NIH-LH-B9 (A A) and calf serum (A A) were added to the plates in various concentrations and 12 h later the cells were pulsed for 16 h with [3H]thymidine. Determinations of [3H]thymidine incorporated into DNA were done as described in the text. The control values were 1200 ± 130 cpm. NIH-FSH-S9 and insulin from 1 ng to 1 fig/ml gave 1200 ± 250 cpm. Every point was done in A) was tested triplicate. Standard deviation did not exceed 10% of the mean. (B) Same as (A), but the EGF (A in a different experiment since, due to its potency, it was necessary to go to lower concentrations than with FGF. Control values were 3800 ± 280 cpm. Every point was done in triplicate. Standard deviation did not exceed 10% of the mean.
Measurement of cell growth Cells were plated at a density of 3000 cells per 6 cm dish in 5 ml F12 with 10% calf serum. Sixteen hours later the media were renewed with fresh F12 containing 10% calf serum or F12 containing 1% calf serum with or without FGF or EGF added at a final concentration of 100 ng/ml. The medium was renewed every other day. Triplicate plates were trypsinized every other day and the cells counted with a Coulter counter.
Results Comparison of the effects of FGF and EGF on the initiation of DNA synthesis by granulosa cells The effects of FGF and EGF on the initiation of DNA synthesis in granulosa cells were investigated by adding increasing concentrations of FGF and EGF to resting cells (in 1% calf serum) and measuring the incorporation of [3H ]thymidine from 12 through 48 h later. Both FGF and EGF were potent stimulants of the initiation of DNA synthesis (Fig. 1A). With FGF, the minimal effective dose was 0.05 ng/ml (4 x 10"12M), and a halfmaximal response was obtained at 0.5 ng/ml (4 x 1 0 - " M ) . At 5 ng/ml (4 x 10- 1 0 M), a maximal response was observed. EGF was even more potent (Fig. 1A,B): the minimal effec-
tive dose was 1 X 10"5 ng/ml (1.5 x 10~15M), a half-maximal response was obtained at 1 x 10"3 ng/ml (1.5 x 10~13M) and a maximal response was observed at 0.1 ng/ml (1.5 x 10~ 12 M). The level of the maximal response of the cells to either FGF or EGF was the same. High concentration of calf serum (20%) stimulated DNA synthesis to the same extent as did 0.1 ng/ml EGF or 5 ng/ml FGF (Fig. 1A). Since NIH-LH-B9 has been reported to contain FGF-like activity (4,11), we compared its effect to that of purified LH and to that of purified FGF and EGF. One microgram of NIH-LH-B9 per mililiter stimulated, DNA synthesis in granulosa cells as effectively as 0.1 ng/ml FGF or 0.001 ng/ml EGF. In contrast, highly purified LH did not stimulate DNA synthesis. This demonstrates that LH is not mitogenic for granulosa cells and indicates that a contaminant present in some LH preparations must be responsible for the mitogenic effects observed. Neither NIH-FSH-S9 (10 /ig/ml) nor insulin (10 /xg/ml) had any effect on DNA synthesis by granulosa cells. [3H ]Thymidine autoradiography
To determine the percentage of cells induced to synthesize DNA under the experi-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
1112
GOSPODAROWICZ, ILL AND BIRDWELL
Endo • 1977 Vol 100 • No 4
FIG. 2. Granulosa cell autoradiography. 100,000 cells were plated onto Lab Tek chamber slides in F12 with 10% calf serum. The cells were then treated and maintained as described in Fig. 1 and Materials and Methods. (A) control, (B) 10% calf serum, (C) 10 ng/ml FGF, (D) 10 ng/ml EGF, (E) 10/xg/ml NIH-LH-B9, (F) 10/xg/ ml highly purified LH.
mental conditions described above, we stimulated the initiation of DNA synthesis by the addition of various mitogens and then prepared the cultures for [3H]thymidine autoradiography. In the absence of mitogens, 10% of the cells maintained in 1% calf serum for 3 days were labelled with [3H ]thymidine during a 12 h pulse (Fig. 2A). In the presence of 10% calf serum, the labelling index increased to 95% (Fig. 2B). The addition of 10 ng/ml FGF (Fig. 2C) or 10 ng/ml EGF (Fig. 2D) to cells in 1% calf serum produced the same result. In contrast, 10 /ng/ml NIH-LH-B9 (Fig. 2E) was needed to raise the labelling index to such a high level. Highly purified LH (10 fig/ml) did not increase the labelling index above the 10% level seen in the controls (Fig. 2F). These results corroborate those obtained by the quantitative assay for [3H]thymidine incorporation (Fig. 1A,B). Since
EGF and FGF stimulate the whole cell population to initiate DNA synthesis, it precludes that they could have an additive effect at saturating concentrations. Evaluation of the requirement of serum for growth Serum enhanced the growth of cells in the presence of either FGF or EGF. Cells were grown in media containing 0.1, 0.5, 1, 5 or 10% calf serum without or with FGF, or EGF at a concentration of 100 ng/ml. After 10-15 days, the cultures were fixed and stained (Fig. 3). In all cases there was a positive correlation between the amount of growth observed and the serum concentration in the medium. A mitogenic effect of EGF and FGF could be observed at serum concentration as low as 0.1% (60 /ug serum protein/ml). In the complete absence of
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
FGF AND EGF EFFECTS ON GRANULOSA CELLS
1113
SERUM . 1 % .5°/c FIG. 3. Correlation between 10% EGF, FGF and the serum concentration on culture growth. Granulosa cells were plated at 3,000 cells per 6 cm dish as described in Materials and Methods and maintained in the presence of a constant concentration of EGF (100 ng/ml) or FGF (100 ng/ml) and with an increasing concentration of calf serum (from 0.1 to 10%). The medium was renewed every other day. Plates maintained in the presence of 10% and 5% serum were fixed with 10% formalin and stained with 0.1% Giemsa on day 10 while plates maintained in tine presence of 1, 0.5 and 0.1% serum were submitted to the same treatment on day 15.
serum the cells maintained with or without EGF and FGF did not survive. The main difference between the response of the cells maintained in the presence of 0.1% serum versus 1% serum with FGF or EGF was that in 0.1% serum the cells had a doubling time of 3 days while in 1% serum the doubling time was 16-20 h. Effects of FGF, EGF and serum on culture growth The relationship between growth of the granulosa cell cultures and the concentration of EGF, FGF and serum in the culture medium was investigated by counting triplicate cultures, maintained under various culture regimens, on successive days. Cultures in 1% serum alone grew slowly, doubling in cell number every 2-3 days (Fig. 4A). When FGF or EGF (100 ng/ml) was added, the doubling time dropped to 20 h for EGF and 16 h for FGF. Cultures grown with 1% calf serum plus EGF and FGF reached cell densities 20 times higher than those reached with 1% serum alone. This was the same density as that reached in 10% calf serum (Fig. 4A). Therefore, it could be said that 1% calf serum plus either EGF or FGF is as potent for promoting granulosa cell growth as is 10% calf
serum. The cells maintained in 10% calf serum doubled in number every 16 h. The addition of FGF or EGF (100 ng/ml) to 10% serum only slightly accelerated the average cell cycle which, at 16 h, is probably limited by the speed of essential, cellular anabolic processes. However, cells maintained in 10% calf serum plus either FGF or EGF reached a final cell density 3 to 4 times higher than that observed with 10% calf serum alone (Fig. 4A). The effects of FGF and EGF have been seen consistently in 20 cultures. Other regular observations are: a) the final cell densities depended on the serum concentration, b) EGF and FGF have a mitogenic effect in cultures maintained with any concentrations of calf serum, from 0.1% to 10%, and c) the mitogenic effect of FGF differed from that of EGF in two ways: the doubling time of cultures growing in FGF was shorter than that of cultures growing in EGF and the final cell density was higher with 100 ng/ml FGF than with 100 ng/ml EGF. Figures 4A and B indicate the greatest variation we have encountered to date in the final cell density of cultures maintained in the presence of 100 ng/ml FGF or 100 ng/ml EGF. While cells in the presence of FGF or EGF continued to divide for 12
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
1114
Endo • 1977 Vol 100 • No 4
GOSPODAROWICZ, ILL AND BIRDWELL
ICT3
2xlO" ! 2x10"' PROTEINS (ng/ml)
2»IO' 5x10
FIG. 4. Effect of FGF and EGF on the proliferation of granulosa cells. (A) Granulosa cells were plated at 3,000 cells per 6 cm dish as described in Materials and Methods. They were then maintained in the presence of 1% calf serum without (A A) or with 100 ng/ml EGF ( • • ) or FGF ( • •). The proliferation effect of FGF and EGF was compared to that of 10% calf serum (A A) to which 100 ng/ml EGF (O O) or 100 ng/ml FGF ( • • ) were added. Every point was done in triplicate. Standard deviation did not exceed 10% of the mean. (B) Similar to (A), the cells were maintained in 1% calf serum without (D • ) or with 100 ng/ml EGF (A A) or FGF (O O). Every point was done in triplicate. Standard deviation did not exceed 10% of the mean. (C) Effect of increasing concentrations of EGF (O O) and FGF (A A) on the proliferation of granulosa cells. The cells were plated as described in Materials and Methods and maintained in the presence of increasing concentrations of EGF and FGF. At day 7 the cells were trypsinized and counted. Control gave 30,000 ± 1800 cells. Every point was done in triplicate. Arrows show the half-maximal response.
