Development of a long-acting erythropoietin by fusing the carboxyl-terminal peptide of human chorionic gonadotropin beta-subunit to the coding sequence of human erythropoietin

Human erythropoietin (EPO) is a glycoprotein hormone secreted from the kidney and controls red blood cell production. EPO has a wide clinical use in the treatment of anemia associated with renal disease, certain chronic diseases, and anemia related to chemotherapy and radiotherapy. One major issue regarding the clinical use of EPO is its relatively short half-life due to its clearance by glomerular filtration. Thus, the therapeutic protocol used in the treatment of patient-required frequent injections of EPO. To address this issue, we constructed a chimeric gene that contains the sequence of the carboxyl-terminal peptide (CTP) of human chorionic gonadotropin-beta subunit bearing four O-linked oligosaccharide recognition sites and the coding sequence of human EPO cDNA. Fusing the CTP to the carboxyl-terminal of EPO did not affect secretion, receptor binding affinity, or in vitro bioactivity. However, both in vivo potency and half-life of EPO-CTP were significantly enhanced. A single injection dose (660 IU/kg) of EPO wild-type administered once a week had no significant effect on haematocrit levels. However, EPO-CTP administered as 660 IU/kg once a week was effective as well as the same total dose of EPO wild-type administered as 220 IU/kg three times a week. This may emphasize the importance of sustained blood levels rather than total dose of administration for in vivo bioactivity. These data established the rationale for using this chimera as a long-acting EPO analog. The therapeutic efficacy of EPO-CTP analog needs to be established in higher animals and human clinical trials.     

Structure of the human chorionic gonadotropin beta-subunit fragment from pregnancy urine

A major portion of the hCG immunoreactivity detectable in pregnancy urine is derived from a fragment of hCG beta. This lacks the COOH-terminal portion of hCG beta, but retains immunoreactivity with most antibodies raised against the beta-subunit of hCG. To improve clinical measurements of hCG and assess the importance of such fragments in human urine, we have isolated and determined the structure of this molecule. The hCG beta fragment was isolated from a partially purified commercial preparation of hCG (Organon) by gel filtration and immunoaffinity chromatography using monoclonal antibodies. It was found to consist of two polypeptide chains composed of residues beta-(6-40) disulfide-bridged to residues beta-(55-92). It also differs from the beta-subunit of hCG in its carbohydrate structure, lacking sialic acid and having a low but variable amount of galactose. A beta-fragment containing the same two NH2-terminal sequences was also isolated from a single pregnant woman's urine. The two major polypeptides comprising the beta-fragment contain a total of nine half-cystine residues, raising the possibility that a free thiol may exist or that a third undetected disulfide-bridged peptide is present in the intact fragment. However, tests for the presence of a free thiol have been negative. Another intrinsic characteristic of the beta-fragment is the formation of a variable amount of dimer in solutions of neutral pH. beta-fragment will not combine with intact alpha-subunit. Despite the absence of regions beta-(1-5), beta-(41-54), and beta-(93-145), the beta fragment is recognized by the SB-6 antibody and most monoclonal antibodies elicited to the beta-subunit, thus excluding half of the amino acids of the beta-subunit from the epitope(s) where these antibodies bind.     

Fragmentation of the beta-subunit of human chorionic gonadotropin produced by choriocarcinoma

When human chorionic gonadotropin (hCG) was subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis under reducing conditions with dithiothreitol (DTT), a smaller weight material (CTP'), in addition to the beta-subunit, could be detected by Western blot analysis using antiserum for hCG beta-carboxy-terminal peptide (CTP). The CTP' band was much more apparent with urinary hCG from a patient with choriocarcinoma than with that from normal pregnant women. Second-dimensional electrophoresis of the choriocarcinoma hCG (c-hCG) after reduction with DTT indicated that the CTP', Mr 25,000, was released from the beta-subunit. The carbohydrate structure of the CTP' was analyzed by affinity with lectin-peroxidase on a nitrocellulose membrane. The CTP' did not interact with Concanavalin A, but exhibited strong interaction with both RCA120 and Arachis hypogaea after removal of sialic acid, indicating that it was released as a fragment containing an O-linked sugar chain as was found in the hCG beta carboxy-terminal portion. Western blot analysis using the antisera for hCG alpha, hCG beta, and hCG beta-CTP showed that the CTP' contains not only the carboxy-terminal portion but also a part of the internal (core) portion of the beta-subunit molecule. This dissociation of the c-hCG beta was further supported by the presence of a faster moving component (FMC) which may correspond to the NH2-terminal side counterpart. The desialylated FMC could be detected by Concanavalin A and RCA120 but not by Arachis hypogaea, indicating that it contains N-linked rather than O-linked sugar chains. The FMC does not contain any of the epitopes for the antisera examined in Western blot. These results indicate that the beta-subunit of the choriocarcinoma urine hCG has an unusual site which is dissociated into two components of Mr 25,000 (CTP') and Mr 18,000 (FMC) by DTT reduction.     

