| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
1
DiaSorin Inc., Stillwater, MN 55082.
2
Northwestern Medical School and Children's Memorial
Hospital, Division of Nephrology and Mineral Metabolism, Chicago, IL
60611.
3
Mayo Clinic and Foundation, Division of
Endocrinology/Metabolism, Rochester, MN 55902.
a Address correspondence to this author at: DiaSorin Inc., P.O. Box 285, 1990 Industrial Blvd., Stillwater, MN.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: We generated goat antibodies and N- and C-terminal Oc. The former was used on solid phase (polystyrene beads), and the latter was used as the tracer in an IRMA.
Results: The assay was linear with no cross-reactivity to Oc(1–43), total imprecision (CV) of <10%, and recovery of 100% ± 10%. Assay values for intact Oc in EDTA plasma samples were unchanged at 18–25 °C for 6 h. Values for intact Oc in serum, EDTA plasma, and heparin plasma samples did not change after storage on ice for 8 h. Serum samples from patients with various conditions were stored at -70 or -135 °C for up to 5 years and yielded z-scores comparable to an Oc(1-43) IRMA for all conditions except for renal failure. In renal failure, the Oc(1–43) assay values were increased, whereas the intact assay values were in the reference interval.
Conclusion: Decreases in Oc assay values are inhibited by calcium chelation, and slowed by reduced temperatures. The described assay for intact Oc allows improved specificity for bone compared with an assay for Oc(1–43).© 1999 American Association for Clinical Chemistry
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The serum concentration of Oc been shown to correlate with the extent of metabolic bone diseases characterized by increased bone turnover (8)(9)(10)(11)(12)(13). Additionally, Oc can be a specific marker for bone formation when formation and resorption are not balanced (10).
Several efforts have been made to improve the clinical interpretation
of assay results through (a) the use of human Oc calibrators
and anti-human Oc antibodies rather than bovine components,
(b) the reporting of patient values as a ratio to
healthy subject values in that assay, and (c)
validation of assays to determine the primary sequences to which
capture and tracer antibodies bind (14). The extent to which
immunoreactivity is affected by the tertiary structure of Oc is an
important difference among existing Oc assays (15). The
tertiary structure of Oc has been proposed to contain two
calcium-binding 
-helices surrounding a disulfide-stabilized
ß-turn. Lack of the disulfide bond between
Cys23-Cys29 or a
decarboxylation of 
-carboxy-glutamic acid has been shown to disrupt
the tertiary structure of Oc (16). If changes in tertiary
structure affect antibody affinity, immunoassay results may be affected
by heterogeneous carboxylation or by the introduction of protease
inhibitors that themselves bind calcium. In addition, some groups have
reported that the freeze-thaw stability of Oc varies, whether in serum
or buffer (17), suggesting that the immunoreactivity of
antibodies can be affected by other physical factors.
We therefore explored these issues as we analytically validated an assay for intact Oc assay. After the analytical validation, we examined samples from clinically characterized patients to evaluate the potential of the assay for clinical application.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
preparation of the solid phase
Rabbit -goat antibody was adsorbed to a polystyrene bead
(Hoover Precision), with the N-terminally directed antibody
captured by the rabbit -goat antibody and the beads blocked with
normal horse serum to prevent nonspecific binding.
preparation of the reporter antibody
The goat anti-Oc(43–49) was radiolabeled with
125I using the chloramine T method. Briefly,
tyrosine residues were activated using chloramine T oxidation in the
presence of 125I. The chloramine T was quenched
with an excess of sodium metabisulfite. Free 125I
was separated from labeled antibody by size-exclusion HPLC.
calibrators and controls
Oc(1–49) purified from human bone was sent blindly to two
independent laboratories for quantitative amino acid analysis. The
results obtained agreed within 4%. Antigen vials were assigned a
concentration based on this quantitative amino acid analysis, and
calibrators were prepared in 0.05 mol/L borate containing 5 g/L bovine
serum albumin. The controls were built from the same antigen stock, but
the control stock and dilution to control concentration were made in
0.05 mol/L borate containing 100 mL/L normal human serum.
wash buffer
The wash buffer was 5 mL/L Tween 80 in 0.1 mol/L
phosphate-buffered saline.
assay procedure
Calibrator, control, or unknown samples (20 µL) were pipetted
into the bottoms of borosilicate or polystyrene tubes.
