0013-7227/91/1286-3259$03.00/0 Endocrinology Copyright © 1991 by The Endocrine Society
Vol. 128, No. 6 Printed in U.S.A.
Immunoreactive and Bioactive Isoforms of Human Thyrotropin* ISABELLE SERGI, MARIE-JEANNE PAPANDREOU, GABRIELLA MEDRI, COLETTE CANONNE, BERNARD VERRIER, AND CATHERINE RONIN Laboratoire de Biochimie, URA 1179 CNRS (I.S., M.-J.P., C.C., C.R.), and U 270INSERM Faculte de Medecine Secteur Nord (B. V.), 13326 Marseille Cedex 15, France; and the Istituto di Scienze Endocrine, Unwersitd di Milano, Ospedale Maggiore IRCCS (G.M.), Milan, Italy
ABSTRACT. Isoforms of intrapituitary human TSH were separated by gel isoelectrofocusing, and their immunoreactivity analyzed by subsequent immunoblotting using polyclonal and monoclonal antibodies. Under these conditions, TSH polymorphism could be resolved as seven major isoforms (pi 8.6, 8.3,8.0, 7.5, 7.0, 6.5, and 6.0) by both silver staining of the gels and binding to anti-TSH polyclonal antibodies. The distribution pattern of these forms appeared totally distinct from that of individual TSHa (pi 8.8, 8.4, 8.2, 7.6, 7.4, 6.8, 6.6, 5.8, and 5.4) and TSH0 (pi 8.7, 8.1, 7.2, 6.8, 6.2, and 5.8) subunits. While most anti-TSH polyclonal antibodies recognized neutral and alkaline isoforms of TSH (pi 8.6, 8.3, 8.0, 7.5, 7.0, 6.5, and 6.0) through j8 determinants, they displayed a variable potency to
T
HE PITUITARY glycoprotein hormones LH, FSH, and TSH are heterodimers consisting of noncovalently associated a- and 0-subunits. The a-subunit is common to these hormones as well as to the placental CG within a given animal species, while the /3-subunits are distinct and confer immunological and biological specificity to the respective dimers (1). It is now well known that gonadotropins exist in multiple molecular forms that vary in number and relative abundance according to their presence in the various biological fluids as well as to the sex and physiological status of the individual. Over the past decade, several laboratories have characterized various isoforms of both LH and FSH from nonmammalian and mammalian species, including rodents, primates, and man (2). Hormone polymorphism has been disclosed by several techniques, such as gel chromatography (3), chromatofocusing (4), and isoelectric focusing (IEF) (5). It was observed that FSH and LH preparations exhibit charge heterogeneity Received November 16,1990. Address all correspondence and requests for reprints to: Dr. Catherine Ronin, Laboratoire de Biochimie, Faculte de Medecine Secteur Nord, Boulevard P. Dramard, 13326 Marseille Cedex 15, France. * This work was supported by grants from the Assocation pour la Recherche contre le Cancer and the Institut National de la Sante et de la Recherche Medicale (CRE 900505).
bind acidic forms of the hormone (pi 5.8, 5.5, 4.8, and 4.5), in contrast to anti-TSHa antisera, which enlighted the broadest spectrum of isoforms. Monoclonal antibodies of various specificities largely reproduced this distribution, indicating that at least five distinct epitopes are coexpressed in the neutral and alkaline forms of TSH, but only two are expressed in the acidic ones. All of the forms were found to induce cAMP production and stimulate growth of FRTL-5 rat thyroid cells, although neutral forms proved to be definitely less potent than the others. We therefore, conclude that TSH isoforms differ in the expression of both their immunoreactive and bioactive domains and that the bioactive/immunoreactive ratio is not an accurate index for the biopotency of the hormone. (Endocrinology 128: 3259-3268, 1991)
and, therefore, may be separated as an average of seven isoforms in animals such as the rat for the former (5) or
the pig for the latter (6). They have been reported to differ from each other not only in their isoelectric point (pi), but also in their relative proportion, receptor binding, biological activity, and plasma half-life. The overall pattern of gonadotropin polymorphism also appears to vary according to the specific endocrine status of the animal; sex steroids and the hypothalamic stimulatory factor, GnRH, seem to act in concert at the pituitary level to influence both the glycosylation and secretion of the gonadotropins and, therefore, in turn, to affect their biological function at the target gonad cells. On a molecular basis, it has been generally observed that in humans, acidic isoforms of FSH are more abundant and display longer circulatory half-lives and in vivo biological activity but reduced receptor-binding activity than less negatively charged forms (7). This was also shown to apply to human LH, for which these changes could be related to a variable sialic acid content (8), indicating that for both hormones, each isoform may contain different glycosylation patterns. However, no carbohydrate (CHO) structure of the various glycoforms has been reported so far to restrict their difference in biopotency to glycosylation, and it may be postulated that each form can also
3259
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 03:30 For personal use only. No other uses without permission. . All rights reserved.
