Proc. Nat. Acad. Sci USA
Vol. 73, No. 3, pp. 847-851, March 1976 Cell Biology
Sex differences in binding of human growth hormone to isolated rat hepatocytes (hormone receptors/prolactin/estrogen effect)
MICHAEL B. RANKE, CHARLES A. STANLEY, DAVID RODBARD*, LESTER BAKER, ALFRED BONGIOVANNI, AND JOHN S. PARKS Department of Endocrinology, The Children's Hospital of Philadelphia, Philadelphia, Pennsylvania 19104; and * Reproduction Research Branch, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland 20014
Communicated by George B. Koelle, December 19,1975
ABSTRACT Since liver is a target for growth hormone action, binding of 125I-labeled human growth hormone to enzymatically isolated rat hepatocytes was studied. Specific binding was shown with hepatocytes from both male and female animals. There was a single class of receptors for human growth hormone on cells from males (affinity constant, Ka = 1.16 X 109 liters/mole; sites per cell, q = 6200). In males, bovine growth hormone was almost as potent as human growth hormone in displacing bound 123I-labeled human growth hormone, while ovine prolactin was about 1000 times less potent. Cells from female rats bound more 25I-labeled human growth hormone than cells from males. The cells from females contained at least two classes of receptors for human growth hormone. The receptor of highest affinity had the same affinity for human growth hormone as the single receptor found in males (Ka = 0.96 X 109 liters mole). However, there were three to four times as many of these receptors per cell in females (q = 21,000). In females, bovine growth hormone and ovine prolactin were both about 20 times less potent than human growth hormone. Treatment of male rats with estrone produced cells that show the same binding characteristics as females. These results indicate that human growth hormone binds to a somatogenic receptor in hepatocytes from male rats. In females and estrogen-treated males, the receptors that bind human growth hormone recognize lactogenic as well as somatogenic properties. This suggests that the lactogenic and growth-promoting effects of human growth hormone in the rat are mediated by different receptors.
fects of estrone treatment in male animals on binding of 125I-labeled hGH were studied, leading to models for growth hormone receptors in the rat liver. MATERIALS AND METHODS Animals. Sprague-Dawley rats, obtained from Charles River, Inc. and weighing 150-250 g, were starved overnight prior to preparation of isolated hepatocytes. Estrone-treated male rats were given 50 .ug of estrone diluted in corn oil (8) by subcutaneous injection for 10 days. Hepatocytes. Suspensions of isolated hepatocytes were prepared by enzymatic digestion of perfused liver with collagenase (Worthington, crude bacterial collagenase, Type I) using the method of Berry and Friend (9) modified by Krebs et al. (10). The yield of cells from one 200-g rat was about 20 ml of 10-15% (vol/vol) suspension or about 200 X 106 cells. The suspension consisted largely of single isolated cells with a few clumps of 2 to 10 cells. Viability, as judged by exclusion of 0.4% Trypan blue, was 90-95%. For radioreceptor studies the cells were centrifuged and resuspended in TrisHC1 buffer (pH 7.35), consisting of 25 mM Tris-HCl, 120 mM NaCl, 5 mM KCl, 1.2 mM MgSO4, 1 mM EDTA, 10 mM dextrose, 15 mM Na-acetate, and 0.1% bovine serum albumin. Hormones. Human growth hormone (NIH HS 1523D) was used for iodination. A preparation of hGH from Kabi/ AS/Stockholm (lot no. 1620-58383) was used for unlabeled hormone. This preparation was shown to be equal in potency to the NIH preparation by radioimmunoassay