days (Fig. 4A), a marked decrease in the rate of proliferation was observed by day 8 with cells maintained with EGF but not with FGF (Fig. 4B). This slowdown resulted in a final cell density 3-fold higher in the presence of FGF than with EGF. In most cases a 2-fold difference was observed.
centrations of FGF was higher than that obtained with saturating concentrations of EGF. This may reflect the longer average cell cycle of cells maintained with EGF. These values demonstrate that FGF and EGF are among the most potent mitogens known for granulosa cells.
Effect of various concentrations of FGF and EGF on the proliferation of granulosa cells Since EGF and FGF stimulated the initiation of DNA synthesis at very low concentrations, we also determined the minimal concentration of mitogen required to obtain a significant increase in cell number and the concentration required for a maximal increase in cell number. EGF was active at 2 X 10~4 ng/ml (3 x 10~14M) and saturation was observed at 0.2 ng/ml (3 x 10~"M) (Fig. 4C). The half-maximal response was observed at 4 x 10"3 ng/ml (6 x 10~13M). FGF was active at 0.02 ng/ml (1.5 x 10~12M) and saturation was achieved at 2 ng/ml (1.5 x 10~10M). The half-maximal response was observed at 0.4 ng/ml (3 x 10-uM) (Fig. 4C). The cell density obtained at saturating con-
Cell morphology Cells maintained in 1% calf serum with neither EGF nor FGF appeared strikingly different from those maintained with EGF or FGF. Cells in the absence of either EGF or FGF were very large and contained single nuclei with 1-3 nucleoli (Fig. 5A). Under phase contrast a granular perinuclear region was visible surrounded by a broad expanse of cytoplasm with prominent longitudinal ridges. The ridges were also visible in fixed and stained cells (Fig. 5C). In the presence of FGF or EGF the cells were much smaller and tightly packed with discrete granules within the cytoplasm (Fig. 5B,D). Discrete lipid inclusions became apparent staining with Oil Red O. The cells did not overgrow each other at confluency but remained in a monolayer (Fig. 5B,D).
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
FGF AND EGF EFFECTS ON GRANULOSA CELLS
1115
FIG. 5. Morphological appearance of granulosa cells maintained in the absence and in the presence of FGF or EGF. (A) Granulosa cells maintained in the absence of EGF and FGF. Cells were plated as described in the text and maintained in the presence of 1% serum for 7 days. (Phase contrast optics 150x). (B) Granulosa cells maintained in the presence of 1% serum and FGF (100 ng/ml) for 7 days. (Phase contrast optics 150x). Similar results were obtained with cells maintained in the presence of 1% serum and EGF (100 ng/ml). (C) Same as (A), but the cells were fixed with 10% formalin and stained with 0.1% Giemsa (60x). (D) Same as (B), but the cells were fixed with 10% formalin and stained with 0.1% Giemsa (60x).
Ultrastructure of cultured granulosa cells grown in the presence of EGF or FGF Figure 6 shows electron micrographs of granulosa cells grown as monolayers in the absence of EGF or FGF (Fig. 6A) and in the presence of 100 ng/ml of EGF (Fig. 6B) or FGF (Fig. 6C). The cytoplasm of these cells is well differentiated and has ultrastructural features similar to those observed by Crisp and Channing (27) for monkey granulosa cells. Many free ribosomes are
found scattered throughout the cytoplasm, often in the form of rosettes, and both microfilaments and microtubules are evident. The mitochondria are often quite elongated. A few large, darkly-stained bodies probably lysosomes are found in the cytoplasm. The Golgi complexes are quite conspicuous and are sometimes found closely associated with the granular (or rough) endoplasmic reticulum, which is very prominent in many of the cells examined. Probably the most characteristic ultrastruc-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
1116
GOSPODAROWICZ, ILL AND BIRDWELL
Endo • 1977 Vol 100 • No 4
FIG. 6. Electron micrographs of granulosa cells grown in the absence of EGF or FGF (A and D), and in the presence of EGF (B) or FGF (C). Cells were grown in 1% serum with or withoug FGF or EGF (100 ng/ml). The presence of the growth factors has no effect on the infrastructure of these granulosa cells. Figure 6D shows a portion of the cytoplasm filled with granular endoplasmic reticulum consisting of very dilated cisternae (arrow), which was a feature common to all of the granulosa cells examined (i.e., in the presence or absence of EGF and FGF). gc-Golgi complex; m—mitochondria; mf— microfilaments; arrows—granular endoplasmic reticulum. 25,000x.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
FGF AND EGF EFFECTS ON GRANULOSA CELLS
1117
FIG. 7. Lipid content of granulosa cells grown in the presence or absence of EGF. (A) Cells grown in the presence of 1% serum: P—polysomes; N—nucleus; M — mitochondria; ER—endoplasmic reticulum. 17,000x. (B) Cells grown in the presence of 1% serum and 100 ng/ml of EGF: LI, lipid inclusions. 9,000x. Similar morphology was observed with cells grown in the presence of 1% serum and 100 ng/ml of FGF.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
1118
GOSPODAROWICZ, ILL AND BIRDWELL
tural feature of these cells is that the granular endoplasmic reticulum consisted of membrane-bound cisternae which were very dilated, sometimes as large as 4 fim in diameter. In some cells (not depicted here) the cytoplasm is almost completely packed with the dilated granular endoplasmic reticulum. This type of granular endoplasmic reticulum has been found in granulosa cells examined in vivo (28). The major difference between cells grown in the absence of mitogens and cells grown in their presence was in the lipid content of the cells. Cells grown without added mitogens and in low serum (1%) had very few lipid granules (Fig. 7A), while cells grown in the presence of mitogens and in low serum (1%) had a high content of lipid granules (Fig. 7B). Discussion Using granulosa cells maintained in tissue culture (1-3) others have shown that trophic hormones such as LH or FSH are involved in the maturation process leading to luteinization, but are not involved, at least as far as LH is concerned, in the control of granulosa cell proliferation (29). This also applies to the proliferation of luteal cells (the progeny of granulosa cells) since it is inhibited by LH in vitro while cell differentiation is enhanced (30). From such evidence one could conclude that granulosa cells are committed to differentiation and, when maintained in tissue culture under the right conditions (in the presence of gonadotropins), will stop dividing and become luteal cells. Another hypothesis is that growth factors exist for these cells which are similar to those which induce the proliferation of granulosa cells in vivo during the early differentiation of follicles when the cells are not yet under gonadotropin control. This possibility prompted us to look at the effect of growth factors such as EGF and FGF on granulosa cells. Such agents must take a prominent position regarding the control
Endo • 1977 Vo! 100 • No 4