ISOFORMS OF HUMAN TSH
3260
present subtle changes in its three-dimensional protein structure that account for the observed properties. Much less information is available for TSH, especially that of human origin. Although the biosynthesis, secretion, and regulation of its bioactivity have been widely investigated for this hormone and shown to depend upon CHO processing, its physicochemical and structural characteristics are poorly known. Four biologically active forms of bovine TSH had been isolated from anterior pituitary glands by horizontal IEF (9) in the pH region of 8.2-8.8. Additional species were detected in the normal rat pituitary, distributed over a wider range of pi from 2.5-8.7 (10). The content and distribution of individual isoforms are differentially affected by circulating thyroid hormones (11) or TRH (12). In both cases, the amount of the more acidic forms of TSH appeared to be preferentially regulated compared to that of their more alkaline counterpart. Using zone electrophoresis, it was also shown that human TSH may be composed of up to 20 components likely to differ by their net charge (13), independently of the sex and age of the individual (14). In all experiments TSH was estimated by conventional RIAs, assuming that all isoforms react equally well with polyclonal antibodies (pAbs). Previous work from this laboratory has shown that several antigenic determinants of human TSH are glycosylation-dependent. Using monoclonal antibodies (mAbs), we found that deglycosylated TSH is no longer recognized by anti-TSH/3 antibodies, but is still fully bound by anti-TSHa antibodies (15). In this process, the a-linked CHO chains are more determinant than the 0 one, as if the conformation of the /3-subunit of TSH were under the control of the a-subunit. In an attempt to understand how expression of the functional domains of human TSH is under the control of glycosylation, we separated the various isoforms of the intrapituitary hormone and compared their respective bioactivities to their recognition by antibodies of various specificity.
Materials and Methods Materials Polyacrylamide gel solution was obtained from Pharmacia (Piscataway, NJ), IEF standards from Bio-Rad (Richmond, CA), ampholines from Serva, and polyvinylidene difluoride (PVDF) membranes from Millipore (Bedford, MA). Goat antirabbit immunoglobulins conjugated to alkalin phosphatase were purchased from Promega (Madison, WI), and Steptavidin conjugated to alkalin phosphatase was obtained from Jackson (Bar Harbor, ME). Coon's 12 was obtained from Sigma. All other reagents were of analytical grade. Hormones Human TSH (immuological potency, 5.2-5.5 IU/mg in terms of First International Reference Preparation (IRP-1) Standard;
Endo»1991 Voll28«No6
contaminations by LH in RIA, 2.5-2.8%; by FSH, 1%), LH and a- and /3-subunits were purchased from UCB Co. (Liege, Belgium). Antibodies Rabbit antisera were either purchased from UCB Co. or homemade using the TSH and subunits from UCB as antigens and conventional immunizing procedures. Antibodies directed against the IRPs (First International Standard) were kindly provided by the National Pituitary Agency (Bethesda, MD). mAbs were a generous gift from Dr. M. Benkirane (Immunotech Co., Marseille, France). Free a-subunit was measured using the commercial kit from Biomerica (Newport, CA). IEF separation Vertical ( 8 x 7 cm) polyacrylamide minigels (acrylamidebisacrylamide ratio, 4.5%) were prepared in the presence of ampholines (pH 4-9). Samples were dissolved in 20% ampholines-water before running. Separation at equilibrium was achieved within 1 h, as described previously (16) and monitored using prestained IEF standards. The final pH gradient was measured every 0.5 cm immediately at the end of the run using a contact electrode. Each blot is representative of three to six distinct experiments. When necessary, fractions (2-mm slice) of the IEF gel were eluted overnight in PBS containing 0.1% BSA, and the released material was further assessed for immunoreactivity or before addition to the cells. Silver staining of the gels