0013-7227/91/1282-1129$03.00/0 Endocrinology Copyright © 1991 by The Endocrine Society

Vol. 128, No. 2 Printed in U.S.A.

Carbohydrate Modifications Transform Human Chorionic Gonadotropin into a Potent Stimulator of Adenosine 3',5'-Monophosphate and Growth Responses in FRTL-5 Thyroid Cells* RUDOLF HOERMANNt, HENRY T. KEUTMANN, AND SYED M. AMIR* Charles A. Dana Research Institute, Harvard Thorndike Laboratory, Department of Medicine, Beth Israel Hospital, and the Endocrine Unit, Department of Medicine, Massachusetts General Hospital (H.T.K.), Harvard Medical School, Boston, Massachusetts 02215

ABSTRACT. To delineate the role of carbohydrate moiety in the expression of in vitro thyrotropic activity of hCG, its variants that lacked sialic acid residues or the entire carbohydrate moiety on one or both subunits were prepared. They along with intact hCG were then tested for their abilities to bind to TSH receptor and stimulate cAMP production and growth responses in FRTL5 cells. The removal of sialic acid from either one or both subunits sharply increased the [125I]bovine TSH binding-inhibiting activity of hCG in the receptor assay. Among the variants tested, desialylated hCG (as-hCG) was the most potent inhibitor, followed by a-as-/3, as-a-0, and hCG in that order. With respect to their abilities to stimulate cAMP generation in the cells, the activities of all desialylated hCG variants were markedly higher than that of hCG itself, and as in the receptor assay, as-hCG was the most potent stimulator tested. At the same concentration (100 jig/ml), as-a-/3, a-as-/3, and hCG were approximately only 57%, 46%, and 27% as active as as-hCG in activating cAMP production. The findings in the growth assay were entirely consistent with those noted in the cAMP response assay. Among the deglycosylated hybrids examined, a-dg-j8 was the


CG IS A placental glycoprotein hormone with a mol wt of approximately 38,000 (1). It is comprised of two dissimilar subunits, a and j8, with mol wt of 16,000 and 22,000, respectively. In the a-subunit, carbohydrate constitutes 30% of the protein weight, of which 27.4% is sialic acid. Similarly, the /3-subunit contains 36% carbohydrate, of which about 28% is sialic acid. The sialic acid residues as well as the entire carbohydrate moiety of hCG have been shown to be essential for the full expresReceived August 13,1990. Address requests for reprints to: Dr. Rudolf Hoermann, Department of Medicine II, Klinikum Grosshadern, Marchioninistrasse 15, D-8000 Munich 70, West Germany. • This work was supported in part by Grant CA-31218 from the NIH (Bethesda, MD). t Supported by the Deutsche Forschungsgemeinschaft, West Germany. ^Present address: National Institutes of Health, Westwood Building, Bethesda, Maryland 20892.

most effective in activating cAMP release. An equivalent dose (200 Mg/ml) was about 1.7, 2.7, and 3.8 times as active as dghCG, dg-a-0, and hCG, respectively. The deglycosylated variants of hCG showed a similar pattern of activity in growth response and cAMP accumulation assays. In both cases, a-dg-j8 was the most potent stimulator, while hCG was the least active. No significant difference between the potencies of dg-a-|8 and a-dg/8 was discerned in the receptor assay; however, deglycosylated hCG (dg-hCG) was sharply more active. Results of these studies strongly suggest that the optimal expression of in vitro thyrotropic activity of hCG variants in FRTL-5 cells may require an a-subunit, with intact carbohydrate, in combination with the deglycosylated /3-subunit. Further, they demonstrate that modification of the hCG carbohydrate moiety alone can transform it from a weak agonist to a potent stimulator of in vitro thyrotropic bioactivity. These results also provide further evidence to support the notion that the degree of expression of thyrotropic activity is strongly affected by the species in which the studies are performed. (Endocrinology 128: 1129-1135,1991)

sion of its in vitro and in vivo gonadotropic activities (25). Desialylated hCG (as-hCG) preparations, while largely devoid of in vivo gonadotropic activity, retained both the receptor binding and almost all of the in vitro steroidogenic activity, as measured in rat testis. Deglycosylated hCG (dg-hCG) has also been shown to be fully active in the hCG receptor binding assay. However, unlike as-hCG, it had little or no activity in in vitro or in vivo bioassays for gonadotropic activity (3, 4). Most of the studies related to the carbohydrate structure-function relationship of hCG have focused on its activity as a gonadotropin. The role of its carbohydrate in respect to its thyrotropic activity has received only scant attention. The few studies, including our own, that have addressed this question have shown that the carbohydrate moiety of hCG greatly influences its binding affinity for receptors in the human thyroid as well as its ability to inhibit the TSH-induced adenylate cyclase 1129

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.



response therein (5-7). In the present studies we have employed FRTL-5, a line of cultured rat thyroid cells, to study the role of carbohydrate in hCG receptor-binding ability, cAMP induction, and growth-promoting response. The findings suggest that carbohydrate plays a remarkably important role in regulating these activities, and that, in some respects, the responses induced in FRTL-5 cells by hCG are entirely distinct from those observed in human thyroid tissue.