of cellular proliferation since it became evident tht trophic hormones such as LH do not have mitogenic activity (29,30) in vitro and others, such as ACTH, are antimitotic (31,32). Our results indicate that EGF and FGF are by far the most potent mitogenic agents for granulosa cells yet found. The mitogenic effect of EGF on granulosa cells differs from its effect on human fibroblasts (33) and lens epithelial cells (34). The molar concentration of EGF that induces a halfmaximal response for the initiation of DNA synthesis in granulosa cells (1.5 x 10~ 13 M) is 500-fold lower than for human fibroblasts (7 x 10-nM (33)) and 5000-fold lower than for lens epithelium (7 x 10~ 10 M (34)). The dose of EGF producing a half-maximal response of granulosa cells for the initiation of DNA synthesis (1.5 x 10" 13 M) was lower than for cell proliferation (6 x 10" 1 3 M). This can be best explained by the observation of others that a lower number of binding sites needs to be occupied to promote initiation of DNA synthesis than to promote cell division (33). With human fibroblasts, for example, 20% occupancy of EGF binding sites is sufficient to effect a full biological response when the initiation of DNA synthesis is considered, but the cells will not divide unless a higher binding occupancy is attained (33). Besides requiring lower concentrations, the response of granulosa cells to EGF differed from the responses of human fibroblasts and lens epithelial cells in other ways. The responses of human fibroblasts and lens epithelium to EGF depend on the cell density, the concentration of serum in the medium, and the length of time the cells are left without renewal of the medium. Under ideal conditions EGF has as great a mitogenic effect as serum, but slight variations in the assay procedure can result in a diminution of the effect of EGF to a level lower than that of serum. In contrast, the response of sparse cultures (100 cells/cm2) of granulosa cells to EGF is not dependent on the length of time the cells are maintained
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
FGF AND EGF EFFECTS ON GRANULOSA CELLS without renewal of the medium; EGF is always as effective a mitogen as serum. For granulosa cells, as for human fibroblasts (9), the magnitude of the response of EGF depends on the serum concentration in the medium. However, while with fibroblasts maintained in low serum (0.35%) the response to EGF is small (20% that of optimal serum concentration) (9), with granulosa cells maintained in low serum (0.1%) the cells respond to EGF by proliferating actively. The only major difference observed is that the cell cycle will be longer in low serum than in high serum concentrations. The observation that primary cultures of granulosa cells, in contrast to other cell types, survived quite well in low serum can be best explained by their origin. In vivo granulosa cells are maintained in an avascular medium and this could account for their lesser dependence on serum than cells derived from richly vascularized organs. Thus, since granulosa cells can be induced to divide in a virtually serumless condition (60 (j.g of serum protein/ml) by very small amounts of mitogens, they should be an ideal cell type with which to study the mechanisms by which cells are induced to proliferate in vitro. The differences between the responses of granulosa cells and human fibroblasts (35) to FGF are generally the same as the differences between the responses of the two cell types of EGF. EGF is a more potent mitogen for granulosa cells than is FGF since lower molar concentrations of EGF induce a half-maximal response in the initiation of DNA synthesis assay (250-fold less than FGF) as well as in the increase in cell number assay (50-fold less than FGF). Because the slope of the FGF dose-response curve is steeper than that of the EGF curve, the differences in potency between the two mitogens are more marked at low concentrations than at levels near saturation. Although both EGF and FGF can stimulate the proliferation of granulosa cells as well as fibroblasts, this does not mean that
1119
the proliferation of all mesoderm-derived cells are under FGF and EGF control. EGF does not stimulate the proliferation of adrenal cortex cells (36) or that of vascular endothelial cells (14), and has only a marginal effect on vascular smooth muscle cells (14). In contrast, FGF is a strong mitogenic agent for all these cells. It remains to be seen whether or not FGF and EGF have similar effects on the proliferation of granulosa cells in vivo. Factors involved in follicular growth are poorly understood. It seems that during the early stage of ovarian development in rats the marked granulosa cell proliferation which takes place is independent of gonadotropin since it is not suppressed by daily administration of rat gonadotropin antibodies (37). Studies by Pedersen (38) and others (39) have indicated that the DNA labelling index of granulosa cells could be affected by gonadotropins but this depends on the age of the follicle. Gonadotropins are only active at a late stage of follicular development. Estrogens are also involved in the control of proliferation of granulosa cells and exert a direct effect on the growing follicles (40) since, in hypophysectomized animals, they increase the number of small and mediumsized follicles (40). Thus, the regulation of follicular development and, consequently, the control of granulosa cell proliferation seem to be extremely complex and may depend on the age of the individual as well as on the endocrine composition at a given time. However, since there is a general consensus that the factors regulating the transformation of primordial follicles into growing follicles do not involve the intervention of gonadotropin (41,42), this phase of follicular development should be an ideal period in which to study the possible in vivo effect of EGF or FGF on proliferation of granulosa cells. Acknowledgments We thank Drs. J. Moran and A. Mescher for helpful discussions.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
1120
GOSPODAROWICZ, ILL AND BIRDWELL
References 1. Channing, C. P., Recent Prog Horm Res 26: 589, 1970. 2. Channing, C. P., Endocrinology 87: 49, 1970. 3. Channing, C. P., In McKems, K. W. (ed.), The Gonads, Appleton-Century-Crofts, New York, 1966, p. 245. 4. Gospodarowicz, D., Nature 249: 123, 1974. 5. Cohen, S.J Biol Chem 237: 123, 1974. 6. Bynny, R. L., D. N. Orth, S. Cohen, and E. S. Doyne, Endocrinology 95: 776, 1974. 7. Cohen, S., and C. R. Savage, Recent Prog Horm Res 30: 551, 1974. 8. Cohen, S., and G. Carpenter, Proc Natl Acad Sci USA 72: 1317, 1975. 9. Cohen, S., G. Carpenter, and K. Lembach, Adv Metab Disord 8: 265, 1975. 10. Gregory, H., Nature 257: 325, 1975. 11. Gospodarowicz, D . J Biol Chem 250: 2515, 1975. 12. Gospodarowicz, D., J. Moran, and H. Bialecki, In Growth Hormone and Related Peptides, Excerpta Med, Int. Congr. 381, Elsevier, New York, 1976, p. 141. 13. Gospodarowicz, D., and J. Moran, Proc Natl Acad Sci USA 71: 4648, 1974. 14. Gospodarowicz, D., J. Moran, and D. Braun,/ Cell Physiol 1976 (In press). 15. Gospodarowicz, D., and A. Mescher,/ Cell Physiol 1976 (In press). 16. Jones, K. L., and J. Allison, Endocrinology 97: 359, 1975. 17. Ross, R., and J. A. Glomset, J Cell Biol 1976 (In press). 18. Gospodarowicz, D., J. Weseman, J. Moran, and J. LindstromJ Cell Biol 70: 395, 1976. 19. Westennark, B., and A. Wasteson, Adv Metab Disord 8: 85, 1975. 20. Gospodarowicz, D., and H. H. Handley, Endocrinology 97: 102, 1975. 21. Gospodarowicz, D., P. Rudland, J. Lindstrom, and K. Benirschke, Adv Metab Disord 8: 301, 1975. 22. Savage, C. R., and S. CohenJ Biol Chem 247: 7609, 1972.