A radioimmunoassay for the core fragment of the human chorionic gonadotropin beta-subunit

We developed a RIA for the beta-core fragment of hCG that is excreted in the urine of pregnant women and some patients with cancer. We purified beta-core from crude commercial preparations of hCG (of which beta-core is a major constituent) derived from pregnancy urine and used this purified beta-core material to immunize rabbits. One antiserum (RW25) was particularly useful in that a RIA with purified beta-core as both radioligand and reference preparation had high sensitivity for beta-core detection and low cross-reactivity with other hCG-related molecules and glycoprotein hormones. The cross-reactivities of purified hCG (CR125), hCG beta (CR125), and hCG alpha (CR125) preparations were in each instance less than 3 x 10(-3) (wt/wt). The cross-reactivities of purified pituitary glycoprotein hormones were in each instance less than 2 x 10(-4) (wt/wt). Using the RW25 RIA, virtually all beta-core immunoreactivity in pregnancy urine eluted from Sephadex G-100 in a position coincident with that of purified beta-core. Urine from men and nonpregnant women contained very low levels of beta-core immunoreactivity (less than 6.5 micrograms/L), while urine from pregnant women and patients with testicular cancer or other neoplasms had levels of beta-core immunoreactivity ranging as high as 26,000 micrograms/L. We conclude that the improved specificity of our beta-core RIA will facilitate studies of the physiology and cancer biology of beta-core molecules.     

Processing of O-linked glycosylation in the chimera consisting of alpha-subunit and carboxyl-terminal peptide of the human chorionic gonadotropin beta-subunit is affected by dimer formation with follicle-stimulating hormone beta-subunit

hCG, LH, FSH, and TSH are a family of heterodimeric glycoprotein hormones that contain a common alpha-subunit, but differ in their hormone-specific beta-subunits. hCGbeta is unique among beta-subunits due to a carboxyl-terminal peptide (CTP) bearing four O-linked oligosaccharides. We previously reported that there were differences in O-glycosylation between two chimeras consisting of alpha-subunit and CTP, i.e. a variant with CTP at the N-terminal region (Calpha) and another analog with CTP at the C-terminus (alphaC) of the alpha-subunit. To address whether O-glycosylation is influenced by the heterodimer formation, Calpha and alphaC were expressed alone or with FSHbeta-subunit in Chinese hamster ovary cells. The O-linked glycosylation was assessed by continuous labeling with [(35)S]methionine/cysteine, immunoprecipitation with anti-alpha or anti-FSH serum, serial digestion with endoglycosidase-F and neuraminidase, and sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The decrease in molecular weight of dimeric chimeras digested with endoglycosidase-F was greater in Calpha than that in alphaC after treatment with neuraminidase, revealing that both chimeras have different numbers of sialic acids on O-linked carbohydrates. By treating with endoglycosidase-F, the dimeric alphaC migrated faster than its free form, whereas the mobility difference between assembled and unassembled forms of Calpha was very little. These data indicate that processing of O-glycosylation is affected by the backbone polypeptide chain(s).     

