Calcium-specific immunoassays for factor IX: reduced levels of antigen in patients with vitamin K disorders

Polyclonal rabbit anti-factor IX antisera were fractionated to establish solid-phase immunoassays recognizing calcium-dependent and non-calcium-dependent epitopes. The assays were greater than 99.9% specific for factor IX and sensitive to 0.05 U/dl plasma or 2 ng/ml purified factor IX. For the calcium-dependent fraction, an absolute requirement of divalent metal ions was found, and Sr(II), Mn(II), and Mg(II) could substitute for Ca(II). On immunoblots of reduced, electrophoresed factor IXa, the 125I-calcium-dependent antibody fraction bound to the amino-terminal light chain. Plasma sampled from 13 patients receiving warfarin and one with cephalosporin-related vitamin K deficiency had a mean level of calcium-dependent factor IX antigen of 22 U/dl, comparable to the 24 jU/dl average of factor IX procoagulant activity; these two results were highly correlated. Antigen levels determined by either the polyclonal or a monoclonal, non-calcium-dependent anti-factor IX assays ranged from 1.7-fold to 6.0-fold greater than the corresponding levels of factor IX procoagulant activity or calcium-dependent antigen level for each subject's plasma. The difference reflects inactive, circulating factor IX. In contrast, factor IX antigen levels determined by an assay using a monoclonal, calcium-dependent anti-factor IX were from one half to one thirteenth as much as those measured by the polyclonal, calcium-dependent immunoassay. The disparity between results of calcium-dependent assays suggests that some Gla residues near the amino terminus of factor IX are relatively less important for normal procoagulant function of factor IX than others, are more sensitive to the effects of vitamin K antagonism or deficiency, and are important for the epitope recognized by this particular calcium-dependent, monoclonal antibody.     

Measurement of human growth hormone (hGH) using a rapid immunochemiluminometric assay

A two-site immunochemiluminometric assay for human growth hormone has been developed based on the use of chemiluminescent acridinium ester labelled monoclonal antibodies and a magnetisable particle - polyclonal antibody solid-phase. The assay has an incubation time of 1 h at room temperature and rapid separation and quantitation stages. The sensitivity of detection is 0.12 mU/l and the working range is 0.88 to greater than 100 mU/l for inter-assay CVs less than or equal to 15%. The assay is convenient both technically and clinically since the wide range of growth hormone levels seen in growth hormone deficiency and acromegaly can be quantified accurately in a single test without the need for repeated sample dilution.     

Cross-reactivity of monomeric and dimeric biosynthetic human growth hormone in commercial immunoassays

Commercial kits give different measurements for concentrations of growth hormone (GH, somatotropin) in serum. Most notably, a two-site monoclonal-antibody-based immunoradiometric assay (IRMA) from Hybritech routinely yields lower values than do conventional RIAs in which polyclonal antibodies are used. We used purified dimeric biosynthetic human GH as a model compound to investigate the specificity of five commercial immunoassays for size variants of GH. In all five assays, biosynthetic monomeric GH was significantly more potent than pituitary-derived standard GH supplied with the kits. Dimeric GH was significantly less potent than monomer in four of the five assays, and cross-reactivities varied more than fivefold, from 15% to 84%. Using three commercial kits selected for their specificity for dimeric GH, we measured GH in serum samples from 18 normal adults. The mean GH concentrations in serum--0.7 (Hybritech, IRMA), 1.8 (Diagnostic Products, RIA), and 3.1 (Cambridge, RIA) micrograms/L--differed significantly, but in the same rank order as that obtained in the experiments on dimer cross-reactivity.     

Standardization with synthetic 22-kDa monomer human growth hormone reduces discrepancies between two monoclonal immunoradiometric assay kits.

Discrepancies among different methods for assaying human growth hormone have been described in various studies. The two major sources of discordant results are the heterogeneity of the antibodies and the different standardization bases used by the assay manufacturers. We propose standardizing assays with 22-kDa biosynthetic monomer human growth hormone diluted with the diluents supplied by the kit manufacturers. In a study of two monoclonal immunoradiometric assays (Hybritech, specific for the 22-kDa monomer; Sorin, recognizing also a 20-kDa variant hormone), standardization with 22-kDa monomer human growth hormone reduced by 63% the differences in results for 44 serum samples from children. The use of 22-kDa human growth hormone as a common standard, highly pure and easily available in large quantities, could help limit the interpretative problems in growth diagnostics.     