was performed after denaturation in 50% MeOH, as previously reported (17). Immunoblotting techniques The gels were electroblotted on PVDF membranes using a semidried Millipore system as recommended by the manufacturer. After transfer, saturation of the membranes was carried out in a TBS-Tween buffer [10 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 0.05% Tween-20] for 1 h at room temperature. Adding blocking, proteins such as BSA and casein did not improve saturation of the blots. Primary antibodies at the appropriate dilution in the TBS-Tween buffer were allowed to bind overnight under moderate agitation. After extensive washing, secondary antibodies were reacted for 2 h, and alkaline phosphatase activity revealed using tetrazolium blue and 5bromo-4-chloro-3-indoxylphosphate in 0.1 M Tris-HCl, pH 9.5, buffer containing 0.1 M NaCl and 5 mM MgCl2 for 10 min at room temperature. The reaction was stopped by adding 10 mM EDTA. TSH RIA TSH (~50 tiCi/ng) was iodinated with Iodogen, as described previously (18). The iodinated ligands (50 X 103 cpm) were incubated with or without unlabeled IRP-1 hormone or IEF fractions in the presence of anti-IRP TSH (final dilution, 1:2,500,000) or anti-IRP TSHa (final dilution, 1:180,000) in a final volume of 300 ^1 50 mM PBS buffer, pH 7.5, containing 0.1% BSA overnight at 4 C. Nonimmune rabbit serum (100 /il; dilution, 1:100) and horse rabbit immunoglobulins (100 n\; dilution, 1:4) were then added for 2 h at 4 C. Immune complexes were finally precipitated with 6% polyethylene glycol 6000 in a
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 03:30 For personal use only. No other uses without permission. . All rights reserved.
ISOFORMS OF HUMAN TSH final volume of 1 ml and recovered by centrifugation at 10,000 X g for 15 min. All assays were performed in duplicate, and the nonspecific binding was 0.5% of the total radioactivity. TSH bioassays TSH biological potency was assessed in an heterologous bioassay using a cloned rat thyroid cell line, FRTL-5. Cells were routinely cultured in Coon's F-12 medium supplemented with 5% fetal calf serum (FCS) and six hormones [bTSH (10 mU/ml), insulin (10 jug/ml), transferrin (5 Mg/ml), hydrocortisone (10 nM), somatostatin (10 ng/ml), and glycyl-L-histidylL-lysine acetate (10 ng/ml)] and growth factors, as previously described (19). Incorporation of [me£/iy/-3H]thymidine into DNA was performed using cells seeded in 24-well plates (5 X 10" cells/0.5 ml) in Coon's F-12 medium containing 5% FCS and five hormones (FCS-5H), but omitting TSH. Four days later, the culture medium was replaced by 0.5 ml of the same medium containing the IEF fraction to be assayed and [methyl3 H]thymidine (1 AtCi/ml). The plates were then incubated for 2 days at 37 C in 5% CO2, after which the cells were carefully washed with ice-cold PBS. Then, 7.5% trichloroacetic acid was added for 20 min at 4 C before the cells were solubilized in 0.1 N NaOH-1% sodium dodecyl sulfate (SDS). After 1-h incubation at room temperature, cell-associated radioactivity was counted in a Packard liquid scintillation counter (Downers Grove, IL). The results are expressed as the mean of quadruplicate determinations. The half-maximal effect of TSH on cell growth is routinely achieved for 10 ng TSH/well, while a plateau is reached above 50 ng TSH/well. cAMP production was measured under the following conditions. FRTL-5 cells were grown at confluence in Coon's F-12 medium containing 5% FCS and six hormones, then cells were refed with the same medium, omitting TSH, for 5-10 days. On the day of the experiment, the cells were carefully washed by a hypotonic medium containing 5.4 mM KC1, 1.3 mM CaCl2, 0.4 mM MgSO4, 0.4 mM Na2HPO4, 0.44 mM KH2PO4, 5.5 mM glucose, 0.5 mM isobutylmethylxanthine, and 0.4% BSA in a 10-mM HEPES buffer, pH 7.4. Samples of TSH were added to the wells in this medium, and the incubation was performed for 1 h at 37 C. Cell culture media were then rapidly recovered at 4 C, and extracellular cAMP was measured in triplicate, using the RIANEN [125I]cAMP RIA kit (DuPont, Dreiech, Germany).
