Development of a Novel Immunoradiometric Assay Exclusively for Biologically Active Whole Parathyroid Hormone 1–84: Implications for Improvement of Accurate Assessment of Parathyroid Function

Authors


Abstract

We developed a novel immunoradiometric assay (IRMA; whole parathyroid hormone [PTH] IRMA) for PTH, which specifically measures biologically active whole PTH(1–84). The assay is based on a solid phase coated with anti-PTH(39–84) antibody, a tracer of125I-labeled antibody with a unique specificity to the first N-terminal amino acid of PTH(1–84), and calibrators of diluted synthetic PTH(1–84). In contrast to the Nichols intact PTH IRMA, this new assay does not detect PTH(7–84) fragments and only detects one immunoreactive peak in chromatographically fractionated patient samples. The assay was shown to have an analytical sensitivity of 1.0 pg/ml with a linear measurement range up to 2300 pg/ml. With this assay, we further identified that the previously described non-(1–84)PTH fragments are aminoterminally truncated with similar hydrophobicity as PTH(7–84), and these PTH fragments are present not only in patients with secondary hyperparathyroidism (2°-HPT) of uremia, but also in patients with primary hyperparathyroidism (1°-HPT) and normal persons. The plasma normal range of the whole PTH(1–84) was 7–36 pg/ml (mean ± SD: 22.7 ± 7.2 pg/ml, n = 135), whereas over 93.9% (155/165) of patients with 1°-HPT had whole PTH(1–84) values above the normal cut-off. The percentage of biologically active whole PTH(1–84) (pB%) in the pool of total immunoreactive “intact” PTH is higher in the normal population (median: 67.3%; SD: 15.8%; n = 56) than in uremic patients (median:53.8%; SD: 15.5%; n = 318; p < 0.001), although the whole PTH(1–84) values from uremic patients displayed a more significant heterogeneous distribution when compared with that of 1°-HPT patients and normals. Moreover, the pB% displayed a nearly Gaussian distribution pattern from 20% to over 90% in patients with either 1°-HPT or uremia. The specificity of this newly developed whole PTH(1–84) IRMA is the assurance, for the first time, of being able to measure only the biologically active whole PTH(1–84) without cross-reaction to the high concentrations of the aminoterminally truncated PTH fragments found in both normal subjects and patients. Because of the significant variations of pB% in patients, it is necessary to use the whole PTH assay to determine biologically active PTH levels clinically and, thus, to avoid overestimating the concentration of the true biologically active hormone. This new assay could provide a more meaningful standardization of future PTH measurements with improved accuracy in the clinical assessment of parathyroid function.

INTRODUCTION

THE INVENTION and evolution of immunoassays measuring human parathyroid hormone (PTH; parathyrin) has provided us with a better understanding of the biological and biochemical nature of this polypeptide hormone and a better tool for the clinical diagnosis and monitoring of the diseases related to primary hyperparathyroidism (1°-HPT), secondary hyperparathyroidism (2°-HPT), and hypoparathyroidism.(1–5) Circulating PTH is immunochemically heterogeneous and the midregional/C-terminal PTH fragments are known to be significantly accumulated in some disease conditions, for example, chronic renal failure.(6) Prior competitive immunoassays for PTH detect a mixture of different PTH fragments as well as the whole biologically active PTH(1–84); hence, these assays have not accurately assessed the level of circulating biologically active hormone and the function of the parathyroid glands. Because the whole or complete molecule of PTH(1–84) is the major circulating form of the serum biologically active hormone, which is capable of binding and activating the PTH-1 receptor on kidney and bone, the primary goal of developing and using intact PTH sandwich assays was to measure biologically active PTH(1–84) exclusively.(3)