Materials and Methods

Endo • 1991 Voll28«No2

Warren's technique (10). hCG or its individual subunits were deglycosylated by treatment with anhydrous hydrogen fluoride (Natheson, East Rutherford, NJ) for 1-3 h at 0 C, as described previously (4). Preparation of hCG variants Hybrid desialylated or deglycosylated hCG variants1 were prepared by incubation of an intact subunit with either the desialylated or deglycosylated complementary subunit in equimolar concentrations in 100 mM NH4HCO3 for 72 h at 4 C. Final purification of the variants was carried out by chromatography on Bio-Gel P-100.

Hormone and enzyme preparations

Receptor binding assays

Purified preparations of hCG (12,800 IU/mg) and its subunits were supplied by the Center for Population Research, NICHHD. hCG was also purified (-10,000 IU/mg) from a crude hCG preparation, donated by Ayerst Research Laboratories (New York, NY), according to previously described methods (1). Purified hCG was dissociated by treatment with 10 M urea, and the subunits were isolated as described by Morgan et al. (8). In some experiments, the DEAE-Sephadex column chromatography employed by these researchers was omitted; instead, the subunits were chromatographed twice on a column of Bio-Gel P-100, preequilibrated with 100 mM ammonium acetate, pH 4.5. The fractions corresponding to each subunit were pooled, dialyzed, and lyophilized. Highly purified preparations of bovine TSH (bTSH) employed for radioiodination were a kind gift of Dr. John G. Pierce, University of California, Los Angeles. Partially purified bTSH preparations employed in other studies were purchased from Armour Pharmaceutical Co. (Chicago, IL). Neuraminidase attached to Sepharose (type VI-A) was purchased from Sigma Chemical Co. (St. Louis, MO); [3H]thymidine and 1251Na were purchased from New England Nuclear (Boston, MA). RIA kits for measurement of cAMP concentrations were purchased from Immunonuclear Corp. (Stillwater, MN).

Both bTSH and hCG were radioiodinated with iodine-125 using chloramine-T, according to techniques described previously (11). The binding of [125I]bTSH to the cells was measured according to the techniques described by Tramontano and Ingbar (12). Before the assay, the medium was replaced by KrebsRinger bicarbonate buffer in which sucrose was substituted for NaCl in a concentration (280 mM) sufficient to maintain tonicity of the solution (modified KRB). The cells were then washed twice with modified KRB buffer (0.5 ml), after which [125I]bTSH (10,000 cpm; 10"12 M)2 and increasing concentrations of either unlabeled bTSH or hCG variants were added in 0.25 ml modified KRB. The incubation was performed at 4 C for approximately 20 h. Thereafter, the medium was removed, the cells were washed with PBS (0.5 ml) and solubilized in 1 N NaOH (0.5 ml), and their 125I content was determined. Nonspecific binding, measured in samples that contained 0.5 IU unlabeled bTSH, was subtracted from total binding to obtain specific binding. A rat testicular particulate fraction for hCG receptor assay was obtained by home genization and differential centrifugation techniques, as previously described (5). The [125I]hCG (~10~12 M; see Footnote 2) was incubated with rat testicular fraction (40 Mg protein/0.3 ml) in the presence and absence of varying concentrations of unlabeled hCG or one of its variants at 22 C for approximately 16 h in 20 mM Tris-Cl, pH 7.5, containing 0.5% BSA, as described previously (5). Nonspecific binding, determined in the presence of 200 IU unlabeled hCG, was subtracted from total binding.