as well as in a radioreceptor assay using cultured human lymphoid cells (11). Other hormones used were: bovine growth hormone (bGH) (NIH GH-B ; 0.92 IU/mg; lactogenic activity not specified); oPRL (NIH PS 10; 25 IU/mg; growth hormone activity less than 0.01 IU/mg); porcine insulin (Lilly, lot no. 615-108253-1081); porcine glucagon (Lilly, lot no. 2580234B-167-1); and bovine parathormone (Dr. Howard Rasmussen, PC 72 53-65). All dilutions were made in TrisHCI buffer. lodination of hGH. 125I-Labeled hGH was prepared by a modification of the chloramine T method of Hunter and Greenwood (12). The reaction mixture consisted of 1 mCi of carrier-free Na'25I (Nuclear Chicago Corp.), 5 ,g of hGH, 0.01 ml of 1.0 M Na2PO4 - buffer, pH 7.4, and 4 ,g of chloramine T in a total volume of 0.034 ml. The reaction was stopped after 15 sec by addition of 8 ug of Na2S2O5 and 0.2 ml of 5% bovine serum albumin solution. Labeled hGH was separated from free 125I by gel filtration on Sephadex G-75 (column size: 0.5 X 12 cm) using Tris-HCI buffer. The specific activity of the 125I-labeled hGH used was 120 ,Ci/,ug. Further purification of the labeled hormone was achieved
The effects of growth hormone on skeletal growth appear to be mediated by somatomedins (1). When 125I-labeled human growth hormone (hGH) is injected in rats in vivo, much of the activity is localized in the liver (2). Somatomedin is produced when rat liver is perfused with growth hormone (3, 4). These findings suggest that liver is a target organ for growth hormone. The first step in the action of polypeptide hormones is binding to specific receptors on the surface of target cells. Recently, binding of 125I-labeled hGH to cell membranes prepared from livers of rabbits, rats, and other species has been reported (5-8). Binding to rat liver membranes could only be demonstrated using female or estrogen-treated male animals. In these membrane preparations, ovine prolactin (oPRL) was identical to hGH in displacing bound labeled hGH (8). It was therefore concluded that hGH binds to a lactogenic receptor in rat liver. The present study was carried out to examine the binding characteristics of hGH to rat liver using enzymatically isolated rat hepatocytes. Differences in males and females and efAbbreviations: hGH and bGH are, respectively, human and bovine growth hormone; oPRL is ovine prolactin; ED50, dose at which displacement is half-maximal. 847
Cell Biology: Ranke et al.
Proc. Nat. Acad. Sci. USA 73 (1976) 32 z
F 24 z
0 z 20 w 0 w
C 20 ui 0.
0 0' a
a-e a IT r 10
: a w m w
2 4 x
8 10 12 14
FIG. 1. Effects of incubation conditions on binding of 1251-labeled hGH to isolated rat hepatocytes. (A) Effect of time and temperature. The values depicted represent combined data from two experiments with hepatocytes from female animals. Hepatocytes (2.2 and 5.3 X 106/ml) were incubated with 125I-labeled hGH (0.7 ng/ml) with and without an excess of hGH (8 ,ug/ml). The ordinate represents the amount of specifically bound labeled hGH as a percentage of the maximal binding observed. (B) Effect of pH. Hepatocytes (2.2 X 106/ml) from a female animal were incubated with 125I-labeled hGH (0.7 ng/ml) with and without an excess of hGH (8 Ag/ml) for 120 min at 220. The ordinate represents the amount of specifically bound labeled hGH as a percentage of the maximal binding observed. (C) Effect of cell concentration. Hepatocytes (2.3 to 11.8 X 106/ml) of a male animal were incubated with 1251labeled hGH with and without an excess of hGH (8 ,g/ml) for 120 min at 220. The ordinate represents the amount of labeled hGH specifically bound as a percentage of the total labeled hGH in the system.
