CLIN. CHEM. 39/7, 1369-1374 (1993)

CLINICAL CHEMISTRY, Vol. 39, No. 7, 1993 1369

Development and Evaluation of an Osteocalcin ChemiluminoimmunoassayPai C. Kao,”2 B. Lawrence Riggs,3 and Patricia G. Schryver1

We developed a competitive chemiluminoimmunoassayof osteocalcin that is similar to radioimmunoassay but

uses acridinium-ester-labeled antigen instead of 125,

labeled osteocalcin; a second antibody immobilized onplastic beads is used to separate free and bound frac-tions. There was good correlation of the new chemilumi-noimmunoassay (y) with a polyclonal antiserum (R102)radioimmunoassay (x) used in many previous clinicalstudies (r = 0.96, y = 0.968x + 2.69, S = 0.029, n =

86). The new assay recognized both the intact and thesmall fragment of osteocalcin in plasma and detecteddecreases of them (total, -47%) after a 24-h infusion ofparathyroid hormone. Patientswith primary hyperparathy-roidism had increased concentrations of intact osteocal-cm. Children had higher concentrations of osteocalcinthan adults did. Healthy women had greater osteocalcinconcentrations at ages 50-70 years than earlier. Inversecorrelations of bone mineral density and osteocalcin werefound in healthy women and in women with osteoporosis.

IndexIng Terms: bone osteoporosis reference valuesage-related effects

An ideal biochemical marker of bone formation wouldbe a matrix constituent of bone specifically synthesizedby osteoblast8 at a rate proportional to collagen synthe-sis, with a fraction released into the circulation in anamount proportional to the amount of the fractionincorporated into the matrix. None of the currentlyavailable biochemical markers of bone formation meetall these criteria, but osteocalcin is the most satisfactorymarker (1). Ost.eocalcin (bone-Gla-protein) is a bonematrix protein of 49 amino acid residues (molecularmass 5800 Da) containing three y-carboxyglutamic ac-ids (Gla) by posttranslational carboxylation of glutamicacid at positions 17, 21, and 24 (2-4). Osteocalcin issynthesized mainly in bone by osteoblasts, and 70% isreleased into the circulation. Only a small amount issynthesized in dentin, but none in the liver or otherorgans. Thus, osteocalcin is a specific biochemicalmarker of bone formation (1).

Bone formation can be measured by histomorphome-try on the bone biopsy specimen, prernarked by tetracy-cline double-labeling, similar to measuring tree growth

1Department of Laboratory Medicine and Patholo’, and3 Di-vision of Endocrinoloi, Metabolism, and Internal Medicine, MayoClinic and Mayo Foundation, Rochester, MN.

2Mdress for correspondence: Mayo Clinic, 200 First St. SW,Rochester, MN 55905.

Received December 9, 1992; accepted February 9, 1993.

between rings. Histomorphometry is accurate but inva-sive; it is also time-consuming, requiring 12 to 23 days.

Measurement of serum osteocalcin is both noninvasiveand sensitive; bone formation rates determined by his-tomorphometry of bone with tetracycline double-label-ing correlated well with serum osteocalcin measured byradioimmunoassay (RIA) (r = 0.7) with polyclonal anti-serum R102 (5, 6). Patients with corticosteroid-inducedosteoporosis who were treated with sodium fluoride,which increases bone formation, had a marked increasein serum osteocalcin values but showed no change inserum alkaline phosphatase activity (1). Serum osteo-calcin has also been demonstrated to be useful formonitoring the effects of specific treatment for bonediseases (6), for determining high- and low-turnoverbone disease of renal osteodystrophy, and for monitoringbone formation and remodeling (5).

