A simple, rapid immunometric assay for determination of functional and growth hormone-occupied growth hormone-binding protein in human serum.

We present a sensitive time-resolved fluorometric immunofunctional assay (TR-FIA) for direct quantitation of functional growth hormone-binding protein (GHBP), using an immunoassay kit for growth hormone (GH-DELFIA). In addition to the immobilized GH antibody, one monoclonal antibody against GHBP was used. This anti-GHBP was labelled with the chelate of europium. The assay was performed in one step. The detection limit for GHBP was 0.044 nmol L-1 (NBS + 3 SD). The calibration curve was linear in the interval 0.11-8.03 nmol L-1. Average intra-assay coefficient of variation (CV) was 3.44%. Average interassay CV at GHBP concentrations 0.563 nmol L-1 and 1.40 nmol L-1 were 12% and 6.3% respectively. Analytical recovery in serum ranged from 76% to 127% with a mean of 101 +/- 3.6%. Serum GHBP in 102 normal subjects ranged from 0.513 to 3.772 nmol L-1 and was positively related to body mass index (P < 0.001). In growth hormone-deficient sera GHBP was higher than in control subjects (1.751 +/- 0.179 nmol L-1 and 1.257 +/- 0.140 nmol L-1 respectively, P < 0.001). Acromegalic patients had lower levels of GHBP than controls (0.946 +/- 0.251 and 1.234 +/- 0.144 nmol L-1 respectively, P = 0.005). This assay also allowed detection of GH-complexed GHBP in serum. These results were in agreement with theoretical values calculated from the measured GH and the functional GHBP concentrations. Results were compared with data obtained by a recently reported, validated ligand immunofunctional assay (LIFA), which is fundamentally different. There was a significant linear relationship between the results from the two assays (r = 0.89, P = 0.001). The slope of the regression line was 0.65. In conclusion, this new convenient GHBP TR-FIA provides a sensitive and precise method for detecting total GHBP as well as complexed GHBP in human serum, and allows easy processing of large numbers of samples.

The growth hormone (GH)-binding protein cloned from human IM-9 lymphocytes modulates the down-regulation of GH receptors by 22- and 20-kilodalton human GH in IM-9 lymphocytes and the biological effects of the hormone in Nb2 lymphoma cells.

The cDNA coding for the 246-amino acid long N-terminal extracellular portion of the human (h) GH receptor, corresponding to the circulating GH-binding protein (hGHBP), was cloned by polymerase chain reaction from human IM-9 lymphocytes. The cDNA sequence was identical to that reported for human liver and placenta and demonstrated alternative splicing of exon 3. The protein with the exon 3-encoded domain was expressed and secreted in glycosylated form from baby hamster kidney (BHK) cells, purified to homogeneity, and sequenced; the amino acid sequence was identical to that predicted from liver cDNA. The cloned hGHBP competed in a dose-dependent fashion for binding of 125I-labeled 22-kilodalton (kDa) hGH, and at higher concentrations for binding of 125I-labeled 20-kDa hGH, to IM-9 lymphocytes. hGHBP decreased the association rate of [125I]hGH to the cells without decreasing the dissociation rate. hGHBP blocked the down-regulation of GH receptor in IM-9 cells by both 22- and 20-kDa hGH. hGHBP also blocked the binding of [125I]hGH to PRL receptors on Nb2 lymphoma cells and the effect of the hormone on thymidine incorporation. Binding of both 22- and 20-kDa hGH to the binding protein was demonstrated directly by immunoprecipitation with monoclonal antibody 263. The present work thus establishes the identity of the IM-9 human GHBP from those of liver and placenta, and demonstrates its ability to bind both 22- and 20-kDa hGH with good affinity and to block their biological actions mediated though both somatogenic and lactogenic receptors. The modulation of receptor down-regulation by the BP may be a relevant facet of its physiological role.

Characterization of the three 125I-iodination isomers of human insulin-like growth factor I (IGF1).

