CLIN. CHEM. 37/2, 277-281 (1991)
Partial Purification of Endogenous Digitalis-like Compound(s) in Cord Blood
SllvanaBalzan,’SergloGhlone,’PascalBlver,2PsoloGazzettl,’ and Umberto MontaII3
Increasing evidence indicates the presence of endogenous digitalis-like compound(s) in human body fluids. In
this preliminary report, we describe a study of the partial
purification by HPLC of these compounds in the plasma of
neonates (who have particularly high concentrations of
this substance) and adults. Plasma samples from neonates (cord blood) and adults, lyophilized and extracted
with methanol, were applied on a 300 x 3.9 mm C18Nova
Pak column and eluted with a mobile phase of acetonitrile/
methanol/water (17/17/66 or 14/14/72 by vol) and, after
30 mm, with 100% methanol. We assayed eluted fractions
for inhibitory activity of soRb uptake and for digoxin-like
immunoreactivity.
The elution profile revealed a first peak
of inhibitoryactivity of soRbuptake at the beginning of the
chromatography; another peak was eluted with the 100%
methanol. The two peaks also cross-reacted with antidigoxin antibodies. Because the second peak could possibly reflect the nonspecific interference of various lipophilic compounds, we focused our attention on the first
peak. For these fractions dose-response curves for 86Rb
uptake and for displacement of digoxmnwere parallel,
respectively, to those of ouabain and digoxmn,suggesting
similarities of digoxmn-like immunoreactive substance to
cardiac glycosides. Similar chromatographic profiles were
also obtained for plasma from adults, suggesting that the
endogenous glycoside-like compound(s) in the neonate
may be the same as those in the adult.
Additional Keyphrases: seRb uptake
e,ythrocytes
digoxinlike immunoreactivity
chromatography, liquid
pediatric
chemistry
cardiacglycosides
radloimmunoassay
.
.
.
.
.
.
Several investigators
have reported the presence in
human body fluids of endogenous circulating glycoside-
like substance(s)that may inhibit the activity of the cell
sodium/potassium
pump (1-14). Increased
circulating concentrations
of these inhibitor(s) have
been found in pregnancy (3), in neonates (4, 5), in
arterial
hypertension
(6, 7), and in hepatic (8) and
cardiac insufficiency
(9), which suggests that the compound(s) could have a role in these states. These substances may also interact with several anti-digoxin
antisera (1-5). However, neither the structure nor the
site of production of these substances is clear. In addition, whether these inhibitors represent various compounds in plasma and (or) to what extent they are the
results of the nonspecificity of the assaysused is not yet
clear. The different methods used for extracting and
membrane
‘C.N.R. Institute
Italy.
2Department
of Clinical Physiology, via Savi n. 8, Pisa,
of Neonatology, and 3lnstitute of Biological
Chemistry, University of Pisa, Pisa, Italy.
Received February 16, 1990; accepted November21, 1990.
purifying
these substances could partly explain the
controversial results in the literature.
We describe a method for partial purification of the
substances that cross-react with anti-digoxin antisera
and inhibit soRbuptake in plasma from neonates (cord
blood) and adults.
Materials and Methods
Preparation
of samples. Heparinized blood samples,
obtained from neonates (cord blood) and from normal
adults, were placed on ice, centrifuged within 1 h, and
stored at -20 #{176}C.
None of the subjects studied had
cardiac, hepatic, or renal diseases, and all were normotensive.
Plasma was subsequently treated and assayed within
one week according to the following procedure: We
lyophilized 10-40 mL of plasma, suspended this in 20
mL of methanol, and centrifuged the sample at 15 000 x
g for 5 mm at 4#{176}C.
After removing and reserving this
supernate, we resuspended the precipitate in 5 mL of
methanol and centrifuged it again at 15000 X g for 5
ruin. This procedure was repeated three times, and the
supernates
were pooled and evaporated under reduced pressure; The extract obtained was then redissolved in 10 mL of doubly distilled water and purified on
Sep Pak C18 cartridges (Millipore, Bedford, MA) as
previously described (15). We dried the eluate under
reduced pressure, reconstituted it in 2-4 mL of the
HPLC mobile phase of acetothtril&methanol/water (17/
17/66 or 14/14/72 by vol; see below), centrifuged for 10
mm at 10000 x g, and filtered the supernate through
0.5-zm (pore size) ifiters (Waters, Milford, MA).
