ANALYTICAL
BIOCHEMISTRY
202,71-75
(1992)
Characterization of Human Follicle-Stimulating Binding to Human Granulosa Cells by an lmmunoenzymological Method Franck Perrotin,* Jacques Lansac,?
Dominique Royere,j’p’ and Jean-Pierre Miih*
Martine
Roussie,*
Hormone
Yves Combarnous,$
*Laboratoire de Biochimie, INSERM U 316, CHU, 37044 Tours, France; 7 Unite’ ok Biologie de la Reproduction, Departement Gynkcologie-Obste’trique, CHU Bretonneau, Fact&e’ de M&deck, 2 Boulevard TonnelIk, 37044 Tours Cedex, France; and SINRA, URA CNRSHNRA, Station de Physiologie de la Reproduction, Nouzilly, 37 380 Monnaie, France
Received
July
3, 1991
An original, nonradiometric method has been developed for studying the binding parameters of native follicle-stimulating hormone (FSH) to its specific receptors in human ovarian granulosa cells. After binding and washing of the cells, hFSH was desorbed from its receptors and quantitatively measured by a specific enzyme immunoassay (EIA) in which nonspecific binding was estimated in the presence of an excess of equine chorionic gonadotropin (eCG/PMSG), which binds to human FSH receptors but does not interfere in the hFSH EIA. This method makes use of native nonmodifled hFSH molecules (in contrast to radiometric methods) and permits direct estimation of the bindingparameters (I& and total number of sites). The Kd of hFSH for its human granulosa receptors measured by this technique (4.8 + 0.3 X 10-l”M) is close to that determined by other methods. However, we found a total number of specific FSH receptors per granulosa cell (1 to 6 X 104) higher than that reported by others by Scatchard analysis of competition dose-response curves in radioreceptor assays. The method is also sensitive enough to measure the in vivo occupancy of receptors by endogenous hFSH, which was found to be less than 6% in women undergoing hormonal treatment for in vitro fertilization. D 19S2 Academic Press, Inc.
hormones and their receptors (5,6). In particular equine follicle-stimulating hormone was reported to have a biological potency higher than that observed for rat FSH on rat Sertoli cells in culture as well as ovine FSH on ovine Sertoli cells in culture (6). Thus such data argued for the interest of evaluating binding/biologic activities under homologous conditions. However, only a few data concerning human folliclestimulating hormone (hFSH)2 binding on homospecific target cells have been established (7,8). Studies using in vitro and in vivo autoradiographic techniques suggest that rat, ovine, or human granulosa cells are able to specifically bind iodinated FSH, and that only granulosa cells possess such a specific high-affinity receptor within the ovary (9-11). In addition, 1251-labeled iodoFSH binding to these cells increases during the course of follicular growth (10,12-14). However, little is known about the quantitative binding of hFSH to human granulosa cells and the previous binding studies using radiolabeled hFSH are time-consuming and difficult to adapt to small quantities of cells (8,15). Consequently, the purpose of the present study was to set up a convenient and precise method for measuring hFSH binding to human granulosa cells obtained during the course of oocyte retrieval for in vitro fertilization. MATERIALS
Many reports have already focused on the possible discrepancy between immunological activity and bioactivity of glycoproteic hormones, particularly for human follicle-stimulating hormone (l-4), with particular attention to the charge of the glycoprotein and its carbohydrate moiety. Other reports have documented the interspecies differences in the interactions of glycoprotein 0003-2697/92 $3.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
Patient
AND selection
cells were obtained
METHODS and ovulation
induction.
from the follicular
Granulosa aspirates of 36
r To whom correspondence should be addressed. ’ Abbreviations used: hFSH, human follicle-stimulating hormone; eCG, equine chorionic gonadotropin; PBS, phosphate-buffered saline; PMSG, pregnant mare serum gonadotropin; RRA, radioreceptor assay; EIA, enzyme immunoassay; LH, luteinizing hormone. 71
72
PERROTIN
women undergoing in vitro fertilization attempts between November 89 and April 90. The patients’ ages ranged from 25 to 41 years (32.2 +- 4.2 years, mean + SD). Endometriosis was the sole exclusion criteria. Multiple follicular development was induced by combined therapy of gonadotropin-releasing hormone analog (0.1 mg Decapeptyl, Ipsen) and menotropins (Neopergonal, Serono). Oocyte retrieval was done within 34 or 36 h after hCG (10,000 IU Gonadotrophine Chorionique Endo, Organon) injection. Cells and cell-free membranes. After oocyte-cumulus-corona complex recovery, follicular aspirates from different women were pooled and centrifuged (15Og, 10 min). After supernatant removal, cell pellets were washed with phosphate-buffered saline (PBS, Gibco, France) and then resuspended in PBS. Granulosa cells were partly separated from erythrocytes by centrifugation over a Percoll gradient using a modification of a previously described method (16). The cell suspension was layered onto a discontinuous Percoll (Sigma Chimie France) gradient (20,30, and 40%, v/v, respectively) in Ham’s FlO medium (Gibco, France). After centrifugation (45Og, 7 min), granulosa cells were recovered from the intermediary layer. After washing in PBS, cells were adjusted to one million/ml, broken by freezing in liquid nitrogen, and stored at -20°C until assay. Binding studies. After rapid thawing at 37”C, cell membranes were separated from cytoplasmic debris by centrifugation (11,5OOg, 15 min). The pellets were washed with a 10 mM Tris buffer (pH 7.4) containing 1% bovine serum albumine (Sigma Chimie, France) and, after centrifugation, the supernatants were removed and the pellets were used for binding studies. Steadystate binding studies were conducted overnight at 4°C in a final volume of 0.2 ml of the above buffer, using membranes from 1 X lo5 to 2 X lo5 granulosa cells in suspension with the indicated concentrations of highly pituitary purified hFSH (UCB-Bioproducts, France). At the end of the incubation, unbound ligand was separated from bound ligand by centrifugation (11,5OOg, 15 min). The supernatants were assayed for the unbound hFSH fraction. The cell pellets were washed with the medium and hFSH bound to the cell surface was removed using a modification of a previously described method (17). Briefly, 0.2 ml of ice-cold 50 mM glytine * HCl/lOO mM NaCl, pH 3.0, was added and the cells were incubated for 3 min at 4°C. After neutralization by 2 N NaOH, the tubes were centrifuged (11,5OOg, 15 min) and the supernatants were assayed for the bound fraction of hFSH. hFSH assay. hFSH was measured by an immunoenzymometric assay (Seronodiagnostic, France). Two anti-hFSH monoclonal antibodies directed against distant epitopes and conjugated with fluorescein and alkaline phosphatase, respectively, were used. After in-
ET AL.