by affinity chromatography. Aliquots of labeled hormone (0.5 ,ug) were applied to a small column (0.3 X 2 cm) containing anti-hGH antibody (Dr. M. McGillivray: R 37-38 920-71) linked to Sepharose 4B. After 2 hr at room temperature the column was washed with 10 ml of Tris.HCI buffer and the 125I-labeled hGH was eluted with 6 M guanidineHGI, 0.2% bovine serum albumin, pH 3. Labeled hormone in 0.5 ml of this eluate was separated from salts by gel filtration on Sephadex G-75, as above. Aliquots of the labeled hormone purified weekly by this procedure allowed the use of material from one iodination for 6 weeks. Incubation Procedure. Incubations were carried out in polypropylene tubes (12 X 75 mm, Falcon no. 2063). The incubation mixture consisted of 0.5 ml of cell suspension, 0.05 ml of 125I-labeled hGH containing approximately 0.5 ng of hGH, and 0.05 ml of unlabeled hGH, all in Tris-HCI buffer. The mixture was shaken for 15 min every 30 min. At the end of the incubation period, 3 ml of cold Tris.HCI buffer were added and the cells were immediately sedimented by centrifugation (40) at 1500 X g for 4 min. After the supernatant was decanted, the tip of the tubes containing the cell pellet was cut with a razor blade; radioactivity bound to the pellet was measured in a gamma counter (Searle, model 1185). Nonspecific binding was defined as counts bound in the
10 20 40 100 200 400 hGH IN INCUBATION (ng/ml)
FIG. 2. Displacement of 125I-labeled hGH from isolated rat hepatocytes in a representative female (0), male (0), and estronetreated male (X) animal. Hepatocytes (5 X 106/ml in the female; 8.3 X 106/ml in the male; and 4.9 X 106/ml in the estrone-treated male) were incubated with 1251-labeled hGH (0.7 ng/ml) for 120 min at 220. The ordinate represents the radioactivity bound to the hepatocytes as a percentage of the total in the system.
presence of excess hGH (5 jig per tube). Dose response curves were established in each experiment using quadruplicate samples at each dose level of unlabeled hormone. The
values were corrected for nonspecific binding. Results were calculated using the "Radioimmunoassay Data Processing Package" of Rodbard and Faden (13).
RESULTS Optimization of assay conditions As shown in Fig. 1, binding of 125I-labeled hGH to isolated rat hepatocytes was time and temperature dependent. Maximum specific binding at 150, 22°, and 370 was reached by 120 min and remained stable for an additional 120 min. Specific binding was greatest at 220. Little variation in binding was noted between pH 7.1 and 7.6. Specific binding of labeled hGH was directly proportional to the cell concentration over a range of 2 to 12 X 106 cells per ml. To test degradation of labeled hormone during incubation with hepatocytes, identical amounts of 125I-labeled hGH were incubated with 5 X 106 hepatocytes and with equal volumes of buffer alone for 120 min at 220. Aliquots of the supernatants were used for binding to fresh hepatocytes and an excess of anti-hGH antibody. After preincubation with hepatocytes, binding to fresh cells was 96% of the control and binding to anti-hGH antibody was 98% of the control. Displacement of 125I-labeled hGH from hepatocytes by unlabeled hormones Hepatocytes from males, females, and estrone-treated males demonstrated specific binding of 1251-labeled hGH. As shown in Fig. 2, the initial binding was similar in females and estrone-treated males, but higher than in males. There was significant displacement of bound '25I-labeled hGH by 2.9 ng/ml (1.47 X 10-10 M) of unlabeled hGH. Porcine insulin, porcine glucagon, and bovine parathormone, at concentrations up to 1.7 jig/ml, failed to displace bound labeled hGH.
Cell Biology: Ranke et al.
Proc. Nat. Acad. Sci. USA 73 (1976)
102 lo, 104 103 HORMONE IN INCUBATION (ng/ml)
FIG. 3. Specificity of binding of 125I-labeled hGH to hepatocytes from male rats. The abscissa denotes the concentration of hormone-hGH (0), bGH (X), and oPRL (0)-in the incubation mixture. The ordinate represents the ratio of the amount of labeled hGH bound at a given hormone concentration to the amount of labeled hormone bound in absence of nonlabeled hormone. The points for hGH represent mean values of six experiments; for bGH and oPRL the means of two experiments.