125I-labeled anti-Oc(43–49) antibody (300 µL)
was added to all test tubes. The test tubes were vortex-mixed, and
beads were dispensed into each tube. The tubes were incubated for
2 h at 18–25 °C and 190 ± 10 rpm (2.0-cm diameter orbit
shaker), and then aspirated and washed three times with assay wash
buffer. Test tubes were counted in a gamma counter. Radioactivity was
directly proportional to the concentration of Oc. The radioactivity
(cpm) of the calibrators was used to construct the calibration curve
using a spline-smoothed curve.
specificity and suppression of assay binding
Oc peptides were added to the assay E calibrator (200 µg/L), and
these samples were assayed as unknowns. Reduced cpm in the samples with
added Oc fragments compared with a control E calibrator indicated that
the fragment was bound by one of the component antibodies of the
reported assay.
imprecision
The interassay CV was assessed with three pools of with below
(hypo), within (normal), and above (hyper) the estimated health-related
reference interval.
patient population for dilution study
Sera were collected over a 2-week period, during routine testing,
from seven children (age range, 7–14 years) with chronic renal
insufficiency from either glomerular (n = 2) or
tubulo-interstitial (n = 5) diseases in whom the residual renal
function was 25–70% of normal. An additional 15 sera were collected
over a 3-week period, during routine testing, from children (age range,
6–18 years) referred to the clinical practice of one of the
investigators (C.B.L.) for suspected disorders of bone metabolism, but
in whom no pathologic process was identified subsequently. Samples were
stored at -20 °C until assayed, and they were thawed and stored on
ice during the assay. Samples were refrozen within 60 min, and no
samples were frozen and thawed more than twice. The samples from adults
with renal failure were collected before routine hemodialysis and had
increased parathyroid hormone; all of these adult samples were
obtained anonymously from the Minneapolis regional kidney dialysis
program with informed consent.
dilution linearity
Samples from children with renal insufficiency were used to
validate dilution linearity. These samples were chosen for the high
content of Oc and Oc breakdown products. The clinical samples were all
diluted with calibrator matrix, and all dilutions were serial
dilutions.
z-SCORE ANALYSIS OF PATIENT
POPULATION
For evaluation of the assay, fasting state serum samples were
obtained from healthy young women [n = 30; age, 32.0
± 2.8 years (mean ± SD)]; healthy elderly women (n = 30;
age, 74.2 ± 3.4 years); estrogen-treated postmenopausal women
(n = 15; age, 73.5 ± 3.6 years); hypoparathyroid subjects
(n = 5; age, 61.6 ± 12.9 years); untreated osteoporotic
women (n = 15; age, 65.0 ± 4.5 years); patients with
Paget disease (n = 5; age, 75.4 ± 13.8 years);
hyperparathyroid subjects (n = 5; age, 54.6 ± 16.8 years);
and patients with renal failure (n = 5; age, 66.4 ± 6.1
years). All of these samples were obtained from the Mayo Clinic with
informed consent.
comparison of z-SCORES
Two immunoassays for Oc from CIS Biointernational were compared
side by side with the reported assay. The CIS-ELSA-OSTEO assay measures
the fragment Oc(1-43) in addition to the intact molecule. The
CIS-ELSA-OST-NAT assay measures intact Oc exclusively. The healthy
young and elderly women had bone mineral density values within 2 SD of
the age-matched mean, no history of fracture, and were not on
medication known to affect calcium metabolism. The disease subjects
were all classified using standard clinical criteria.
sample stability analysis
Samples were obtained from healthy male and female donors 21–60
of age from the DiaSorin Inc., employee population with informed
consent. Multiple EDTA plasma, heparin plasma, serum separator, and
serum tubes were collected. To obtain t = 0 samples, one EDTA
whole blood and one heparin whole blood aliquot was centrifuged in a
4 °C Eppendorf 5414 microcentrifuge to separate cellular material.
