| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
1
Department of Clinical Biochemistry and
2
Department of Medicine, Renal Research Laboratory, St. Bartholomew's Hospital Centre for Clinical Research, London ECIA 7BE, UK.
3
Netria, Department of Chemical Endocrinology, St.
Bartholomew's Hospital, London EC1A 7BE, UK.
a Author for correspondence. Fax 44-171-796-4676; e-mail A.B.Dawnay{at}mds.qmw.ac.uk
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Key Words: indexing terms: chromatography, reversed-phase • antibody specificity
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The physiological role of OLC is unknown. Ouabain has a strong affinity for the integral membrane sodium–potassium pump, Na+,K+-ATPase (15). Blaustein (16) suggested that the binding of ouabain to this pump leads to numerous physiological and pathophysiological events, including an increase of calcium release on cell activation. Ouabain at pathophysiological concentrations (1 nmol/L) substantially increased intracellular calcium concentrations in cultured rat vascular smooth muscle cells (17). OLC is thought to have a role in the pathogenesis of essential hypertension (18)(19), and concentrations of OLCs are above normal in subjects with essential hypertension (9)(20).
The notion of an endogenous OLC has met with skepticism (21)(22). OLC biosynthesis, release, and sites of production are little understood. Lewis et al. (23) indicated that the immunoreactive compound extracted from human plasma was not ouabain because it did not coelute with pure ouabain on HPLC and Doris et al. (24) found only small amounts of OLC in a minority of the healthy volunteers studied. However, Di Bartolo et al. (11) detected OLC in human plasma from newborns, which coelutes with authentic ouabain and, like ouabain, inhibits 86Rb uptake; moreover, the addition of anti-ouabain antibodies reverses this inhibitory effect. Investigations with ouabain and OLC by Zhao et al. (25), using HPLC and circular dichroism, suggest that OLC is very similar to but not identical with authentic ouabain. Differences in antibody specificity for OLC may account for some of these conflicting results.
Few assays have been described in the literature for the detection of endogenous ouabain. Most tend to be labor-intensive RIAs with tritiated ouabain (9)(24)(26)(27)(28)(29). The ELISAs described by Harris et al. (6) and Gomez-Sanchez et al. (30) have many procedural advantages over the RIA methods, but have not been described in detail. We describe here a novel enzyme immunoassay (EIA) for endogenous OLC, giving full details of its development and of an extensive HPLC study showing that OLC in human plasma coelutes with authentic ouabain under two different chromatographic conditions.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Reagents.
Ouabain octahydrate, dipotassium hydrogen phosphate,
erythritol, 1,6-hexanediamine, sodium metaperiodate, horseradish
peroxidase (HRP; Type II), 3,3',5,5'-tetramethylbenzidine (TMB), and
bovine serum albumin (BSA; Fraction V) were from Sigma Chemical Co.,
Poole, Dorset, UK. Sodium hydrogen carbonate, sodium dihydrogen
orthophosphate, sodium acetate, citric acid, disodium hydrogen
orthophosphate, Triton X-100, Tween 20, Brij 35, 2.5 mol/L sulfuric
acid and concentrated hydrochloric acid (all AnalaR grade), 300 g/L
hydrogen peroxide (AristaR), methanol, trifluoroacetic acid (TFA),
acetonitrile, dimethyl sulfoxide (all Far UV or Spectrosol grade), and
thin-layer chromatography plates (0.2 mm silica on aluminum) were from
Merck, Lutterworth, Leic, UK. Donkey anti-rabbit IgG-coated microtiter
plates were from Netria, London, UK; Freund's adjuvant from Difco
Lab., Detroit, MI; and PD10 gel columns from Pharmacia Biotech, St.
Albans, Herts, UK. Other chemicals were from Aldrich Chemical Co.,
Gillingham, Dorset, UK.
Solutions.
Phosphate-buffered saline (PBS), pH 7.4, consisted
of 20 mmol/L Na2HPO4, 5 mmol/L
NaH2PO4, and 150 mmol/L NaCl. Assay diluent
buffer (ADB) for the EIA consisted of 40 mmol/L
Na2HPO4, 10 mmol/L
NaH2PO4, 0.5 mL/L Tween 20, and 1 g/L gelatin,
pH 7.4. EIA substrate solution consisted of 100 mmol/L sodium acetate,
2 mmol/L citric acid, 1.3 mmol/L hydrogen peroxide, pH 6.0, in AnalaR
water. TMB reagent was stored at 42 mmol/L in dimethyl sulfoxide and
was diluted 100-fold in substrate solution just before addition.
Calibrators.
Because no pure human OLC was available,
solutions of ouabain octahydrate ranging from 40 to 10 000 pmol/L were
prepared in ADB and stored at -20 °C for use as calibrators.