Radioimmunoassay of the serum free beta-subunit of human chorionic gonadotropin in trophoblastic disease

We developed a RIA specific for the free beta hCG employing anti-beta hCG monoclonal antibody 1D12. This RIA was highly sensitive to free beta hCG; the minimum detectable concentration was 0.4 ng/ml. alpha hCG, LH, beta LH, and FSH had little effect in the assay; the cross-reactivity of hCG was about 4%. Using this RIA, we measured serum free beta hCG concentrations in 38 normal pregnant women and 72 untreated patients with 3 types of trophoblastic disease: hydatidiform mole (n = 15), invasive mole (n = 29), and choriocarcinoma (n = 28). All of these samples were simultaneously assayed for hCG by RIA. In normal pregnant women, serum hCG changed as pregnancy progressed, but serum free beta hCG was not detected at any time. In contrast, serum free beta hCG was measurable in the majority of patients with trophoblastic disease. Strong correlations were found between the concentration of free beta hCG and that of hCG in each type of trophoblastic diseases. The mean free beta hCG to hCG ratio was lowest for hydatidiform mole and highest for choriocarcinoma, and the difference between the ratios in these 2 groups was statistically significant. Serial measurements in 7 patients with trophoblastic disease failed to reveal remarkable changes in the free beta hCG to hCG ratio throughout their clinical course. We conclude that the production of free beta hCG increases with the immaturity of the trophoblastic cell, and the degree of differentiation of trophoblastic cells may be reflected by the free beta hCG to hCG ratio.     

Structural and functional studies of the tryptic core of the human chorionic gonadotropin beta-subunit

The beta-subunit of hCG was digested with trypsin to produce a modified form of the subunit for structure-function and immunological studies. After digestion of hCG beta with trypsin, the residual disulfide-linked core was isolated and found to be lacking the carboxy-terminal peptide (residues 115-145) and to contain bond cleavages between residues 2-3, 43-44, 74-75, and 95-96. The locations of these bond cleavages within the disulfide-bridged core were identified by isolation of the following peptides after reduction and S-carboxymethylation of the trypsin beta-core: beta 1-43, beta 3-43, beta 44-74, beta 44-95, beta 75-95, and beta 96-114. The circular dichroic spectrum of the tryptic beta-core over the wavelength region of about 200-320 nm was similar to that of the native subunit. In addition, the tryptic beta-core retained nearly full immunopotency in both polyclonal and monoclonal competitive RIAs and could combine with complementary native alpha-subunit. The hybrid, composed of the tryptic beta-core and native alpha, was purified and displayed a molar potency of about 0.1% relative to intact hCG in both a radioreceptor assay and an adenylate cyclase assay. Thus, the hybrid retained little biological activity. Although the extensive bond cleavages in the tryptic beta-core did not appear to change its secondary and tertiary structure sufficiently to significantly alter the circular dichroic spectrum, the immunoreactivity, or the capability to combine with its alpha-subunit complement, the biological functional integrity of the tryptic beta-core-containing hybrid was essentially abolished. Hence, the tryptic beta-core provides a useful derivative for detailed structure-function studies aimed at defining the necessary determinants for subunit association, receptor binding, and subsequent biological actions.     

Combination of the chorionic gonadotropin free beta-subunit with alpha.

The rate-limiting event for combination of hCG alpha- and beta-subunits in JAR choriocarcinoma cells is the rate of disulfide bond formation in the beta-subunit. This is accompanied by a conformational change that produces a combination-competent form of the beta-subunit. The combination reaction, however, is incomplete, and 50% of the synthesized beta molecules remain uncombined (free). In addition, 70% of biosynthetically labeled free beta is degraded in the cell. Possible explanations for incomplete dimer formation include 1) biochemical differences between the free and combined beta-subunits that limit combination of free beta, and 2) an inefficient combination reaction due to low intersubunit affinities or limiting concentrations of combination-competent subunits within the cell. To examine whether the biochemical differences between free and combined beta-subunits that we have previously observed affect the combined beta-subunits that we have previously observed affect the combination competence of free beta, free and dimer beta-subunits were purified from the culture medium and lysates of JAR cells and examined for their ability to combine with alpha purified from pregnancy urine in an in vitro combination assay. Secreted free and dimer beta obtained from culture medium combined to the same extent with urinary alpha. Although the combination efficiencies were lower for the intracellular forms, the free and dimer beta-subunits purified from cell lysates also combined to the same extent with urinary alpha. Thus, biochemical differences that exist between the beta forms do not prevent combination of free beta with alpha in an in vitro combination assay. To examine the second possibility, we speculated that if high concentrations of hCG subunits remained in the rough endoplasmic reticulum (ER) for extended periods of time, the extent of dimer formation would increase in the cell. To increase the residence time of hCG subunits in the ER, JAR cells were treated with carbonyl cyanide trifluoromethoxyphenylhydrazone, an agent that inhibits the translocation of hCG subunits from the ER to the Golgi. Treatment of cells with trifluoromethoxyphenylhydrazone in long and short term pulse-chase labeling studies did not result in an increase in the extent of dimer formation. Thus, the subunit combination reaction in JAR cells may be incomplete due to subtle conformational differences in the free beta-subunit; however, these differences do not inhibit the combination of the free beta-subunit in vitro.     