cubation with the samples, “magnetic” sheep antifluorescein was added to recover the complex and the quantitation was made by measuring the alkaline phosphatase activity associated to the precipitate, by adding a specific substrate for the enzyme and quantifying its reaction product. Enzymatic reaction was stopped by addition of a chelating agent and sodium hydroxyde. The absorbance measured at 492 or 550 nm was directly proportional to the level of hFSH. Determination of nonspecific binding. Nonsaturable (nonspecific) hFSH binding was determined in the presence of an excess (1000 IU/ml, i.e., 0.1 mg pure PMSG/ ml) of crude pregnant mare serum gonadotropin (PMSG 1119, UCB-Bioproducts, France) (18). Nonspecific binding was substracted from total binding to determine specific binding (19). DNA assay. In order to avoid some bias with regard to the cell number, DNA concentration was calculated using the Sarkosyl extraction method, then RNase action (Sigma R9005), and then spectrophotometric measurement at 260 and 280 nm. Sensitivity was 0.05 pg DNA for l/1000 O.D. variation. Calculations. Binding parameters were calculated with a nonlinear regression using a sigmoid curve (Graph PAD, ISI, Philadelphia, PA). The calculated parameters were least detectable dose, maximal response dose values, ED50, and slope of the dose-response curve. RESULTS
After the incubation of hFSH with granulosa cells, one can measure bound hFSH after washing of the cells and subsequent desorption of the hormone. Alternately, the percentage binding can be calculated from the difference between total initial hormone concentration and free hormone concentrations after incubation with the cells. The first method requires that desorption of hFSH from its receptors be accomplished without denaturation of the hormone that would prevent its measurement in the EIA. 1. Preliminary Studies Assay of hFSH. The method used in this study was previously performed in plasma and serum. We therefore assayed the absence of interference with the buffer used in the incubation medium. No significant difference between the concentrations of a solution of hFSH prepared with or without Tris buffer was found (data not shown). Acid release of surface-bound hFSH. A critical aspect of this study was the use of a 3-min exposure of the cells to ice-cold 50 mM glycine - HCl/lOO mM NaCl, pH 3.0, to remove surface-bound hFSH. Preliminary experiments
IMMUNOENZYMOLOGICAL
TABLE
CHARACTERIZATION
OF
FSH
73
BINDING
1
hFSH Concentrations (PM) Measured following an Incubation with 50 mM Glycine, 100 mM NaCl, pH 3, buffer (means f SD; n = 3) Control
hFSH 0 10 100 1,000 10,900
Acid
treated
0 8.3 92 979 9627.7
0 + 0.6 f 1.9 + 9.5 + 14.3
8.1 86.5 993.2 9699.6
+ 0.5 -+ 1.3 + 6.7 + 10.8
showed no alteration of hFSH as revealed in the hFSH EIA after acidic treatment (Table 1). Direct and indirect calculation of bound hFSH. In several preliminary experiments we compared the saturation of FSH receptors by hFSH by measuring either desorbed hormone or the difference between the initial and final concentrations of free hormone. The results (not shown) were in excellent agreement (r = 0,98). Determination of nonspecific binding. In order to determine the nonspecific binding of hFSH to human granulosa cells, it was necessary to use an hormone able to compete with it for its specific receptors but not in the hFSH EIA. eCG/PMSG fulfill these requirements since increasing amounts of this hormone inhibit the binding of hFSH to its receptors but do not show any cross reaction in the hFSH EIA. During preliminary experiments, we observed no significant difference between assayed concentrations of hFSH (0, 316, and 1000 PM) with or without increasing concentrations of eCG (from 0 to 2 FM). Figure 1 shows that 2 PM PMSG was sufficient to
1000
10000
hFSH1 (pM) FIG. 2. by an the be
Saturation of hFSH binding sites on human granulosa cells increasing concentrations of hFSH in the presence or absence of excess of PMSG (2 PM). The initial points of the curve measured in presence of PMSG allowed the nonspecific binding curve (NS) to drawn by extrapolation.
totally inhibit the specific binding of hFSH to its receptors at suboptimal conditions (560 PM). Figure 2 shows the dose-response curve of the saturation of granulosa cell receptors by hFSH in the presence or absence of eCG/PMSG. The saturation of specific binding sites is reached with 1 nM hFSH in the absence of eCG/PMSG. At concentrations higher than 1 nM hFSH, 2 PM PMSG is unsufficient to totally inhibit the specific binding of the human hormone. Therefore the parameters of hFSH nonspecific binding were determined with hFSH concentrations lower than 1 nM using 2 PM eCG/PMSG. 2. Effect of Endogeneous Hormones on hFSH Binding Parameters
izo-
+ loo"
20 -
+ +
+
0 20
200
2000
eCG
FIG. 1.
100
I PMSG
1 20000
(nM)
Deplacement of hFSH from its binding sites by increasing concentrations of eCG/PMSG. The concentration of hFSH was suboptimal (560 PM) with regard to the saturation data. Results are expressed in terms of percentage of total binding.