Since 1987, commercially available “intact” PTH assays have greatly increased assay sensitivity and simplified the assay procedures for PTH measurement. However, the clinical use of these intact PTH assays is still fraught with challenges. For example, intact PTH levels frequently overestimate the presence and severity of parathyroid-mediated osseous abnormalities in uremic patients.(7–9) In addition, interlaboratory discordances of PTH values arose when different intact PTH kits from different manufacturers were used. One of the explanations could be that different paired antibodies with different specificities are used to form the sandwich assay for intact PTH. Indeed, recent studies have revealed that there are circulating non-(1–84) PTH fragments that interfere significantly with intact PTH measurements obtained from commercial assays in uremic patients.(10,11) One of these studies using high-performance liquid chromatography (HPLC) and different intact PTH assays has found that more than 30% of total immunoreactive intact PTH is comprised of non-(1–84) PTH fragments in this group of patients. Therefore, those intact PTH assays are not truly intact specific and still measure a mixture of the biologically active whole PTH(1–84) and large PTH fragments that show similar hydrophobicity as synthetic PTH(7–84).(10)

It is our opinion that an optimal immunoassay for PTH should measure only the clinically significant, biologically active form of PTH, which is capable of binding to the G protein-linked PTH receptors,(12,13) which initiates signal transductions in the intracellular biochemical process resulting in the regulation of calcium metabolism. In addition to its specificity,(14,15) this optimal PTH assay should be sensitive, to allow diagnosis of hyperparathyroidism(16,17); easy to perform; and of high performance in assay characteristics. To meet these goals for assaying PTH, we developed a whole PTH(1–84) immunoradiometric assay (IRMA) using a PTH(39–84) region-specific polyclonal capture antibody and a PTH(1–4) highly specific polyclonal label antibody. With these antibodies, this assay is restricted to measure only the authentic whole PTH(1–84) without any cross-reaction with the high levels of non-(1–84) PTH fragments found in patient samples. Clinical studies have shown that this specific whole PTH(1–84) assay unexpectedly provides a unique tool for the diagnosis of patients with parathyroid diseases. In studies with this new whole PTH IRMA and HPLC fractionated clinical samples, we clearly show that previously described non-(1–84) PTH fragments are aminoterminally truncated polypeptides and these PTH fragments are significantly present not only in uremic patients but also in patients with 1°-HPT and normal persons. Moreover, we further show that the ratio of full-length PTH(1–84) to aminoterminally truncated PTH fragments is significantly variable from patient to patient with HPT.

MATERIALS AND METHODS

Chemicals and reagents

Most chemicals were of reagent grade and were purchased from Sigma (St. Louis, MO, USA). Synthetic PTH(1–84) was from Peninsula Laboratories, Inc. (Belmont, CA, USA). Synthetic peptides of PTH(7–84), PTH(44–68), PTH(53–84), and PTH(39–84) were purchased from Bachem (Torrance, CA, USA). [Tyr34]PTH(1–34)amide {PTH(1–34)}, [Tyr34]PTH(2–34)amide {PTH(2–34)}, [Tyr34]PTH(3–34)amide {PTH(3–34)}, [Tyr34]PTH(4–34)amide {PTH(4–34)}, [Tyr34]PTH(5–34)amide {PTH(5–34)}, and [Tyr34]PTHrP(1–34)amide {PTHrP(1–34)} fragments were synthesized by the Massachusetts General Hospital Polymer Core Facility (Boston, MA, USA). Cyanogen bromide-activated Sepharose 4B was purchased from Pharmacia (Uppsala, Sweden). One liter of 0.01 M phosphate-buffered saline (PBS; pH 7.4) contained 0.23 g sodium dihydrogen phosphate, 1.2 g disodium hydrogen phosphate, and 8.5 g sodium chloride. One liter 0.1 M glycine hydrochloride buffer (pH 2.5) contained 8.76 g sodium chloride. Assay wash buffer was 0.01 M PBS (pH 7.4) with 0.01% Triton X-100. Nichols intact PTH IRMA kit was purchased from Nichols Institute Diagnostics (San Juan Capistrano, CA, USA).

Standards and controls for the whole PTH IRMA were prepared by adding synthetic PTH(1–84) to a normal human serum that did not show any detectable PTH level with the intact PTH assay. The concentrations of the standard set were 0, 10, 16, 46, 165, 700, and 2300 pg/ml. All standards and controls were aliquoted, lyophilized, and stored at 2–8°C.