HPLC
chromatography.
We injected 100-p.L samples,
obtained as described above, onto a 300 x 3.9 mm C18
Nova Pak column (Waters) coupled to an HPLC instrument fitted with a variable-wavelength detector set at
220 nm (UvIDEc 100-V; Jasco International
Co., Tokyo,
four
Japan).
Chromatographic
runs were performed under
isocratic conditions, with a mobile phase of acetothtril&
methanol/water (17/17/66 by vol) at a flow rate of 1
mL/min. For each run, we collected 16 fractions, then
changed the mobile phase to 100% methanol, and collected five subsequent fractions. In a further set of
experiments, the mobile phase was changed slightly
(14/14/72 by vol) for better resolution of the early fractions. In some cases, to increase the material for the
assays, we pooled the fractions from two or more chro-
matographic runs that had the same absorbance profile
and the same retention time. The fractions were dried
under reduced pressure and stored at -20 #{176}C
for the
assays.
Eiythrocyte
uptake of 86Rb. We used the method of
CLINICALCHEMISTRY, Vol. 37, No. 2, 1991 277
et al. (16) with some modifications (17). We
obtained erythrocytes
immediately before the experiments from fresh blood samples from healthy donorsnot
receiving any drugs. The blood samples were transferred to an ice-water bath and, after separation from
plasma by centrifugation at 4#{176}C
(3000 X g), were
washed three times by centrifligation
with MgCl2 (110
mmol/L) at 4#{176}C,
resuspended in Ringer solution (containing, per liter, 9 g of NaC1, 0.24 g CaCl2, and 0.2 g of
NaHCO3, pH 7.4) to give a hematocrit of 50%, and added
in 25-1.tL aliquots to the HPLC-eluted fractions reconstituted in 100 L of Ringer solution. All cell suspensions were pre-incubated for 4 h at 37#{176}C
in a shaking
water bath. At the end of the pre-incubation, we added
20 1zL of nsRb solution, containing 1-2 pCi of 86RbCl
(Amersham International, Bucks., U.K.; 1 Ci/L, 1 CiJg),
to each tube. The incubation was continued for 1 h and
stopped by washing the cells at 4#{176}C
(except for the vials
to be used for determining total radioactivity) three
times with 2 mL of isotonic saline (NaC1, 154 mmol/L).
The radioactive rubidium taken up by the erythrocytes
was determined with a gamma counter, and inhibition
of the seRb uptake in erythrocytes was calculated as a
percentage, assuming as 100% the pump activity in the
presence of the Ringer solution devoid of column-fraction samples.
Aronson
Dose-response curves were set up for ouabain (con-
centrations ranging from 0.1 nmol/L to 1 pmol/L) and
for the early HPLC peak (mobile phase: acetonitrile/
methanol/water, 14/14/72 by vol) obtained by drying
and diluting the fractions pooled from several elutions
of extracts from cord blood and corresponding to a range
of 0.16-5.7 mL of the original pre-extraction plasma.
RIA. HPLC fractions from cord blood and adult
plasma, reconstituted with 400 uL of phosphate buffer
(pH 7.4, 0.15 mol/L), were assayed for endogenous
digoxin-like
immunoreactive
substance (DLIS) by RIA
as previously
from 7.8 to 1000 pg/O.4 mL and the volumes of original
neonate plasma from 1.6 to 33 mL.
Results
Figure 1 shows the HPLC elution profile of a cordblood extract (corresponding to 1 mL of original cord
blood)chromatographed with the acetomtrile/methanoll
water (17/17/66 by vol) mobile phase, and then 100%
methanol. The seRb uptake revealed two significant
peaks of inhibitory activity: the first in fraction 2 (55%
inhibition) and the second in fraction 19 (64% inhibition), when the mobile phase was changed to 100%
methanol. Fractions 2 and 19 also cross-reacted with
anti-digoxin
antibodies, and another significant immunoreactive peak was present in fraction 8.