Figure 3 shows the saturation of hFSH-specific receptors when endogeneous hormones was not previously desorbed. When no hFSH is added in vitro to the membranes, bound hFSH was found to be low (2-6% of the available sites). In order to eliminate this low quantity of endogeneous hFSH, before measuring the saturation of receptors by added hFSH, we tested whether acidic treatment had an influence on the subsequent saturation parameters. However, when the acidic desorption treatment was performed before the in vitro incubation with hFSH, a decrease in both the affinity and the number of sites was observed (Table 2). The parameters calculated from these data were Kd = 4.8 + 0.3 X lo-” M (mean + SD; n = 4) and the estimated number of binding sites (B,) = 1 to 6 X lo* per cell (Table 2). DISCUSSION
In this paper, we describe a new specific and sensitive method for measuring hFSH binding to human granu-
PERROTIN
10
100 1000 [hFSHl (PM
10000
FIG. 3. hFSH binding to human granulosa cells. Specific binding was plotted as a sigmoid curve with a semilogarithmic scale. The parameters with 3 = 0.998 were as follows: A = bottom = 0.75418; B = top = 51.329; C = log (EC50) = 2.8806; D = hill slope = 2.3283. Each point was calculated as the mean (*SD) of four experiences (each in triplicate).
losa cell receptors. It allows us to follow hormone binding to its receptors in vitro and also to evaluate the occupancy of these receptors by endogeneous hFSH. The principle of our technique is to let hFSH bind to its receptors and, after washings to desorb it, to measure its concentration in a specific and sensitive EIA. Desorption of gonadotropins from their receptors implies treatment with acid buffer. Since the (Yand B subunits of gonadotropins dissociate at acid pH, it must be checked whether this treatment permits the desorption of hFSH from its receptors without promoting its dissociation. The data in Table 1 have clearly shown that desorption treatment does not lead to denaturation of hFSH. A second important aim of the present method was to measure correctly the nonspecific binding of hFSH to membranes or cells. This was achieved by using the equine placental gonadotropin eCG, which has been found to compete fully with hFSH for its receptors (l&20) without interfering in the hFSH EIA. It can be seen in Figs. 1 and 2 that hFSH exhibits an apparent affinity for human granulosa FSH receptors 3- to 500fold higher than that of eCG/PMSG. Such an apparent ratio might be deduced from the eCG/hFSH concentration ratio at the half-saturation of the binding sites. Equine CG has been shown to bind both LH and FSH receptors in many species but not in its own species, where it exhibits only LH activity. In the sheep, pig, rat, and camel, the affinity of eCG for the testicular FSH receptors is three to five times lower than that of the corresponding homologous FSHs (21-23). The advantage of this method is that there is no need for modified hFSH (radioiodinated or conjugated to an enzyme) and that it permits a direct assessment of the
ET AL.
occupation of receptors. In RRA, the Kd and number of sites can be measured either by saturation with radiolabeled hFSH or by competition between a fixed dose of ‘251-labeled hFSH and graded amounts of native hormone. The binding parameters are calculated by Scatchard plot but this approach has several limitations: first, it assumes that the radiolabeled FSH exhibits a behavior identical to that of native hormone. Second, the concentrations of free and bound molecules are indirectly calculated from the radioactivities associated with the total and bound molecules, respectively. It has been consistently observed that iodinated gonadotropins are only partly (20-50s) able to bind to their receptors in excess, indicating that they must be partly denatured. Moreover, the calculation of [B] and [B] l[F] implies a number of assumptions that are not generally demonstrated. Even if the proportion of iodinated hormone that is not able to bind is taken in account in the calculation of [B], the calculation of [F] is very imprecise since [B] is generally much lower than [T] and the errors on [F] and even more on [B]I[F] will be important. However, the direct measurement of the saturation of receptors by the native hormone overcomes all these limitations. By this approach, we obtained a Kd value for the binding of hFSH to human granulosa cells (4.8 X 10-l’ M) that was consistent with those obtained in this and other species by RRA (rat, 13,18,20,22,26, human, 15; bovine, 24 (female), 28 (male); porcine, 27). Only one study has been performed on human ovarian cells of the corpus luteum (4,ll). However, the methodological conditions were reported to make the discrimination between specific and nonspecific binding difficult. In particular the affinity constant was evaluated with pig granulosa homogenates. Thus our data to our knowledge are the first quantitative data for hFSH on human granulosa cells. In contrast, we found a number of binding sites (1 to 6 x lo4 per cell or 2 to 12 fmol/pg DNA) that is higher than those previously reported in other species (rat, 13, 18,20,22,26; human, 15; bovine, 24 (female), 28 (male); porcine, 27). Such a discrepancy raises two main questions about the role if any of the biological as well as methodological conditions in our assay. First, granulosa cells were obtained from preovulatory follicles after su-
TABLE Binding
& (M) B,. (sites/cell)
Characteristics Endogenous
2
of hFSH, before and after bFSH Desorption
Without desorption of endogenous hFSH
With desorption of endogenous hFSH
4.8 +- 0.3 X 10-l’ 10 to 60 x 1oa
10 f 2 x lo-i0 4to12Xl@
IMMUNOENZYMOLOGICAL
perovulation
treatment.
