Search

Free alpha molecules isolated from pregnancy, as well as highly purified reference preparations of hCG alpha subunit (CR119 or CR123), stimulated the release of prolactin from human decidual cells in culture. The amount of prolactin secreted during a 24 h incubation was concentration-dependent over a range of increasing doses of alpha from 0.2 to 20 ng/ml with an ED50 of about 1.6 ng/ml. These concentrations are well within the physiologic maternal serum free alpha levels which average 350 ng/ml during the third trimester of pregnancy. Incubation of decidual cells with a reference preparation of intact hCG (CR123) at a concentration of 260 ng/ml resulted in stimulated secretion of prolactin, however, the observed stimulation could be attributed to contamination of the preparation with free alpha or dissociated hCG alpha subunit. Purified hCG beta subunit had no stimulatory activity on the decidual cell culture. The effect of alpha subunit on the stimulated release of prolactin was not due to a generalized stimulation of protein synthesis and secretion since no increase was observed in the release of 35S-labeled proteins compared to controls. In addition, the observed increase in prolactin secretion was not due to a toxic effect of the alpha subunit since there was no visible effect on cell viability, and the cellular enzymes, LDH and alkaline phosphatase, were not detected in the culture medium. Addition of exogenous hCG alpha subunit to primary cultures of human trophoblast cultures did not result in stimulated release of human placental lactogen. We conclude that free alpha molecules of pregnancy stimulate release of prolactin from human decidual cells in culture. These results suggest a novel role for free alpha in the paracrine regulation of decidual prolactin secretion.

The carbohydrate compositions of pregnancy-derived hCG alpha (dissociated from intact hCG) and free alpha-subunit were analyzed using a combination of chemical analysis, lectin affinity chromatography, and glycosidase sensitivity. For direct compositional analysis, parallel samples were hydrolyzed in trifluoroacetic acid and analyzed for sialic acid and neutral sugars without prior derivatization. Separation of the monosaccharides was achieved by HPLC on a Dionex CarboPac column eluted at high pH, and the resolved monosaccharides were quantified by pulsed amperometric detection. The amounts of sugar that were found relative to peptide indicated the presence of two N-linked oligosaccharides per molecule on both hCG alpha and free alpha. Free alpha contained 2.5-fold higher amounts of sialic acid and galactose as well as a higher amount of N-acetylglucosamine than did hCG alpha. Free alpha also contained a 6-fold higher amount of fucose than did hCG alpha (1.2 vs. 0.2 residues of fucose/molecule). Serial fractionation of intact hCG alpha and free alpha molecules by lectin affinity chromatography indicated that virtually all of the hCG alpha-subunits contained at least one Concanavalin-A (Con-A)-binding site, whereas as many as 32% of the free alpha molecules could not bind to Con-A. Chromatography on Lens culinaris (Lch) resulted in 12% binding of hCG alpha and approximately 72% binding of free alpha (80-85% of the Con-A-bound free alpha and 47-48% of the Con-A-nonbound free alpha bound to Lch). Endoglycosidase-H (endo-H) treatment of hCG alpha released a portion of the oligosaccharides. The endo-H-released material appeared to be a monoantennary hybrid based on DEAE-binding properties and carbohydrate composition. In contrast to hCG alpha, free alpha was completely resistant to endo-H treatment. Incubation of endo-H-resistant hCG alpha with glycopeptidase-A resulted in the release of two components, which could be separated into monoantennary and biantennary fractions on the basis of size and charge. The collective data suggest that hCG alpha contains primarily monoantennary hybrid oligosaccharide structures and relatively little fucose. In contrast, free alpha contains primarily multiantennary oligosaccharide structures, and most of the free alpha molecules contain at least one oligosaccharide with fucose attached to the asparagine-linked N-acetylglucosamine residue.

Pregnancy urine contains large quantities of hCG, free beta-subunit, free alpha-subunit, and a population of fragments of beta-subunit known as beta-core. This beta-core population, which can account for as much as 70% of the total beta-immunoreactivity in pregnancy urine, is of interest as both a normal metabolite of pregnancy and a potential marker for malignancy. We have purified the beta-core fragment from pregnancy urine (P-core) and have characterized it with respect to size and carbohydrate composition. P-Core was purified by chromatography on Sephadex G-100, Concanavalin A (Con A)-Sepharose, DEAE-Sephacel, and Sephadex G-75 (superfine). The purified P-core has an apparent mol wt of 17,500 and 17,000, as determined by gel filtration on Sephadex G-75 (superfine) and sodium dodecyl sulfate-polyacrylamide gel electrophoresis under nonreducing conditions, respectively. The sialic acid content of P-core was assayed chemically and was less than 0.07 mumol sialic acid/mumole P-core. For comparison to P-core, we have prepared a trypsin fragment of beta-subunit that retains the beta-core immunological determinant recognized by SB6 antiserum and lacks the carboxy-terminal immunological determinant. We have designated this beta-core molecule as T-core (tryptic fragment of beta-subunit) to distinguish it from the beta-core molecule that we have purified from pregnancy urine (i.e. P-core). Most of the P-core and T-core molecules bind to Con A (84% and 86%, respectively). The Con A-bound material was used for subsequent characterizations. Neither P-core nor T-core binds to DEAE using conditions under which intact hCG beta binds to DEAE. A variety of agarose-bound lectins were used to further investigate the carbohydrate nature of the Con A-bound P-core and T-core molecules. The lectin binding data indicate that the antennae on P-core do not contain appreciable sialic acid or galactose, in contrast to the antennae on T-core, which contain both. We conclude that P-core, the naturally occurring beta-core fragment in pregnancy, has been processed to a form in which nearly all of the sialic acid and galactose residues are removed, but the Con A-binding site (consisting of the core sugars) and most of the core fucose are retained.