Goat anti-PTH(39–84) polyclonal antibody coated onto 5/16-in polystyrene beads (Hoover Precision Products, Sault Ste. Marie, MI, USA) were used as the solid phase. The antibody was prepared by affinity purification. Briefly, synthetic PTH(39–84) peptide was conjugated covalently to Sepharose 4B gel using the manufacturer's suggested procedures by mixing the gel with the peptide at room temperature for 16 h. The peptide-bound Sepharose 4B gel was transferred to a chromatography column and the packed column was washed and equilibrated with 0.01 M PBS. Goat anti-PTH(39–84) antiserum was loaded onto the column. Unbound protein and other matrix components were washed away using 0.01 M PBS and the specific goat anti-PTH(39–84) polyclonal antibody was eluted with 0.1 M glycine hydrochloride buffer. The eluted polyclonal antibody was neutralized and stored at 2–8°C. The purified goat anti-PTH(39–84) polyclonal antibody was attached physically onto the surface of the polystyrene beads by means of passive absorption.(5,18) The beads were blocked by Scancoat (Scantibodies Laboratory, Santee, CA, USA) and finally dried at room temperature. These antibody-coated beads were then stored at 2–8°C and were ready for assay use.

125I-PTH(1–4) region-specific polyclonal antibody was used as the assay signal antibody. This antibody also was affinity-purified by the same procedure as described previously. The chloramine T method was used for the iodination of this most N-terminal PTH-specific antibody. A PD-10 column was used for the separation of the125I-labeled antibody from the free iodine. Selected fractions of labeled antibody were pooled and diluted using 0.01 M sodium phosphate-based buffer approximately to 300,000 disintegrations per minute (dpm) per 100 μl. This solution was the final tracer to be used in the whole PTH IRMA.

IRMA for whole PTH(1–84)

A single incubation step IRMA specific for the whole PTH(1–84) was developed and optimized with the previously mentioned assay reagents. Briefly, 200 μl of assay standards, controls, and patient samples were pipetted into appropriately labeled 12 mm × 75 mm polypropylene test tubes. One hundred microliters of125I-labeled PTH(1–4)-specific antibody tracer solution and one goat anti-PTH(39–84) polyclonal antibody-coated bead were added to all test tubes. The immunochemical reaction was conducted at room temperature with shaking at 170 rpm for 18–22 h. During this assay incubation period, the immunochemical reaction forming the sandwich of {solid-phase goat anti-PTH(39–84) antibody}-{whole PTH(1–84)}-{125I-goat anti-PTH(1–4) antibody} takes place in correlation with the amount or concentration of whole PTH(1–84) in the test sample. All beads in the test tubes except the total count tube were washed with the wash solution, and the radioactive signals from each bead were counted for 1 minute using a gamma scintillation counter (ISO-Data, Palatine, IL, USA). The data were processed and calculated using nonlinear regression data reduction software.

Chromatographic separations

Sep-Pak Plus C18 cartridges (Waters Chromatographic Division, Milford, MA, USA) were used for the extraction of PTH from serum samples derived from single individuals or pools from up to 10 individuals among uremic patients, 1°-HPT patients, and normal persons. One cartridge was used for each 3 ml of serum and extracted volumes varied between 12 and 25 ml depending on the PTH concentration.(19) The eluted samples from the cartridges were first evaporated with nitrogen and then the residual volume was freeze-dried. All extracted samples were then reconstituted with 2 ml of 0.1% trifluoroacetic acid and chromatographed on a C18 μ-Bondapak analytical column (3.8 × 200 mm; Waters Chromatographic Division) using a noncontinuous linear gradient of acetonitrile (15–50% in 1.0 g/liter trifluoroacetic acid). After evaporation and freeze-drying, each 1.5-ml fraction was reconstituted to 1 ml with 0.7% bovine serum albumin (BSA) in H2O. Both the whole PTH IRMA and the Nichols PTH IRMA were used to determine the PTH values in each fractionated sample. The recovery of intact PTH throughout all these procedures was 109 ± 10% in normal individuals, 70 ± 14% in renal failure patients, and 108 ± 4% in 1°-HPT.