Such a stimulation
CHARACTERIZATION
regimen
3.
might, through an enhancedsteroidogenesis,increase the number of hFSH binding sites as reported in the rat (14,20). However, it seems rather unlikely that such an increase of estradiol concentration would explain the 10 times more binding sites observed in our study. Another major factor concerns our methodological approach. All the previously reported studies were based on the displacement of radioiodinated hormone with unlabeled hormone with an indirect assessment (Scatchard plots) of the binding characteristics. We assume that the use of native hFSH and the direct method of calculation make the present method more reliable than RRA competition curves analyzed by Scatchard plot. The origin of the granulosa cells also raises the question of the receptor occupancy by the in uiuo administrated gonadotropins. Previous studies have shown that hormones undergo rapid dissociation from their receptors at low pH. Therefore the removal of in uiuo bound fraction of hFSH was obtained by acidic treatment prior to binding study. With regard to the amount of hFSH released (2-6% of the total binding) we may not strictly assume the similarity between these in uiuo occupied receptors and those that we have further evaluated. For this purpose we analyzed the binding characteristics after such an acidic treatment. However, this was revealed to be somewhat deleterious with a decrease in both affinity constant and number of binding sites. In contrast, Ascoli demonstrated that an acidic treatment removed all the functional hFSH and had no deleterious effect on hFSH binding (17). Finally this report presents the first direct in vitro assessment of a specific receptor for human hFSH on the human granulosa cells after superovulation treatment. The pleomorphism of the gonadotropin preparations utilized for ovulation induction (29), as well as the dissimilarities observed using heterospecific models (30), argue for the development of a homospecific evaluation of the binding/biological activities ratio for gonadotropins. Such an evaluation is currently under study.
REFERENCES L. (1981)
106,17-27. Padmanabhan, V., Lang, L. L., Sonstein, J., Kelch, R. P., and Beitins, I. Z. (1988) J. Clin. Endocrinol. Metab. 67, 465-471. 5. Dahl, K. D., Papkoff, H., and Hsueh, A. J. W. (1989) Gen. Comp. Enakrinol. 73,368-373. 4.
6. Monet-Kuntz, (1991) Acta I. Berman, (1985)
C., Guillou,
Endocrinol.
M.
F., Fontaine,
I., Anand-Srisvatava,
Mol. Cell. Endocrinol.
8. Bramley, T. A., Stirling, Baird, D. T. (1987) J.
Y.
126,86-92. M.
B., and
Sairam,
M.
R.
42, 49-57.
D., Swanston,
I. A., Menzies,
G. S., and
Endocrinol. 113,305-315. A. R. (1973) Adv. Exp. Med. Biol. 36,365-378. D., and de Reviers, M. M. (1989) J. Reprod. Fertil. 85,
9. Midgley, 10. Monniaux, 151-162. 11. Shima,
I., and Combamous,
(Copenhagen)
K., Kitayama,
S., and Nakano,
R. (1987)
Obstet. Gynecol.
69,800-806. 12. Richards, J. S., Ireland, J. J., Rao, M. C., Bernath, G. A., Midgley, A. R., and Reichert, L. E. (1976) Endocrinology 99, 1562-1570. 13. Louvet, J. P., andvaitukaitis, J. L. (1976) Endocrinology99,758764. 14. Ireland, 883.
J. J., and Richards,
15. Bramley, McNeilly, 327. 16. Golos,
T. A., Stirling, A. S., and Baird, T. G., Soto,
J. S. (1978)
Endocrinology
D., Swanston, D. T. (1987)
J. Endocrinol.
E. A., Tureck,
102,876-
I. A., Menzies,
R. W., and Strauss,
G. S.,
113,317J. F. (1985)
J. Clin. Endocrinol. Metab. 61, 633-638. 17. Ascoli, M. (1982) J. Biol. Chem. 257, 13,306-13,311. 18. Knecht,
M., Ranta,
T., Katz,
M. S., and Catt,
K. J. (1983)
Endocri-
nology 112,1247-1255. 19. Garnier, J. (1980) J. Physiol. 76, 189-197. 20. Knecht, M., K. J. (1984)
Darbon,
J. M.,
Ranta,
T., Baukal,
Endocrinology 115,41-49. Y. (1989) in Structure-Functions
21. Combarnous, Gonadotropins, pp. 81-93, Raven Press, 22. Guillou, F., and Combarnous, Y. (1983) 755,229-236. 23. Combarnous, Y., Guillou, F., and Martinat, ogy 115,1821-1827.
New
A. J., and Relationships York.
B&him.
G. F., and Ryan,
of
Biophys. Acta
N. (1984) Endocrinol-
24. Inskeep, E. K., Braden, T. D., Lewis, P. E., Garcia-Winder, and Niswender, P. E. (1988) Biol. Reprod. 38.587-591. 25. Ford, K. A., and LaBarbera, A. R. (1988) Endocrinology 2374-2381. A., Erickson,
Catt,
M.,
123,
K. J. (1976)
Endocrinology
R. (1989)
Endocrinology
98,56-64.
This study was supported by Grant 1032 Rl from University of Tours and a grant from Serono Laboratories (Paris, France). FSH immunoenzymoassays kits were kindly donated by Serono Laboratories (Paris, France).
1. Wide,
Blum, W. F., Riegelbauer, G., and Gupta, D. (1985) J. Endocrinol.
26. Nimrod,
ACKNOWLEDGMENTS
75
OF FSH BINDING
Metub. 56, 682-688. B., and DicfaIuzy, E. (1982) J. Endocrinol.
J. Clin.
2. Zaidi, A. A., Froysa, 92,195-204.
Enuixrinol.
27. Nishimori,
K., Yamoto,
M., and Nakano,
124,2659-2665. 28. Branca, A. A., Sluss, P. M., Smith, R. A., and Reichert, L. E. (1985) J. Biol. Chem. 260,9988-9993. 29. Ulloa-Aguirre, A., Espinoza, R., Damian-Matsumara, P., and Chappel, S. C. (1988) Hum. Reprod. 3,491-501. 30. Hsueh, A. J. H., Bicsak, T. A., Jia, X., Dahl, K. D., Fauser, B. C. J. M., Galway, A. B., Czekala, N., Pavlou, S. N., Papkoff, H., Keene, J., and Boime, I. (1989) Recent Prog. Horm. Res. 46,209277.