FRTL-5 cells The present studies were carried out in FRTL-5 cells, a line of follicular cells that is derived from Fischer rat thyroids. These cells have been demonstrated to retain the functional characteristics of the thyroid gland in vivo during a period of at least 3 yr (9). These characteristics include iodide uptake and thyroglobulin synthesis. Desialylation and deglycosylation of hCG subunits Desialylation of hCG or its subunits (10 mg/ml) was effected by treatment with insolubilized neuraminidase (600 mlU/ml) in 0.1 M sodium acetate buffer, pH 5.6, at 37 C for 30 min. (5). The enzyme was removed by centrifugation, and the medium was applied to a column of Bio-Gel P-100, preequilibrated with 100 mM NH4HCO3. as-hCG or its subunits, which eluted as single protein peaks, were pooled and lyophilized. The sialic acid content of hCG and the subunits was determined both before and after treatment with neuraminidase according to

Measurement of cAMP release from FRTL-5 cells For the present studies, FRTL-5 cells were grown in Coon's Modified Ham's F-12 Medium that contained 5% calf serum as well as insulin (10 Mg/ml), transferrin (5 Mg/ml), and bTSH (10 mlU/ml), designated as 3H-medium. The cells were plated in 24-well Costar plates (2.0 X 106 cells/well; Costar, Cam1 The following abbreviations are used: a-as-/3 or a-dg-/3, intact asubunit combined with asialo or deglycosylated /3-subunit; as-a-/3 or dg-a-/3, asialo or deglycosylated a-subunit combined with intact /?subunit. 2 It was assumed that purified bTSH and hCG have bioactivities of 40 and 13,400 IU/mg, respectively. The mol wt values employed for calculation of molarity were: hCG, 38,000; bTSH, 28,000.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.


ACTIVITY OF hCG VARIANTS IN FRTL-5 CELLS bridge, MA) in 3H-medium, allowed to grow to confluence, washed once with RPMI medium containing 0.25% BSA, and then reincubated for a week in the hormone-free medium. For cAMP measurement, FRTL-5 cells were washed with KrebsRinger bicarbonate buffer containing 0.1% BSA (KRB-BSA), and the washed cells were incubated with the test agents in KRB-BSA (0.5 ml) containing 1 mM methylisobutylxanthine and 1 mg/ml glucose. The incubation was performed at 37 C for 30 min in an atmosphere of 95% O2-5% CO2. In control experiments, the incubation medium contained all of the ingredients except test samples. The cAMP measurements were carried out directly on an aliquot of the incubation medium. Measurement of pHJthymidine incorporation The cells employed for measurement of [3H]thymidine incorporation in the cell DNA were plated in 3H-medium from which bTSH was omitted. After incubation for 4-7 days, the cells were washed once with RPMI (0.5 ml; the medium), and then incubated with bTSH or one of the hCG variants in 0.5 ml medium at 37 C for 36 h in an atmosphere of 95% O2-5% CO2. The control wells contained no test specimen but all other reagents. After approximately 36 h, [3H]thymidine (2.5 fid in 0.25 ml medium) was added to each well, and the samples were incubated for an additional 5 h. After this time, the medium was removed, and the cells were successively washed with cold PBS and trichloroacetic acid (0.5 ml each). Finally, 1 N NaOH (0.5 ml) was added to each well, the suspension was thoroughly mixed, and an aliquot was counted in a /3-scintillation counter.

Results hCG recombinants Intact hCG and its a- and /3-subunits contained 10.1%, 6.4%, and 11.4% sialic acid by weight, respectively. Measurement of residual sialic acid in as-hCG and its subunits indicated that more than 90% of this sugar, as originally present, had been removed. The recombination of the asialo subunit with the complementary intact or asialo subunit proceeded readily, and the asialo derivatives were obtained in high yield (80-85%). The as-hCG preparations obtained by either recombination of as-a- and as-0-subunits or desialylation of intact hCG were indistinguishable in the hCG receptor assay. Therefore, in all subsequent studies only the preparation obtained by the latter technique was employed. The dg-hCG and its a- and /3-subunits employed in the present studies were similar to those that were characterized in detail previously (4). Thus, the treatment with hydrogen fluoride had removed approximately 80% of the carbohydrate from the a-subunit and 66% from the /?-subunit. The principal residual sugar in the case of dg-a-subunit was JV-acetylglucosamine, while in the case of the dg-/3-subunit, besides a substantial amount of Nacetylglucosamine (44%), almost all of the JV-acetylgalactosamine also remained attached. The dg-a and intact j8 combined to a greater extent than did intact a and dg-