It is well known that the MCR of proteins can be influenced by their carbohydrate structure; i.e. the presence of terminal galactose on proteins results in uptake by hepatic receptors for galactose-terminated glycoproteins. Thus, a protein with its galactose residues covered or removed exhibits a far longer half life in plasma than one with its galactose residues exposed. The free alpha-subunit of human CG (hCG) has been shown to have a different carbohydrate composition than does the alpha-subunit dissociated from the intact hormone. In our laboratory, analysis of alpha-subunits isolated from pregnancy urine indicated that the alpha-subunit dissociated from hCG (hCG alpha) contains primarily monosialylated oligosaccharides, whereas the free alpha-subunit contains more than one sialic acid per oligosaccharide. This difference in the degree of sialylation prompted us to examine the clearance rates of these two subunits. Accordingly, free alpha and hCG alpha were purified by affinity chromatography and labeled with 125I. The labeled subunits were injected iv into rats and serum samples were removed at various time intervals over a 2-h period. The amount of [125I]alpha-subunit remaining in the serum was determined by immunoprecipitation using an antiserum to hCG alpha. The disappearance curves for the two subunits were indistinguishable and could be analyzed by a biexponential model. The t 1/2 of the faster component was 5 min, while the t 1/2 of the slower component was 74 min. In order to determine whether or not terminal galactose was present on either the hCG alpha or the free alpha-subunit, the labeled molecules were subjected to lectin column chromatography using Ricin or peanut lectin linked to agarose. Both of these lectins bind galactose but have different specificities with respect to the penultimate sugar. Both subunit preparations contained only minor amounts of material which could bind or either lectin. However, after desialylation, both hCG alpha and free alpha bound extensively to Ricin, indicating the presence of penultimate galactose residues in both. We conclude that terminal galactose residues are not present on the oligosaccharides of either hCG alpha or free alpha-subunits, and that the difference observed in the sialic acid contents of the two subunits does not affect their rates of clearance.

The glycoprotein hormone hCG and a free alpha-subunit are secreted by trophoblastic cells during pregnancy. We have purified the alpha-subunit of hCG and the free alpha-subunit population from pregnancy urine. Dissociation of hCG with 10 M urea at 37 C, followed by chromatography on DEAE-Sephacel, resulted in separation of alpha- from beta-subunit, as hCG alpha does not bind to DEAE in the presence of urea, while beta-subunit does bind. Similar treatment of the free alpha population resulted in fractionation into two populations, a nonbinding fraction of free alpha and a population which was retained by DEAE in the presence of urea (free alpha 2). The three populations, hCG alpha, free alpha, and free alpha 2, were further purified by affinity chromatography using anti-alpha antisera linked to Sepharose. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of the preparations showed that hCG alpha consisted of a single component with an apparent mol wt of 22,000, while free alpha and free alpha 2 consisted of multiple components. Radioactive labeling of sialic acid by limited periodate oxidation and NaB[3H]4 reduction resulted in a higher specific activity for free alpha than for hCG alpha, suggesting that free alpha contains more sialic acid per immunoreactive molecule than does alpha dissociated from hCG. [3H]hCG alpha, but not 3H-labeled free alpha, was able to combine with native hCG beta-subunit. The oligosaccharide moieties were released from the different labeled subunits by alkaline hydrolysis and then compared with respect to Concanavalin A (ConA)-binding and DEAE-binding properties. Most of the oligosaccharides from dissociated hCG alpha bound to ConA-Sepharose (72%), while less material from free alpha (40%) and even less from free alpha 2 (25%) were capable of binding to ConA. DEAE chromatography of the oligosaccharides suggested that hCG alpha contained primarily monosialylated forms (greater than 60%), while free alpha and alpha 2 contained primarily (greater than 70%) di- and trisialylated forms. Thus, the ConA and DEAE binding data indicated that the oligosaccharide contents of free alpha and free alpha 2 were quite different from that of hCG alpha. These results also suggest that some of the oligosaccharide structures contained on hCG alpha and most of those on free alpha and free alpha 2 differ substantially from the structures that have been previously described.