Results Separation of TSH isoforms We first investigated the distribution of TSH native isoforms as a function of their isoelectric point by direct silver staining of the gels and analyzed their reactivity toward pAbs after electrotransfer. Figure 1 shows the comparative separation of TSH and individual subunits under both conditions. As can be seen in Fig. 1A, silver staining identified a highly purified preparation of TSH as seven major bands exhibiting different pi values, ranging from 5.5-8.6 (lane 1), whereas a- and /3-subunits were less heterogeneous, as
3261
they displayed one major species of 8.8 for a (lane 2) and two closely related species of pi 8.7 and 8.1 for /? (lane 3). Alkaline pi values for the subunits were largely expected, as they originate from the hormone by acid dissociation. Such treatment probably results in a concomittant loss of sialic acid, which is highly labile under low pH conditions. The silver-staining pattern observed for TSH and its subunits is likely to represent the species quantitatively predominant in the intrapituitary preparation. Figure IB shows the recognition of TSH isoforms by an anti-TSH antiserum as a function of hormone concentration. It appeared that, as found by silver staining, as many as seven bands could be visualized, with pi values ranging from 6-5-8.6 and comigrating with the species stained with silver (lanes 4-7). As little as 250 ng TSH could be separated and detected on immunoblots, whatever the specificity of the antibody or the hormone material analyzed. In Fig. 1C, is presented the corresponding separation of the individual subunits of TSH revealed by their respective antibodies. It appears that in contrast to silver staining, several bands of neutral to acidic pis could be attributed to both a-subunit, for which repetitive doublets were observed (pi 8.8, 8.4-8.2, 7.6-7.4,6.8-6.6, and 5.8-5.4; lane 8), and /3-subunit (pi 8.7, 8.1, 7.2, 6.8, 6.2, and 5.8; lane 10). These forms putatively represent minor components of the subunit preparation differring in their CHO content or unknown proteolytic fragments. As expected, none of the antibodies used was able to bind the opposite subunit (lanes 9 and 11). These findings indicate that the distribution pattern of the hormone is quite distinct from that of its subunits. This rules out as well a putative recognition of free subunits which could overlap with that of TSH isoforms. Indeed, our TSH preparation proved to be highly purified by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and silver staining, and revealed the presence of less than 2% of free a by a specific immunoassay (data not shown). This amount is far below what can be detected by the current immunoblotting technique (see Fig. IB). Specificity of antibody recognition To assess whether antibodies to TSH cross-react with gonadotropins on blots as they can do in conventional immunoassays, we compared the respective distribution and immunoreactivity of LH and TSH in IEF gels (Fig. 2).As expected, both anti-TSHa and anti-LHa antibodies reacted well with all isoforms of TSHa (lanes 1 and 2) as well as with those of LHa (lanes 3-4). Anti-LH and anti-LHa antibodies revealed a virtually identical distribution pattern when tested on the gonadotropin (lanes 4-5). Anti-TSH antibodies proved to be specific for TSH (lane 6) and unable to bind LH (lane 7), indicating that
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 03:30 For personal use only. No other uses without permission. . All rights reserved.