Parathyroid hormone (PTH) in vitro increased bonedegradation or resorption and released calcium frombone tissue (7). Infusion of PTH for 24 h decreased serumosteocalcin in normal women (8), even though patientswith chronic diseases of primary and secondary hyper-parathyroidism had increased osteocalcin (9, 10). ThePTH in vitro incubation and in vivo infusion experimentsfurther demonstrated that osteocalcin is a marker of boneformation, not bone resorption. The heterogeneity ofcirculating osteocalcin was demonstrated by gel-ifitra-tion chromatography, which yielded intact osteocalcinand a small fragment of the protein (11). The decrease inserum osteocalcin after PTH infusion may occur as adecrease of the intact osteocalcin, of the low-Mi osteocal-cm fragment, or of both. This dynamic change wasdetected also by separation of serum with a gel-filtrationcolumn, followed by RIA measurement with a monocle-nal antibody that recognized both intact osteocalcin andthe low-Mr fragment. After Pm infusion, both intactosteocalcin and the fragment were decreased, but thefragment to a lesser degree. Because antiserum .R102,which recognized the intact osteocalcin, detected thedecrease of only intact osteocalcin, the results of themonoclonal antibody assay were greater than that of thepolyclonal assay by -1.5 times (11). To be useful, an

osteocalcin assay should be able to detect the change incirculating osteocalcin after PTH infusion.

The purposes of this study were to characterize thenew chemiluminoimmunoassay (CIA) by comparing itwith the RIA with antiserum R102, to determinewhether the new assay also recognizes both the intactand the 1OWMr fragment of osteocalcin, and to detect

1370 CUNICAL CHEMISTRY, Vol. 39, No. 7, 1993

any changes in their concentrations after 24-h infusionof PTH.

Materials and MethodsPreparation of Osteocalcmn Tracer

Bovine osteocalcin, 50 g (purchased from BiomedicalTechnologies, Inc., Stoughton, MA), was reconstitutedwith 200 j.L of 0.1 moIJL phosphate buffer, pH 8.0. Thereconstituted osteocalcin solution was transferred to avial containing 0.2 mg of N-hydroxysuccinunude acridi-mum ester (London Diagnostics, Eden Prairie, MN) andallowed to react for 15 miii with occasional mixing.Then, lysine HC1 (100 L of a 10 gIL solution) was addedto the reaction vial for an additional 15 mm. Bothlabeling and lysine quenching reactions were at roomtemperature. The labeled solution was injected into ahigh-pressure liquid chromatograph equipped with aC18 column (4.6 x 250 mm, 10-im particle; Alltech,Deerfield, IL) and eluted with a gradient of acetomtrilein 1 g’L trifluoroacetic acid solution, from 0/100 to 60/40(by vol), over 60 miii; fractions of the effluent werecollected at 1 mL/min. Aliquots (5 L) from each frac-tion were taken to determine chemiluminescent emis-sion in a luminometer (Magic Lite-Il; Ciba Corning,Park Ridge, IL). The third peak of chemiluminescence,which gave the highest binding with bovine osteocalcunantibody (Biomedical Technologies, Inc.), was pooledand stored at 4#{176}Cfor use as tracer.

Purtflcation and Immobilization of Second Antibody

Goat antibody to rabbit gamma globulin (IgO) waspurified with a rabbit IgG affinity column, which wasprepared by conjugating 1.0 mg of rabbit IgG to 1.0 g ofcyanogen bromide-activated Sepharose 4B gel (SigmaChemical Co., St. Louis, MO). Goat antiserum (2.0 mL)was applied to the affinity column and washed with 0.1mol/L sodium phosphate buffer containing 0.5 mol ofNaC1 and 1 mL of Triton X-100 per liter, pH 7.4. Theserum that passed through the column was reappliedand then washed with an additional 50 mL of the same0.1 mol/L phosphate buffer, pH 7.4. The column wasfurther washed with 10 mL of 0.01 mol/L phosphatebuffer, pH 6.0. The antibodies that were immunoad-sorbed on the IgG affinity column were eluted with HC1solution, pH 2.0. Fractions of 1.0 mL were collected andneutralized with 150 tL of 0.1 moIIL phosphate buffer,pH 8.0. The protein content of the antibodies in eachfraction was measured by the protein determinationmethod of Lowry et al. (12).