Samples

One hundred and thirty-five normal human EDTA-plasma and serum samples were obtained from healthy laboratory staff members or donors, with an age ranging from 20 to 62 years (mean ± SD: 42 ± 12.6 years). Three hundred and eighteen patient samples of EDTA plasma (frozen/thawed once) were obtained from uremic patients with ongoing dialysis. The serum samples were collected and allowed to clot for approximately 30–40 minutes at room temperature and then centrifuged at 4°C. EDTA-plasma blood was collected into EDTA sample collection tubes (Becton Dickinson, Franklin Lakes, NJ, USA) and immediately centrifuged at 4°C. The separated EDTA plasma and serum samples were stored at −20°C until used. One hundred and sixty-five samples (111 serum and 54 EDTA-plasma) from patients with surgically proven 1°-HPT were obtained from −70°C sample banks.

A stability study of whole PTH(1–84) in clinical samples was conducted with EDTA plasma, heparinized plasma, and serum. All three types of samples were drawn from three blood donors at the same time. One of the individuals was a patient with 1°-HPT, the other two were normal persons. Samples from only one of the normal persons, who had an original whole PTH(1–84) value of 9 pg/ml, were spiked with synthetic PTH(1–84) to an approximate level of 100 pg/ml. For this study the serum was obtained after routine blood clotting at room temperature for 30 minutes and centrifuged at 2–8°C for 10 minutes; for both EDTA-plasma and heparinized plasma the whole blood was placed immediately into an ice bath and centrifuged at 4°C. All samples were pooled, aliquoted at a 2-ml quantity, and incubated in 2-ml quantities at both room temperature and 2–8°C for 0–72 h, and frozen at −20°C until measured.

RESULTS

Performance characteristics of the whole PTH IRMA

Calibration curve and precision:

An IRMA for whole PTH(1–84) was developed and optimized using the assay procedure described previously. A typical whole PTH IRMA standard curve is shown in Fig. 1. The affinity-purified antibodies used in the assay, either as capture antibody or as125I-labeled antibody, ensured the strong immunoreaction of antigen-antibody binding and low background of 526 ± 86 cpm (mean ± SD) for six iodinations. The intra-/interassay precision was determined by assaying two control samples with whole PTH(1–84) concentrations of 32 pg/ml and 340 pg/ml either by performing 60 replicate measurements in the same assay or in 40 different assays. The within-run variation was 6.1% and 2.3% and the between-run variation was 8.9% and 2.9%. No high-dose “hook” effect was observed after the addition to test samples of synthetic PTH(1–84) up to 20,000 pg/ml.

Figure FIG. 1..

A typical calibration curve obtained with the IRMA for whole PTH(1–84) as described in the Materials and Methods section. Data are expressed as means ± SD of triplicate measurements and are represented directly by the radioactivity (cpm × 1000).

Analytical sensitivity:

The assay detection limit was determined to be 1.0 pg/ml, which was the lowest measurable concentration of PTH value distinguishable from zero. It was determined by measuring the assay standard zero 22 times in the same assay and the value corresponding to the counts of 2 times of SD above the mean of the zero standard. This assay sensitivity was confirmed by validating with three independent production batches of the whole PTH reagents.

Linearity and analytical recovery:

Three patient serum samples with PTH concentrations over 60 pg/ml were diluted 1:2, 1:4, and 1:8 with the assay zero standard. The percent recovery was determined after measurement of the diluted samples. Satisfactory assay linear recoveries of 93–112% were observed within the assay measurement range of 1.0–2300 pg/ml, respectively. Sample spiking recovery was determined by adding two different amounts of PTH into three patient serum samples with known whole PTH(1–84) values. The percentage of sample spike recovery was calculated following the assay of the spiked samples in comparison with the expected value. Recoveries from 99.3 to 113% were observed.

Analytical specificity and interference:

Assay specificity to synthetic PTH(7–84) was studied by comparing this whole PTH IRMA with the Nichols intact PTH IRMA. Nearly 100% cross-reaction to this fragment was observed with the Nichols intact PTH assay, but no cross-reaction was detected with this newly developed whole PTH IRMA even at a PTH(7–84) concentration of 10,000 pg/ml (Fig. 2). The whole PTH IRMA also showed no cross-reaction to other PTH fragments, such as PTH(1–34), PTH(39–84), PTH(44–66), and PTH(53–84).

Figure FIG. 2..