Samples.
Blood was drawn from the antecubital vein and
collected in 10-mL tubes containing EDTA. Plasma was separated
immediately by centrifugation (16 °C, 1400g, 10 min) and
stored in 1.2-mL aliquots at -20 °C. Samples were collected in
accordance with the ethical requirements of our institution.
immunogen synthesis
The method used was a modified version of the protocol described
by Harris et al. (6), the major difference being that we
used an oxime intermediate to facilitate the conjugation, whereas
Harris et al. directly coupled ouabain dialdehyde to
BSA–C6 by using sodium cyanoborohydride.
Oxidation of ouabain.
We added 1 mL of distilled water and 670
µL of acetone to 100 mg of ouabain octahydrate, warming the solution
to 30–40 °C until all the ouabain dissolved. When the solution had
cooled to ambient temperature, we added 75 mg of sodium metaperiodate.
We then slowly added, over a 2-h period, 75 mg of
K2HPO4 to maintain pH between 7 and 8. The
reaction was monitored by thin-layer chromatography [developing
solvent, dichloromethane/methanol (3/1 by vol); visualization under
UV]. The reaction was complete at 2 h, as demonstrated by the
presence of a new single product spot (Rf 0.8) and the absence of
starting material (Rf 0.4). To terminate the reaction, we added 8 mg of
erythritol and stirred for 2 h. We removed volatile contaminants
by rotary evaporation and dissolved the product in a mixture of
methanol and dichloromethane (5 mL each). To this we added magnesium
sulfate to remove any remaining water and filtered the solution through
sintered glass with suction. Removal of the solvent by rotary
evaporation yielded 87 mg of product, ouabain dialdehyde, as a white
powder.
Formation of ouabain carboxymethyloxime.
We dissolved ouabain
dialdehyde (87 mg) in 5 mL of ethanol and then added 750 µL of
distilled water and 73 mg of carboxymethoxylamine hemihydrochloride. We
added triethylamine (89 µL) and monitored the reaction by thin-layer
chromatography (developing solvent as before) for 80 min. We judged the
reaction to be complete by the absence of starting material (Rf 0.8)
and the presence of a single product spot (Rf 0.1). To this reaction
mixture we added pyridine (400 µL) to prepare for the addition of
derivatized BSA.
Conjugation of BSA to hexanediamine.
We dissolved 1 g of
hexanediamine (C6) in 10 mL of distilled water and adjusted
to pH 8 with concentrated hydrochloric acid. BSA (260 mg) was then
added and allowed to dissolve. Conjugation proceeded after addition of
300 mg of ethylcarbodiimide hydrochloride (EDC) in small portions every
15 min over a 6-h period. This mixture was stirred at room temperature
for an additional 18 h and then dialyzed for 4 days against six
1-L changes of 30 mmol/L K2HPO4, pH 8.
Coupling ouabain carboxymethyloxime to BSA–C6.
We
added with gentle stirring 40 mg of the BSA–C6 to the
solution of ouabain carboxymethyloxime, and then added EDC: 30 mg over
4 h, an additional 15 mg over 1 h, and a final 15 mg. After
the mixture had been stirred at room temperature for 18 h, we
dialyzed it for 4 days against five changes of 1 L of distilled water
and finally freeze-dried the final product to produce the desired
immunogen (49.4 mg) as an off-white powder.
enzyme label synthesis
Oxidation of ouabain.
We used sodium periodate to oxidize
ouabain to the aldehyde as described above for the preparation of
immunogen.
Conjugation to HRP.
We added 390 µL of 0.1 mol/L
NaHCO3 reagent to 5.5 mg of ouabain dialdehyde, followed by
10 µL of an aqueous solution of sodium cyanoborohydride (0.2 mol/L).
We immediately added the resulting mixture to 5 mg of freeze-dried HRP,
purified according to Tijssen and Kurstak (31), and
incubated the mixture in the dark for 4 h. To remove free ouabain
from the ouabain–HRP conjugate, we used a PD10 gel-filtration column
(prepacked with 8.5 mL of Sephadex G-25 and equilibrated with 50 mL of
0.1 mol/L NaHCO3). After loading the reaction mixture onto
the column, we added 1 mL of NaHCO3 and discarded this
first eluate. We repeated this with an additional 1 mL of
NaHCO3, again discarding the eluate. After adding another
200 µL of NaHCO3, we eluted the product (detected by the
color of HRP) with 1.5 mL of NaHCO3, collecting the eluate
in a glass vial. We diluted the product to 50 mL with 50 mmol/L sodium
phosphate buffer, pH 7.4, containing 10 g/L BSA and 20 g/L mannitol,
and freeze-dried this in 1-mL aliquots. We resuspended one of these
aliquots in 2 mL of ADB and stored it in 50-µL aliquots at
-20 °C. Just before assay, we diluted each aliquot 10 600-fold in
ADB to give the working solution.