BIOCHEMISTRY
202,71-75
(1992)
Characterization of Human Follicle-Stimulating Binding to Human Granulosa Cells by an lmmunoenzymological Method Franck Perrotin,* Jacques Lansac,?
Dominique Royere,j’p’ and Jean-Pierre Miih*
Martine
Roussie,*
Hormone
Yves Combarnous,$
*Laboratoire de Biochimie, INSERM U 316, CHU, 37044 Tours, France; 7 Unite’ ok Biologie de la Reproduction, Departement Gynkcologie-Obste’trique, CHU Bretonneau, Fact&e’ de M&deck, 2 Boulevard TonnelIk, 37044 Tours Cedex, France; and SINRA, URA CNRSHNRA, Station de Physiologie de la Reproduction, Nouzilly, 37 380 Monnaie, France
Received
July
3, 1991
An original, nonradiometric method has been developed for studying the binding parameters of native follicle-stimulating hormone (FSH) to its specific receptors in human ovarian granulosa cells. After binding and washing of the cells, hFSH was desorbed from its receptors and quantitatively measured by a specific enzyme immunoassay (EIA) in which nonspecific binding was estimated in the presence of an excess of equine chorionic gonadotropin (eCG/PMSG), which binds to human FSH receptors but does not interfere in the hFSH EIA. This method makes use of native nonmodifled hFSH molecules (in contrast to radiometric methods) and permits direct estimation of the bindingparameters (I& and total number of sites). The Kd of hFSH for its human granulosa receptors measured by this technique (4.8 + 0.3 X 10-l”M) is close to that determined by other methods. However, we found a total number of specific FSH receptors per granulosa cell (1 to 6 X 104) higher than that reported by others by Scatchard analysis of competition dose-response curves in radioreceptor assays. The method is also sensitive enough to measure the in vivo occupancy of receptors by endogenous hFSH, which was found to be less than 6% in women undergoing hormonal treatment for in vitro fertilization. D 19S2 Academic Press, Inc.
hormones and their receptors (5,6). In particular equine follicle-stimulating hormone was reported to have a biological potency higher than that observed for rat FSH on rat Sertoli cells in culture as well as ovine FSH on ovine Sertoli cells in culture (6). Thus such data argued for the interest of evaluating binding/biologic activities under homologous conditions. However, only a few data concerning human folliclestimulating hormone (hFSH)2 binding on homospecific target cells have been established (7,8). Studies using in vitro and in vivo autoradiographic techniques suggest that rat, ovine, or human granulosa cells are able to specifically bind iodinated FSH, and that only granulosa cells possess such a specific high-affinity receptor within the ovary (9-11). In addition, 1251-labeled iodoFSH binding to these cells increases during the course of follicular growth (10,12-14). However, little is known about the quantitative binding of hFSH to human granulosa cells and the previous binding studies using radiolabeled hFSH are time-consuming and difficult to adapt to small quantities of cells (8,15). Consequently, the purpose of the present study was to set up a convenient and precise method for measuring hFSH binding to human granulosa cells obtained during the course of oocyte retrieval for in vitro fertilization. MATERIALS
Many reports have already focused on the possible discrepancy between immunological activity and bioactivity of glycoproteic hormones, particularly for human follicle-stimulating hormone (l-4), with particular attention to the charge of the glycoprotein and its carbohydrate moiety. Other reports have documented the interspecies differences in the interactions of glycoprotein 0003-2697/92 $3.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
Patient
AND selection
cells were obtained
METHODS and ovulation
induction.
from the follicular
Granulosa aspirates of 36
r To whom correspondence should be addressed. ’ Abbreviations used: hFSH, human follicle-stimulating hormone; eCG, equine chorionic gonadotropin; PBS, phosphate-buffered saline; PMSG, pregnant mare serum gonadotropin; RRA, radioreceptor assay; EIA, enzyme immunoassay; LH, luteinizing hormone. 71
72
PERROTIN
women undergoing in vitro fertilization attempts between November 89 and April 90. The patients’ ages ranged from 25 to 41 years (32.2 +- 4.2 years, mean + SD). Endometriosis was the sole exclusion criteria. Multiple follicular development was induced by combined therapy of gonadotropin-releasing hormone analog (0.1 mg Decapeptyl, Ipsen) and menotropins (Neopergonal, Serono). Oocyte retrieval was done within 34 or 36 h after hCG (10,000 IU Gonadotrophine Chorionique Endo, Organon) injection. Cells and cell-free membranes. After oocyte-cumulus-corona complex recovery, follicular aspirates from different women were pooled and centrifuged (15Og, 10 min). After supernatant removal, cell pellets were washed with phosphate-buffered saline (PBS, Gibco, France) and then resuspended in PBS. Granulosa cells were partly separated from erythrocytes by centrifugation over a Percoll gradient using a modification of a previously described method (16). The cell suspension was layered onto a discontinuous Percoll (Sigma Chimie France) gradient (20,30, and 40%, v/v, respectively) in Ham’s FlO medium (Gibco, France). After centrifugation (45Og, 7 min), granulosa cells were recovered from the intermediary layer. After washing in PBS, cells were adjusted to one million/ml, broken by freezing in liquid nitrogen, and stored at -20°C until assay. Binding studies. After rapid thawing at 37”C, cell membranes were separated from cytoplasmic debris by centrifugation (11,5OOg, 15 min). The pellets were washed with a 10 mM Tris buffer (pH 7.4) containing 1% bovine serum albumine (Sigma Chimie, France) and, after centrifugation, the supernatants were removed and the pellets were used for binding studies. Steadystate binding studies were conducted overnight at 4°C in a final volume of 0.2 ml of the above buffer, using membranes from 1 X lo5 to 2 X lo5 granulosa cells in suspension with the indicated concentrations of highly pituitary purified hFSH (UCB-Bioproducts, France). At the end of the incubation, unbound ligand was separated from bound ligand by centrifugation (11,5OOg, 15 min). The supernatants were assayed for the unbound hFSH fraction. The cell pellets were washed with the medium and hFSH bound to the cell surface was removed using a modification of a previously described method (17). Briefly, 0.2 ml of ice-cold 50 mM glytine * HCl/lOO mM NaCl, pH 3.0, was added and the cells were incubated for 3 min at 4°C. After neutralization by 2 N NaOH, the tubes were centrifuged (11,5OOg, 15 min) and the supernatants were assayed for the bound fraction of hFSH. hFSH assay. hFSH was measured by an immunoenzymometric assay (Seronodiagnostic, France). Two anti-hFSH monoclonal antibodies directed against distant epitopes and conjugated with fluorescein and alkaline phosphatase, respectively, were used. After in-
ET AL.