Characterization of assay specificity for two PTH IRMAs [top, Nichols intact PTH IRMA; bottom, whole PTH IRMA; solid-circle, PTH(1–84); open-circle, PTH(7–84)].

Evaluating the specificity of tracer antibodies

The specificities of the two125I-labeled antibodies from the Nichols intact PTH IRMA and this new whole PTH IRMA were compared. Calibrators with a constant PTH(1–84) concentration of approximately 440 pg/ml were determined by both assays with increasing amounts (from 0 to 100,000 pg/ml) of coincubated aminoterminal PTH analogues. In the Nichols intact PTH IRMA, specific binding of125I-labeled tracer antibody to PTH(1–84) was reduced progressively by increasing concentrations of PTH(1–34), PTH(2–34), PTH(3–34), PTH(4–34), and PTH(5–34). In the whole PTH IRMA, in contrast, the bound signal of125I-labeled antibody was only competitively inhibited by PTH(1–34). No binding reduction could be determined by increasing concentrations of PTH(2–34), PTH(3–34), PTH(4–34), and PTH(5–34) (Fig. 3). Increasing concentrations of PTHrP(1–34) had no inhibitory effect on the125I-labeled antibodies in both assays.

Figure FIG. 3..

Characterization of two tracer antibodies used in the Nichols intact PTH IRMA (top) and the whole PTH IRMA (bottom). Data are expressed as means ± SD of duplicate measurements and are represented by percentage changes from the original uninhibited antibody binding.

Assay validations using chromatographic fractionated samples

Figure 4 shows the two different immunoreactive PTH profiles with HPLC fractionated samples from 1 normal person, 1 patient with 1°-HPT, and one patient with 2°-HPT caused by chronic renal failure. The elution position of PTH(1–84) and of PTH(7–84), a prototype of those circulating non-(1–84) PTH fragments, also is indicated. Two immunoreactive peaks were detected in samples from all three groups using the Nichols intact PTH IRMA; the first peak corresponded to the aminoterminal truncated PTH with similar hydrophobicity and elution position as PTH(7–84) and the second one to the immunoreactive PTH(1–84), whereas, only one major immunoreactive peak corresponding to the elution position of PTH(1–84) was detected in all three samples using the newly developed whole PTH IRMA. Results of all HPLC runs are summarized in Table 1. There was a good agreement between the results of whole/intact PTH ratio and the amount of PTH(1–84) obtained by planimetric evaluation of the intact PTH HPLC profiles in the populations studied.

Table Table 1.. Comparison of HPLC Profile Results with Whole/Intact PTH Ratios in Normal Individual, Renal Failure, and 1°-HPT Patients
original image
Figure FIG. 4..

HPLC profiles of immunoreactive PTH present in serum of a normal individual, a 1°-HPT patient, and a hemodialysis patient. Profiles were analyzed using the Nichols intact PTH IRMA and the whole PTH IRMA. Results are expressed as a percentage of the total immunoreactivity. A peak distinct from PTH(1–84) is detected by the intact PTH assay but not by the whole PTH(1–84) assay.

Sample stability for the whole PTH(1–84) measurement

The stability of whole PTH(1–84) was studied as follows: (1) in serum, EDTA plasma, and heparinized plasma; and (2) at 2–8°C and at room temperature (RT). The results indicated that: (a) whole PTH(1–84) in EDTA plasma and heparinized plasma is stable (<5% degradation) at 2–8°C or RT for at least 24 h; and (b) whole PTH(1–84) in serum, however, is only stable for 6 h at RT (>10% degradation) and for about 24 h at 2–8°C (Fig. 5). Additionally, a study of four times sample freeze/thaw showed that both serum and EDTA plasma were relatively stable with a <5% decrease in immunoreactivity.

Figure FIG. 5..

Sample stability for the whole PTH(1–84) measurement. Data are expressed as means ± SD of duplicate measurements and are represented by percentage changes from the original concentrations.

Assay correlation and clinical evaluation

The normal range of whole PTH(1–84) was found to be 7–36 pg/ml (mean ± SD: 22.7 ± 7.2 pg/ml; n = 135) for EDTA plasma.