Solving the problem of antibody interference in commercial "sandwich"-type immunoassays of carcinoembryonic antigen

We evaluated the effect of human anti-murine antibodies (HAMA) on apparent concentrations of carcinoembryonic antigen (CEA) as measured in serum with commercial enzyme immunoassay (EIA) kits manufactured by Abbot ("two-step" double monoclonal antibody assay), Roche, and Hybritech (room-temperature protocol). In sera from patients given parenteral murine monoclonal antibody for experimental diagnostic and immunotherapy studies, HAMA titers were determined with Immunomedics' "ImmuSTRIP HAMA-EIA" kit reagents. "True" CEA titers were established by using the ImmuCEA/MA-EIA and heat-extraction to destroy HAMA before assay for CEA. The concordance of the ImmuCEA/MA assay with the Abbott and Roche CEA EIAs was established with sera from normal individuals and from patients who had not received parenteral injections of murine monoclonal antibody. At high (100 mg/L) concentrations of HAMA, false-positive results were observed with all three kits. The Hybritech and Roche assays were more sensitive to interference by HAMA than was the Abbott CEA-EIA, false-positive results being observed at HAMA concentrations between 1 and 10 mg/L. Similar sensitivity of the three kits to interference by primate anti-MAb sera was demonstrated. Use of primate anti-MAb sera to create controls with HAMA activity and of analyte is recommended to evaluate MAb assays for potential HAMA interference and for use to devise methods to eliminate HAMA interference.     

Time-resolved fluorometric sandwich immunoassay for human growth hormone in serum and urine

A specific and sensitive time-resolved fluorometric sandwich immunoassay for human growth hormone (hGH) in serum and urine is described. A monoclonal anti-hGH IgG1 (5802)-coated polystyrene ball was incubated with serum or dialyzed urine and subsequently with europium ion-labeled monoclonal anti-hGH IgG1 (5801). Europium ion bound to the polystyrene ball was measured by time-resolved fluorometry. The detection limit of hGH was 0.3 pg/tube, which was 15-fold higher than that by sandwich enzyme immunoassay using horseradish peroxidase as label. The assay range of serum hGH was 15-15,000 ng/liter using 20 microliters of serum. The assay range of urinary hGH was 2-2,000 ng/liter using 150 microliters of dialyzed urine. The detection limits of hGH in serum and urine by this immunoassay were satisfactory for measuring hGH levels in serum and urine of healthy children and in serum of healthy adults and higher levels but not in urine of healthy adults and in serum and urine of patients with hGH deficiency.     

Enzyme-linked immunosorbent assay to measure apolipoproteins AI and B secreted by a human hepatic carcinoma cell line (Hep G2).

We describe an enzyme-linked immunosorbent assay (ELISA) to measure apolipoproteins AI and B secreted by Hep G2 cells and in cell homogenates. These assays utilize commercially available polyclonal antibodies, affinity-purified to improve their specificity, thereby achieving a dramatic increase in the sensitivity of the assay. These affinity-purified antibodies were also more sensitive than a series of monoclonal antibodies tested. We achieved a sensitivity of 0.4 ng in the apo AI assay, and a sensitivity of 5 ng in the apo B assay. By these methods, we measured secretion rates by Hep G2 cells of 358 +/- 41 ng/mg cell protein/hr for apo B and 137 +/- 8 ng/mg cell protein/hr for apo AI. These assays also allowed the measurement of intracellular apolipoproteins and thus can be used to facilitate investigations of human lipoprotein metabolism in cell culture systems.     

Radioimmunological determination of insulinlike growth factors I and II in normal subjects and in patients with growth disorders and extrapancreatic tumor hypoglycemia.