To coat plastic beads by passive adsorption, we diluted1 mg of the immunoaffinity-purifled goat anti-rabbitIgG in 200 mL of 0.1 mol/L sodium carbonate buffer, pH9.7, then mixed this solution with 1000 polystyrenebeads (Clifton Plastics, Clifton Heights, PA) and incu-bated for 2.5 h at room temperature on an orbital shakerat 100 rpm. To block any empty adsorption sites, wepoured the beads plus the incubation solution into aBuchner funnel, washed them with 1.0 L of doublydistilled water, then transferred the beads into 200 mLof 0.1 mol/L phosphate buffer containing glycine HC1,

100 g/L, pH 7.4, and incubated at room temperature onan orbital shaker at 100 rpm for an additional 2.5 h. Weused glycine for blocking because the larger-Me bovineserum albumin had a tendency to remove antibody fromthe beads. Finally, we washed the beads with 1 L ofdoubly distilled water and stored them at 4#{176}C.

ChemiluminoimmunoassayAcridiniuni-ester-labeled osteocalcin was diluted to

125 000 relative light units per 100 L with assay buffer

(0.01 mol/L phosphate buffer containing, per liter, 0.123mol of NaC1, 25 mmol of EDTA, 1 g of bovine serumalbumin, and 1 mL of Tween 20, pH 7.4). Lyophilizedosteocalcin antiserum (Biomedical Technologies, Inc.,500 tubes/vial) specific for intact osteocalcin, with therecognition site at the carboxyl terminus, was reconsti-tuted with 1.0 mL of assay buffer and further diluted50-fold with assay buffer. For the assay, we incubated100 L of labeled osteocalcun, 25 4 of sample orstandard, and 1004 of diluted antibody plus 1004 ofassay buffer overnight at room temperature. Then, afteradding a bead immobilized with second antibody, weshook the sample on an orbital rotator at 150 rpm for 3h at room temperature. After washing the bead threetimes with water, we measured the chemiluminescenceon the bead with an automated luminometer. The cali-bration curve was prepared with bovine osteocalcin atconcentrations of 0.39, 0.78, 1.56, 3.12, 6.25, 12.5, 25.0,50.0, and 100.0 g/L. This assay required EDTA, 25mmolIL, in the assay buffer to function properly; theRIA with antiserum R102, on the other hand, requiredtrace amounts (2 mmol/L) of calcium ion in its assaybuffer.

Specimens

Specimens from apparently healthy women, womenwith osteoporosis, and patients with chronic renal fail-ure were aliquoted and stored at -70 #{176}C.Differentaliquots of the specimens were assayed by the RJA withpolyclonal antibody R102 from rabbit and by CIA.Aliquots of serum from a healthy subject who hadparticipated in previous studies (11) of 24-h infusion ofPTH (1-34) were stored at -70 #{176}Cand used in thegel-filtration study. Specimens from patients with pri-mary hyperparathyroidism or hypercalcemia associatedwith malignancy were residual samples that had beencollected for routine determination of PTH and PTH-related peptide. Specimens from healthy adults wereobtained from volunteers who consented to participatein a Mayo Clinic normal value study. Specimens fromchildren were from the University of Louisville TwinsStudy program.

Other Measurements

The RIA of osteocalcin involved use of polyclonalantiserum R102, rabbit anti-bovine osteocalcin (9). In-tact PTH was also determined by an immunochemilu-minometric assay (13, 14), C-terminal PTH by an(15), and PTH-related peptide by an extraction N-ter-minal RIA (16). Bone mineral density was determined

-j

IS..

zz0-J

0‘UI-.U,0

50 60 70 80FRACTION,1HL

Ostsocalcln, 1ig/L

Added Observed % recovery Clinical StudiesOsteocakin values in normal adults and children. We

109 also measured osteocalcin in normal men and women100 from the third through eighth decades of age by CIA

_______(Table 5) and in children from ages 23 to 108 months

CLINICAL CHEMISTRY, Vol. 39, No. 7, 1993 1371

at the mid-radius and distal radius with a 125J absorp-tiometric technique (17).