Endo • 1977 Vol 100 • No 4
23. Papkoff, H., D. Gospodarowicz, A. Candiotti, and C. H. Li, Arch Biochem Biophys 111: 431, 1975. 24. Pierce, J. G., and M. E. Carsten, In Werner, S. C. (ed.), Thyrotropin, Charles C Thomas, Springfield, 111., 1963, p. 216. 25. Channing, C. P., and F. Ledwitz-Rigby, In Hardman, J. G., and B. W. O'Malley (eds.), Methods of Enzymology, vol. 39, Academic Press, New York, 1975, p. 183. 26. Gospodarowicz, D., and J. Moran, Exp Cell Res 90: 379, 1975. 27. Crisp, T. M., and C. P. Channing, Biol Reprod 7: 55, 1972. 28. Blanchette, E. J.J Cell Biol 31: 501, 1966. 29. McNatty, K. P., and R. S. Sowers, J Endocrinol 66: 391, 1975. 30. Gospodarowicz, D., and F. Gospodarowicz, Endocrinology 96: 458, 1975. 31. Masui, H., and L. D. Garren, Proc Natl Acad Sci USA 68: 3206, 1971. 32. Ramachandran, J., and A. T. Suyama, Proc Natl Acad Sci USA 72: 113, 1975. 33. Hollenberg, M. D., and P. Cuatrecasas, J Biol Chem 250: 3845, 1975. 34. Hollenberg, M. D., Arch Biochem Biophys 171: 371, 1975. 35. Gospodarowicz, D., and J. S. Moran,/ Cell Biol 6: 451, 1975. 36. Gospodarowicz, D., C. Ill, P. Hornsby, and G. Gill, Endocrinology 1976 100: 1080, 1977. 37. Eshkol, L., and B. Lunenfeld, In Saxena, B., C. J. Belring, and H. M. Gandy (eds.), Gonadotropins, Wiley, New York, 1972, p. 335. 38. Pedersen, T., Ada Endocrinol (Kbh) 62: 117,1969. 39. Ryle, M.J Reprod Fertil 19: 87, 1969. 40. Croes-Buth, S., F. J. A. Paesi, and S. E. de Jongh, Ada Endocrinol (Kbh) 32: 399, 1959. 41. Greenwald, G. S., In Geiger, S. R. (ed.), Handbook of Physiology, vol. IV, section 7, part 2, American Physiological Society, Washington, D.C., 1974, p. 293. 42. Blandau, R. J., In Greep, R. O., and L. Weiss, Histology, ed. 3, Academic Press, New York, 1973, p. 768.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
ABSTRACT. Despite numerous studies on the effects of gonadotropins on ovarian cells in tissue culture, the factors controlling the proliferation of granulosa cells in vitro remain unknown. We have examined the effect of fibroblast growth factor (FGF) and epidermal growth factor (EGF) on granulosa cell proliferation in vitro in an attempt to clarify their possible roles in the control of ovarian development. FGF and EGF both stimulate DNA synthesis in resting populations of granulosa cells. The halfmaximal response for this effect with FGF was observedat4 x lO^'MandwithEGFat 1.5 x 10-'3M. Autoradiography demonstrated that the whole cell population initiated DNA synthesis in the presence of either EGF or FGF, thus precluding an additive effect of the two mitogens. When cells were maintained at low density (100 cells/cm2) in the presence of low serum (1%) they divided with a doubling time of 48-72 h, but addition of either EGF or FGF accelerated their proliferation. The doubling time observed in the presence of FGF was 16 h versus 20 h with EGF and the final cell density reached in the presence of EGF or FGF was 20 times that of cells maintained in the presence of 1% calf serum alone. In the presence of 10% serum, granulosa cells had a doubling time of 24 h and the
G
RANULOSA cells maintained in tissue culture have been used extensively to study the agents and mechanisms involved in luteinization (1). However, little attention has been given to agents controlling the proliferation of these cells. While gonadotropins (LH and FSH) are believed to influence granulosa cell proliferation in vivo, they have been shown to have little or no effect on the growth of these cells
in vitro (1-3). In the few instances where granulosa cells have been observed to proliferate after the Received August 18, 1976. Supported by grants from the National Institutes of Health (No. 08180 and 11082), The American Cancer Society (VC 194), and the Rockefeller Foundation. * Present address: Cancer Research Institute, University of California School of Medicine, San Francisco, Ca. 96163.
addition of gonadotropins to the culture medium (2,3), the response may have been due to a potent mitogenic agent, distinct from known pituitary hormones, that is a common contaminant of gonadotropin preparations (4). Recently, several factors have been shown to stimulate the proliferation of a variety of cells in tissue culture. Two of these are epidermal growth factor (EGF) and fibroblast growth factor (FGF). EGF is found at a high concentration in the submaxillary gland of the adult male mouse (5). However, it is not produced solely in the submaxillary gland since extirpation of the gland, while lowering the level of EGF in plasma, does not make it disappear completely (6). First demonstrated as a mitogen for epidermal cells (7), EGF was later shown to stimulate fibroblast prolifera-
1108 The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
FGF AND EGF EFFECTS ON GRANULOSA CELLS tion as well (8,9) and to have a primary structure similar to that of urogastrone (10). FGF has been isolated from the brain and pituitary of mammals (11,12). First shown to be a mitogenic agent for fibroblasts (13), it was later found to promote the proliferation of a wide variety of mesoderm-derived cells (12). FGF has been shown to stimulate the division of vascular endothelial cells (14), cornea endothelial cells (15), chondrocytes (12,16), smooth muscle cells (17), myoblasts (18), glial cells (19), adrenal cortex cells (12,20), and amphibian regeneration cells (21). Since granulosa cells are also derived from the mesoderm and the mitogenic agents for these cells are unknown, we investigated the effect of FGF, as well as EGF, on these cells in vitro. We found that both EGF and FGF are potent mitogens for these cells. Materials and Methods Materials Fibroblast growth factor (FGF) was purified from bovine pituitary glands (11) and brains (12) as previously described. Both pituitary and brain FGF yield single bands on polyacrylamide gel electrophoresis atpH 4.5. No bands are observed at pH 8.5. EGF was purified as described by Savage and Cohen (22) From the submaxillary glands of adult male, Swiss Webster mice that had been given daily sc injections of testosterone propionate (1 mg per animal) for 8 days. The final preparation yielded a single band on polyacrylamide gel electrophoresis at pH 8.5 and, like the preparation of Savage and Cohen (22), three amino acids—lysine, alanineandphenylalanine—were absent from the final preparation. The biological activity of EGF, as measured by the stimulation of the initiation of DNA synthesis in human foreskin fibroblasts, was equal to that of a reference preparation from Dr. S. Cohen, Vanderbilt University. LH was purified from bovine pituitary glands by the method of Papkoffet al. (23), and further purified by diethylaminoethyl cellulose chromatography to remove contaminating thyroid stimulating hormone, as described by Pierce and Carsten (24). The biological activity as measured by the ovarian ascorbic acid depletion
1109