cubation with the samples, “magnetic” sheep antifluorescein was added to recover the complex and the quantitation was made by measuring the alkaline phosphatase activity associated to the precipitate, by adding a specific substrate for the enzyme and quantifying its reaction product. Enzymatic reaction was stopped by addition of a chelating agent and sodium hydroxyde. The absorbance measured at 492 or 550 nm was directly proportional to the level of hFSH. Determination of nonspecific binding. Nonsaturable (nonspecific) hFSH binding was determined in the presence of an excess (1000 IU/ml, i.e., 0.1 mg pure PMSG/ ml) of crude pregnant mare serum gonadotropin (PMSG 1119, UCB-Bioproducts, France) (18). Nonspecific binding was substracted from total binding to determine specific binding (19). DNA assay. In order to avoid some bias with regard to the cell number, DNA concentration was calculated using the Sarkosyl extraction method, then RNase action (Sigma R9005), and then spectrophotometric measurement at 260 and 280 nm. Sensitivity was 0.05 pg DNA for l/1000 O.D. variation. Calculations. Binding parameters were calculated with a nonlinear regression using a sigmoid curve (Graph PAD, ISI, Philadelphia, PA). The calculated parameters were least detectable dose, maximal response dose values, ED50, and slope of the dose-response curve. RESULTS
After the incubation of hFSH with granulosa cells, one can measure bound hFSH after washing of the cells and subsequent desorption of the hormone. Alternately, the percentage binding can be calculated from the difference between total initial hormone concentration and free hormone concentrations after incubation with the cells. The first method requires that desorption of hFSH from its receptors be accomplished without denaturation of the hormone that would prevent its measurement in the EIA. 1. Preliminary Studies Assay of hFSH. The method used in this study was previously performed in plasma and serum. We therefore assayed the absence of interference with the buffer used in the incubation medium. No significant difference between the concentrations of a solution of hFSH prepared with or without Tris buffer was found (data not shown). Acid release of surface-bound hFSH. A critical aspect of this study was the use of a 3-min exposure of the cells to ice-cold 50 mM glycine - HCl/lOO mM NaCl, pH 3.0, to remove surface-bound hFSH. Preliminary experiments
IMMUNOENZYMOLOGICAL
TABLE
CHARACTERIZATION
OF
FSH
73
BINDING
1
hFSH Concentrations (PM) Measured following an Incubation with 50 mM Glycine, 100 mM NaCl, pH 3, buffer (means f SD; n = 3) Control
hFSH 0 10 100 1,000 10,900
Acid
treated
0 8.3 92 979 9627.7
0 + 0.6 f 1.9 + 9.5 + 14.3
8.1 86.5 993.2 9699.6
+ 0.5 -+ 1.3 + 6.7 + 10.8
showed no alteration of hFSH as revealed in the hFSH EIA after acidic treatment (Table 1). Direct and indirect calculation of bound hFSH. In several preliminary experiments we compared the saturation of FSH receptors by hFSH by measuring either desorbed hormone or the difference between the initial and final concentrations of free hormone. The results (not shown) were in excellent agreement (r = 0,98). Determination of nonspecific binding. In order to determine the nonspecific binding of hFSH to human granulosa cells, it was necessary to use an hormone able to compete with it for its specific receptors but not in the hFSH EIA. eCG/PMSG fulfill these requirements since increasing amounts of this hormone inhibit the binding of hFSH to its receptors but do not show any cross reaction in the hFSH EIA. During preliminary experiments, we observed no significant difference between assayed concentrations of hFSH (0, 316, and 1000 PM) with or without increasing concentrations of eCG (from 0 to 2 FM). Figure 1 shows that 2 PM PMSG was sufficient to
1000
10000
hFSH1 (pM) FIG. 2. by an the be
Saturation of hFSH binding sites on human granulosa cells increasing concentrations of hFSH in the presence or absence of excess of PMSG (2 PM). The initial points of the curve measured in presence of PMSG allowed the nonspecific binding curve (NS) to drawn by extrapolation.
totally inhibit the specific binding of hFSH to its receptors at suboptimal conditions (560 PM). Figure 2 shows the dose-response curve of the saturation of granulosa cell receptors by hFSH in the presence or absence of eCG/PMSG. The saturation of specific binding sites is reached with 1 nM hFSH in the absence of eCG/PMSG. At concentrations higher than 1 nM hFSH, 2 PM PMSG is unsufficient to totally inhibit the specific binding of the human hormone. Therefore the parameters of hFSH nonspecific binding were determined with hFSH concentrations lower than 1 nM using 2 PM eCG/PMSG. 2. Effect of Endogeneous Hormones on hFSH Binding Parameters
izo-
+ loo"
20 -
+ +
+
0 20
200
2000
eCG
FIG. 1.
100
I PMSG
1 20000
(nM)
Deplacement of hFSH from its binding sites by increasing concentrations of eCG/PMSG. The concentration of hFSH was suboptimal (560 PM) with regard to the saturation data. Results are expressed in terms of percentage of total binding.