Gel-Filtration Column Separation

Aliquots (1 mL) of serum from healthy individuals

before and after 24-h infusion of PTH, or from patientswith surgically proven primary hyperparathyroidism,were applied to a 115 x 1 cm Sephacryl S-300 (Pharma-cia, Uppsala, Sweden) gel-filtration column and eluted

with 0.1 mol/L ammomum bicarbonate, pH 8.0. Frac-tions (1.0 mL) were collected and freeze-dried and thenreconstituted with 0.25 mL of assay buffer (11). Theosteocalcin in each fraction was measured by CIA. Thecolumn was calibrated with three markers. Blue dex-tran (2 x 106 Da) marked the void volume, 125I-labeledosteocalcun marked the peak of intact osteocalcin, andpotassium ferricyanide marked the salt peak. Peakareas of osteocalcin before and after PTH infusion weremeasured with a digital planimeter (Tamaya, Tokyo,Japan).

Results

Characteristics of CIA

Analytical recovery of added osteocalcun with the newCIA ranged from 94% to 109% (Table 1); linearity ofdilutions was good (Table 2). Interassay variations oflow- (4.9 ± 0.74 gfL; n = 34 assays), medium- (8.0 ±

0.96 ig/L; n = 36), and high- (21 ± 2.1 g/L; n = 35)concentration control pools were 15%, 12%, and 10%,respectively. Intraassay variation was 6% (8.0 ± 0.48gfL; n = 40 determinations). All these quality-controlstudies were done with a single preparation of acridin-ium-ester-labeled osteocalcin within 6 months. The per-centage of binding of the labeled osteocalcun was be-tween 45% and 50% of the total counts; after 6 months,the percentage of binding decreased to 35%. We esti-

mated the stability of acridinium-ester-labeled osteocal-cm to be about 6 months, about the same length of“shell-life” as other acridinium ester labels (13).

The new CIA recognized intact osteocalcin and the1ow-M fragment of osteocalcin after separation of se-rum by gel filtration.. It also detected the decrease ofintact osteocalcin and the osteocalcin fragment after a24-h infusion of PTH. The relative peak area after PTHinfusion decreased from 640 to 342 arbitrary area units,a decrease of 47% (Figure 1).

Table 1. AnalytIcal Recovery of Added BovineOsteocalcin by CIA

0 4 ...

7.5 12.5 11.515 19 19

30 32 34

Table 2. Unearlty of DIlutlon In CIA

Dilutionfactor

Osteocalcin, iaJL

Observed Expected

Nondiluted 19 ...

1:2 8.6 9.51:4 4.8 4.81:8 2.8 2.41:16 1.0 1.2

Fig. 1. Decrease of intact and fragment osteocalcin after 24-hinfusion of PTHSerum samples from a normal lndMdual before (#{149})and after (0)24-h Infusionof PTh were applied separately onto a Sephacr1 S-300 gel-filtration column(115 x 1 cm); 1-mLfractionswere eluted and collected. Each fraction wasfreeze-dried, reconstituted, and assayed for osteocalcin by CIA. The columnwas markedby bluedextran(2 x 10 Da), ‘#{176}I-labeledosteocalcin, and salt(potassium ferdcyanide).Relativeareasbeforeandafter 24-h Infusionof PTHwere measured with a Tamaya digital planimeter. Theywere 640 and 342,respectively, a decrease of 47%

Comparison with AlA

Osteocalcin in 86 patients was measured with CIAand compared with the results with RIA involvingantiserum R102 (r = 0.955). The comparison was madein three groups: normal women, women with osteoporo-sis, and patients with chronic renal failure (Table 3).Osteocalcin values by CIA were 1.3 times greater thanthose by RIA in normal and osteoporotic women. Inpatients with chronic renal failure, the values with thetwo assays were close (Table 3). A steady increase ofosteocalcin in normal women was found by both CIAand RIA for the sixth, seventh, and eighth decades ofage. The CIA values for these age groups were 1.2, 1.3,and 1.5 times more than the RIA values, respectively(Table 4).