Modifications of carbohydrate structures of hCG, such as deglycosylation or desialylation, have been shown to reduce the biological activity of the hormone derivatives in vivo. We posed the question of whether deglycosylated hCG (dg-hCG) and desialylated hCG (ds-hCG) would behave as agonists at the LH/CG receptor in the primate in vivo, as this would bear on their potential clinical utility as LH/CG agonists or antagonists. Thus, we administered large doses (approximately 3 nmol) of highly purified dg-hCG, ds-hCG, hCG, or normal saline as a rapid iv injection to adult male cynomolgus monkeys (n = 3/group). Mean areas under the curves of plasma T over the first 6 h achieved with dg-hCG and ds-hCG were about 5-fold, significantly (P less than 0.05) greater than that in the saline controls and not significantly (P greater than 0.05) different from that in hCG-injected animals. Despite comparable plasma T responses in the first 6 h, mean plasma concentrations of ds-hCG, dg-hCG, and hCG differed dramatically among the groups. Plasma ds-hCG and dg-hCG levels were undetectable by 15 and 180 min, respectively, while the mean plasma hCG level was more than 2.10 nmol/L at 360 min. These data indicate that 1) dg-hCG is a full agonist at the LH/CG receptor in the primate in vivo, despite having minimal intrinsic activity in the rat Leydig cell adenyl cyclase assay and being able to near-completely antagonize hCG action therein; and 2) ds-hCG is a full agonist in the monkey in vivo, capable of stimulating a full testicular response over 6 h, despite being cleared from the circulation in 15 min. We conclude that the signal transduction system at the monkey LH/CG receptor is capable of achieving full steroidogenesis despite dramatically shortened exposure to stimulus or exposure to a stimulus with markedly reduced adenyl cyclase-stimulating activity in vitro.

We devised an in vivo biological assay for ovarian growth inhibiting activity to examine extracts of human pregnancy urine for the presence of ovarian growth inhibiting factor. Diethylstilbestrol (DES) capsules were implanted sc in immature hypophysectomized female rats; FSH was injected sc with or without test substance for 5 days. Rats with unstimulated ovaries were implanted with blank capsules and given the vehicle without FSH. Twenty four hours after the last injection, the ovaries were removed and weighed. The ovarian growth inhibition of the ovarian weight gain achieved in rats treated with DES and FSH. Crude commercial human CG (hCG) preparations, extracted from pregnancy urine, were chromatographed on Sephadex G-100, and the fractions were tested for ovarian growth inhibiting activity. The peak of ovarian growth inhibiting activity was found in fractions eluting from the column in the mol wt range of 12,000-20,000. Ovarian growth inhibiting activity was heat sensitive, not extracted by ether, and precipitated by acetone. Ovarian growth inhibiting activity was stable in acid at pH 2, but was inactivated by digestion with trypsin. The ovarian growth inhibiting activity was purified by chromatography on Concanavalin A-Sepharose and diethylaminoethyl-Sephacel. The active material contained hCG alpha, hCG beta, and beta-carboxyterminal peptide-immunoreactivity and its inhibiting activity could be removed from solution by immunoadsorption with antisera specific for hCG beta. The ovarian growth inhibiting activity was further purified on an anti-hCG alpha-immunoglobulin G affinity column. The activity was eluted from the affinity column at low pH, and eluted material contained all of the immunodeterminants of hCG. Virtually identical dose-response curves of ovarian inhibition were obtained using equivalent doses of beta-carboxyterminal peptide immunoreactivity of purified inhibitor and purified hCG (CR123). The inhibiting activity reached plateau of 80-90% at doses of 50-100 ng hCG/rat. Upon rechromatography on Sephadex G-100, the ovarian growth activity that was pooled from fractions corresponding to the 12,000-20,000 mol wt range was recovered in fractions corresponding to the elution position of hCG. We conclude that the low mol wt inhibiting activity observed in the crude pregnancy extracts is due to hCG and that hCG is a very potent inhibitor of FSH/DES stimulation of ovarian growth.

beta-Core is a major component of the hCG-related molecules found in pregnancy urine. We previously have purified the beta-core molecule and have deduced portions of its carbohydrate structure based on lectin binding data. In the present study we used recently developed technology to determine the carbohydrate composition of beta-core and hCG beta (CR119). For direct compositional analysis, parallel samples were hydrolyzed in trifluoroacetic acid and analyzed for sialic acid and neutral sugars without prior derivatization. Separation of the monosaccharides was achieved by HPLC on a Dionex CarboPac column eluted at high pH, and the resolved monosaccharides were quantified by pulsed amperometric detection. The amounts of sugar that were found relative to peptide indicated the presence of two N-linked oligosaccharides per molecule on both beta-core and hCG beta. hCG beta contained additional sugars consistent with the presence of four O-linked oligosaccharides. Compared to hCG beta, beta-core contained negligible sialic acid, galactose, or N-acetylgalactosamine. The compositional data suggest that beta-core does not contain N-acetylglucosamine at the nonreducing end of the molecule, whereas the trimannosyl-chitobiose core is apparently intact at both glycosylation sites, consistent with the ability of the molecule to bind to Concanavalin-A. Comparable fucose contents and abilities of beta-core and hCG beta to bind to Lens culinaris indicate a similar extent of fucosylation on the internal N-acetylglucosamine in both molecules. We propose that the N-linked oligosaccharides on beta-core closely resemble the underlying N-linked structures of hCG beta with the antennary sialic acid, galactose, and N-acetylglucosamine removed.