Figure 3 shows the saturation of hFSH-specific receptors when endogeneous hormones was not previously desorbed. When no hFSH is added in vitro to the membranes, bound hFSH was found to be low (2-6% of the available sites). In order to eliminate this low quantity of endogeneous hFSH, before measuring the saturation of receptors by added hFSH, we tested whether acidic treatment had an influence on the subsequent saturation parameters. However, when the acidic desorption treatment was performed before the in vitro incubation with hFSH, a decrease in both the affinity and the number of sites was observed (Table 2). The parameters calculated from these data were Kd = 4.8 + 0.3 X lo-” M (mean + SD; n = 4) and the estimated number of binding sites (B,) = 1 to 6 X lo* per cell (Table 2). DISCUSSION
In this paper, we describe a new specific and sensitive method for measuring hFSH binding to human granu-
PERROTIN
10
100 1000 [hFSHl (PM
10000
FIG. 3. hFSH binding to human granulosa cells. Specific binding was plotted as a sigmoid curve with a semilogarithmic scale. The parameters with 3 = 0.998 were as follows: A = bottom = 0.75418; B = top = 51.329; C = log (EC50) = 2.8806; D = hill slope = 2.3283. Each point was calculated as the mean (*SD) of four experiences (each in triplicate).
losa cell receptors. It allows us to follow hormone binding to its receptors in vitro and also to evaluate the occupancy of these receptors by endogeneous hFSH. The principle of our technique is to let hFSH bind to its receptors and, after washings to desorb it, to measure its concentration in a specific and sensitive EIA. Desorption of gonadotropins from their receptors implies treatment with acid buffer. Since the (Yand B subunits of gonadotropins dissociate at acid pH, it must be checked whether this treatment permits the desorption of hFSH from its receptors without promoting its dissociation. The data in Table 1 have clearly shown that desorption treatment does not lead to denaturation of hFSH. A second important aim of the present method was to measure correctly the nonspecific binding of hFSH to membranes or cells. This was achieved by using the equine placental gonadotropin eCG, which has been found to compete fully with hFSH for its receptors (l&20) without interfering in the hFSH EIA. It can be seen in Figs. 1 and 2 that hFSH exhibits an apparent affinity for human granulosa FSH receptors 3- to 500fold higher than that of eCG/PMSG. Such an apparent ratio might be deduced from the eCG/hFSH concentration ratio at the half-saturation of the binding sites. Equine CG has been shown to bind both LH and FSH receptors in many species but not in its own species, where it exhibits only LH activity. In the sheep, pig, rat, and camel, the affinity of eCG for the testicular FSH receptors is three to five times lower than that of the corresponding homologous FSHs (21-23). The advantage of this method is that there is no need for modified hFSH (radioiodinated or conjugated to an enzyme) and that it permits a direct assessment of the
ET AL.
occupation of receptors. In RRA, the Kd and number of sites can be measured either by saturation with radiolabeled hFSH or by competition between a fixed dose of ‘251-labeled hFSH and graded amounts of native hormone. The binding parameters are calculated by Scatchard plot but this approach has several limitations: first, it assumes that the radiolabeled FSH exhibits a behavior identical to that of native hormone. Second, the concentrations of free and bound molecules are indirectly calculated from the radioactivities associated with the total and bound molecules, respectively. It has been consistently observed that iodinated gonadotropins are only partly (20-50s) able to bind to their receptors in excess, indicating that they must be partly denatured. Moreover, the calculation of [B] and [B] l[F] implies a number of assumptions that are not generally demonstrated. Even if the proportion of iodinated hormone that is not able to bind is taken in account in the calculation of [B], the calculation of [F] is very imprecise since [B] is generally much lower than [T] and the errors on [F] and even more on [B]I[F] will be important. However, the direct measurement of the saturation of receptors by the native hormone overcomes all these limitations. By this approach, we obtained a Kd value for the binding of hFSH to human granulosa cells (4.8 X 10-l’ M) that was consistent with those obtained in this and other species by RRA (rat, 13,18,20,22,26, human, 15; bovine, 24 (female), 28 (male); porcine, 27). Only one study has been performed on human ovarian cells of the corpus luteum (4,ll). However, the methodological conditions were reported to make the discrimination between specific and nonspecific binding difficult. In particular the affinity constant was evaluated with pig granulosa homogenates. Thus our data to our knowledge are the first quantitative data for hFSH on human granulosa cells. In contrast, we found a number of binding sites (1 to 6 x lo4 per cell or 2 to 12 fmol/pg DNA) that is higher than those previously reported in other species (rat, 13, 18,20,22,26; human, 15; bovine, 24 (female), 28 (male); porcine, 27). Such a discrepancy raises two main questions about the role if any of the biological as well as methodological conditions in our assay. First, granulosa cells were obtained from preovulatory follicles after su-
TABLE Binding
& (M) B,. (sites/cell)
Characteristics Endogenous
2
of hFSH, before and after bFSH Desorption
Without desorption of endogenous hFSH
With desorption of endogenous hFSH
4.8 +- 0.3 X 10-l’ 10 to 60 x 1oa
10 f 2 x lo-i0 4to12Xl@
IMMUNOENZYMOLOGICAL
perovulation
treatment.