Table 3. Comparison of CIA and RIA wIth R102Antiserum

Ost.ocalcln, 1&g/L

Table 6. Osteocalcin Values In Children, by AgeOst.ocalcln, pg/L

0.79 0.0001

Normalwomen(n = 30)

Mean ± SDRange

Osteoporoticwomen

(n = 36)Mean ± SDRange

Chronicrenal failure(n = 20)

Mean ± SDRange

All (n = 86)Mean ± SDRange

Age, months No. Mean ± SD Rang.P 23-36 27 17±11 6-40

37-48 49 21±10 3-3749-60 25 20 ± 9 7-3361-72 12 26±10 10-4288-108 10 25±11 12-43

CIA

9.9 ± 3.011.6 ± 4.215.3 ± 4.712.2 ± 4.5

CIA RIA

12.2 ± 4.5 9.3 ± 1.96-22 6.5-13.6

12.5 ± 6.2 9.9 ± 2.2 0.82 o.000i the correlations were less significant (r = -0.27 and5.6-36 6.5-17 -0.308, respectively, both P >0.05) (Figure 2).

Primary hyperparathyroidism. Patients with surgi-cally proven primary hyperparathyroidism had variable

36.2±26 34.5 ± 26.7 0.96 0.0001 (normal to high) osteocalcin values (mean ± SD, 21 ±247-100 5.5-104 pgfL; n = 40). Their osteocalcin values showed no

correlation (r = 0.05) with the concentrations of intact18.1 ± 16.6 15.4 ± 16.6 0.96 0.0001 PTH but showed good correlation with the concentra-

5.6-100 5.5-104______________________________ tions of C-terminal PTH (r = 0.83, P = 0.0001) (Table 7).After gel-column separation, we found an increasedpeak for intact osteocalcin in the plasma of patientsTable 4. ComparIson of CIA Values with RIA Values in

Normal Women, According to Age with primary hyperparathyroidism. The peak for thesmall fragment of osteocalcin was hardly noticeable,

Ostsocalcln, ,/L (mean ± SD) perhaps because of the increased concentration of intact

Ags, years No. RIA r P osteocalcin. An example is shown in Figure 3.50-59 10 8.2 ± 1.8 0.85 0.02 Hypercalcemia associated with malignancy. Six pa-60-69 10 9.3 ± 1.5 0.65 0.04 tients with cancer and hypercalcemia were studied: one70-79 10 10.5 ± 1.9 0.71 0.02 had breast cancer, two had prostate cancer, two had islet50-79 30 9.3 ± 1.9 0.78 0.0001 cell or pancreas cancer, and one had parathyroid carci-

noma. Osteocalcin, PTH, and PTH-related peptide weremeasured. High osteocalcin concentrations (mean ± SD,

Table 5. Osteocaicin Values in Normal Adults, by Age 28 ,agfL ± 27; n = 6) were found in one of the patientsOsteocalcin, iQJL with prostate cancer, one of the patients with islet cell

As., yw. No. Mean ± R. cancer, and the patient with parathyroid carcinoma. AllWomen except the patient with parathyroid carcinoma had