To study the correlation and difference between whole PTH(1–84) and conventional intact PTH levels in normal persons, 56 normal human EDTA plasma samples were measured at the same time with two different PTH assays, the newly developed whole PTH IRMA and the Nichols intact PTH IRMA. All the samples had measurable whole PTH(1–84) values. There were also measurable PTH values in all normal samples using the Nichols intact PTH IRMA. However, all intact PTH values measured by the Nichols PTH IRMA were higher than the whole PTH(1–84) values (Table 2) revealing an average of about 33% PTH fragments being comeasured with PTH(1–84) by intact PTH assay. Paired Student's t-test showed a significant difference (p < 0.0001) between the two sets of PTH values with these two PTH IRMAs (Fig. 6, bottom). The correlation of these two groups of PTH values also was calculated (r = 0.923; slope = 1.456).

Table Table 2.. Comparison of Intact PTH Values, Whole PTH(1–84) Values, and pB% {(whole PTH Value/Intact PTH Value) × 100%} in Patients with Uremia and Surgically Proven Primary HPT with Normal Persons
original image
Figure FIG. 6..

Assay correlation studies of 56 normal persons (bottom; open circle, Nichols intact PTH IRMA; solid diamond, whole PTH IRMA) and from 7 artificial samples containing only whole PTH(1–84) (top). Data are expressed as means of duplicate measurements.

To ensure that this difference of PTH values was only caused by the specific antibody-antigen binding and not caused by differences in assay matrix or calibrators, different amounts of synthetic PTH(1–84) were spiked into several normal human sera with nondetectable PTH levels and measured with the previously mentioned two PTH assays. The result showed these two assays detect PTH(1–84) equally (r = 0.999; slope = 1.04; Fig. 6, top).

Human PTH values from a sample group of 318 uremic patients with ongoing hemodialysis also were determined with these two assays. The results showed that the PTH values displayed a heterogeneous distribution pattern in normal, below-normal, and elevated levels using both assays. The mean and median for the whole PTH(1–84) in this group also differed significantly from that obtained with the Nichols intact PTH assay (p < 0.0001; paired Student's t-test; Table 2). Figure 7 shows the correlation comparison of these two assays in the uremic patient group (r = 0.977; slope = 1.482). Samples from 165 patients with surgically confirmed 1°-HPT with parathyroid adenomas also were measured using the whole PTH IRMA (mean ± SD: 116.7 ± 129.6 pg/ml) and the Nichols intact PTH IRMA (mean ± SD: 200.3 ± 208.9 pg/ml). An effective differentiation of this patient group from normal persons was observed (Fig. 8). The overall clinical diagnostic sensitivity with a single sample PTH measurement was 93.9% (155/165) using whole PTH IRMA and 91.5% (151/165) using Nichols intact PTH IRMA.

Figure FIG. 7..

Assay correlation study of 165 1°-HPT samples (top) and 318 uremic samples (bottom) using the Nichols intact PTH IRMA and the whole PTH IRMA. Paired Student's t-test was used for p value calculation.

Figure FIG. 8..

Scatterplot of whole PTH(1–84) values in healthy controls and various patient groups. Shaded area indicates the plasma normal range (7–36 pg/ml) of whole PTH(1–84). The y axis is expressed by log2 scale. The whole PTH levels of 10 1°-HPT patients were located in the upper normal range and the overall diagnostic sensitivity was 93.9% (155/165).

The ratios of whole PTH to intact PTH or percentage of biologically active PTH(1–84) (pB%) to the total immunoreactive intact PTH were calculated for all 318 uremic patients and 165 1°-HPT patients. The results display an almost Gaussian distribution pattern from 20% to >90% in both patient groups (Fig. 9). This inconsistent pB% may be the result of variations in peripheral clearance of PTH or the glandular secretion of PTH(1–84) and its fragments.(20) This finding further indicates that currently available intact PTH values could not assess accurately the parathyroid function of patients.

Figure FIG. 9..

Histogram showing the frequency distribution of the pB% in the pool of the total immunoreactive intact PTH value in patients with 1°-HPT (n = 165, top) and 2°-HPT (n = 318, bottom) of uremia.