Serum levels of immunoreactive insulinlike growth factors (IGF) I and II were determined by a modified IGF I and a new IGF II radioimmunoassay in normal children and adults, and in patients with acromegaly, isolated growth hormone deficiency, and extrapancreatic tumor hypoglycemia. Serum samples were gel filtered by a simple routine procedure at acidic pH to dissociate and separate IGF from the IGF carrier protein. Mean immunoreactive IGF I levels (+/- SD; corrected for crossreactivity of IGF II) were 193 +/- 58 ng/ml in normal adult subjects, 712 +/- 245 ng/ml in acromegalic patients and 24 +/- 14 ng/ml in patients with isolated growth hormone deficiency. The lack of growth hormone alone, irrespective of an otherwise normal hormonal status, appears to be responsible for the drastic decrease of IGF I levels. Oversecretion of growth hormone does not increase the levels of immunoreactive IGF II: mean levels (+/- SD; corrected for crossreactivity of IGF I) in normal and acromegalic subjects are virtually identical (647 +/- 126 and 641 +/- 189 ng/ml, respectively). Apparently, normal growth hormone levels stimulate IGF II production already maximally. However in growth hormone deficiency immunoreactive IGF II is significantly decreased (252 +/- 99 ng/ml). Thus, IGF II, like IGF I, is growth hormone dependent. But in contrast to IGF I, the growth hormone dependence of IGF II seems to become apparent only at subnormal growth hormone levels. In normal children IGF I is age dependent: it is low in newborn cord sera (51 +/- 20 ng/ml) and gradually rises into the adult range with increasing age. At the onset of and during puberty mean IGF I levels lie above prepubertal values. In contrast, IGF II levels in normal children are independent of age and pubertal stage beyond the first year of life, whereas newborns have significantly lower IGF II values. Hypoglycemia resulting from extrapancreatic tumors is not associated with increased immunoreactive IGF I or II levels. IGF I is decreased in most of the sera (mean level +/- SD:56 +/- 39 ng/ml) whereas IGF II lies in the normal range (556 +/- 195 ng/ml).     

Evaluation of five high-sensitivity American thyrotropin assays

We compared five commercial immunometric kits for thyrotropin (TSH) manufactured in the United States--Abbott, Diagnostic Products, Hybritech, Nichols, and Becton Dickinson--and one sensitive "in-house" radioimmunoassay for their ability to monitor TSH levels in patients with hyperthyroidism and during thyroid hormone suppression therapy. With these five methods, we measured TSH concentrations is serum specimens from the following categories of patients: euthyroid (N = 99), hyperthyroid (N = 51), hypothyroid (N = 50), and thyroid cancer and nodular goiter on thyroid hormone suppression therapy (N = 33). The immunometric methods were similar in assay sensitivity and precision; the in-house radioimmunoassay was less sensitive than all other assays studied. The simultaneous correlation of TSH values obtained from all diagnostic categories was more than 90% among the assays studied. The immunometric assays clearly distinguished hyperthyroid from euthyroid patients and quantitated TSH levels in the subnormal range.     

Immunoradiometric assay of carcinoembryonic antigen with use of avidin-biotin labeling

In evaluating an enzyme-linked immunoassay of carcinoembryonic antigen (CEA) we found that the IgG fraction of polyclonal anti-CEA antibodies (DAKO) bound very well to the walls of polypropylene test tubes. We therefore developed an immunoradiometric CEA assay based on this binding of polyclonal anti-CEA antibody. We biotinylated a commercially available monoclonal antibody (Hybritech) and bound this to the CEA-anti-CEA bound to the tube wall. For detection we used 125I-labeled streptavidin. In comparison with several immunoassays for CEA this system offered several advantages such as greater linearity of the standard curve (from 0 to 74 micrograms/L), a steeper dose-response curve, and smaller coefficients of variation in the clinically useful range. This assay system may be used for other large molecules, so that only one tracer, the 125I-labeled streptavidin, has to be labeled; thus the technique seems suited for several different assays.     

Two ELISA's for measurement of protein S, and their use in the laboratory diagnosis of protein S deficiency