Preparation and characterization of antibodies to the urinary fragment of the human chorionic gonadotropin beta-subunit

hCG is a glycoprotein hormone composed of two dissimilar subunits (alpha and beta) and is normally synthesized by trophoblastic tissue. Its measurement by immunoassay is widely employed as a test for pregnancy, but can be complicated by cross-reactivity with human (h) LH. Immunoassays based on the beta-subunit of hCG have been employed to decrease this cross-reactivity with hLH, but when these assays are used with urine specimens, the antibodies employed also detect a fragment of hCG beta, which can lead to significant differences in measurement. To overcome these problems, we have developed a series of monoclonal antibodies to the beta fragment of hCG recovered from pregnancy urine. Some of the antibodies that bind to this beta fragment are directed to a region of hCG beta that is different from the epitopes recognized by antibodies raised against the intact beta-subunit. The new epitopes available in the hCG beta fragment form the basis for novel immunoassays. These beta fragment antibodies are used in conjunction with other antibodies, directed to different epitopes of the hormone, to produce a series of immunoradiometric assays that can discriminate among intact hormone, free hCG beta, and hCG beta fragment. The hCG beta fragment antibodies described herein have affinities between 10(9) and 10(11) M-1 for the beta fragment and exhibit varying degrees of discrimination between the hCG beta fragment, the beta-subunits of hCG and hLH, and intact hCG and hLH.     

Single-chain human chorionic gonadotropin analogs containing the determinant loop of the beta-subunit linked to the alpha-subunit

Human chorionic gonadotropin (hCG) is a member of the family of glycoprotein hormones containing a common alpha-subunit and distinct beta-subunits that confer hormonal specificity. hCG binds to the relatively large ectodomain of the human luteinizing hormone receptor (hLHR), a member of the G protein-coupled receptor superfamily, leading to increased intracellular production of cAMP. Using protein engineering, two miniaturized versions of hCGbeta have been separately fused to the N-terminus of the alpha-subunit to give N-des[1-91]hCGbeta-alpha-C and N-des[1-91,110-114]hCGbeta-alpha-C, i.e. fusion proteins of the hCGbeta determinant loop (extended to include the complete seat belt and carboxy-terminal peptide) coupled to the alpha-subunit. Bioactivity of these single-chain gonadotropin analogs was assessed in two systems following transient transfections into HEK 293 cells and subsequent cAMP measurements. In one, each mini-beta-alpha cDNA was fused to that of hLHR and transfected into cells to create yoked miniaturized hCG-hLHR complexes; in the other, the cDNA of each single chain mini-beta-alpha was co-transfected with that of hLHR in an effort to produce non-covalent miniaturized hCG-hLHR complexes. Using yoked hCG-hLHR and hLHR as positive and negative controls respectively, expression of each mini-hCG-hLHR complex was confirmed using antibody and ligand binding assays. The two mini-hCGs led to minimal activation of hLHR, suggesting weak intrinsic activity of the mini-beta-alpha fusion proteins. These results suggest that potent agonists and antagonists will require the presence of other portions of hCGbeta in addition to the determinant loop/seat belt.     

A human chorionic gonadotropin-specific antiserum against synthetic peptide analogs to the carboxyl-terminal peptide of its beta-subunit.