When the corresponding individual fractions, obtained by HPLC elutions of extracts from cord blood,
were pooled to increase the concentration of the active
substancesup to fourfold, only fraction 2 was associated
with increasing inhibition of 86Rbuptake. On the other
hand, when the active fraction eluted with 100% methanol was used, we observed hemolysis when the concentration was increased by >1.5 times. Using the other
fractions, we observed no clear trend but rather a high
variability of inhibitions, between -30% and +30%.
To obtain a resolution of the early fractions, we
chromatographed an extracted sample (correspondingto
1 mL of original cord blood), using a mobile phase of
acetonitrile/methanol/water
(14/14/72 by vol), and collected 15 fractions of 1 mL (Figure 2). Both profiles of
immunoreactivity and inhibitory activity of soRb uptake revealed the presence of a single peak, eluted with
a retention time of 4.7 mm, that showed a dose-dependent inhibition of 86Rb uptake and of ‘251-labeled
digoxin displacement from antisera. The dose curves
described (18). For our radioimmunoas-
say, we used a solid-phase system (antibody-coated test
tubes) for separation of the free/bound moiety and i251
labeled digoxin as tracer (“Spak” Digoxin RIA kit; BYK,
Gulden, the Netherlands). The antibody was prepared
by injecting a conjugate of digoxin and bovine serum
albumin into goats. This antiserum cross-reacts with
digitoxin by 35% but negligibly (<0.01%) with testosterone, progesterone, cortisone, and aldosterone (19). For
preparing the standard curve, we dissolvedpure digoxin
(Wellcome Laboratories, Pomezia, Italy) in pyridine,
then diluted this with phosphate buffer (pH 7.4, 0.15
molIL) containing 40 g of human serum albumin and 20
mmol of sodium azide per liter. We incubated 0.4 mL of
standard solution or reconstituted fractions with 0.5 mL
of tracer solution in the test tubes for 12 h at 4 #{176}C.
After
incubation, we aspirated the supernate and counted for
1-2 mm the radioactivity remaining in each tube, using
a well-type gamma counter. The mean sensitivity of our
RIA was 5.16 (SD 1.1) pg/0.4 mL.
Dose-response curves were set up similarly to that for
ouabain (see above). Digoxin concentrations ranged
278 CLINICALCHEMISTRY, Vol. 37, No. 2, 1991
I
I
p
I
4
4
t_.
I_
a
4
-I
4
4
yI_
4_I.)
Fig. 1. Elution profile by reversed-phase HPLC of endogenous
digitalis-like compounds
(DLIS) from cord-blood plasmaextract
The extract (1 mL of original
plasma) was fractionated under isocratl#{231}
conditions, with a mobile phase of acetonitflle/methanol/water(17/17/66by
vol)at a flow rateof 1 mL/min over 30 mm. The mobile phase was then
changedto 100% methanol (a,mn. Absorbance was monitored at 220 nm
(-). Each of the 21 fractionscollected was dried underreducedpressureand
assayedfor theinhibitoryactivityof #{176}Rb
uptake(l2) and forDLIS(). The
shaded areas (top) represent the fractions with Rb uptake or DLIS concentration differentfrom the values found in the presence of onlyRingersolution
120
100
C
60
0
0
60
x
0
#{149}0)
C
40
20
0
1
10
100
1000
Digox in (pg/0 .4 mL)
1
1)
100
1000
HPLC early peak collection
4
8
Ti,.
8
O
)2
(II..)
6
C
a
Tim.
were dried under reduced pressure and assayed for the inhibitory
activity
of
Rb uptake (t3) andforDLIS()
were parallel to those of ouabain for 66Rb uptake (Figure 3) and of digoxin for radioimmunoassay (Figure 4).
Chromatographies with adult plasma extracts (corresponding to 1.5 mL of original plasma), performed as
120
100
80
V
60
0.
.0
0
a)
(jiL) (.)