Protein S is a vitamin K-dependent coagulation factor involved in the regulation of the anticoagulant activity of activated protein C (APC). Heterozygosity for hereditary protein S deficiency is considered as a risk factor for the development of thrombotic disease. To measure the plasma concentration of protein S, two enzyme-linked immunosorbent assays (ELISA) have been developed: the ELISA-p (using polyclonal antibodies both as catching and tagging antibodies) and the ELISA-m (using polyclonal antibodies as catching antibodies and a monoclonal antibody as tagging antibody). The performance of these assays has been compared with that of an immunoradiometric assay (IRMA) for total protein S. Results obtained with both ELISA's in plasma's of healthy volunteers, patients with hereditary protein S deficiency type I and patients using oral anticoagulant treatment correlate very well with those of the IRMA: r = 0.974 (n = 57) for ELISA-p and r = 0.966 (n = 57) for ELISA-m. The intra- and inter-assay variation (n = 6) were calculated to be 5% and 11% for the ELISA-p and 4% and 10% for the ELISA-m. Both ELISA's can be used for the measurement of total protein S, but also for the measurement of free protein S and C4b-BP-PS complexes, after PEG precipitation. The ELISA-m was further used to measure free protein S in healthy volunteers (n = 40), patients with a protein S deficiency type I (n = 20) and patients with unexplained familial thrombosis (n = 76). In the first two groups measured and calculated free protein S levels were very similar. In the third group 7 patients were identified with total protein S in the normal range, but with measured free protein S and the ratio between the measured free protein S and calculated free protein S below the lower limits of the normal range.     

Sensitive calcitonin measurement by two-site immunometric assays: implications for calcitonin screening in nodular thyroid disease

The aim of this study was to investigate the impact of analytical aspects on the clinical usefulness of calcitonin (CT) measurement. In a retrospective analysis, CT levels measured by a polyclonal immunometric assay (Scantibodies Laboratory, CA, USA) were evaluated in various clinical situations. CT in newly diagnosed medullary thyroid cancer (MTC) (n = 20) ranged from 15.5-87130 pg/ml (median 661 pg/ml). Levels >10 pg/ml were seen in 7.3% of 314 patients with benign nodules, 48.9% of 45 patients with impaired kidney function, 97.7% of 87 patients on hemodialysis, 30.2% of 43 patients after renal transplantation, and in 71.0% of 31 patients with critical illnesses. Subgroups of patients were reevaluated by two monoclonal immunometric assays specific for mature CT. CT levels measured by the monoclonal immunometric assays were highly correlated to the polyclonal assay results in MTC patients, but were significantly different with a lower incidence of elevated levels in patients with renal disease and critical illnesses. In conclusion, highly sensitive assays with cut-off values of 10 pg/ml or below are mandatory for CT screening in nodular thyroid disease. The specificity of CT measurement in patients with renal disease and critical illnesses is higher with monoclonal assays specific for monomeric CT. These methodological aspects have to be regarded if CT measurement is used for decision making in nodular thyroid disease.     

Standardization of an enzymometric assay for apolipoprotein A-I by using mixtures of monoclonal antibodies)

For the standardization of human plasma apolipoprotein A-I assay two well characterized monoclonal antibody mixtures were used to develop a sandwich immunoenzymometric assay. The monoclonal antibody mixture 1 (A05-A17-A30) in solid phase technique was selected on the basis of its higher binding capacity of [125I]HDL (41 ng per well) compared to polyclonal antibody (23 ng per well). The epitopes recognized by monoclonal antibody mixture 1 are surface antigenic sites of apolipoprotein A-I expressed on native HDL as determined by competitive inhibition of labeled HDL. The peroxidase conjugated monoclonal antibody mixture 2 (A03-A05-A17-A51) was selected on the basis of its ability to bind to apolipoprotein A-I captured by monoclonal antibody mixture 1. For this, we used the 125I-labeled monoclonal antibodies. Under optimized assay conditions, the immunoenzymometric assay is precise (intra- and inter-assay coefficients of variations 5.4% and 9.2% respectively). It yields plasma apolipoprotein A-I values that correlate highly with those obtained with polyclonal antibody (r = 0.96). So the use of well characterized monoclonal antibody mixtures reacting only to surface antigenic sites of apolipoprotein A-I present on native lipoprotein may provide the possibility of standardization of apolipoprotein A-I measurement.     

Immunoradiometric assay of parathyrin with polyclonal and monoclonal region-specific antibodies.