ISOFORMS OF HUMAN TSH
3262 TSH TSH-
Vol. 128, No. 6 Printed in U.S.A.
Immunoreactive and Bioactive Isoforms of Human Thyrotropin* ISABELLE SERGI, MARIE-JEANNE PAPANDREOU, GABRIELLA MEDRI, COLETTE CANONNE, BERNARD VERRIER, AND CATHERINE RONIN Laboratoire de Biochimie, URA 1179 CNRS (I.S., M.-J.P., C.C., C.R.), and U 270INSERM Faculte de Medecine Secteur Nord (B. V.), 13326 Marseille Cedex 15, France; and the Istituto di Scienze Endocrine, Unwersitd di Milano, Ospedale Maggiore IRCCS (G.M.), Milan, Italy
ABSTRACT. Isoforms of intrapituitary human TSH were separated by gel isoelectrofocusing, and their immunoreactivity analyzed by subsequent immunoblotting using polyclonal and monoclonal antibodies. Under these conditions, TSH polymorphism could be resolved as seven major isoforms (pi 8.6, 8.3,8.0, 7.5, 7.0, 6.5, and 6.0) by both silver staining of the gels and binding to anti-TSH polyclonal antibodies. The distribution pattern of these forms appeared totally distinct from that of individual TSHa (pi 8.8, 8.4, 8.2, 7.6, 7.4, 6.8, 6.6, 5.8, and 5.4) and TSH0 (pi 8.7, 8.1, 7.2, 6.8, 6.2, and 5.8) subunits. While most anti-TSH polyclonal antibodies recognized neutral and alkaline isoforms of TSH (pi 8.6, 8.3, 8.0, 7.5, 7.0, 6.5, and 6.0) through j8 determinants, they displayed a variable potency to
T
HE PITUITARY glycoprotein hormones LH, FSH, and TSH are heterodimers consisting of noncovalently associated a- and 0-subunits. The a-subunit is common to these hormones as well as to the placental CG within a given animal species, while the /3-subunits are distinct and confer immunological and biological specificity to the respective dimers (1). It is now well known that gonadotropins exist in multiple molecular forms that vary in number and relative abundance according to their presence in the various biological fluids as well as to the sex and physiological status of the individual. Over the past decade, several laboratories have characterized various isoforms of both LH and FSH from nonmammalian and mammalian species, including rodents, primates, and man (2). Hormone polymorphism has been disclosed by several techniques, such as gel chromatography (3), chromatofocusing (4), and isoelectric focusing (IEF) (5). It was observed that FSH and LH preparations exhibit charge heterogeneity Received November 16,1990. Address all correspondence and requests for reprints to: Dr. Catherine Ronin, Laboratoire de Biochimie, Faculte de Medecine Secteur Nord, Boulevard P. Dramard, 13326 Marseille Cedex 15, France. * This work was supported by grants from the Assocation pour la Recherche contre le Cancer and the Institut National de la Sante et de la Recherche Medicale (CRE 900505).
bind acidic forms of the hormone (pi 5.8, 5.5, 4.8, and 4.5), in contrast to anti-TSHa antisera, which enlighted the broadest spectrum of isoforms. Monoclonal antibodies of various specificities largely reproduced this distribution, indicating that at least five distinct epitopes are coexpressed in the neutral and alkaline forms of TSH, but only two are expressed in the acidic ones. All of the forms were found to induce cAMP production and stimulate growth of FRTL-5 rat thyroid cells, although neutral forms proved to be definitely less potent than the others. We therefore, conclude that TSH isoforms differ in the expression of both their immunoreactive and bioactive domains and that the bioactive/immunoreactive ratio is not an accurate index for the biopotency of the hormone. (Endocrinology 128: 3259-3268, 1991)
and, therefore, may be separated as an average of seven isoforms in animals such as the rat for the former (5) or
the pig for the latter (6). They have been reported to differ from each other not only in their isoelectric point (pi), but also in their relative proportion, receptor binding, biological activity, and plasma half-life. The overall pattern of gonadotropin polymorphism also appears to vary according to the specific endocrine status of the animal; sex steroids and the hypothalamic stimulatory factor, GnRH, seem to act in concert at the pituitary level to influence both the glycosylation and secretion of the gonadotropins and, therefore, in turn, to affect their biological function at the target gonad cells. On a molecular basis, it has been generally observed that in humans, acidic isoforms of FSH are more abundant and display longer circulatory half-lives and in vivo biological activity but reduced receptor-binding activity than less negatively charged forms (7). This was also shown to apply to human LH, for which these changes could be related to a variable sialic acid content (8), indicating that for both hormones, each isoform may contain different glycosylation patterns. However, no carbohydrate (CHO) structure of the various glycoforms has been reported so far to restrict their difference in biopotency to glycosylation, and it may be postulated that each form can also
3259
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 03:30 For personal use only. No other uses without permission. . All rights reserved.