assay was 2.75 U/mg (95% confidence limits 2.1 to 3.7) compared with the standard NIH-LH-B9. The standard was a gift from the NIH Endocrine Study Section, Bethesda, Md. Partially purified preparations of FSH (NIHFSH-S9) were obtained from the NIH Endocrine Study Section, Bethesda, Md. Crystalline bovine serum albumin (BSA) was from Schwarz/Mann and insulin was from Sigma. [3H]Thymidine and Liquifluor solution were from New England Nuclear. Tissue culture media (F12, 199, Dulbecco's modified Eagle's medium [DME] were from GIBCO. Gentamicin was from Schering. Tissue culture dishes were from Falcon Plastics and Lab Tek tissue culture chambers (2 chambers per slide) were from Scientific Products. Methods Others (1-3) have reported the cultivation of granulosa cells at high densities in medium 199. However, since we were concerned with growth at low cell densities, we screened several media and found that media 199 and F10 did not support the survival of the granulosa cells at low density (100 cells/cm2). After a few days the cells became enlarged, vacuolated, and unresponsive to mitogens. In contrast, Ham's F12 and DME supported cell survival under the same conditions. Since F12 is routinely used in our laboratory for the culture of adrenal cortical cells, we chose to use that medium, supplemented with 10% calf serum and 50 /u-g/ml gentamicin, to culture granulosa cells. Establishment of primary cell cultures Ovaries obtained from 1- 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. Medium-sized follicles (1 cm diameter) were slit with a scalpel as described by Channing et al. (25). The follicular fluid was removed with a Pasteur pipette and the granulosa cells within were removed by gently rinsing the inner wall of the follicle with a Pasteur pipette containing 0.1 ml F12 medium. The rinsings from 5-10 follicles were pooled in a sterile centrifuge tube containing 5 ml F12 medium. Two milliliters of the eel suspension medium was then transferred to a 10 cm plastic tissue culture dish and 15 ml F12 with 10% calf serum
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
1110
GOSPODAROWICZ, ILL AND BIRDWELL
was added. The dish was then incubated in a humidified CO2 incubator at 37 C. After 18 h the medium was exchanged with fresh medium to remove unattached cells and the dish was left at 37 C for 3 more days. Measurement of the initiation ofDNA synthesis Stock plates containing granulosa cells were trypsinized with a thin film of 0.25% trypsin in phosphate-buffered saline. After the cells began to detach, they were harvested by suspending the cells in 5 ml F12 with 10% calf serum. The number of cells in the suspension was determined by Coulter counting. Aliquots containing 30,000 cells each were distributed in 3.5 cm dishes with 2 ml F12 with 10% calf serum added. The dishes were incubated for 16 h at 37 C after which the medium was removed, the cells were washed once with F12 and 2 ml F12 with 1% calf serum were added to each dish. The cells were further incubated for 72 h at 37 C. Fifty microliter aliquots of various concentrations of the different mitogens dissolved in 0.5% crystalline BSA in F12 (to reduce the non-specific adsorption of mitogens to vessel walls) were added to the dishes. Twelve hours later [3H]thymidine was added to the dishes as already described (26) at a final concentration of 1 /xCi/ml. After 16 h, the medium was removed, the cells were lysed by the addition of 2 ml of 0.5N NaOH per dish, and the dishes were further incubated for 8 h at 37 C. DNA was precipitated by the addition of cold trichloroacetic acid at a final concentration of 30%. The precipitates were collected on glass fiber filters, washed 3 times with 15% trichloroacetic acid, once with ethanol and then dried. To determine the radioactivity incorporated into DNA, the filters were counted in a liquid scintillation counter (Nuclear Chicago Mark I) with a 4% Liquifluor solution in toluene. The efficiency of [3H]thymidine counting was 48%. The assay was done in triplicate. Autoradiography For autoradiography, 100,000 cells were plated onto Lab Tek 2 chamber slides in 2 ml F12 with 10% calf serum. Sixteen hours later the medium was changed to 2 ml F12 with 1% calf serum and the cells were left at 37 C for 72 h. The different mitogens were then added as already described for measurement of the initia-
Endo • 1977 Vol 100 • No 4
tion of DNA synthesis. Twelve hours later, 1 fiCi of [3H]thymidine per ml was added. The cells were pulsed for 12 h at 37 C after which the slides were washed 3 times in phosphate-buffered saline, immersed in .10% formalin for 20 min, washed in distilled water, and allowed to dry at room temperature. The slides were then dipped in Ilford L4 nuclear emulsion, diluted 1:3 with distilled water, allowed to dry in a vertical position, and placed at 4 C in light-tight slide boxes containing desiccant. After 7 days, the autoradiographs were developed in Kodak D 19, fixed in Kodak Rapid-Fix, rinsed thoroughly, stained with hematoxylin, dehydrated and mounted. Morphology of the cells To study the morphology of the cells maintained in the presence of various concentrations of serum with and without FGF, the granulosa cells were plated at 3000 cells per 6 cm dish in 5 ml F12 with 10% calf serum. Sixteen hours later the medium was removed and the cells were washed once with F12. Then, F12 containing 10, 5, 2.5, 1, 0.5, or 0.1% calf serum was added to the cells with or without EGF or FGF at a concentration of 100 ng/ml. The cells were maintained at 37 C and the media were renewed every other day. When clones were visible to the naked eye, photographs were taken using a Nikon phase contrast microscope. The plates were fixed with 10% formalin and stained with 1% Giemsa for light photomicroscopy in order to count and measure the clones. Electron microscopy Cells were grown in tissue culture dishes as already described and fixed in situ with 2.5% glutaraldehyde-O.lM cacodylate (pH 7.2) for 20 min at room temperature. After washing with the same buffer, the cells were postfixed in 2% OsO4 in 0.1M cacodylate (pH 7.2) for 20. min at room temperature, stained with 2% uranyl acetate in 3% ethanol, dehydrated with ethanol and finally embedded in Epon 812 on the plate to give a layer 1-2 mm thick. When the Epon was cured, the plastic was peeled away from the Epon and areas of interest were cut from the Epon mold and glued to Epon blocks. Thin sections were cut parallel to the monolayer, stained with lead citrate and examined in an Hitachi HU-12 electron microscope.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
FGF AND EGF EFFECTS ON GRANULOSA CELLS
1111
FIG. 1. Initiation of DNA synthe213 sis in subcultures of granulosa a. it cells in response to increasing concentrations of EGF, FGF and serum. (A) Granulosa cells were plated at 30,000 cells per 3.5 cm dish in Ham's F12 me2 | ,.0001 .001 .01 .1 I 10 KX) KXX) | 0 -3 0-4 io-3 K)- O- 0.51 5 I 0 5 0 O 0 dium with 10% calf serum. PROTEINS CONCENTRATION(ng/irt) PROTEINS (ng/ml) Twenty-four hours later the medium was removed and replaced with F12 with 1% calf serum. ~ The cells were left for 3 days before sample additions. FGF ( • • ) , EGF ( • • ) , highly purified LH ( • • ) , NIH-LH-B9 (A A) and calf serum (A A) were added to the plates in various concentrations and 12 h later the cells were pulsed for 16 h with [3H]thymidine. Determinations of [3H]thymidine incorporated into DNA were done as described in the text. The control values were 1200 ± 130 cpm. NIH-FSH-S9 and insulin from 1 ng to 1 fig/ml gave 1200 ± 250 cpm. Every point was done in A) was tested triplicate. Standard deviation did not exceed 10% of the mean. (B) Same as (A), but the EGF (A in a different experiment since, due to its potency, it was necessary to go to lower concentrations than with FGF. Control values were 3800 ± 280 cpm. Every point was done in triplicate. Standard deviation did not exceed 10% of the mean.
Measurement of cell growth Cells were plated at a density of 3000 cells per 6 cm dish in 5 ml F12 with 10% calf serum. Sixteen hours later the media were renewed with fresh F12 containing 10% calf serum or F12 containing 1% calf serum with or without FGF or EGF added at a final concentration of 100 ng/ml. The medium was renewed every other day. Triplicate plates were trypsinized every other day and the cells counted with a Coulter counter.