Such a stimulation
CHARACTERIZATION
regimen
3.
might, through an enhancedsteroidogenesis,increase the number of hFSH binding sites as reported in the rat (14,20). However, it seems rather unlikely that such an increase of estradiol concentration would explain the 10 times more binding sites observed in our study. Another major factor concerns our methodological approach. All the previously reported studies were based on the displacement of radioiodinated hormone with unlabeled hormone with an indirect assessment (Scatchard plots) of the binding characteristics. We assume that the use of native hFSH and the direct method of calculation make the present method more reliable than RRA competition curves analyzed by Scatchard plot. The origin of the granulosa cells also raises the question of the receptor occupancy by the in uiuo administrated gonadotropins. Previous studies have shown that hormones undergo rapid dissociation from their receptors at low pH. Therefore the removal of in uiuo bound fraction of hFSH was obtained by acidic treatment prior to binding study. With regard to the amount of hFSH released (2-6% of the total binding) we may not strictly assume the similarity between these in uiuo occupied receptors and those that we have further evaluated. For this purpose we analyzed the binding characteristics after such an acidic treatment. However, this was revealed to be somewhat deleterious with a decrease in both affinity constant and number of binding sites. In contrast, Ascoli demonstrated that an acidic treatment removed all the functional hFSH and had no deleterious effect on hFSH binding (17). Finally this report presents the first direct in vitro assessment of a specific receptor for human hFSH on the human granulosa cells after superovulation treatment. The pleomorphism of the gonadotropin preparations utilized for ovulation induction (29), as well as the dissimilarities observed using heterospecific models (30), argue for the development of a homospecific evaluation of the binding/biological activities ratio for gonadotropins. Such an evaluation is currently under study.
REFERENCES L. (1981)
106,17-27. Padmanabhan, V., Lang, L. L., Sonstein, J., Kelch, R. P., and Beitins, I. Z. (1988) J. Clin. Endocrinol. Metab. 67, 465-471. 5. Dahl, K. D., Papkoff, H., and Hsueh, A. J. W. (1989) Gen. Comp. Enakrinol. 73,368-373. 4.
6. Monet-Kuntz, (1991) Acta I. Berman, (1985)
C., Guillou,
Endocrinol.
M.
F., Fontaine,
I., Anand-Srisvatava,
Mol. Cell. Endocrinol.
8. Bramley, T. A., Stirling, Baird, D. T. (1987) J.
Y.
126,86-92. M.
B., and
Sairam,
M.
R.
42, 49-57.
D., Swanston,
I. A., Menzies,
G. S., and
Endocrinol. 113,305-315. A. R. (1973) Adv. Exp. Med. Biol. 36,365-378. D., and de Reviers, M. M. (1989) J. Reprod. Fertil. 85,
9. Midgley, 10. Monniaux, 151-162. 11. Shima,
I., and Combamous,
(Copenhagen)
K., Kitayama,
S., and Nakano,
R. (1987)
Obstet. Gynecol.
69,800-806. 12. Richards, J. S., Ireland, J. J., Rao, M. C., Bernath, G. A., Midgley, A. R., and Reichert, L. E. (1976) Endocrinology 99, 1562-1570. 13. Louvet, J. P., andvaitukaitis, J. L. (1976) Endocrinology99,758764. 14. Ireland, 883.
J. J., and Richards,
15. Bramley, McNeilly, 327. 16. Golos,
T. A., Stirling, A. S., and Baird, T. G., Soto,
J. S. (1978)
Endocrinology
D., Swanston, D. T. (1987)
J. Endocrinol.
E. A., Tureck,
102,876-
I. A., Menzies,
R. W., and Strauss,
G. S.,
113,317J. F. (1985)
J. Clin. Endocrinol. Metab. 61, 633-638. 17. Ascoli, M. (1982) J. Biol. Chem. 257, 13,306-13,311. 18. Knecht,
M., Ranta,
T., Katz,
M. S., and Catt,
K. J. (1983)
Endocri-
nology 112,1247-1255. 19. Garnier, J. (1980) J. Physiol. 76, 189-197. 20. Knecht, M., K. J. (1984)
Darbon,
J. M.,
Ranta,
T., Baukal,
Endocrinology 115,41-49. Y. (1989) in Structure-Functions
21. Combarnous, Gonadotropins, pp. 81-93, Raven Press, 22. Guillou, F., and Combarnous, Y. (1983) 755,229-236. 23. Combarnous, Y., Guillou, F., and Martinat, ogy 115,1821-1827.
New
A. J., and Relationships York.
B&him.
G. F., and Ryan,
of
Biophys. Acta
N. (1984) Endocrinol-
24. Inskeep, E. K., Braden, T. D., Lewis, P. E., Garcia-Winder, and Niswender, P. E. (1988) Biol. Reprod. 38.587-591. 25. Ford, K. A., and LaBarbera, A. R. (1988) Endocrinology 2374-2381. A., Erickson,
Catt,
M.,
123,
K. J. (1976)
Endocrinology
R. (1989)
Endocrinology
98,56-64.
This study was supported by Grant 1032 Rl from University of Tours and a grant from Serono Laboratories (Paris, France). FSH immunoenzymoassays kits were kindly donated by Serono Laboratories (Paris, France).
1. Wide,
Blum, W. F., Riegelbauer, G., and Gupta, D. (1985) J. Endocrinol.
26. Nimrod,
ACKNOWLEDGMENTS
75
OF FSH BINDING
Metub. 56, 682-688. B., and DicfaIuzy, E. (1982) J. Endocrinol.
J. Clin.
2. Zaidi, A. A., Froysa, 92,195-204.
Enuixrinol.
27. Nishimori,
K., Yamoto,
M., and Nakano,
124,2659-2665. 28. Branca, A. A., Sluss, P. M., Smith, R. A., and Reichert, L. E. (1985) J. Biol. Chem. 260,9988-9993. 29. Ulloa-Aguirre, A., Espinoza, R., Damian-Matsumara, P., and Chappel, S. C. (1988) Hum. Reprod. 3,491-501. 30. Hsueh, A. J. H., Bicsak, T. A., Jia, X., Dahl, K. D., Fauser, B. C. J. M., Galway, A. B., Czekala, N., Pavlou, S. N., Papkoff, H., Keene, J., and Boime, I. (1989) Recent Prog. Horm. Res. 46,209277.