The plasmas were aspirated and stored at -70 °C for 13 samples
within 10 min of drawing and within 15 min for all samples. Because of
the difficulty of obtaining and freezing serum within 10–15 min and
the parallel nature of degradation between heparin plasma and serum,
the heparin plasma t = 0 sample value was used as the t = 0
sample value to model degradation in the serum stability analysis.
The remaining EDTA whole blood and heparin whole blood tubes were allowed to settle for 60 ± 20 min at 2–8 °C and 18–25 °C before centrifugation at 760g and collection of plasma. Samples were centrifuged at the temperature designated for that experimental condition. All plasmas were separated by the 2 h time point, and subsequent incubation occurred without cellular material present. The serum was allowed to clot at 18–25 °C for 30 ± 10 min, then was stored at 2–8 °C and 18–25 °C for an additional 30 ± 15 min before centrifugation at the appropriate temperature for that experimental condition. Serum was separated by the 2 h time point, and subsequent incubation occurred without cellular material present.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The limit of detection (2 SD above the value for nonspecific binding) in three separate assays with three lots of tracer at 8 days postiodination was 0.03–0.13 µg/L, with a mean (± SD) of 0.07 ± 0.05 µg/L. The reported limit of detection for the assay is 0.2 µg/L.
imprecision
Samples were run as duplicates in 10 different assays by six
different technicians over 20 days. The interassay imprecision (CV) was
7.1–9.5%. The mean recovery of Oc at 10, 30, 60, and 110 µg/L added
to patient sera, assayed in duplicate in three assays by two
technicians, was 90–110%.
Linearity was evaluated with serial dilutions with the zero calibrators of sera from pediatric (n = 10) and adult (n = 12) renal patients. Renal samples were chosen because they contain high concentrations of both Oc and Oc breakdown products. The undiluted values ranged from 3.4 to 72 µg/L. The geometric mean values of observed/expected results at dilutions of 1:2, 1:4, and 1:8 were 104%, 104%, and 106%, respectively.
Synthetic human Oc(1–43, GLA21,24) produced no
detectable cross-reactivity up to 11 000 µg/L. The lack of
suppression of the Oc(9–16) and Oc(30–43) peptides and the
suppression of the Oc(1–16), Oc(37–49), and Oc(30–49), shown in Fig. 1
, demonstrated that the capture antibody binds primarily to
Oc(1–10) and the tracer binds in the region of Oc(43–49). The assay
recognized (1–49, GLA17,21,24), (1–49,
GLA21,24), and (1–49,
GLU17,21,24) equally (data not shown).
|
Matched serum, EDTA plasma, and heparin plasma specimens from
healthy subjects were assayed after storage in an ice bath (2–8 °C)
for 8 h and at 18–25 °C for 6 h (Fig. 2
). Storage on ice for up to 8 h had no effect on
results for serum or plasma. A significant (P
<0.0001) change was observed after storage for 2 h for heparin
plasma and serum at 18–25 °C, but the change was not significant
even at 6 h for EDTA plasma. Values for matched serum and heparin
plasma samples were not statistically different from each other during
storage at 18–25 °C, and thus the rates of change of assay values
were not different in heparin plasma compared with serum. Values for
EDTA plasma samples were slightly lower than heparin plasma values at
t = 0, which could be attributed to a modest calcium-dependent
matrix effect. Addition of as little as 2 mmol/L EDTA to the
calibrators decreased the cpm, suggesting that the EDTA directly or
indirectly inhibits antibody binding in the described assay. No
significant changes were observed over time in EDTA plasma values in
this experiment; however, changes were observed in both heparin plasma
and serum, suggesting that calcium could be involved in the observed
decreases in the amount of immunoreactive Oc that could be assayed over
time. Changes in plasma sample values were also explored in a
separate condition where the intact cellular material was simply
allowed to settle at 2–8 °C and 18–25 °C. We did not test
samples with disrupted cellular materials. These additional conditions
yielded no difference in degradation rates compared with separated
plasma (data not shown). Repeated freeze-thaw cycles of plasma yielded
sample value imprecision within the magnitude of the interassay
precision.