DISCUSSION

The present report describes for the first time an immunoassay that measures only the biologically active whole PTH(1–84) without any cross-reactivity to PTH fragments, although current intact PTH immunoassays have been used and presumed to be specific for intact PTH for over 10 years. One study evaluated serum intact PTH levels in conjunction with histological analyses of iliac crest bone biopsy specimens.(7) It was found that serum intact PTH assays overestimate the presence and severity of PTH-mediated osseous abnormalities associated with uremia. Although at that time the reason for this overestimation was not elucidated, it might have been explained partially by this work in combination with recent studies.(10,11,21) It has been shown that the commercially available intact PTH assays measure both PTH(1–84) and non-PTH(1–84) fragments that are present in significant concentrations in the blood of uremic patients.(10) Therefore, these intact PTH assays are not truly intact PTH specific and the term “intact” is used inaccurately.

The specificity studies of the tracer antibody show that the newly developed anti-PTH(1–4) antibody is truly aminoterminal PTH specific. In fact, it is directed at the first amino acid of the aminoterminal polypeptide (Fig. 3), therefore, being able to bind to PTH(1–34) but not PTH(2–34), -(3–34), -(4–34), and -(5–34). By contrast, the tracer antibody used in Nichols intact PTH IRMA is broadly PTH(1–34) specific and, therefore, cross-reacts with PTH(2–34), -(3–34), -(4–34), and -(5–34). It is the specificity of the tracer antibody used in this new whole PTH assay that ensures that this unique assay only detects the full-length PTH(1–84) without cross-reaction to any aminoterminally truncated PTH fragments. In theory, this assay also could detect carboxy-terminally slightly truncated PTH fragments, which should be the same for other intact PTH assays. Using the commercially available synthetic aminoterminally truncated PTH fragment, PTH(7–84) other intact PTH assays (Incstar, Diagnostic System Laboratory, Diagnostic Product Corp.) show variable cross-reactivity of 60–80% from assay to assay.(22) The whole PTH IRMA was thoroughly designed and developed in a coated bead format and single incubation step. It is easy to perform and presents a clinically adequate measurement range of 1–2300 pg/ml with acceptable assay performance characteristics, including linearity, sample spiking recovery, and intra-/interassay precision.

The study of the chromatographically fractionated serum samples from normal population and patients with either 1°-HPT or 2°-HPT further shows that there are two forms of PTH or immunoreactive peaks detected by the Nichols intact PTH IRMA. The first immunoreactive peak corresponds to non-PTH(1–84) fragments migrating on HPLC to a similar position as PTH(7–84) and the second peak corresponds to the full-length PTH(1–84).(10) However, when the same samples were measured with the whole PTH IRMA, only one immunoreactive peak was detected corresponding to the full-length PTH(1–84). Comparing the specificity of the antibodies used in these two assays, it is quite obvious that the non-PTH(1–84) corresponds to aminoterminally truncated PTH fragments. Moreover, these HPLC fractionated patient sample measurements further show that these aminoterminally truncated polypeptides are present in significant amounts not only in uremic patients, but also in the normal population and in patients with 1°-HPT (Fig. 4; Table 1). The exact molecular structure of these PTH fragments should be further determined by isolating and analyzing their amino acid sequences using pools of patient serum samples.

The correlation study of whole PTH IRMA to Nichols intact PTH IRMA from samples that contain only synthetic PTH(1–84) indicates that the two assays are nearly equivalent in their detection of PTH(1–84) (Fig. 6, top). However, when clinical samples from a normal population group and patients with 1°-HPT or 2°-HPT were used for the study, significant differences with higher intact than whole in the absolute PTH values were found (p < 0.0001, paired t-test) in all three groups (Figs. 6 and 7; Table 2).