Fig. 4. Dose-response curves of 125l-Iabeled digoxin displacement
from antisera for digoxin (0) and for the early peak of neonate
plasma (#{149})
Materialfrom the early peak was dilutedin Ringer solution, and aliquots
ranging from 20 to 400 p1 (from1.6 to 33 mL of original plasma) were assayed
describedabove, gave inhibitory profiles similar to those
for cord blood (data not shown). The inhibitory values of
the two peaks were respectively 40% and 55%.
(m,n4
Fig. 2. Elution profile by reversed-phaseHPLC of endogenous
digitalis-likecompounds(DLIS) from cord-bloodplasmaextract
Absorbancewas monitored at 220 nm (-). Extract (1 mL of original plasma)
wasfractionatedInto 15 fractionsover15 mm under isocraticcondItions,with
a mobile phase ofacetonitrile/methanol/water
(14/14/72by vol).Thefractions
a)
(a)
40
20
DIscussIon
Several laboratories
are currently
engaged in the
study of endogenous cardiac glycoside-like compounds
that may inhibit the activity of the cell membrane
sodium/potassium
pump. Although different groups of
molecules-e.g., steroid-like compound(s), some peptides, fatty acids, and phospholipids-have been claimed
to have digitalis-like activity, no unique substance has,
at present, been completely purified, chemically characterized, and identffied (10-14) except for a steroidal
dienolide in toad skin (20). Here we describe our preliminary data on the purification of endogenous digitalislike factor(s) in neonatal plasma, which represents an
enriched sourceof these compounds.In fact, particularly
high values of a factor with potential digitalis-like
activity have been reported by us and others in samples
from neonates (4, 15, 17). Because of the high and
continuous ingestion of sodium from the amniotic fluid,
an inhibitor of the sodium pump may be of physiological
importance in newborns to maintain high sodium excretion by reducing tubular sodium reabsorption (4,22).
Several studies on HPLC separation of endogenous
digitalis-like factors in the plasma and urine of adults
partially purified inhibitors of the sodium pump as a series of unsaturated
fatty acids and
lysophospholipids (2, 11,23) or as a polar digitalis-like
factor, resembling ouabain in polarity, nonpeptidic nature, and molecular mass (24-27). On the other hand,
few studies of HPLC separation of endogenousdigitalislike factors in newborns have been done (5,28,29).
In our studies we observed the presence of (at least)
two digitalis-like activity fractions, corresponding to
one or more polar (peak I) and nonpolar (peak II)
have
0
;o-1o io
io_8
Ouabain (mol/L)
1
10
i’0_6
(0)
100
(j.iL) (.)
Fig. 3. Dose-response
curves for effect on ssRb uptakeof ouabain
(0) and oftheearlypeakofactivity obtained fromHPLC separations
of cord-blood plasmaextracts(#{149})
Materialfrom the early peak was dilutedin Ringersolution,and ahquots
ranging from 2 to70 L (from0.16 to 5.7 mL oforiginalplasma) wereassayed
HPLC early peak collection
identified
CLINICALCHEMISTRY, Vol. 37, No. 2, 1991 279
substances.
Gault et al. (28), in agreement with our results, also
reported the presence of a polar digoxin-like immunoreactive peak (appearing in the first two fractions of
HPLC procedure) in cord blood. Seccombe et al. (29)
recently reported the presence of several immunoreactive peaks obtained by preparative HPLC fractionation
of cord blood but, in contrast to our results, were unable
to demonstrate significant inhibition of seRb uptake.
Whether the activity present in the first peak reflects
one or more substances is uncertain. The relatively
broad peak obtained by the second HPLC step could
suggest that more than one substance is present in this
peak. On the other hand, for the first peak the dose-
response curve, both for 86Rb uptake and for digoxin
displacement, paralleled respectively that of ouabain
and digoxin. This finding, albeit not definite proof,
suggests the presence of a compound that shares similarities to cardiac glycosides.