20-29 13 8.6 ± 1.5 5.8-li suppressed to subnormal concentrations (<1.0 pmol/L)30-39 25 7.8 ± 2.1 2.5-1 2.5 of intact PTH, as measured by the immunochemilumi-40-49 19 9.2 ± 2.7 4.4-15 nometric assay. The patient with parathyroid carci-50-59 13 g.g ± 2.7 6.1-17 noma had a highly increased concentration of PTH (2560-69 10 11.6 ± 4.2 6.3-19 pmol/L; normal, <5 pmol/L) and a normal concentration70-79 10 15.3 ± 4.7 6.0-22 of PTH-related peptide (2.2 pmol/L; normal, 2-5 pmollMen L). Of the five remaining patients, four had above-20-29 15 10.6 ± 2.3 7.2-15 normal and one had normal concentrations of PTH-30-39 20 9.3 ± 2.5 5.0-15 related peptide (mean ± SD, 10.7 ± 10.1 pmol/L). All40-49 23 8.7 ± 2.8 2.5-15 the patients had above-normal values for total serum50-59 6 6.8 ± 3.2 2.0-10 calcium (mean ± SD, 136 ± 16 mgfL; normal, 89-10160-69 2 5.9 ± 1.0 5.2-6.6 mgfL).

70-79 1 10Discussion

1372 CUNICAL CHEMISTRY, Vol. 39, No. 7, 1993

(Table 6). The osteocalcin concentrations in childrenshowed a steady increase with age, and were higher (P= 0.0001) than those in normal adults.

Comparison with bone mineral density. The bonemineral densities of 30 normal women and 36 womenwith osteoporosis were measured. The densities of bothgroups showed significantly inverse correlations withosteocalcin values measured by CIA (r = -0.41 and-0.360, respectively, both P <0.05); by RIA, however,

Using a nonradioactive acridinium ester tracer in-creased the shelf-life of tracer without losing assaysensitivity and also, as an environmental concern, re-duced the amount of radioactive waste generated from aclinical laboratory. To prevent light blocking by thesecond-antibody precipitation pellet and to improve thereproducibility of measurement, we modified the sec-ond-antibody precipitation method by using immobi-lized second antibody on plastic beads to separate freeand bound fractions. Washing the bead-immobilized

1.5

1.4

1.3

1.2

1.1a01

.9

.8

.7

I .4

1.3

1.2

& .1

01

.9

.7

0 RIA0

00

0

0 00

0

0

o RIA0

o 0 0

o 0

o0

o0

7 8 9 10 11 12 13 14OSTEOCALCIN, NGIML

O CIA0

O 0

000 g

0O 0

0

6 8 10 12 14

OSTEOCALCIN, NC/MI

1.

16 18

a

0

4 6 0 10 12 14 16 18 20 22 24 10 15 20 25 30 35 40

OSTEOCALCIN, NO/MI OSTEOCALCIN, NG/ML

Fig. 2. Bone mineral density (BMD) was Inverselycorrelad with serum osteocalcin concentration In 30 normal women (left panels) and 36women with osteoporosis (right panels)The Inversecorrelationsof BMDwith AlA (R102)and CIA were as follows (r, P): (left,t: -0.27,0.16; left,bottom:-0.41,0.02; rIght, t: -0.31,0.07; rIght,bottom: -0.36, 0.03

60 70 80FRACTION. 1ML

Fig. 3. Circulating osteocalcin in patientswith primary hyperparathy-roldismSerum, 1.0 mL, from a patient with surgically proven primary hyperparathy-roidismwas separatedby SephacrI S-300 column (115 xl cm). Osteocalcinin each fraction was measuredby CIA,as describedin FIg. 1. Notethelackofthe noticeablepeak for fragment osteocalcin around fractIon 80, as shown Inthe normal lndMdual In Fig. 1, pethapsovershadowedbythehighconcentra-tion of intact osteocalcin

NORMAL FEMALE OSTEOPOROTIC FEMALE

CUNICAL CHEMISTRY, Vol. 39, No. 7, 1993 1373

Table 7. Osteocalcin, Intact PTh, and C-Terminal PTHValues In Patients wIth Primary Hyperparathyroldism1

Mean SD p

Osteocalcin, ig/L 21 24Intact PTh, pmol/L 8 4.8 0.06 0.812C-terminalPTH,mmol/L 180 318 0.83 0.0001

a Of the 40 patIents with primaiyhyperparathyroidism,19 had Intact PTh

measured with the CIA and 21 had C-terminal Pill measuredwith an RL&

second antibody was easier than washing the pellets ofthe second-antibody precipitation and obviated thequenching of the chemilununescence by pellets.