A chemically synthesized Na-acetyl-triacontapeptide analogous to the carboxyl-terminal region (residues 116--145) of hCG beta-subunit was conjugated to bovine thyroglobulin. This synthetic antigen was used to immunize five rabbits; one of these rabbits (H-114) generated a useable antiserum. The binding properties of this antiserum were compared extensively to a well characterized antiserum (H-93) against the natural tricosaglycopeptide (residues 123--145) of desialylated hCG beta-subunit conjugated to bovine serum albumin. The two antisera showed nearly identical immunological specificities and sensitivities when examined in a conventional RIA system. These similarities included: 1) antibody recognition sites residing at the carboxyl-terminal pentadecapeptide region; 2) binding neither [125I]iodo-hLH nor [125I]iodo-oLH; and 3) lack of neutralizing capability against the biological activity of hCG in vivo. However, synthetic pentadeca or longer peptides are twice as potent in inhibiting 125I]iodo-hCG binding to H-114 than to H-93. Synthetic peptides may provide a superior means of generating antisera capable of distinguishing hCG from human LH, since they can be obtained more easily in large quantities than analogous purified natural fragments and because peptides of synthetic origin avoid the risk of contamination from strongly antigenic regions of the native molecules. Such hCG-specific antisera may have extensive application in RIA systems and fertility control, where selective binding to hCG but not human LH is desired.     

The central role of human chorionic gonadotropin in the formation of human placental syncytium

A number of trophoblast products, including human chorionic gonadotropin (hCG), can increase the formation of human placental syncytium through the differentiation of mononuclear cytotrophoblasts. The present study investigated the central role of hCG in this process by using antisense receptor phosphorothioate oligodeoxynucleotides (ODNs). Culturing cytotrophoblasts with the hCG/LH receptor antisense, but not sense, ODN resulted in a significant decrease in receptor protein levels and inhibited spontaneous as well as exogenous hCG induced increase in differentiation. The hCG/LH receptor antisense ODN also inhibited epidermal growth factor (EGF), TGF-alpha, leukemia inhibitory factor (LIF), but not 8-bromo-cAMP, induced increases in differentiation, suggesting that hCG is required for EGF, TGF-alpha and LIF, but not for the cAMP actions. Although antisense EGF receptor and LIF receptor ODNs inhibited EGF and LIF induced increase in differentiation, respectively, they were ineffective against hCG, suggesting that they use separate pathways, but they both converge on a common pathway requiring the hCG actions. Mechanism of action studies revealed that EGF treatment activates its receptors and MAPK, both of which are required for EGF to increase the differentiation, cAMP levels and activate protein kinase A. In summary, our results demonstrate that hCG is an autocrine and paracrine regulator that is required for EGF, TGF-alpha, and LIF, but not for cAMP to increase human placental syncytium formation. Direct activation of protein kinase A seems to bypass the hCG pathway, perhaps by targeting genes associated with the differentiation.     

Occurrence of fragmentation of free and combined forms of the beta-subunit of human chorionic gonadotropin.

In an attempt to further study various fragments of free and combined forms of hCG beta present in biological fluids, we performed one- and two-dimensional sodium dodecyl sulfate-polyacrylamide gel electrophoresis, followed by Western immunoblotting using antipeptide antibodies directed to the hCG beta-(111-116) portion (monoclonal antibody FB12) antiserum to the hCG beta(8-16) portion or antiserum which was specific for fragments ending at residue 47. Results observed in a crude preparation of urinary hCG demonstrated that in addition to the carboxyl-terminal part of the reduced hCG beta nicked subunit (beta NS) [hCG beta-(48-145)], three other fragments of mol wt 18,000 (F1), 16,500 (F2), and 12,000 (F3) were detectable after cleavage of disulfide bonds. Both the immunoreactivity pattern and peptide sequencing revealed that the F1 fragment was constituted of the hCG beta-(1-47) sequence, whereas the F2 fragment comprised the 6-47 portion. We then studied the beta NS in urine from either pregnant women or four patients with choriocarcinomas. Results showed that both hCG and the free beta-subunit contained beta NS. Furthermore, free hCG beta present in those urine samples appeared to be extensively, if not totally, nicked. Results observed in urine were confirmed using separation of hCG from its beta-subunit by a two-step chromatography procedure, identification of hCG and hCG beta immunoreactive peaks by specific monoclonal immunoradiometric assay, and analysis of resulting preparations by one-dimensional electrophoresis under reducing conditions, followed by Western immunoblotting with FB12. This latter protocol was also used to investigate the presence of beta NS in sera of four patients with choriocarcinoma tumors. In those sera, hCG appeared to be nicked. This study demonstrates that the beta-subunit of hCG is modified by multiple fragmentations.     