It is unclear whether the second peak represents also
a truly interacting compound (e.g., a nonpolar steroid)
or rather reflects the sum of minor nonspecific effects of
several lipophilic compounds eluted from the column
with 100% organic phase, which could exert an inhibitory action on seRb uptake through a detergent effect on
the cells and interfere with the RIA assay by sequestration of the label. Further studies are needed to clarify
this point.
Our method, in view of the relatively small volume of
plasma required to detect inhibition peaks, could be
used for the partial purification of endogenous glycoside-like compound(s) in several human fluids. In fact,
the extraction procedure we used, compared with the
dialyzation techniques used by others (23-24,26), could
allow a higher recovery of endogenous glycoside-like
compound(s)by detaching the moiety bound to proteins,
because this substance reportedly is predominantly
(90%) bound to plasma proteins (30). However, the high
variability of our seRbmethod hinders detection of other
fractions with smaller but nonetheless real inhibition.
Whether
digitalis-like
factor(s) in the plasma of neo-
nates is identical to that in adults remains uncertain
and cannot be established until these compounds are
unequivocally
identified. However, our preliminary
finding of similar elution proffles supports this hypothesis.
References
1. Masugi F, Ogihara T, Hasegawa T, Kuinahara Y. Ouabain-like
and non-ouabain-like factors in plasma with essential hypertension. Clin Exp Hypertens [A] 1987A9:1233-42.
2. Kelly BA, O’Hara DS, Canessa ML, Mitch WE, Smith 1W.
Characterization
of digitalis like factors in human plasma. J Biol
Chem 1985;260:11396-405.
3. Graves SW, Valdes R, Brown BA, Knight AB, Craig HR.
Endogenous digoxrn immunoreactive substance in human pregnancies. J Clin Endocrinol Metab 1984;58:748-61.
4. Morris JF, McEachern MD, Poston L, Smith MJ, Hilton PJ.
Evidence for an inhibitor of leukocyte sodium transport
in the
serum of neonates. Clin Sci 1987;73:291-7.
5. Diainandis EP, Papanastasiou-Diamandi A, Soldin SJ. Digoxin
immunoreactivity in cord and maternal serum and placental
extracts: partial characterization
of immunoreactive
sub8tances by
280 CLINICAL CHEMISTRY, Vol. 37, No. 2, 1991
performance
liquid chromatography
and inhibition
of
NaK-ATPase.
Clin Biochem 1985;18:48-54.
6. Hamlyn JM, Ringel RB, Schaeffer J, et al. A circulating
inhibitor of(Na/K)
ATPase associated with essential hypertension. Nature (London) 1982;300:650-2.
7. Graves SW, Williams GH. An endogenous ouabain-like factor
associated with hypertensive
pregnant women. J Clin Endocrinol
high
Metab 1984;59:1070-4.
8. Nanji AA, Greenway DC. Falsely raised plasma digoxin concentrations in liver disease. Br Med J 1985;287:432-3.
9. Shilo L, Adawi A, Solomo G, Shenkman L. Endogenous digoxinlike immunoreactivity
in congestive heart failure. Br Med J
1987;295:415-6.
10. Ng LL, Hockaday
TDR. Non-esterifled
late human leukocyte sodium pump
activity.
acids may reguClin Sci 1986;7:737-
fatty
42.
11. Tamura M, Kuwano H, Kinoshita T, Inagami T. Identification
of linoleic and oleic acids as endogenous Na,K-ATPase
inhibitors from acute volume expanded dogs. J Biol Chem 1985;
260:9672-7.
12. Gruber A, WhitackerJM, Buckalew VM. Endogenous
digitalis-like substance in plasma of volume expanded dogs. Nature
(London) 1980;287:743-5.
13. La Bella FS, Bihier I, TempletonJ, Kim RS, Hnatowich M,
Rohrer D. Progesterone
derivatives
that bind to the digitalis
receptor: effects on Na,K
ATPase and isolated tissue. Fed Proc
1985;44:2806-11.
14. Hnatowich M, LaBella F. Endogenous digitalis like factors:
in vitro comparison of biological and immunological activities of
peptide and steroid candidates. Eur J Pharmacol 1985;106:56775.