ISOFORMS OF HUMAN TSH
3260
present subtle changes in its three-dimensional protein structure that account for the observed properties. Much less information is available for TSH, especially that of human origin. Although the biosynthesis, secretion, and regulation of its bioactivity have been widely investigated for this hormone and shown to depend upon CHO processing, its physicochemical and structural characteristics are poorly known. Four biologically active forms of bovine TSH had been isolated from anterior pituitary glands by horizontal IEF (9) in the pH region of 8.2-8.8. Additional species were detected in the normal rat pituitary, distributed over a wider range of pi from 2.5-8.7 (10). The content and distribution of individual isoforms are differentially affected by circulating thyroid hormones (11) or TRH (12). In both cases, the amount of the more acidic forms of TSH appeared to be preferentially regulated compared to that of their more alkaline counterpart. Using zone electrophoresis, it was also shown that human TSH may be composed of up to 20 components likely to differ by their net charge (13), independently of the sex and age of the individual (14). In all experiments TSH was estimated by conventional RIAs, assuming that all isoforms react equally well with polyclonal antibodies (pAbs). Previous work from this laboratory has shown that several antigenic determinants of human TSH are glycosylation-dependent. Using monoclonal antibodies (mAbs), we found that deglycosylated TSH is no longer recognized by anti-TSH/3 antibodies, but is still fully bound by anti-TSHa antibodies (15). In this process, the a-linked CHO chains are more determinant than the 0 one, as if the conformation of the /3-subunit of TSH were under the control of the a-subunit. In an attempt to understand how expression of the functional domains of human TSH is under the control of glycosylation, we separated the various isoforms of the intrapituitary hormone and compared their respective bioactivities to their recognition by antibodies of various specificity.
Materials and Methods Materials Polyacrylamide gel solution was obtained from Pharmacia (Piscataway, NJ), IEF standards from Bio-Rad (Richmond, CA), ampholines from Serva, and polyvinylidene difluoride (PVDF) membranes from Millipore (Bedford, MA). Goat antirabbit immunoglobulins conjugated to alkalin phosphatase were purchased from Promega (Madison, WI), and Steptavidin conjugated to alkalin phosphatase was obtained from Jackson (Bar Harbor, ME). Coon's 12 was obtained from Sigma. All other reagents were of analytical grade. Hormones Human TSH (immuological potency, 5.2-5.5 IU/mg in terms of First International Reference Preparation (IRP-1) Standard;
Endo»1991 Voll28«No6
contaminations by LH in RIA, 2.5-2.8%; by FSH, 1%), LH and a- and /3-subunits were purchased from UCB Co. (Liege, Belgium). Antibodies Rabbit antisera were either purchased from UCB Co. or homemade using the TSH and subunits from UCB as antigens and conventional immunizing procedures. Antibodies directed against the IRPs (First International Standard) were kindly provided by the National Pituitary Agency (Bethesda, MD). mAbs were a generous gift from Dr. M. Benkirane (Immunotech Co., Marseille, France). Free a-subunit was measured using the commercial kit from Biomerica (Newport, CA). IEF separation Vertical ( 8 x 7 cm) polyacrylamide minigels (acrylamidebisacrylamide ratio, 4.5%) were prepared in the presence of ampholines (pH 4-9). Samples were dissolved in 20% ampholines-water before running. Separation at equilibrium was achieved within 1 h, as described previously (16) and monitored using prestained IEF standards. The final pH gradient was measured every 0.5 cm immediately at the end of the run using a contact electrode. Each blot is representative of three to six distinct experiments. When necessary, fractions (2-mm slice) of the IEF gel were eluted overnight in PBS containing 0.1% BSA, and the released material was further assessed for immunoreactivity or before addition to the cells. Silver staining of the gels