Human chorionic gonadotropin beta-subunit affects the folding and glycosylation of alpha-cys mutants

Human chorionic gonadotropin (hCG) is a member of a family of heterodimeric glycoprotein hormones that contain a common alpha-subunit but differ in their hormone-specific beta-subunits. Both subunits have five and six disulfide bonds, respectively, which consist of cystine knot structure. It is evident from numerous studies that the structure of beta-subunits is rigid, whereas that of alpha-subunit is flexible and can be molded by a beta-subunit. Previously, we reported that secreted forms of a mutants where either cysteine residue in the disulfide bond 7-31 or 59-87 was converted to alanine contained a disulfide-linked homodimer in addition to a monomer. To study whether the hCGbeta-subunit affects the conformations of alpha mutants, alpha-subunits lacking either the 7-31 or 59-87 disulfide bond were expressed with wild-type (WT) hCGbeta in Chinese hamster ovary cells, and homodimer formation and glycosylation of dimerized alpha-subunit were assessed by continuous labeling with [35S]methionine/cysteine, immunoprecipitation with anti-alpha or -hCGbeta serum, digestion with endoglycosidase-H or -F, and sodium dodecyl sulfate-polyacrylamide gel electrophoresis in a non-reducing condition. Our data showed that a homodimer was not observed in the half-Cys mutants except one, where cysteine at position 7 was converted to alanine, in the presence of beta-subunit. This finding indicated that hCGbeta-subunit rescued the a half-Cys mutants from the formation of intermolecular disulfide-linked homodimer by preferentially combining with the alpha mutants. In both free WT and all mutants treated with endoglycosidase-H, no or faint bands were recognized as the same migration as seen in endoglycosidase-F treatment. Even in the endoglycosidase-H sensitive cases, the amount of sensitive alpha-subunits was less than 5% of total alpha-subunits. In contrast to free alpha-subunits, distinct endoglycosidase-H sensitive bands were seen in both WT and mutants, although the ratio was various. We concluded that hCGbeta-subunit affects the folding and glycosylation of the alpha-subunit mutants.     

Isoimmunization against human chorionic gonadotropin with conjugates of processed beta-subunit of the hormone and tetanus toxoid.

The immunogenicity of the conjugate prepared from "processed" beta-subunit of human chorionic gonadotropin (choriogonadotropin, HCG) and tetanus toxoid has been studied in animals and a human subject. The conjugate elicited the formation of high-affinity (Ka = 10(9)-10(11) M-1) anti-HCG and anti-tetanus antibodies. On primary immunization, the antibody, response lasted for several months. Repeat injection of the conjugate in the declining phase of antibody titers produced a booster response without a lag period. The antibodies reacted with the beta-subunit of HCG and the complete HCG molecule but were devois of significant crossreactivity with human growth hormone, placental lactogen, follicle-stimulating hormone, thyroid-stimulating hormone, and luteinizing hormone at tonic and surge levels. The antibodies were competent for neutralizing the biological activity of HCG in the mouse uterine weight gain assay, the ventral prostate weight gain assay, and the radioligand assay for binding of 125I-labeled HCG to receptors on corpus luteum. HCG (5000 international units) administered to an immunized subject was completely bound by circulating antibodies. Administration of HCG (in contrast to conjugate) was without booster effect on anti-HCG titers.     

A monoclonal antibody against a synthetic peptide is specific for the free native human chorionic gonadotropin beta-subunit

A totally synthetic molecule (109-145 peptide) analogous to the beta-subunit carboxyl terminus was used as an antigen in the development of antibodies by the hybridoma technique. A monoclonal antibody (702 D7) specifically recognized the free native beta-human CG (beta hCG). 702 D7 was of the immunoglobulin G1 subclass and was directed against an antigenic site localized in a 10-amino acid sequence (109-118) or less. The recognition of an epitope located in the 109-118 region could explain the specific recognition of beta hCG observed with 702 D7, in contrast to monoclonal antibodies directed against a 118-145 region with a recognition of both beta hCG and whole hCG, as observed with a second monoclonal antibody (1032) to synthetic peptide. Immunohistochemical results and preliminary data obtained from the immunoradiometric assay show that 702 D7 provides a clinical tool for the detection of free beta-subunit secretion even at low concentrations, and could allow the study of this subunit or its metabolites produced by normal and tumoral cells.