15. Balzan S, Ghione 5, Clerico A, Montali U. Correlation between
endogenous digoxin like immunoreactivity
and 3H ouabain displacement on erythrocyte
membranes in extracts of human
plasma. Clin Biochem 1986;19:311-4.
16. Aronson JK, Grahame-Smith
DG, Hallis KF, Hibble A,
Wigley
F. Monitoring
digoxin therapy. Plasma concentrations
and in vitro assay of tissue response.
Br J Clin Pharmacol
1977;4:213-21.
17. Balzan S, Montali U, Genovesi-Ebert A, Biver P, Fantoni M,
Ghione S. Comparison between endogenous digoxin-like immunoreactivity and
uptake by erythrocytes
in extractsof human
plasma. Clin Sci 1989;77:375-81.
18. Balzan 5, Clerico A, Del Chicca MG, Montali U, Ghione S.
Digoxin-like immunoreactivity
in normal human plasma and
urine, as detectedby a solid-phase radioimmunoassay.
Cliii Chem
1984;30:450-1.
19. Clerico A, Ghione S, Balzan S, Del Chicca MG. Digoxin like
immunoreactive substance: nonspecific interference or a new hormone? J NucI Med Allied Sci 1984;28:297-301.
20. Lichstein D, Kachalasky 5, Deutch J. Identification
of a
ouabain-like
compound in toad skin and plasma as bufodienolide
derivate. Life Sci 1986;38:1261-70.
21. Goto A, Yamada K, Ishii M, et al. Purification and characterization of human urine-derived digitalis-like factor. Biochem Biophys Res Commun 1988;154:847-53.
22. Ebara H, Suzuki S, Nagashima K, Shimano S. Kuroume T.
Digoxin-like immunoreactive substances
in urine and serum of
preterm and term infants: relationship to renal excretion of sodium. Eur J Pharmacol 1986;106:567-75.
23. Hamlyn JM, SchendenJA, Zyren J, Baczynsky) L. Purification and characterization of digitalis like factors from human
plasma.Hypertension 1987;10:I-71-7.
24. Dasgupta A, Msalik S, Yeo KT, SanduP, Ahmad S. Kenny M.
Two novel endogenousdigoxin-like immunoreactive
substances
isolated
from human plasma ultraflltrate.
BiochemBiophys Res
Commun 1987;148:623-8.
25. Goto A, Yamada K, Ishii M, et a!. Existence
of a polar
digitalis-like
factor in mammalian
hypothalamus.
Biochem Biophys Res Commun 1989;161:953-8.
26. Hamlyn JM, Harris DW, Clark M, Rogowski AC, White RJ,
Ludens JH. Isolation and characterization
of a sodium pump
inhibitor from human plasma. Hypertension 1989;13:681-9.
27. Hamlyn JM, Harris DW, LudensJH. Digitalis-like activity in
human plasma. J Biol Chem 1989;264:7395-404.
28. Gault MH, LongerichL, Dawe M, Vasdev SC. Combined liquid
chromatography/radioimmunoassay
with improved specificity for
serum digoxun. Cliii Chem 1985;31:1272-7.
29. Seccombe DW, Pudek MB, Humphries KH, et al. A study into
the nature and organ source of digoxin-like
immunoreactive
sub-
in the perinatal period. Biol Neonate 1989;56:136-46.
30. Valdes R, Graves SW, Brown BA, Landt M. Endogenous
substance in newborn infants causing false positive digoxin measurements. J Pediatr 1983;102:947-50.
stance(s)
CLIN. CHEM. 37/2, 281-284 (1991)
Osteocalcin Concentrations in Plasma Prepared with Different Anticoagulants
MIchael J. Power,1 Bernadette
O’Dwyer’
EugeneBreen,2 and PatrickF. Fottreil”3
We investigated the effects on plasma osteocalcin concentrations of different anticoagulants used to collect the
blood samples. Plasma osteocalcin concentrations measured by enzyme immunoassay and radioimmunoassay
are influenced by the nature of the anticoagulants used.