The clinical significance of this newly developed whole PTH IRMA was shown in three population groups. The normal range of whole PTH(1–84) was 7–36 pg/ml for samples of EDTA plasma. Samples of EDTA plasma are preferred for whole PTH measurement because the hormone appears to be more stable in EDTA plasma than in the serum (Fig. 5). There is an unexpected distinction in whole PTH(1–84) levels of patients with 1°-HPT from the normal population with an overall diagnostic sensitivity of 93.9% (n = 165) in this study. A diagnostic sensitivity of 91% also was found with Nichols intact PTH IRMA in this study. However, Kao et al.(23) evaluated 361 patients with surgically proven 1°-HPT in whom intact PTH had been determined with an immunochemiluminometric assay and found 45 patients to have an intact PTH value below the upper limit of normal. Endres et al.(24) also reported that only 21 of 29 cases of 1°-HPT had values above the normal level when the Nichols Allégro intact PTH assay was used. These early studies indicated a diagnostic sensitivity of intact PTH assay of about 72.4–87.5% only. Most recently, Silverberg et al.(25) reported a prospective clinical validation using whole PTH assay, Nichols intact PTH assay, and a midregional PTH competitive assay. In her study, a well-defined group of patients with mild 1°-HPT was chosen and the clinical diagnostic sensitivities were 96% for whole PTH assay, 76% for intact PTH assay, and 54% for a midregional PTH assay. Significant statistical differences were found between each assay in this study. Whole PTH(1–84) values from 318 uremic patients displayed a heterogeneous distribution pattern with both normal and elevated levels.

This study has shown that there is no consistent percentage of aminoterminally truncated PTH fragments (Fig. 9; Table 2). It is inconsistent percentage of aminoterminally truncated PTH fragments among patients with HPT that could easily give rise to two previously unforeseen major problems in the clinical decisions based on available intact PTH assays for evaluating the function of the parathyroid glands. First, because most intact PTH assays have >60% cross-reaction(10) to the PTH fragments and the ratio of whole PTH/intact PTH or the pB% is not consistent even in patients in the same disease condition, the parathyroid function will always be overestimated and inconsistently estimated in different degrees by intact PTH assays measuring both the full-length whole PTH(1–84) and its aminoterminally truncated fragments. Second, because of the significantly different molar rates of cross-reactivity of commercially available intact PTH assays, interlaboratory discordance of PTH levels have been observed from the use of different intact PTH assays. Theoretically, the aminoterminally truncated PTH fragment is a naturally produced polypeptide, which is able to bind to PTH/PTH-related protein (PTHrP) receptors. One preliminary in vivo study with parathyroidectomized rats showed an 80% decreased calcemic response for a 1:1 molar ratio of infused PTH(7–84) and PTH(1–84) compared with PTH(1–84) alone.(26) The biological importance of these aminoterminally truncated fragments that have been shown to act as PTH antagonist or inhibitor appears to regulate eventually the sensitivity of PTH/PTHrP receptors and warrants further investigation. These PTH fragments also could be ligands for a thus far unisolated receptor for the carboxy-terminal part of PTH. However, whether this receptor plays a role in the regulation of calcium metabolism is not known.(27)

In summary, a novel IRMA was developed that only detects biologically active whole PTH(1–84) without cross-reaction to the aminoterminally truncated PTH fragments. The assay uses only a single incubation procedure. The PTH(1–84) specificity of the new assay was defined by tracer antibody evaluation, cross-reactivity experiments, and measurements of HPLC fractionated patient samples. With this whole PTH IRMA, we first showed that previously described non-(1–84) PTH fragments(10) should be aminoterminally truncated. The presence of these aminoterminally truncated PTH fragments was shown not only in uremic patients, but also in 1°-HPT patients and normal persons. Moreover, the percentage concentration of the biologically active whole PTH(1–84) in the pool of total immunoreactive intact PTH is significantly variable from patient to patient, even in patients with the same type of HPT and, thus, it is impossible to interpret biologically active PTH levels with current intact PTH assays. The new whole PTH IRMA is clinically significant in differentiating patients with 1°-HPT and 2°-HPT from the normal population in measuring PTH(1–84) exclusively. Because of the immunological heterogeneity of circulating PTH, this new assay model could be applied as a more meaningful and standardized method for the measurement of biologically active and hence clinically significant PTH.

Acknowledgements

The authors acknowledge the contributions of John Van Duzer, Jim Killion, Carolyn Costlow, and Damon Cook for their invaluable assistance. M.R.J. was supported by grants from Deutsche Forschungsgemeinschaft (JO 315/1-1 and JO 315/1-2).

Ancillary