was performed after denaturation in 50% MeOH, as previously reported (17). Immunoblotting techniques The gels were electroblotted on PVDF membranes using a semidried Millipore system as recommended by the manufacturer. After transfer, saturation of the membranes was carried out in a TBS-Tween buffer [10 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 0.05% Tween-20] for 1 h at room temperature. Adding blocking, proteins such as BSA and casein did not improve saturation of the blots. Primary antibodies at the appropriate dilution in the TBS-Tween buffer were allowed to bind overnight under moderate agitation. After extensive washing, secondary antibodies were reacted for 2 h, and alkaline phosphatase activity revealed using tetrazolium blue and 5bromo-4-chloro-3-indoxylphosphate in 0.1 M Tris-HCl, pH 9.5, buffer containing 0.1 M NaCl and 5 mM MgCl2 for 10 min at room temperature. The reaction was stopped by adding 10 mM EDTA. TSH RIA TSH (~50 tiCi/ng) was iodinated with Iodogen, as described previously (18). The iodinated ligands (50 X 103 cpm) were incubated with or without unlabeled IRP-1 hormone or IEF fractions in the presence of anti-IRP TSH (final dilution, 1:2,500,000) or anti-IRP TSHa (final dilution, 1:180,000) in a final volume of 300 ^1 50 mM PBS buffer, pH 7.5, containing 0.1% BSA overnight at 4 C. Nonimmune rabbit serum (100 /il; dilution, 1:100) and horse rabbit immunoglobulins (100 n\; dilution, 1:4) were then added for 2 h at 4 C. Immune complexes were finally precipitated with 6% polyethylene glycol 6000 in a
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 03:30 For personal use only. No other uses without permission. . All rights reserved.
ISOFORMS OF HUMAN TSH final volume of 1 ml and recovered by centrifugation at 10,000 X g for 15 min. All assays were performed in duplicate, and the nonspecific binding was 0.5% of the total radioactivity. TSH bioassays TSH biological potency was assessed in an heterologous bioassay using a cloned rat thyroid cell line, FRTL-5. Cells were routinely cultured in Coon's F-12 medium supplemented with 5% fetal calf serum (FCS) and six hormones [bTSH (10 mU/ml), insulin (10 jug/ml), transferrin (5 Mg/ml), hydrocortisone (10 nM), somatostatin (10 ng/ml), and glycyl-L-histidylL-lysine acetate (10 ng/ml)] and growth factors, as previously described (19). Incorporation of [me£/iy/-3H]thymidine into DNA was performed using cells seeded in 24-well plates (5 X 10" cells/0.5 ml) in Coon's F-12 medium containing 5% FCS and five hormones (FCS-5H), but omitting TSH. Four days later, the culture medium was replaced by 0.5 ml of the same medium containing the IEF fraction to be assayed and [methyl3 H]thymidine (1 AtCi/ml). The plates were then incubated for 2 days at 37 C in 5% CO2, after which the cells were carefully washed with ice-cold PBS. Then, 7.5% trichloroacetic acid was added for 20 min at 4 C before the cells were solubilized in 0.1 N NaOH-1% sodium dodecyl sulfate (SDS). After 1-h incubation at room temperature, cell-associated radioactivity was counted in a Packard liquid scintillation counter (Downers Grove, IL). The results are expressed as the mean of quadruplicate determinations. The half-maximal effect of TSH on cell growth is routinely achieved for 10 ng TSH/well, while a plateau is reached above 50 ng TSH/well. cAMP production was measured under the following conditions. FRTL-5 cells were grown at confluence in Coon's F-12 medium containing 5% FCS and six hormones, then cells were refed with the same medium, omitting TSH, for 5-10 days. On the day of the experiment, the cells were carefully washed by a hypotonic medium containing 5.4 mM KC1, 1.3 mM CaCl2, 0.4 mM MgSO4, 0.4 mM Na2HPO4, 0.44 mM KH2PO4, 5.5 mM glucose, 0.5 mM isobutylmethylxanthine, and 0.4% BSA in a 10-mM HEPES buffer, pH 7.4. Samples of TSH were added to the wells in this medium, and the incubation was performed for 1 h at 37 C. Cell culture media were then rapidly recovered at 4 C, and extracellular cAMP was measured in triplicate, using the RIANEN [125I]cAMP RIA kit (DuPont, Dreiech, Germany).