The dose-response curves for hGH, bGH, and oPRL using hepatocytes from males, females, and estrogen-treated males are shown in Figs. 3, 4, and 5, respectively. Both bGH (30 Ag/ml) and oPRL (177 ,ug/ml) would completely displace specifically bound 125I-labeled hGH in all three groups of animals. Computer analysis of these results using a logit-logarithm transformation showed that the curves for bGH and hGH were parallel in males, but were not parallel in females and estrogen-treated males. The curves for oPRL were not parallel to those for hGH in any of the three groups. This invalidated the use of parallel line potency estimates. Instead, potencies were expressed as dose levels resulting in halfmaximal displacement of bound 125I-labeled hGH (ED5o). As shown in Table 1, the ED50 for hGH was lower in males (22 ng/ml) than in females (40 ng/ml, P < 0.001). Males and females also differed in the potencies of bGH (48 compared to 990 ng/ml) and of oPRL (32,000 compared to 750 1.0'
.9 .7' 0
10 102 103 104 10i HORMONEIN INCUBATION (ng/ml)
FIG. 5. Specificity of binding of 1251-labeled hGH to hepatocytes from estrone-treated male rats. The abscissa denotes the
concentration of hormone-hGH (-), bGH (X), and oPRL (0)-in the incubation mixture. The ordinate represents the ratio of the amount for labeled hGH bound at a given hormone concentration to the amount for labeled hormone bound in absence of nonlabeled hormone. The points of hGH represent mean values of four experiments; for bGH and oPRL, the values from one experiment.
ng/ml). In estrone-treated males, the potencies of the three hormones were similar to those found in females. Scatchard plot analysis Shown in Fig. 6 are Scatchard plots (14) for hGH binding to hepatocytes from a male, a female, and an estrone-treated male rat. The plot is linear in the male, indicating the presence of a single class of receptors. In the females and estrone-treated males, curvature of the plots indicates the presence of two or more classes of receptors. The Scatchard plots from each of the experiments in six males, eight females, and four estrone-teated males were calculated and analyzed using the computer program previously described (13). This program permits a choice, based on residual variance and objective F-test, among three models for the best fit of the data: a two-parameter model corresponding to a single class of noninteracting binding sites; a three-parameter model corresponding to one class of saturaTable 1. Relative potency of hGH, bGH, and oPRL in competing with 1251-labeled hGH for binding to isolated rat hepatocytes: Dose at which displacement is half-maximal
106 105 HORMONE IN INCUBATION (ng/ml) FIG. 4. Specificity of binding of 125I-labeled hGH to hepatocytes from female rats. The abscissa denotes the concentration of hormone-hGH (-), bGH (X), and oPRL (0)-in the incubation mixture. The ordinate represents the ratio of the amount of labeled hGH bound at a given hormone concentration to the amount of labeled hormone bound in absence of nonlabeled hormone. The points for hGH represent mean values of eight experiments; for bGH and oPRL the means of four experiments. 102
48.1 32,000 (40.5; 60.3) (26,000; 40,000)
32.12t (17.2-60.0) (4)
* Mean (ng/ml); in parentheses, 95% limits and number of experiments. Means and 95% confidence limits calculated from logarithms to approximate a normal distribution. t For hGH: male compared to female, P < 0.001; male compared to estrogen-treated males, not significant; female compared to estrogen-treated males, not significant.
Cell Biology: Ranke et al.
Proc. Nat. Acad. Sci. USA 73 (1976) Table 2. Affinity constants (K) and number of sites* for binding of hGH to isolated rat hepatocytes No. of K x 10-9 expts. (liters/mol) Male
6200t (4600-8200) 21,000t (10,000-42,000) 26,000t
4 1.45t Estrone-treated males (6000-110,000) (0.50-4.18) * Mean and 95% limits are given. Mean and 95% confidence limits calculated from logarithms to approximate a normal distribution. t Differences between groups, not significant. t Males compared to females, P < 0.005; males compared to estrone-treated males, P < 0.005; females compared to estronetreated males, not significant.
TOTAL HORMONE BOUND
FIG. 6. Scatchard plot of data from dose response curves for hGH using hepatocytes from a representative female (0), male (0), and estrone-treated male (X) animal. The concentrations of hepatocytes in the incubation mixtures were 2.1 X 106/ml in the female, 8.2 X 106/ml in the male, and 6.2 X 106/ml in the estronetreated male animal. The ordinate depicts the ratio of bound to free hormone, and the abscissa the total amount of hormone bound to the hepatocytes.