Results Comparison of the effects of FGF and EGF on the initiation of DNA synthesis by granulosa cells The effects of FGF and EGF on the initiation of DNA synthesis in granulosa cells were investigated by adding increasing concentrations of FGF and EGF to resting cells (in 1% calf serum) and measuring the incorporation of [3H ]thymidine from 12 through 48 h later. Both FGF and EGF were potent stimulants of the initiation of DNA synthesis (Fig. 1A). With FGF, the minimal effective dose was 0.05 ng/ml (4 x 10"12M), and a halfmaximal response was obtained at 0.5 ng/ml (4 x 1 0 - " M ) . At 5 ng/ml (4 x 10- 1 0 M), a maximal response was observed. EGF was even more potent (Fig. 1A,B): the minimal effec-
tive dose was 1 X 10"5 ng/ml (1.5 x 10~15M), a half-maximal response was obtained at 1 x 10"3 ng/ml (1.5 x 10~13M) and a maximal response was observed at 0.1 ng/ml (1.5 x 10~ 12 M). The level of the maximal response of the cells to either FGF or EGF was the same. High concentration of calf serum (20%) stimulated DNA synthesis to the same extent as did 0.1 ng/ml EGF or 5 ng/ml FGF (Fig. 1A). Since NIH-LH-B9 has been reported to contain FGF-like activity (4,11), we compared its effect to that of purified LH and to that of purified FGF and EGF. One microgram of NIH-LH-B9 per mililiter stimulated, DNA synthesis in granulosa cells as effectively as 0.1 ng/ml FGF or 0.001 ng/ml EGF. In contrast, highly purified LH did not stimulate DNA synthesis. This demonstrates that LH is not mitogenic for granulosa cells and indicates that a contaminant present in some LH preparations must be responsible for the mitogenic effects observed. Neither NIH-FSH-S9 (10 /ig/ml) nor insulin (10 /xg/ml) had any effect on DNA synthesis by granulosa cells. [3H ]Thymidine autoradiography
To determine the percentage of cells induced to synthesize DNA under the experi-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
1112
GOSPODAROWICZ, ILL AND BIRDWELL
Endo • 1977 Vol 100 • No 4
FIG. 2. Granulosa cell autoradiography. 100,000 cells were plated onto Lab Tek chamber slides in F12 with 10% calf serum. The cells were then treated and maintained as described in Fig. 1 and Materials and Methods. (A) control, (B) 10% calf serum, (C) 10 ng/ml FGF, (D) 10 ng/ml EGF, (E) 10/xg/ml NIH-LH-B9, (F) 10/xg/ ml highly purified LH.
mental conditions described above, we stimulated the initiation of DNA synthesis by the addition of various mitogens and then prepared the cultures for [3H]thymidine autoradiography. In the absence of mitogens, 10% of the cells maintained in 1% calf serum for 3 days were labelled with [3H ]thymidine during a 12 h pulse (Fig. 2A). In the presence of 10% calf serum, the labelling index increased to 95% (Fig. 2B). The addition of 10 ng/ml FGF (Fig. 2C) or 10 ng/ml EGF (Fig. 2D) to cells in 1% calf serum produced the same result. In contrast, 10 /ng/ml NIH-LH-B9 (Fig. 2E) was needed to raise the labelling index to such a high level. Highly purified LH (10 fig/ml) did not increase the labelling index above the 10% level seen in the controls (Fig. 2F). These results corroborate those obtained by the quantitative assay for [3H]thymidine incorporation (Fig. 1A,B). Since
EGF and FGF stimulate the whole cell population to initiate DNA synthesis, it precludes that they could have an additive effect at saturating concentrations. Evaluation of the requirement of serum for growth Serum enhanced the growth of cells in the presence of either FGF or EGF. Cells were grown in media containing 0.1, 0.5, 1, 5 or 10% calf serum without or with FGF, or EGF at a concentration of 100 ng/ml. After 10-15 days, the cultures were fixed and stained (Fig. 3). In all cases there was a positive correlation between the amount of growth observed and the serum concentration in the medium. A mitogenic effect of EGF and FGF could be observed at serum concentration as low as 0.1% (60 /ug serum protein/ml). In the complete absence of
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
FGF AND EGF EFFECTS ON GRANULOSA CELLS
1113
SERUM . 1 % .5°/c FIG. 3. Correlation between 10% EGF, FGF and the serum concentration on culture growth. Granulosa cells were plated at 3,000 cells per 6 cm dish as described in Materials and Methods and maintained in the presence of a constant concentration of EGF (100 ng/ml) or FGF (100 ng/ml) and with an increasing concentration of calf serum (from 0.1 to 10%). The medium was renewed every other day. Plates maintained in the presence of 10% and 5% serum were fixed with 10% formalin and stained with 0.1% Giemsa on day 10 while plates maintained in tine presence of 1, 0.5 and 0.1% serum were submitted to the same treatment on day 15.
serum the cells maintained with or without EGF and FGF did not survive. The main difference between the response of the cells maintained in the presence of 0.1% serum versus 1% serum with FGF or EGF was that in 0.1% serum the cells had a doubling time of 3 days while in 1% serum the doubling time was 16-20 h. Effects of FGF, EGF and serum on culture growth The relationship between growth of the granulosa cell cultures and the concentration of EGF, FGF and serum in the culture medium was investigated by counting triplicate cultures, maintained under various culture regimens, on successive days. Cultures in 1% serum alone grew slowly, doubling in cell number every 2-3 days (Fig. 4A). When FGF or EGF (100 ng/ml) was added, the doubling time dropped to 20 h for EGF and 16 h for FGF. Cultures grown with 1% calf serum plus EGF and FGF reached cell densities 20 times higher than those reached with 1% serum alone. This was the same density as that reached in 10% calf serum (Fig. 4A). Therefore, it could be said that 1% calf serum plus either EGF or FGF is as potent for promoting granulosa cell growth as is 10% calf
serum. The cells maintained in 10% calf serum doubled in number every 16 h. The addition of FGF or EGF (100 ng/ml) to 10% serum only slightly accelerated the average cell cycle which, at 16 h, is probably limited by the speed of essential, cellular anabolic processes. However, cells maintained in 10% calf serum plus either FGF or EGF reached a final cell density 3 to 4 times higher than that observed with 10% calf serum alone (Fig. 4A). The effects of FGF and EGF have been seen consistently in 20 cultures. Other regular observations are: a) the final cell densities depended on the serum concentration, b) EGF and FGF have a mitogenic effect in cultures maintained with any concentrations of calf serum, from 0.1% to 10%, and c) the mitogenic effect of FGF differed from that of EGF in two ways: the doubling time of cultures growing in FGF was shorter than that of cultures growing in EGF and the final cell density was higher with 100 ng/ml FGF than with 100 ng/ml EGF. Figures 4A and B indicate the greatest variation we have encountered to date in the final cell density of cultures maintained in the presence of 100 ng/ml FGF or 100 ng/ml EGF. While cells in the presence of FGF or EGF continued to divide for 12
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
1114
Endo • 1977 Vol 100 • No 4
GOSPODAROWICZ, ILL AND BIRDWELL
ICT3
2xlO" ! 2x10"' PROTEINS (ng/ml)
2»IO' 5x10
FIG. 4. Effect of FGF and EGF on the proliferation of granulosa cells. (A) Granulosa cells were plated at 3,000 cells per 6 cm dish as described in Materials and Methods. They were then maintained in the presence of 1% calf serum without (A A) or with 100 ng/ml EGF ( • • ) or FGF ( • •). The proliferation effect of FGF and EGF was compared to that of 10% calf serum (A A) to which 100 ng/ml EGF (O O) or 100 ng/ml FGF ( • • ) were added. Every point was done in triplicate. Standard deviation did not exceed 10% of the mean. (B) Similar to (A), the cells were maintained in 1% calf serum without (D • ) or with 100 ng/ml EGF (A A) or FGF (O O). Every point was done in triplicate. Standard deviation did not exceed 10% of the mean. (C) Effect of increasing concentrations of EGF (O O) and FGF (A A) on the proliferation of granulosa cells. The cells were plated as described in Materials and Methods and maintained in the presence of increasing concentrations of EGF and FGF. At day 7 the cells were trypsinized and counted. Control gave 30,000 ± 1800 cells. Every point was done in triplicate. Arrows show the half-maximal response.
days (Fig. 4A), a marked decrease in the rate of proliferation was observed by day 8 with cells maintained with EGF but not with FGF (Fig. 4B). This slowdown resulted in a final cell density 3-fold higher in the presence of FGF than with EGF. In most cases a 2-fold difference was observed.
centrations of FGF was higher than that obtained with saturating concentrations of EGF. This may reflect the longer average cell cycle of cells maintained with EGF. These values demonstrate that FGF and EGF are among the most potent mitogens known for granulosa cells.