Analytically and clinically, the performance of thenewly developed CIA was good. The new assay recog-nized both intact osteocalcin and the small fragment ofosteocalcin in normal individuals. It also detected thedecrease in osteocalcin after a 24-h infusion of PTH.This finding indicated that the new assay could detectthe dynamic change of circulating osteocalcin. The newassay compared well (r = 0.955) with a previouslyestablished RIA that used antiserum R102. The CIAresult is about 1.3 times the result obtained with RIA innormal women and in women with osteoporosis. How-ever, in patients with chronic renal failure (secondaryhyperparathyroidism), the results obtained with CIAand RIA were similar. In a patient with hyperparathy-roidism, we found an increase of intact osteocalcin butcould hardly detect the small fragment of osteocalcin,which might have been obscured by the highly in-

1.:

1.1

.9

a

0

.7

.6

creased amount of intact osteocalcin during gel-columnseparation (Figure 3).

Another laboratory demonstrated that a monoclonalRIA recognized both intact osteocalcin and the smallfragment of osteocalcin. The results of monoclonal RIA

1374 CUNICAL CHEMISTRY, Vol. 39, No. 7, 1993

were 1.5 times greater than the results of the RIA withpolyclonal R102, because the latter RIA could not detectthe small fragment of osteocalcin (11). Our currentfinding that CIA results were 1.3 times greater thanRIA results obtained with the same antiserum R102might be for the same reason; i.e., our results werehigher than RIA because our assay also recognized thesmall fragment.

We could not explain the good correlation of circulat-ing osteocalcin values with C-terminal PTH values (r =

0.83, P = 0.0001) combined with a poor correlation withthe biologically active intact PTH (Table 7). We couldonly speculate that both osteocalcin and C-terminalPTH required kidneys for excretion and that intact PTHwas degraded very quickly in the circulation. Values ofC-terminal PTH were extremely skewed; this findingmight suggest variable kidney functions in these indi-viduals. We also studied a group of children, ages 23 to108 months, whose osteocalcin values were higher thanthose in adults. These values could be used as a refer-ence for future studies of growth disorders in children.Unfortunately, we lacked volunteers at puberty age.

During the preparation of this manuscript, assay ofosteocalcin became more important, as shown by therecent publication in Clinical Chemistry of two articleson osteocalcin measurement. One was an avidin-biotinenzyme immunoassay (18) and the other was a radio-immunoassay (19), but neither one used the new CIAtechnology.

In conclusion, we achieved two goals: (a) acridiniumester can be used to prepare CIAs, and the performanceof the new assay compares very well with the estab-lished RIA that uses antiserum R102; (b) the new CIAappears to be as useful as RIA (R102) for determiningbone formation and for monitoring bone turnover rateand thus may reduce use of the invasive bone histomor-phometry test.