ble binding sites in the presence of a second class of binding sites with "infinite" capacity but infinitesimal affinity; and a four-parameter model corresponding to the presence of two separate classes of saturable sites. The first model fits a straight line Scatchard plot, while the second two fit curved lines. In all of the six experiments using males, the two-parameter model gave the best fit. In contrast, for seven of the eight females and three of the four estrone-treated males, a three-parameter model proved a significantly better fit (P < 0.05). This suggests the presence of a single class of binding sites for hGH in males and more than one class of binding sites in females and estrone-treated males. The affinity constants (K) and binding capacities (q) of the primary (high affinity) binding sites for hGH are shown in Table 2. The affinity constant of the single class of receptors in males (1.2 X 109 liters/mole) was the same as that of the primary class of receptors in females (0.96 X 109 liters/ mole) and estrone-treated males (1.4 X 109 liters/mole). The number of these receptors per cell was about four times higher in females (21,000) or in estrone-treated males (26,000) than in males (6200). DISCUSSION
The binding of 125I-labeled hGH to isolated hepatocytes is compatible with the presence of physiologically important hormone receptors on the surface of liver cells. Dose-response curves for hGH in males, females, and estrone-treated males show significant displacement of '25I-labeled hGH at physiologic dose levels. The affinity constants and number of binding sites per cell in all three groups of animals are of
the same order of magnitude found for other polypeptide hormone receptors (11, 15). Hepatocytes prepared from male and female rats show differences in binding of 125I-labeled hGH. Linearity of the Scatchard plots of hGH binding to male heptocytes indicates the presence of a single class of receptors. In rodent bioassays, hGH has been shown to have both growth-promoting (somatogenic) (16) and lactogenic (17) effects. Bovine growth hormone has somatogenic (17) but not lactogenic effects and oPRL has only lactogenic effects (18). Binding of bGH to hepatocytes from male rats is nearly identical to binding of hGH, as indicated by similar potency estimates and parallel dose-response curves for the two hormones. In contrast, oPRL is 1000-fold less potent than hGH. Displacement of 125I-labeled hGH by bGH but not oPRL at physiologic concentrations suggests that membrane receptors on hepatocytes from male rats recognize only the somatogenic properties of hGH. The interactions of hGH with hepatocytes from female rats are considerably more complex. Total binding of hGH is greater than to male hepatocytes. Scatchard plots of hGH binding are nonlinear. This indicates either the presence of two or more classes of receptors with different affinities for hGH or a negative cooperative interaction of the hormone with its receptor, such as has been shown for insulin binding to lymphocytes (19). If the former interpretation is correct, the higher affinity site has the same affinity constant for hGH as the single site in males. Estimates of the number of such receptors per cell are higher than for males. The number and affinity of the secondary class of receptors for hGH cannot be estimated from the present data. With hepatocytes from female rats, bGH and oPRL are approximately equipotent in displacing bound 125I-labeled hGH. In contrast to results using male cells, the ED50 of bGH is 20-fold greater than that of hGH and the displacement curves are not parallel. Ovine prolactin is much more potent in displacing 125I-labeled hGH from female than from male cells. However, the ED50 of ovine prolactin is also 20-fold greater than that of hGH and the displacement curves are not parallel. Since physiologic concentrations of either bGH or oPRL give partial displacement of 125I-labeled hGH, the binding of hGH in females appears to combine both somatogenic and lactogenic properties. The characteristics of 125I-labeled hGH binding to hepatocytes from female rats suggest the presence of two distinct classes of binding sites with differing affinities and specifici-