Effect of various concentrations of FGF and EGF on the proliferation of granulosa cells Since EGF and FGF stimulated the initiation of DNA synthesis at very low concentrations, we also determined the minimal concentration of mitogen required to obtain a significant increase in cell number and the concentration required for a maximal increase in cell number. EGF was active at 2 X 10~4 ng/ml (3 x 10~14M) and saturation was observed at 0.2 ng/ml (3 x 10~"M) (Fig. 4C). The half-maximal response was observed at 4 x 10"3 ng/ml (6 x 10~13M). FGF was active at 0.02 ng/ml (1.5 x 10~12M) and saturation was achieved at 2 ng/ml (1.5 x 10~10M). The half-maximal response was observed at 0.4 ng/ml (3 x 10-uM) (Fig. 4C). The cell density obtained at saturating con-
Cell morphology Cells maintained in 1% calf serum with neither EGF nor FGF appeared strikingly different from those maintained with EGF or FGF. Cells in the absence of either EGF or FGF were very large and contained single nuclei with 1-3 nucleoli (Fig. 5A). Under phase contrast a granular perinuclear region was visible surrounded by a broad expanse of cytoplasm with prominent longitudinal ridges. The ridges were also visible in fixed and stained cells (Fig. 5C). In the presence of FGF or EGF the cells were much smaller and tightly packed with discrete granules within the cytoplasm (Fig. 5B,D). Discrete lipid inclusions became apparent staining with Oil Red O. The cells did not overgrow each other at confluency but remained in a monolayer (Fig. 5B,D).
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
FGF AND EGF EFFECTS ON GRANULOSA CELLS
1115
FIG. 5. Morphological appearance of granulosa cells maintained in the absence and in the presence of FGF or EGF. (A) Granulosa cells maintained in the absence of EGF and FGF. Cells were plated as described in the text and maintained in the presence of 1% serum for 7 days. (Phase contrast optics 150x). (B) Granulosa cells maintained in the presence of 1% serum and FGF (100 ng/ml) for 7 days. (Phase contrast optics 150x). Similar results were obtained with cells maintained in the presence of 1% serum and EGF (100 ng/ml). (C) Same as (A), but the cells were fixed with 10% formalin and stained with 0.1% Giemsa (60x). (D) Same as (B), but the cells were fixed with 10% formalin and stained with 0.1% Giemsa (60x).
Ultrastructure of cultured granulosa cells grown in the presence of EGF or FGF Figure 6 shows electron micrographs of granulosa cells grown as monolayers in the absence of EGF or FGF (Fig. 6A) and in the presence of 100 ng/ml of EGF (Fig. 6B) or FGF (Fig. 6C). The cytoplasm of these cells is well differentiated and has ultrastructural features similar to those observed by Crisp and Channing (27) for monkey granulosa cells. Many free ribosomes are
found scattered throughout the cytoplasm, often in the form of rosettes, and both microfilaments and microtubules are evident. The mitochondria are often quite elongated. A few large, darkly-stained bodies probably lysosomes are found in the cytoplasm. The Golgi complexes are quite conspicuous and are sometimes found closely associated with the granular (or rough) endoplasmic reticulum, which is very prominent in many of the cells examined. Probably the most characteristic ultrastruc-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
1116
GOSPODAROWICZ, ILL AND BIRDWELL
Endo • 1977 Vol 100 • No 4
FIG. 6. Electron micrographs of granulosa cells grown in the absence of EGF or FGF (A and D), and in the presence of EGF (B) or FGF (C). Cells were grown in 1% serum with or withoug FGF or EGF (100 ng/ml). The presence of the growth factors has no effect on the infrastructure of these granulosa cells. Figure 6D shows a portion of the cytoplasm filled with granular endoplasmic reticulum consisting of very dilated cisternae (arrow), which was a feature common to all of the granulosa cells examined (i.e., in the presence or absence of EGF and FGF). gc-Golgi complex; m—mitochondria; mf— microfilaments; arrows—granular endoplasmic reticulum. 25,000x.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
FGF AND EGF EFFECTS ON GRANULOSA CELLS
1117
FIG. 7. Lipid content of granulosa cells grown in the presence or absence of EGF. (A) Cells grown in the presence of 1% serum: P—polysomes; N—nucleus; M — mitochondria; ER—endoplasmic reticulum. 17,000x. (B) Cells grown in the presence of 1% serum and 100 ng/ml of EGF: LI, lipid inclusions. 9,000x. Similar morphology was observed with cells grown in the presence of 1% serum and 100 ng/ml of FGF.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
1118
GOSPODAROWICZ, ILL AND BIRDWELL
tural feature of these cells is that the granular endoplasmic reticulum consisted of membrane-bound cisternae which were very dilated, sometimes as large as 4 fim in diameter. In some cells (not depicted here) the cytoplasm is almost completely packed with the dilated granular endoplasmic reticulum. This type of granular endoplasmic reticulum has been found in granulosa cells examined in vivo (28). The major difference between cells grown in the absence of mitogens and cells grown in their presence was in the lipid content of the cells. Cells grown without added mitogens and in low serum (1%) had very few lipid granules (Fig. 7A), while cells grown in the presence of mitogens and in low serum (1%) had a high content of lipid granules (Fig. 7B). Discussion Using granulosa cells maintained in tissue culture (1-3) others have shown that trophic hormones such as LH or FSH are involved in the maturation process leading to luteinization, but are not involved, at least as far as LH is concerned, in the control of granulosa cell proliferation (29). This also applies to the proliferation of luteal cells (the progeny of granulosa cells) since it is inhibited by LH in vitro while cell differentiation is enhanced (30). From such evidence one could conclude that granulosa cells are committed to differentiation and, when maintained in tissue culture under the right conditions (in the presence of gonadotropins), will stop dividing and become luteal cells. Another hypothesis is that growth factors exist for these cells which are similar to those which induce the proliferation of granulosa cells in vivo during the early differentiation of follicles when the cells are not yet under gonadotropin control. This possibility prompted us to look at the effect of growth factors such as EGF and FGF on granulosa cells. Such agents must take a prominent position regarding the control
Endo • 1977 Vo! 100 • No 4
of cellular proliferation since it became evident tht trophic hormones such as LH do not have mitogenic activity (29,30) in vitro and others, such as ACTH, are antimitotic (31,32). Our results indicate that EGF and FGF are by far the most potent mitogenic agents for granulosa cells yet found. The mitogenic effect of EGF on granulosa cells differs from its effect on human fibroblasts (33) and lens epithelial cells (34). The molar concentration of EGF that induces a halfmaximal response for the initiation of DNA synthesis in granulosa cells (1.5 x 10~ 13 M) is 500-fold lower than for human fibroblasts (7 x 10-nM (33)) and 5000-fold lower than for lens epithelium (7 x 10~ 10 M (34)). The dose of EGF producing a half-maximal response of granulosa cells for the initiation of DNA synthesis (1.5 x 10" 13 M) was lower than for cell proliferation (6 x 10" 1 3 M). This can be best explained by the observation of others that a lower number of binding sites needs to be occupied to promote initiation of DNA synthesis than to promote cell division (33). With human fibroblasts, for example, 20% occupancy of EGF binding sites is sufficient to effect a full biological response when the initiation of DNA synthesis is considered, but the cells will not divide unless a higher binding occupancy is attained (33). Besides requiring lower concentrations, the response of granulosa cells to EGF differed from the responses of human fibroblasts and lens epithelial cells in other ways. The responses of human fibroblasts and lens epithelium to EGF depend on the cell density, the concentration of serum in the medium, and the length of time the cells are left without renewal of the medium. Under ideal conditions EGF has as great a mitogenic effect as serum, but slight variations in the assay procedure can result in a diminution of the effect of EGF to a level lower than that of serum. In contrast, the response of sparse cultures (100 cells/cm2) of granulosa cells to EGF is not dependent on the length of time the cells are maintained