Human insulin-like growth factor I (IGF1) was labeled with 125I and the resulting mixture of iodination isomers was separated by reverse-phase HPLC. Three major radioactive peaks were isolated and identified by sequencing as the expected three monoiodinated species. The ranking of the affinities of the three isomers for the human IGF1 receptor was found to be Tyr24(125I) > Tyr31(125I) >> Tyr60(125I). The Tyr31(125I) isomer was shown to have an affinity similar to that of unlabeled IGF1 and is thus the tracer of choice for IGF1. The tracers were stable upon storage at -20 degrees C for at least 3 months.

Altered ligand specificity of proteolysed insulin-like growth factor binding protein-3.

IGF binding protein-3 (IGFBP-3) undergoes limited proteolysis in human pregnancy serum, altering its electrophoretic mobility and its binding of radioiodinated IGF tracers. IGF-I tracer, monoiodinated on Tyr31, discriminated less than other tracers between intact and proteolysed IGFBP-3. In competitive binding curves using this tracer, intact IGFBP-3 showed comparable reactivity towards IGF-I and analogs with substitutions at Tyr24, Tyr31 or Tyr60. In contrast, proteolysed IGFBP-3 reacted equally with IGF-I and [Ala31]IGF-I (approximately 10-fold lower potency than with intact IGFBP-3), but showed a marked selectivity against [Ser24]IGF-I and [Leu60]IGF-I (approximately 100-fold lower potency than with intact IGFBP-3). The affinity of IGF-I binding to proteolysed, but not intact, IGFBP-3 was increased by the addition of the acid-labile subunit of the IGFBP-3 complex. This study defines more fully the lesion in IGFBP-3 caused by serum proteolysis during pregnancy and demonstrates that tyrosine-substituted IGF-I derivatives are valuable tools in studying IGFBP-3 proteolysis.

Development of an optimized refolding process for recombinant Ala-Glu-IGF-1.

Denatured and reduced N-terminal extended insulin-like growth factor-1 (AE-IGF-1) was purified from Escherichia coli extracts and subjected to in vitro folding. The renaturation process was shown to be a function of the redox potential of the solution. Folding by different methods had no significant effect on the renaturation. A maximal yield of 60% (w/w) was obtained. The folded AE-IGF-1 was enzymatically converted to IGF-1. The major by-product (20% w/w) was identified as scrambled IGF-1. Enzymatic digestion at alkaline and acidic pH suggested two possible disulphide bond arrangements; (i) Cys6-Cys47, Cys18-Cys61, Cys48-Cys52; or (ii) Cys6-Cys52, Cys18-Cys61, Cys47 and Cys48 being in their reduced forms. Energy minimization and molecular modelling suggested that the scrambled IGF-1, having reduced cysteines at positions 47 and 48, was the energetically most stable conformation of the two.

Primary sequences of insulin-like growth factors 1 and 2 isolated from porcine plasma.

Insulin-like growth factors 1 and 2 were purified from porcine plasma. In addition to the determination of their isoelectric points, the primary structures of both proteins were determined, using low microgram quantities of protein, by the versatile combination of time-of-flight plasma desorption mass spectrometry and automated Edman degradation. Porcine insulin-like growth factor 1 was shown to be homologous to both human and bovine proteins; the type 2 growth factor showed one mutation to both human and bovine type 2 proteins.

Confirming the primary structures of insulin-like growth factors 1 and 2 isolated from porcine plasma using mass analysis.

The amino acid sequences of insulin-like growth factors 1 and 2 from porcine plasma have been determined by mass analysis of peptides obtained from tryptic and AspN digests for the former protein and from tryptic and peptic digests for the latter. Complete homology to the human Type 1 protein was indicated by the overlapping peptide map obtained by mass analysis. For the Type 2 protein, a mutation of a Ser in the human protein to an Asn in the porcine was confirmed by the combination of mass analysis and automated protein microsequencing. The studies were performed on 3 micrograms (0.4 nmol) of each protein, indicating the sensitive potential of this method in primary structure-homology studies.

One-chain urokinase-type plasminogen activator from human sarcoma cells is a proenzyme with little or no intrinsic activity.