Results Separation of TSH isoforms We first investigated the distribution of TSH native isoforms as a function of their isoelectric point by direct silver staining of the gels and analyzed their reactivity toward pAbs after electrotransfer. Figure 1 shows the comparative separation of TSH and individual subunits under both conditions. As can be seen in Fig. 1A, silver staining identified a highly purified preparation of TSH as seven major bands exhibiting different pi values, ranging from 5.5-8.6 (lane 1), whereas a- and /3-subunits were less heterogeneous, as
3261
they displayed one major species of 8.8 for a (lane 2) and two closely related species of pi 8.7 and 8.1 for /? (lane 3). Alkaline pi values for the subunits were largely expected, as they originate from the hormone by acid dissociation. Such treatment probably results in a concomittant loss of sialic acid, which is highly labile under low pH conditions. The silver-staining pattern observed for TSH and its subunits is likely to represent the species quantitatively predominant in the intrapituitary preparation. Figure IB shows the recognition of TSH isoforms by an anti-TSH antiserum as a function of hormone concentration. It appeared that, as found by silver staining, as many as seven bands could be visualized, with pi values ranging from 6-5-8.6 and comigrating with the species stained with silver (lanes 4-7). As little as 250 ng TSH could be separated and detected on immunoblots, whatever the specificity of the antibody or the hormone material analyzed. In Fig. 1C, is presented the corresponding separation of the individual subunits of TSH revealed by their respective antibodies. It appears that in contrast to silver staining, several bands of neutral to acidic pis could be attributed to both a-subunit, for which repetitive doublets were observed (pi 8.8, 8.4-8.2, 7.6-7.4,6.8-6.6, and 5.8-5.4; lane 8), and /3-subunit (pi 8.7, 8.1, 7.2, 6.8, 6.2, and 5.8; lane 10). These forms putatively represent minor components of the subunit preparation differring in their CHO content or unknown proteolytic fragments. As expected, none of the antibodies used was able to bind the opposite subunit (lanes 9 and 11). These findings indicate that the distribution pattern of the hormone is quite distinct from that of its subunits. This rules out as well a putative recognition of free subunits which could overlap with that of TSH isoforms. Indeed, our TSH preparation proved to be highly purified by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and silver staining, and revealed the presence of less than 2% of free a by a specific immunoassay (data not shown). This amount is far below what can be detected by the current immunoblotting technique (see Fig. IB). Specificity of antibody recognition To assess whether antibodies to TSH cross-react with gonadotropins on blots as they can do in conventional immunoassays, we compared the respective distribution and immunoreactivity of LH and TSH in IEF gels (Fig. 2).As expected, both anti-TSHa and anti-LHa antibodies reacted well with all isoforms of TSHa (lanes 1 and 2) as well as with those of LHa (lanes 3-4). Anti-LH and anti-LHa antibodies revealed a virtually identical distribution pattern when tested on the gonadotropin (lanes 4-5). Anti-TSH antibodies proved to be specific for TSH (lane 6) and unable to bind LH (lane 7), indicating that
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 03:30 For personal use only. No other uses without permission. . All rights reserved.
ISOFORMS OF HUMAN TSH
3262 TSH TSH-