References1. Delmas PD. Biochemical markers of bone turnover in os-teoporosia. In: Riggs BL, Melton LJ Ill, eds. Osteoporosis: etiology,diagnoeia, and management. New Yoric Raven Press, 1988:297-316.2. Price PA, Poser JW, Raman N. Primary structure of they-carboxyglutamic acid-containing protein from bovine bone. ProcNail Aced Sd USA 1976;73:3374-5.3. Poser JW, Each FS, Ling NG, Price PA. Isolation and sequence

of the vitamin K-dependent protein from human bone: undercar-boxylation of the first glutamic acid residue. J Biol Chem 1980;255:8685-91.4. Nishimoto SK, Price PA. Proof that the carboxyglutamicacid-containing bone protein is synthesized in calf bone: compara-tive synthesis rate and effect of coumadin on synthesis. J BiolChem 1979;254:437-41.5. Charon SA, Delmas PD, Malaya! L, Chavassieux PM, Arlot M,Chapuy M-C, Meunier PJ. Serum bone Gla-protein in renal osteo-dystrophy: comparison with bone histomorphometry. J Clin Endo-crunol Metab 1986;63:892-7.6. Brown JP, Malaval L, Chapuy MC, Dehnas PD, Edouard C,Meunier PJ. Serum bone Gla-protein: a specific marker for boneformation in poatmenopausal osteoporosis. Lancet 1984;i:1091-3.7. Stewart PJ, Stern PH. Vertebral bone resorption in vitro: effectsof parathyroid hormone, calcitonin, l,25-dihydroxyvitamin D3,epidermal growth factor, prostaglandun E2, and estrogen. CalcifTissue mt 1987;40:21-6.8. Riggs BL, Tsai K-S, Mann KG. Effect of acute increases in bonematrix degradation on circulating levels of bone-Gin protein. JBone Miner Rea 1986;1:539-42.9. Duda RJ Jr, O’Brien JF, Katzmann JA, Peterson JM, MannKG, Riggs BL. Concurrent assays of circulating bone Gin-proteinand bone Rlknhine phosphatase: effects of sex, age, and metabolicbone disease. J Cliii Endocrinol Metab 1988;66:951-7.10. Slovik DM, Gundberg CM, Neer RM, Liar JB. Clinical eval-uation of bone turnover by serum osteocalcin measurements in ahospital setting. J Cliii EndocrunolMetab 1984;59:228-30.11. Tracy RP, Andrianorivo A, Riggs BL, Mann KG. Comparisonof monoclonal and polyclonal antibody-based immunoassays forosteocalcin: a study of sourcesof variation in assay results. J BoneMiner Res 1990;5:451-61.12. Lowry OH, Rosebrough NJ, Parr AL, Randall RJ. Proteinmeasurement with the Folin phenol reagent. J Biol Chem 1951;193:265-75.13. Klee GG, Preisaner CM, Schryver PG, Taylor EL, Kao PC.Multisite immunochemiluminometric assay for simultaneouslymeasuring whole-molecule and amino-terminal fragments of hu-man parathyrin. Cliii Chem 1992;38:628-35.14. Kao PC, van Heerden JA, Grant CS, Klee GG, Khosla S.Clinical performance of parathyroid hormone inimunometric as-says. Mayo Cliii Proc 1992;67:637-45.15. Kao PC, Jiang N-S, Klee GG, Purnell DC. Development andvalidation of a new radioimmunoassay for parathyrin (PTH). ClinChem 1982;28:69-74.16. Kao PC, Klee GG, Taylor EL, Heath H Ill. Parathyroidhormone-related peptide in plasma of patients with hypercalcemiaand malignant lesions. Mayo Cliii Proc 1990;65:1399-407.17. Riggs BL, Wahner HW, Seeman E, et al. Changes in bonemineral density of the proximal femur and spine and agungdifferences between the poetmenopausal and senile osteoporosissyndromes. J Clin Invest 1982;70:716-23.18. Jaouhari J, Schiele F, Dragacci S, Tarallo P, Siest J-P, HennyJ, Siest G. Avidin-biotin enzyme immunoassay of osteocalcin inserum or plasma. Cliii Chem 1992;38:1968-74.19. Bouillon R, Vanderschueren D, Van Herck E, Nielsen IlK,Bex M, Heyns W, Van Baelen H. Homologous radioiinmunoassayof human osteocalcin. Clin Chem 1992;38:2055-60.