We have compared the plasminogen activating capacity of one- and two-chain urokinase-type plasminogen activator (u-PA). In a 125I-plasminogen conversion assay in the presence of high amounts of a plasmin inhibitor, one-chain u-PA pretreated with diisopropyl fluorophosphate had no detectable activity, the detection limit corresponding to the activity of a 400-fold lower amount of two-chain u-PA. In coupled assays in which generated plasmin was measured with a synthetic substrate, activity was clearly observed with the one-chain preparation, but the initial rate of plasminogen activation was lower than that of a 250-fold smaller concentration of two-chain u-PA. The coupled assays for one-chain u-PA are self-activating because plasmin catalyzes conversion of one- to two-chain u-PA, and it is not possible to decide whether the low activity of one-chain u-PA observed with this type of assay is intrinsic or due to contaminations. On the basis of these findings and a discussion of previous studies, it is concluded that one-chain u-PA has a variety of properties similar to the one-chain proenzyme forms of other serine proteases and that it should, therefore, be considered as a genuine proenzyme form of u-PA.

Plasminogen activator inhibitor from human fibrosarcoma cells binds urokinase-type plasminogen activator, but not its proenzyme.

An approximately 75% pure form of a human Mr approximately 54,000 plasminogen activator inhibitor from conditioned culture fluid of the fibrosarcoma cell line HT-1080 was obtained by a single step of chromatography on concanavalin A-Sepharose. The inhibitor inhibited human urokinase-type plasminogen activator (u-PA) and tissue-type plasminogen activator, but not plasmin. Rabbit antibodies against this plasminogen activator inhibitor also reacted with a plasminogen activator inhibitor with identical electrophoretic mobility in extracts of human blood platelets, indicating that the HT-1080-inhibitor is of the same type as the inhibitor of blood platelets. As revealed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis followed by fibrin-agarose zymography, incubation of HT-1080-inhibitor with the active form of human u-PA led to the formation of an equimolar sodium dodecyl sulfate-resistant complex between them; in contrast, no complex formation was observed between the inhibitor and the proenzyme form of human u-PA (pro-u-PA). Likewise, using a column of anti-inhibitor antibodies coupled to Sepharose for removal of excess inhibitor and activator-inhibitor complexes, the potential enzymatic activity of pro-u-PA was found to be unaffected by incubation with inhibitor under conditions in which more than 95% of the active u-PA had formed complex with inhibitor.

Plasminogen activators catalyse conversion of inhibitor from fibrosarcoma cells to an inactive form with a lower apparent molecular mass.

Purified approximately 54 kDa plasminogen activator inhibitor from human fibrosarcoma cells was converted to an inactive form with slightly higher electrophoretic mobility by incubation with catalytic amounts of urokinase-type or tissue-type plasminogen activator. Serine proteinase inhibitors and a monoclonal antibody against urokinase-type plasminogen activator inhibited the conversion, indicating that it was caused by plasminogen activator-catalyzed proteolysis. These findings represent the first demonstration of a well-defined protein apart from plasminogen, constituting a substrate for plasminogen activators.

Enzyme-linked immunosorbent assay for human urokinase-type plasminogen activator and its proenzyme using a combination of monoclonal and polyclonal antibodies.

A sensitive and specific enzyme-linked immunosorbent assay (ELISA) for human urokinase-type plasminogen activator (u-PA) and its inactive proenzyme (pro-u-PA) was developed. A monoclonal antibody was used as solid-phase antibody, while rabbit antibodies against human u-PA followed by peroxidase-conjugated third antibody were used for detection of bound u-PA. No reaction was observed with tissue-type plasminogen activator or with a variety of other human proteins. The assay was used for quantitation of u-PA in human urine and in culture fluid from human tumor cells. The recovery of added pro-u-PA was greater than 95%. A good agreement with the results obtained by enzymatic assays was found. The detection limit was less than 0.1 ng per ml, both for u-PA and pro-u-PA. The advantages of the use of ELISA compared with enzymatic assays and radioimmunoassays for quantitation of u-PA and pro-u-PA in biological samples are discussed.