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
FGF AND EGF EFFECTS ON GRANULOSA CELLS without renewal of the medium; EGF is always as effective a mitogen as serum. For granulosa cells, as for human fibroblasts (9), the magnitude of the response of EGF depends on the serum concentration in the medium. However, while with fibroblasts maintained in low serum (0.35%) the response to EGF is small (20% that of optimal serum concentration) (9), with granulosa cells maintained in low serum (0.1%) the cells respond to EGF by proliferating actively. The only major difference observed is that the cell cycle will be longer in low serum than in high serum concentrations. The observation that primary cultures of granulosa cells, in contrast to other cell types, survived quite well in low serum can be best explained by their origin. In vivo granulosa cells are maintained in an avascular medium and this could account for their lesser dependence on serum than cells derived from richly vascularized organs. Thus, since granulosa cells can be induced to divide in a virtually serumless condition (60 (j.g of serum protein/ml) by very small amounts of mitogens, they should be an ideal cell type with which to study the mechanisms by which cells are induced to proliferate in vitro. The differences between the responses of granulosa cells and human fibroblasts (35) to FGF are generally the same as the differences between the responses of the two cell types of EGF. EGF is a more potent mitogen for granulosa cells than is FGF since lower molar concentrations of EGF induce a half-maximal response in the initiation of DNA synthesis assay (250-fold less than FGF) as well as in the increase in cell number assay (50-fold less than FGF). Because the slope of the FGF dose-response curve is steeper than that of the EGF curve, the differences in potency between the two mitogens are more marked at low concentrations than at levels near saturation. Although both EGF and FGF can stimulate the proliferation of granulosa cells as well as fibroblasts, this does not mean that
1119
the proliferation of all mesoderm-derived cells are under FGF and EGF control. EGF does not stimulate the proliferation of adrenal cortex cells (36) or that of vascular endothelial cells (14), and has only a marginal effect on vascular smooth muscle cells (14). In contrast, FGF is a strong mitogenic agent for all these cells. It remains to be seen whether or not FGF and EGF have similar effects on the proliferation of granulosa cells in vivo. Factors involved in follicular growth are poorly understood. It seems that during the early stage of ovarian development in rats the marked granulosa cell proliferation which takes place is independent of gonadotropin since it is not suppressed by daily administration of rat gonadotropin antibodies (37). Studies by Pedersen (38) and others (39) have indicated that the DNA labelling index of granulosa cells could be affected by gonadotropins but this depends on the age of the follicle. Gonadotropins are only active at a late stage of follicular development. Estrogens are also involved in the control of proliferation of granulosa cells and exert a direct effect on the growing follicles (40) since, in hypophysectomized animals, they increase the number of small and mediumsized follicles (40). Thus, the regulation of follicular development and, consequently, the control of granulosa cell proliferation seem to be extremely complex and may depend on the age of the individual as well as on the endocrine composition at a given time. However, since there is a general consensus that the factors regulating the transformation of primordial follicles into growing follicles do not involve the intervention of gonadotropin (41,42), this phase of follicular development should be an ideal period in which to study the possible in vivo effect of EGF or FGF on proliferation of granulosa cells. Acknowledgments We thank Drs. J. Moran and A. Mescher for helpful discussions.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
1120
GOSPODAROWICZ, ILL AND BIRDWELL
References 1. Channing, C. P., Recent Prog Horm Res 26: 589, 1970. 2. Channing, C. P., Endocrinology 87: 49, 1970. 3. Channing, C. P., In McKems, K. W. (ed.), The Gonads, Appleton-Century-Crofts, New York, 1966, p. 245. 4. Gospodarowicz, D., Nature 249: 123, 1974. 5. Cohen, S.J Biol Chem 237: 123, 1974. 6. Bynny, R. L., D. N. Orth, S. Cohen, and E. S. Doyne, Endocrinology 95: 776, 1974. 7. Cohen, S., and C. R. Savage, Recent Prog Horm Res 30: 551, 1974. 8. Cohen, S., and G. Carpenter, Proc Natl Acad Sci USA 72: 1317, 1975. 9. Cohen, S., G. Carpenter, and K. Lembach, Adv Metab Disord 8: 265, 1975. 10. Gregory, H., Nature 257: 325, 1975. 11. Gospodarowicz, D . J Biol Chem 250: 2515, 1975. 12. Gospodarowicz, D., J. Moran, and H. Bialecki, In Growth Hormone and Related Peptides, Excerpta Med, Int. Congr. 381, Elsevier, New York, 1976, p. 141. 13. Gospodarowicz, D., and J. Moran, Proc Natl Acad Sci USA 71: 4648, 1974. 14. Gospodarowicz, D., J. Moran, and D. Braun,/ Cell Physiol 1976 (In press). 15. Gospodarowicz, D., and A. Mescher,/ Cell Physiol 1976 (In press). 16. Jones, K. L., and J. Allison, Endocrinology 97: 359, 1975. 17. Ross, R., and J. A. Glomset, J Cell Biol 1976 (In press). 18. Gospodarowicz, D., J. Weseman, J. Moran, and J. LindstromJ Cell Biol 70: 395, 1976. 19. Westennark, B., and A. Wasteson, Adv Metab Disord 8: 85, 1975. 20. Gospodarowicz, D., and H. H. Handley, Endocrinology 97: 102, 1975. 21. Gospodarowicz, D., P. Rudland, J. Lindstrom, and K. Benirschke, Adv Metab Disord 8: 301, 1975. 22. Savage, C. R., and S. CohenJ Biol Chem 247: 7609, 1972.
Endo • 1977 Vol 100 • No 4
23. Papkoff, H., D. Gospodarowicz, A. Candiotti, and C. H. Li, Arch Biochem Biophys 111: 431, 1975. 24. Pierce, J. G., and M. E. Carsten, In Werner, S. C. (ed.), Thyrotropin, Charles C Thomas, Springfield, 111., 1963, p. 216. 25. Channing, C. P., and F. Ledwitz-Rigby, In Hardman, J. G., and B. W. O'Malley (eds.), Methods of Enzymology, vol. 39, Academic Press, New York, 1975, p. 183. 26. Gospodarowicz, D., and J. Moran, Exp Cell Res 90: 379, 1975. 27. Crisp, T. M., and C. P. Channing, Biol Reprod 7: 55, 1972. 28. Blanchette, E. J.J Cell Biol 31: 501, 1966. 29. McNatty, K. P., and R. S. Sowers, J Endocrinol 66: 391, 1975. 30. Gospodarowicz, D., and F. Gospodarowicz, Endocrinology 96: 458, 1975. 31. Masui, H., and L. D. Garren, Proc Natl Acad Sci USA 68: 3206, 1971. 32. Ramachandran, J., and A. T. Suyama, Proc Natl Acad Sci USA 72: 113, 1975. 33. Hollenberg, M. D., and P. Cuatrecasas, J Biol Chem 250: 3845, 1975. 34. Hollenberg, M. D., Arch Biochem Biophys 171: 371, 1975. 35. Gospodarowicz, D., and J. S. Moran,/ Cell Biol 6: 451, 1975. 36. Gospodarowicz, D., C. Ill, P. Hornsby, and G. Gill, Endocrinology 1976 100: 1080, 1977. 37. Eshkol, L., and B. Lunenfeld, In Saxena, B., C. J. Belring, and H. M. Gandy (eds.), Gonadotropins, Wiley, New York, 1972, p. 335. 38. Pedersen, T., Ada Endocrinol (Kbh) 62: 117,1969. 39. Ryle, M.J Reprod Fertil 19: 87, 1969. 40. Croes-Buth, S., F. J. A. Paesi, and S. E. de Jongh, Ada Endocrinol (Kbh) 32: 399, 1959. 41. Greenwald, G. S., In Geiger, S. R. (ed.), Handbook of Physiology, vol. IV, section 7, part 2, American Physiological Society, Washington, D.C., 1974, p. 293. 42. Blandau, R. J., In Greep, R. O., and L. Weiss, Histology, ed. 3, Academic Press, New York, 1973, p. 768.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 October 2014. at 23:55 For personal use only. No other uses without permission. . All rights reserved.