|
z-SCORE ANALYSIS
The CIS ELSA-OSTEO Oc(1–43) assay and the two intact Oc(1–49)
assays had dissimilar absolute means (Table 1
), but had similar z-scores (Table 2
) for most of the tested conditions. The two intact assays did
not differ significantly in z-scores for any population
tested. For patients with renal failure, the CIS ELSA-OSTEO Oc(1–43)
assay z-score was significantly higher than that of the
Oc(1–49) assays (P <0.0001).
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The N-terminal Oc(1–43) fragment has been shown to circulate in vivo, but its origin is currently unknown. It has been reported in the supernatant of osteoblastic cells, but the study was performed without a control for protease activity in the tissue culture from the matrix itself or from the cells (18). Therefore, it is not clear if some Oc(1–43) is a de novo product or if it is exclusively a breakdown product.
In the absence of convincing evidence that this Oc(1–43) fragment is produced by osteoblasts, it would seem intuitive that the preferred assay for osteoblast activity would be an intact Oc assay that does not measure fragments, the production of which is highly dependent on kidney function. In this retrospective investigation of the classification accuracy of intact Oc measurements of banked serum samples, there appeared to be no improved discrimination between healthy patients and disease states using an IRMA that also recognizes the Oc(1–43) fragment. However, in the renal failure samples, the Oc(1–43) assay values appear significantly increased, whereas the intact Oc values not. This suggests that an intact Oc IRMA without cross-reactivity to breakdown products has greater specificity to bone metabolism in cases of renal insufficiency. Moreover, as assayed by the intact assay, samples were stable under various collection and storage conditions for up to 8 h, and retrospective analysis of frozen serum samples stored up to 5 years provided expected results.
Oc immunoreactivity has repeatedly been shown to be affected by many natural and some artificial phenomena, with decreased immunoreactivity attributed to degradation of Oc. It was reported that protease inhibitors have protective effects on sample stability before freezing in one assay (18), had no effect in the present assay, and decreased the observed immunoreactivity of a third assay during freeze-thaw cycles (19). A control for possible matrix effects in the report of protective effects of a protease cocktail (18) was not reported, although one of the protease inhibitors used is known to bind calcium, which has been shown to affect the tertiary structure (16) and immunoreactivity of Oc in some assays (15). One Oc assay that is said to measure Oc(1–43) as well as the intact molecule nonetheless has problems with sample stability (20); the specificity of that assay must be questioned, however, on the basis of our earlier work (21). Several assays that measure Oc(1–43) in addition to the intact molecule showed increased immunoreactivity in samples stored in liquid form, apparently because the immunoreactivity of a synthetic Oc(1–45) fragment was higher than that of native Oc (15).
Because in vitro degradation of Oc clearly occurs and the reported conclusions vary with the immunoassays used to study it are contradictory, data should be interpreted with caution. Immunoassays can only suggest degradative events. Effects of matrix and tertiary structure must be explored thoroughly and reported. One constraint of the present assay is the reduced cpm caused by EDTA. Conclusions about Oc measurements will be more credible with these additional data. Because the reported immunoreactivity is dependent on so many variables, it appears necessary to assess apparent sample stability and clinical classification accuracy of each assay individually.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-carboxyglutamate in mineralized tissue. Proc Natl Acad Sci U S A 1975;72:3925-3929.
-carboxylated in humans. J Bone Miner Res 1994;9:1989-1987.
[Web of Science][Medline]
[Order article via Infotrieve]
carboxyglutamic acid, serum osteocalcin as bone markers. Studies in osteoporosis and Paget's disease. J Clin Endocrinol Metab 1983;57:1221-1225.
-helical structure in osteocalcin. Biochemistry 1982;21:2538-2547.
[Medline]
[Order article via Infotrieve]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