Proenzyme to urokinase-type plasminogen activator in the mouse in vivo.

We have investigated whether urokinase-type plasminogen activator (u-PA) is present in the mouse in vivo as the proenzyme or as the active enzyme. u-PA in extracts of various murine tissues was of a one-polypeptide chain form with an electrophoretic mobility indistinguishable from purified proenzyme (pro-u-PA), as demonstrated by SDS-polyacrylamide gel electrophoresis under reducing conditions followed by immunoblotting. No 2-chain u-PA was detected in any of the extracts (detection limit 10% of that of one-chain u-PA). In bladder urine more than half of the u-PA was of the one-chain form. Together with previous immunocytochemical studies of the normal murine tissues and studies of the Lewis lung carcinoma, the present results indicate that in these tissues the one-chain proenzyme is the predominant form of u-PA in intracellular stores and for the first time demonstrates that at least in some cases the one-chain form constitutes a sizeable fraction of the u-AP in extracellular fluids in the intact organism.

Immunocytochemical localization of urokinase-type plasminogen activator in Lewis lung carcinoma.

The invasively growing and metastasizing Lewis lung carcinoma consistently contained urokinase-type plasminogen activator (u-PA) enzyme activity. When investigated immunocytochemically with antibodies against u-PA, different parts of individual tumors showed a pronounced heterogeneity in staining intensity. Strong staining was found in areas with invasive growth and degradation of surrounding normal tissue, while other areas were completely devoid of staining. Immunoreactivity occurred both with a perinuclear cytoplasmic localization in tumor cells and associated with apparently extracellular material. SDS PAGE of tumor extracts, under both reducing and nonreducing conditions, followed by immunoblotting, showed only one immunocytochemically stainable band with an electrophoretic mobility corresponding to that of purified proenzyme to u-PA, while no two-chain u-PA was detected. This indicates that the major part of the activator in Lewis lung carcinoma is present as one-chain pro-u-PA.

Distribution of urokinase-type plasminogen activator immunoreactivity in the mouse.

Immunocytochemistry, using rabbit antibodies to a urokinase-type 48-Kdalton Mr mouse plasminogen activator, showed that enzyme immunoreactivity is widely distributed in the normal mouse. Strong staining was obtained in widely disseminated connective tissue cells with a fibroblast-like morphology. Such cells occurred in high numbers in the lamina propria mucosae of the gastrointestinal tract, and in moderate numbers in the connective tissue septa of the pancreas. A few such cells were detected around the larynx, trachea, and bronchi. Immunoreactivity also occurred in epithelial cells of the proximal and distal kidney tubules, the ductus deferens, and in pulmonary pneumocytes. In addition, presumably extracellular staining was seen irregularly along the basement membrane and fibrillar structures in the lamina propria of the small and large intestines. Moreover, decidual cells of the mouse placenta stained strongly, and a moderate staining was observed in epithelial cells of involuting mammary glands, but not in those of noninvoluting glands. No immunoreactivity was observed in endothelial cells. Control experiments included absorption of the antibodies against highly-purified mouse plasminogen activator and the corresponding proenzyme, and the finding of a good correspondence between the number of immunoreactive cells and measurable enzymatic activity determined in adjacent tissue sections. Separation by SDS PAGE followed by immunoblotting revealed only one immunochemically stainable protein band with Mr approximately 48 Kdaltons in extracts from tissues showing immunoreactivity.

Inactive proenzyme to tissue-type plasminogen activator from human melanoma cells, identified after affinity purification with a monoclonal antibody.