In this immunoradiometric assay (IRMA) of parathyrin (PTH) a polyclonal anti-amino-PTH(1-34) is the capture antibody and a radiolabeled monoclonal anti-hPTH(44-68) is the second antibody. Gel filtration of serum from a hyperparathyroid patient yielded only a single peak of PTH, corresponding to the elution position of synthetic PTH(1-84). Healthy elderly individuals (ages 78 +/- 5 y, mean +/- SD, n = 45) had PTH concentrations (21 +/- 13 ng/L) not significantly higher than those from healthy younger (38 +/- 11 y) adults (20 +/- 8 ng/L, n = 94). Assay results agreed well with those obtained with a carboxyl-terminal PTH assay both in normal subjects (r = 0.63, P less than 0.001) and in patients with primary hyperparathyroidism (r = 0.59, P less than 0.001). Both assays equally discriminated patients with surgically confirmed primary hyperparathyroidism from normal individuals, but the PTH(1-84) IRMA also allowed a nearly absolute discrimination between normal subjects and patients with primary hypoparathyroidism (undetectable serum PTH in 18 of 21 cases) and secondary hypoparathyroidism (caused by hypercalcemia that was caused by a malignant tumor, PTH 1.3 +/- 1.3 ng/L, n = 32). Moreover, the PTH(1-84) IRMA is more sensitive (detection limit in serum, 0.8 ng/L) and easier and quicker to perform than the carboxyl-terminal assay.     

Development of a microplate elisa for free psa and psa–act complex in serum

An ELISA on microplate was established for the total serum PSA. We selected the monoclonal antibody for the assay from commercial sources making certain that it reacted with both free PSA and PSA–α₁-antichymotrypsin (PSA–ACT) complex from human serum with similar affinity (so-called “equimolar”). We also chose a test format with polyclonal anti-PSA antibodies coated on the well and monoclonal anti-PSA antibodies for quantification to gain higher test sensitivity. Two different sample volumes from each specimen, 5 and 50 m?l, were used for the assay in order not only to further increase test sensitivity and improve precision at both low and highly elevated PSA concentrations, but also to widen the assay concentration range (0–500 ng total PSA per ml). Using two sample volumes also reduces any hook effect and shortens the turn-around time because repeated determinations are usually required when specimens contain highly elevated PSA concentrations. The use of pooled sera containing approximately 95% PSA–ACT complex and 5% free PSA as a calibrator allows for a close matching of the calibrator with serum specimens in immunoreactivity and PSA composition. Moreover, our assay shows no hook effect up to 15,000 ng/ml. The within-day precision (% CV) in the critical concentration range of 4–12 ng/mL is approximately 5%. The PSA values obtained from this assay correlate well with that of the Hybritech kit (γ = 0.998, slope equals to 1.033), indicating that this kit can replace the Hybritech Tandem E PSA kit for serum PSA determination in clinical laboratories.     

Comparison of radioimmunoassay and immunoradiometric assay for serum prostatic acid phosphatase

We have compared the laboratory performance of immunoradiometric (IRMA) and radioimmunoassay (RIA) methods developed in this laboratory for measurement of serum prostatic acid phosphatase (PAP). The IRMA utilizes a radiolabelled mouse monoclonal anti-PAP and a solid phased rabbit polyclonal anti-PAP. The same rabbit antibody is used in the RIA. The IRMA shows excellent precision over a much wider working range (0.25-1000 micrograms/l) than the RIA (0.73-14.0 micrograms/l), and can be completed in 5 h, while the RIA requires 3 days. Levels in healthy males and in patients with benign prostatic hypertrophy are similar in both assays, upper limits of normal being 1.8 micrograms/l (IRMA) and 4.7 micrograms/l (RIA). The two assay methods correlate very well (r = 0.97) when PAP is measured in serum from prostatic cancer patients, although IRMA results are generally lower than those obtained by RIA. About 20% of patients with non-metastatic prostatic carcinoma had elevated serum PAP, whereas about 80% of those with metastatic disease had raised levels. The diagnostic efficiencies of the RIA and IRMA appeared similar. The value of the IRMA in follow-up and staging remains to be determined.     

One-step sandwich enzyme immunoassay for human laminin using monoclonal antibodies

A one-step sandwich enzyme immunoassay (one step sandwich EIA) for human serum immunoreactive laminin was set up with a pair of monoclonal antibodies prepared against human placental laminin P1 fragment. The assay was characterized by carrying out two immunoreactions simultaneously, laminin P1 fragment reacting with both a monoclonal antibody as a solid phase and a horseradish peroxidase-labeled monoclonal antibody (Fab') against human laminin P1 fragment as conjugate. Sensitivity of the immunoassay was 0.01 ng/well (0.5 microgram/l), and linearity was obtained between 0.01-20 ng/well (0.5-1,000 micrograms/l). The levels of laminin in sera from normal individuals and patients with liver cirrhosis, hepatocellular carcinoma and primary biliary cirrhosis were 103 +/- 15 micrograms/l, 228 +/- 70 micrograms/l, 341 +/- 163 micrograms/l and 232 +/- 93 micrograms/l, respectively. Protein immunoblotting showed that the serum immunoreactive laminin measured by the assay was a fragment with rel mol mass of 200 kDa.     