/?. In the former case, the recombined product was obtained in 80% yield, while in the latter case, the yield was only 50% or less of the expected value. In agreement with previous reports (3-5), the desialylated and deglycosylated hCG analogs exhibited no significant loss of activity, compared to intact hCG, in the hCG receptor assay. Activity in the TSH receptor assay As judged from their abilities to inhibit the binding of [125I]bTSH to FRTL-5 cells, the asialo variants tested proved far more active than intact hCG itself (Fig. 1). Half-maximal inhibition of the binding was affected by a bTSH concentration of 280 /JU/ml. An equivalent degree of inhibition was shown by 2.4, 8, and 24 Mg/ml as-hCG, a-as-/?, and as-a-ft respectively. Thus, as-hCG, a-as-/3, and as-«-/5 had activities equivalent to 116, 35, and 12 mlU/mg bTSH in this assay. The activity of intact hCG was much lower and could not be accurately estimated. Separate intact subunits as well as the as-asubunit of hCG contained no detectable activity. However, the as-j8-subunit showed a significant inhibitory activity which was estimated to be less than 10% that of as-hCG itself (data not shown). Although the deglycosylated analogs, in general, were also much more potent than intact hCG, the differences among the potencies of these were less remarkable than those seen among the desialylated analogs. Thus, a-dg-ft and dg-a-/? showed very similar activities, each causing 50% binding inhibi-

1.000 hCG ADDED (^Q/ml) 0.1



bTSH ADDED (mlU/ml)

FIG. 1. Inhibition of the specific binding of [125I]bTSH to FRTL-5 cells by intact hCG and desialylated hCG variants. The variants (asa-/3, a-as-j8, and as-hCG) are devoid of sialic acid residues in either one or both subunits of hCG. Incubations were carried out for approximately 20 h at 4 C in modified KRB buffer, as described in Materials and Methods. The bound/total ratio was 19% in this experiment. Points in this and subsequent figures represent the means of closely concurring duplicate determinations, and the relative potencies of the agents were confirmed by repeated experiments.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.




tion at approximate concentrations of 140 and 160 tig/ ml, respectively, in these experiments (Fig. 2). A similar binding inhibition was caused by a bTSH concentration of approximately 400 /xlU/ml. Based on these findings, a-dg-P and dg-a-fi possessed activities equivalent to 2.5 and 2.8 mlU/mg bTSH. In comparison, dg-hCG was substantially more active, with an activity equivalent to 28.6 mlU/mg bTSH. cAMP-releasing activity

Endo • 1991 Voll28»No2




The asialo-hCG variants were also tested for their i ability to stimulate cAMP release from FRTL-5 cells at concentrations ranging between 10-100 tig/roX. In all cases, the loss of sialic acid, either partially or fully, markedly increased the stimulatory activity of hCG (Fig. 3). The enhancement of activity after desialylation was 20 50 100 consistently demonstrable at all concentrations tested, hCG Concentration (jig/ml) regardless of whether the sialic acid was removed from FIG. 3. Stimulation of cAMP release in FRTL-5 cells by hCG and its the a-, j8-, or both subunits. In this experiment, as-hCG desialylated variant forms (as-a-/3, a-as-j8, and as-hCG). Cells grown at the highest level increased the medium cAMP concento confluence were incubated in the presence and absence of the hCG tration to about 27 times the basal level. In comparison, forms in Krebs-Ringer bicarbonate buffer containing 0.1% BSA, 1 tag/ bTSH at a concentration of 0.2 mlU/ml caused a 45-fold ml glucose, and 1 mM methylisobutylxanthine for 30 min at 37 C in an increase over the basal cAMP concentration. Although atmosphere of 95% O2-'>% CO2, as described in detail in Materials and Methods. the maximal activity was observed after the removal of sialic acid from both subunits, the loss of sialic acid from the a-subunit appeared to generate a slightly greater a-dgp activity than its removal from the 0-subunit. At the highest concentrations (100 /xg/ml) tested, the responses evoked by as-a-ft a-as-0, and hCG were approximately 57%, 46%, and 27% of that produced by as-hCG itself, respectively. Consistent with these findings were the results obtained with deglycosylated analogs at doses from 20-200





hCG Concentration (


FIG. 4. Stimulation of cAMP release in FRTL-5 cells by hCG and its deglycosylated variant forms (dg-a-ft a-dg-/3, and dg-hCG). The remaining experimental details were similar to those in Fig. 3.


20 dflhCG


Carbohydrate modifications transform human chorionic gonadotropin into a potent stimulator of adenosine 3',5'-monophosphate and growth responses in FRTL-5 thyroid cells.

To delineate the role of carbohydrate moiety in the expression of in vitro thyrotropic activity of hCG, its variants that lacked sialic acid residues ...
935KB Sizes 0 Downloads 0 Views