The human 66 000 mol. wt. plasminogen activator (HPA66; tissue-type plasminogen activator) has been purified from melanoma cells by a one-step affinity method with a monoclonal antibody. HPA66 purified in this way consists mainly of a one-polypeptide chain form with small amounts (15%) of a form containing two polypeptide chains held together by one or more disulphide bridges. The one-chain form was converted to the two-chain form by catalytic amounts of plasmin. During the conversion, the enzyme activity of HPA66, as measured by an [125I]plasminogen conversion assay and with a chromogenic substrate, increased linearly with the percentage of the two-chain form. A linear regression analysis showed that all enzyme activity could be accounted for by the two-chain form, while the one-chain form had no measurable enzyme activity (detection limit approximately 5% of the activity of the two-chain form). Together with previous findings of inactive proenzymes to murine and human approximately 50 000 mol. wt. (urokinase-type) plasminogen activators, these findings indicate that plasminogen activators are generally formed from inactive one-chain proenzymes which are converted to active two-chain enzymes by limited proteolysis, thus demonstrating a third step in a cascade reaction leading to extracellular proteolysis.

Monoclonal antibody to human 66,000 molecular weight plasminogen activator from melanoma cells. Specific enzyme inhibition and one-step affinity purification.

Hybridomas producing a monoclonal IgG1 antibody to a human plasminogen-activating enzyme with an apparent mol. wt. of 66,000 (66 K, HPA66) from human melanoma cells were obtained by fusion of NSI-Ag 4/1 mouse myeloma cells with spleen cells from a mouse immunized with a partially purified preparation of the enzyme. Screening for clones of hybridomas producing antibodies to HPA66 was performed with the impure enzyme preparation. A preliminary screening included enzyme-linked immunosorbent assay and SDS-polyacrylamide gel electrophoresis (SDS-PAGE) followed by immunoblotting; the final identification was based on inhibition of the enzymatic activity of HPA66 which was complete at high antibody concentrations. No inhibition of three other human and murine plasminogen activators or of plasmin was observed. Employing a one-step affinity procedure with the antibody coupled to Sepharose, HPA66 was purified approximately 200-fold from conditioned medium from the melanoma cells with a yield of 79%. The purified HPA66 was homogeneous as evaluated by SDS-PAGE. Electrophoresis under reducing conditions indicated that it consisted of one polypeptide chain. The binding constant between the antibody and 125I-labelled HPA66 was approximately 2.5 x 10(9) l/mol. The antibody did not bind to a variety of other plasminogen activators, including 52-K and 36-K human enzymes and 48-K and 75-K murine enzymes. Previously, a monoclonal antibody against another enzyme was derived by the sole use of enzyme inhibition for screening. The present study represents a modification of this procedure that can be used when antibody-unrelated inhibitors of the enzyme are present in hybridoma culture fluid.

Purification of zymogen to plasminogen activator from human glioblastoma cells by affinity chromatography with monoclonal antibody.

Incorporation of the serine protease active site reagent diisopropyl fluorophosphate (DFP) into a plasminogen activator with an Mr of approximately 52000 released from cultured human glioblastoma cells was strongly enhanced by incubation with plasmin. This observation led to the isolation of an inactive form of the enzyme from serum-free conditioned culture fluid by affinity chromatography on a column of a Sepharose-bound monoclonal antibody raised against urokinase. An 831-fold purification was obtained with a yield of 41%. The purified molecule was homogeneous as evaluated by polyacrylamide gel electrophoresis with sodium dodecyl sulfate (NaDodSO4), having one stainable band under nonreducing as well as reducing conditions with an Mr of approximately 52000. It was unable to activate plasminogen, but catalytic amounts of plasmin converted it into active enzyme. After NaDodSO4-polyacrylamide gel electrophoresis, the active enzyme showed one band under nonreducing conditions, but after reduction, two bands with Mr values of approximately 20000 and 32000 were observed. The active enzyme incorporated [3H]DFP into the approximately Mr 32000 band, while no incorporation was observed into the inactive form. These findings show that the Mr 52000 human plasminogen activator exists in a proenzyme form consisting of a single polypeptide chain that by proteolysis between half-cystine residues is converted into the active enzyme consisting of two chains with molecular weights of approximately 20000 and 32000, the active site being on the latter chain. The results are consistent with the active form of the enzyme being identical with the higher molecular weight form of urokinase, and together with recent observations that a murine plasminogen activator is released from sarcoma virus transformed cells as an inactive proenzyme, they suggest that zymogens to plasminogen activators are of more general occurrence.