Increased prolactin concentration and thyrotropic insufficiency in patients with growth hormone deficiency

An increase in TSH secretion in patients with growth hormone deficiency (GHD) indicates hypothalamic thyrotropic dysfunction. A simultaneous increase in prolactin (PRL) response to TRH is often observed in these patients. In order to find out whether these patients need additional thyroxine treatment, we investigated thyrotropic function in 15 patients with GHD and increased PRL secretion. 3 groups of patients emerged: 1. Four patients were clinically hypothyroid (T42.9 +/- 1.3 micrograms/dl); 2. Four patients developed low T4 levels during growth hormone therapy (T43.5 +/- 0.1 micrograms/dl), although their T4 levels were normal before growth hormone treatment was initiated (T45.7 +/- 0.5 micrograms/dl). 3. Seven patients had T4 levels in the lower normal range (7.0 +/- 0.8 micrograms/dl). The growth velocity of these patients was monitored when thyroxine treatment was given in addition to hGH. Thyroxine administration resulted in suppression of TSH secretion in all patients while PRL concentrations remained elevated. Growth velocity was improved with additional T4 therapy in groups 2 and 3 without undue acceleration of bone age. All patients with GHD and elevated PRL levels should be carefully monitored for thyrotropic dysfunction.     

Reference levels of insulin-like growth factor I in the serum of healthy adults: comparison of four immunoassays.

The measurement of insulin-like growth factor-I (IGF-I) has become an essential tool for diagnosing growth hormone deficiency and acromegaly, as well as for monitoring the efficacy of treatment in these disorders. The latter aspect gains significance in the light of epidemiological studies which indicate a relationship between IGF-I levels and the incidence of certain malignancies. We aimed to evaluate the performance of widely implemented IGF-I assays by testing four representative, commercially available immunoassays. Thus, four parallel determinations of the IGF-I levels of 427 healthy blood donors aged between 18 and 79 years were conducted. Apart from divergent performance criteria, the assays also differed systematically. These differences were, however, linear and of lower magnitude among the lower ranges. We conclude that despite the wide variance among commercially available IGF-I assays, which principally involve assay-specific normative data, each of the implemented assays was robust and thus an appropriate tool in the diagnostic work-up of growth hormone deficiency in adult life, when IGF-I levels are low.     

Two-site assays of bone gla protein (osteocalcin) demonstrate immunochemical heterogeneity of the intact molecule.

We developed a panel of monoclonal antibodies to human bone gla protein (BGP; osteocalcin) peptides that span the linear sequence of the molecule, specifically BGP 1-12 (N-terminal), BGP 15-30 (midregion), and BGP 38-49 (C-terminal). These antibodies were evaluated in various combinations of two-site formats in studies of serum BGP concentrations. For clinical studies, we selected from a panel of antibodies the two most sensitive antibody pairs for the intact molecule (N-C); we also used a polyclonal RIA based on BGP-C. For the two-site format, we used two N-terminal antibodies, 029 and 052, adsorbed to polystyrene beads, and radioiodinated a C-terminal antibody, 663. The standard for each of the assays was purified human BGP. The following BGP serum concentrations (microgram/L, mean +/- SE) were measured with the various assays: by the 029-663 assay, results for normal subjects were 7 +/- 3, for patients with renal failure 25 +/- 8, and for patients with Paget disease 12 +/-4; by the 052-663 assay, the respective results were 22 +/- 4, 44 +/- 12, and 31 +/- 7; by the polyclonal assay, the results were 3 +/- 0.2, 13 +/- 2, and 5 +/- 1. The two intact (N-C) assays were significantly (P < 0.01) correlated (r = 0.94), but their serum values differed by more than twofold in terms of the same BGP standard. The polyclonal assay significantly correlated with each of the intact assays (r = 0.83, 0.77), but it, too, gave different serum values for BGP. These studies demonstrate the immunochemical heterogeneity of circulating BGP, heterogeneity that is manifest even in immunoassays specific for the same region of the molecule.