Cell Biology: Ranke et al. ties. In this model, the first site has a higher affinity for hGH, recognizes the somatogenic properties of hGH, andl recognizes the somatogenic hormone, bGH. This site may be identical to the receptor found in males. The second site has a lower affinity for hGH, recognizes the lactogenic properties of hGH, and recognizes the lactogenic hormone, oPRL. This site is lacking in males but can be induced by estrogen treatment. The fact that oPRL at high concentrations completely displaces '25I-labeled hGH in females, as in males, suggests some weak interaction with the first site. This may represent trace contamination of the oPRL or weak intrinsic activity of the oPRL molecule. A similar weak interaction of bGH with the second site is suggested by the fact that bGH at high concentrations also completely displaces 125I-labeled hGH in females. While the two-site model is consistent with the experimental data, validation of the model and full characterization of the postulated second class of receptors will require experiments using labeled oPRL. The postulated second site for hGH binding to hepatocytes from female and estrogen-treated male rats may be identical to the "lactogenic" receptor reported by Posner et al. (8) for rat liver membranes. However, liver membranes from untreated male rats did not bind '25I-labeled hGH and there was no evidence for a separate site with somatogenic specificity in liver membranes from female rats. The demonstration of "somatogenic" binding sites in rat hepatocytes is consistent with the hypothesis that the liver is a target organ for the growth promoting properties of growth hormones and suggests that the lactogenic and somatogenic effects of hGH in rats are mediated by different receptors. We thank the National Institute of Arthritis, Metabolism, and Digestive Diseases for providing growth hormones and prolactin; Kabi/AS/Stockholm for hGH; Dr. M. McGillivray, Buffalo, N.Y., for anti-hGH antibody; Dr. H. Rasmussen, Philadelphia, Pa., for bovine parathormone; and Dr. J. Galloway, Eli Lilly Laboratories, for insulin and glucagon. The assistance of Martin Z. Weingarten and Dr. Charlotte H. Greene is greatly appreciated. This research was supported by the following grants: Deutsche Forschungsgemeinschaft Ra 232/1-2 (M.R.), USPHS Training Grant no. 259 (C.S.),
Proc. Nat. Acad. Sci. USA 73 (1976)
RCDA 1K04-AM 00005 (J.S.P.), AM 13518, HD 00215, HD 00371, and GM 20138. 1. Van Wyk, J. J., Underwood, L. E., Hintz, R. L., Clemmons, D. R., Voina, S. J. & Weaver, R. R. (1974) Recent Prog. Horm. Res. 30, 259-318. 2. Mayberry, H. E., Van den Brande, J. L., Van Wyk, J. J. & Waddell, W. J. (1971) Endocrinology 88, 1309-1317.. 3. McConaghey, P. & Sledge, C. B. (1970) Nature 225, 12491250. 4. McConaghey, P. & Dehmel, J. (1972) J. Endocrinol. 52, 587588. 5. Tsushima, T. & Friesen, H. G. (1973) J. Clin. Endocrinol. Metab. 37, 334-337. 6. Posner, B. I., Kelly, P. A., Shiu, R. P. C. & Friesen, H. G. (1974) Endocrinology 95,521-531. 7. Kelly, P. A., Posner, B. I., Tsushima, T. & Friesen, H. G. (1974) Endocrinology 96,532-539. 8. Posner, B. I., Kelly, P. A., & Friesen, H. G. (1974) Proc. Nat. Acad. Sci. USA 71, 2407-2410. 9. Berry, M. N. & Friend, D. S. (1969) J. Cell Biol. 43, 506-520. 10. Krebs, H. A., Cornell, N. W., Lund, P. & Hems, R. (1974) in Alfred Benzon Symposium VI, eds., Lundquist, F. & Tygstrup, N. (Academic Press, New York), pp. 718-743. 11. Lesniak, M. A., Gorden, P., Roth, J. & Gavin, J. R., III (1974) J. Biol. Chem. 249,1661-1667. 12. Hunter, W. M. & Greenwood, F. G. (1962) Nature 194, 495496. 13. Rodbard, D. & Faden, V. B. (1975) Radioimmunoassay Data Processing Report PB 246222-4 (National Technical Information Service, Springfield, Va.), 3rd ed. 14. Scatchard, G. (1949) Ann. N.Y. Acad. Sci. 51, 660. 15. Gavin, J. R., III, Gorden, P., Roth, J., Archer, J. A. & Buell, D. N. (1973) J. Biol. Chem. 248,2202-2207. 16. Evans, H. M., Simpson, M. E., Marx, W. & Kibrick, E. (1943) Endocrinology 32, 13-16. 17. Kleinberg, D. L. & Frantz, A. G. (1971) J. Clin. Invest. 50, 1557-1568. 18. Li, C. H. (1972) "Lactogenic hormones," in Ciba Foundation Symposium, eds. Wolstenholme, G. E. W. & Knight, J. (Churchhill Livingstone, London), p. 23. 19. DeMeyts, P., Roth, J., Neville, D. M., Jr., Gavin, J. R., III & Lesniak, M. A. (1973) Biochem. Biophys. Res. Commun. 55, 154-161.