The most significant difference between concentrations
found in plasma and serum was seen with oxalate/fluonde
anticoagulant, which reduced osteocalcin concentrations
to 37.3% of serum values. This is probably related to
increased hemolysis with this anticoagulant compared
with osteocalcin concentrations in plasma prepared with
other anticoagulants. Samples prepared with sodium cit-
rate (0.105 mol/L) or lithium hepann gave values 92.4%
and 83.6% of those obtained with matched serum samples. Osteocalcin concentrations were relatively stable in
plasma and serum at -20 #{176}C
for two freeze/thaw cycles.
In blood from 100 patients there was a good correlation
between osteocalcin concentrations in serum and plasma
(lithium heparin) (r2 = 0.831); the slope and intercept
(±SE) were 0.924 ± 0.04 and 4.92 ± 1.25 pg/L, respectively. However, in 10 patients, serum osteocalcin concentrations were two- to threefold higher than those in
matched plasma samples.
Additional Keyphrasee: calcium-binding protein
bone gla protein
metabolic bone disease
assay
enzyme immunoassay
.
osteoporosis
radioimmuno-
Osteocalcin is a bone-specific protein consisting of 49
amino acids, three of which are ‘y.carboxyglutamic acid
(1). Synthesized by osteoblasts, osteocalcin enters the
blood (2-4), where its concentration
in plasma reflects
the rate of bone formation rather than the rate of bone
resorption (5). Measurements
of plasma osteocalcin are
used in various clinical situations, including metabolic
bone diseases, hyperthyroidism,
liver cirrhosis, renal
disorders, and diseases related to excess glucocorticoids
(6-12). Among the published immunoassays of osteocalcm (13-17) is one from this laboratory, a relatively rapid
enzymoimmunoassay
(EIA) involving use of a monocloDepartments
of’ Biochemistry,
icine, University College Hospital,
University College, and2 MedGalway, Ireland.
3Addresacorrespondence to this author.
Received September 7, 1990; accepted December 12, 1990.
nal antibody (18).
Published ranges of plasma osteocalcin concentrations in controls’ and patients’ samples vary widely (19).
One reason for this is the use of various polyclonal
antisera, which differ in the extent to which they may
recognize some of the multiple forms of osteocalcin in
the circulation. In addition to intact osteocalcin, other
immunoreactive forms in plasma are a high-molecularmass form (100 kDa) (19, 20) and peptides of lower
molecular mass (1000-1500 Da) (21-23).
Here we show that anticoagulants used to prepare
plasma samples influence the immunoreactivity
of osteocalcin and may represent another reason for the wide
range of published osteocalcin values.
MaterIals and Methods
Subjects: Plasma and serum were prepared from the
same blood samples from two groups of patients (A and
B), selected without conscious bias in the outpatient
department of University College Hospital. Group A
consisted of 20 women, ages 30-60 years, who had no
clinical history or symptoms of bone-related disorders or
conditions known to influence bone metabolism; one
patient with anorexia nervosa was under investigation
for osteoporosis. Samples from these subjectswere used
for the studies described in Tables 1-3. Group B consisted of 100 subjects, both men and women: 40, ages
18-38 years, had no clinical history or symptoms of
bone-related disorders; 60, ages 50-90 years, included
patients with hyperparathyroidism, Paget’s disease, Osteoporosis,and renal disease. We used 36 samples from
these 120 subjects, Groups A and B, selected without
consciousbias, for the studies of the stability of osteocalcin.
Serum and plasma:
Immediately
after collection,
portions of each blood sample were put into tubes that
contained various anticoagulants or no anticoagulant
(for serum preparation). Samples were kept at 4#{176}C
until
serum and plasma were separated-usually
after -24
h, but some samples were separated
after
3 h. The
following anticoagulants were used in Vacutainer
Tubes (Becton-Dickinson, Meylan Cedex, France):
EDTA (75 gIL), lithium heparin (14 300 USP units/L),
potassium oxalate/sodium fluoride (10 and 12.5 mg,
respectively, per 5-mL tube, i.e., 2.0 and 2.5 g/L), and
CLINICALCHEMISTRY, Vol. 37, No. 2, 1991 281