antibody production and assessment
We immunized four New Zealand White rabbits by intradermal
injection with an emulsion of 1 mg of BSA–C6–ouabain
in 1 mL of complete Freund's adjuvant, 100 µL of 9 g/L NaCl
solution, and 20 µL of Brij 35, which we vortex-mixed for 30 min just
before immunization. Subsequent immunizations (every 4 weeks) were with
incomplete Freund's adjuvant. Animals were bled and serum was
harvested 11 days after each immunization. We immunized two more
rabbits by the same procedure but with 1 mg of BSA–C6 in
place of the immunogen, to obtain a control antiserum. To determine the
anti-ouabain antibody titers from the immunized rabbit serum, we used
doubling dilutions of rabbit antiserum in ADB and added to duplicate
tubes 200 µL of each antiserum dilution and 100 µL (200 fmol) of
[3H]ouabain (200 fmol/100 µL). All dilutions were
vortex-mixed and then incubated for 21–24 h at 4 °C. To separate
the antibody bound and free fraction, we added 500 µL of
dextran-coated activated charcoal (10 g/L activated charcoal and 1 g/L
dextran in assay buffer) to each tube (except those for total count),
vortex-mixed the sample, and centrifuged them at 4 °C for 15 min at
1400g. We decanted the supernatants into scintillation vials
to which 4 mL of Ecoscint-A (National Diagnostic, Wessle, UK) was
added. Each vial was then capped, shaken, and left in the dark for at
least 90 min before counting radioactivity for 10 min. The titer
(dilution at which binding is 50% of maximum) from the best-responding
rabbit was found to increase with time, reaching a plateau at 1:12 000
after five immunizations. We used this antiserum at a dilution of
1:20 000 in the EIA. Antiserum from rabbits immunized with hapten only
did not show any substantial binding to [3H]ouabain.
sample pretreatment
The method was similar to that described by Harris et al.
(6). Frozen plasma samples were thawed and acidified with
TFA to a final concentration of 1 mL/L, mixed, and centrifuged just
before extraction. The C18 solid-phase extraction columns
were connected to a vacuum manifold (Varian), which we modified by
replacing the metal delivery tips with plastic ones. The columns were
initially activated with 6 mL of 400 mL/L acetonitrile in water
containing 1 mL/L TFA; the solvent was removed by rinsing the column
twice with 3 mL of 1 mL/L TFA in water. We then loaded 1 mL of an
acidified plasma and removed any unbound material by five washes with 3
mL of 1 mL/L TFA in water. Ouabain was eluted in 5 mL of the
acetonitrile/water/TFA solution. We dried the eluate under vacuum
centrifugation at 55 °C and resuspended the residues in 300 µL of
ADB before assay. This process concentrated the original plasma sample
3.3-fold.
Samples for HPLC analysis followed the same protocol as above except that instead of eluting after the sample was loaded and washed through, we loaded an additional 1 mL of plasma onto the column, repeating this procedure until 5 mL of plasma had been loaded. Retained components were then eluted with 5 mL of the acetonitrile/water/TFA solution. The eluate was dried as before and resuspended in 300 µL of 0.5 or 0.2 mL/L TFA, depending on the mobile phase used.
procedures
EIA.
We added (250 µL/well) rabbit anti-ouabain antiserum
diluted 1:20 000 in PBS to microtiter plates precoated with donkey
anti-rabbit IgG. Each plate was incubated at 4 °C in a
constant-humidity chamber. After 2 h (or as long as 7 days) we
washed the plates four times with PBS containing 0.1 mL/L Triton X-100
and dried them on a paper towel by tapping. We then added to each plate
50 µL of calibrator/control/sample, followed by 200 µL of
enzyme-labeled ouabain working solution, and incubated these in the
dark, with shaking, for 2 h at room temperature. We washed each
plate four more times with PBS containing 0.1 mL/L Triton X-100 and
dried it by tapping onto a paper towel before adding 200 µL of TMB
substrate reagent. Color was allowed to develop for 20 min before we
terminated enzyme turnover by adding 50 µL of 2.5 mol/L
H2SO4. Absorbance was measured in a microplate
spectrophotometer (Molecular Devices Vmax) at 450 nm (reference
wavelength, 650 nm). We expressed results for endogenous OLC as pmol/L,
assuming 100% cross-reactivity in the assay, which may or may not be
correct. Therefore, the numerical values for OLC may or may not be
accurate, but are correct as relative measures.
HPLC.
For HPLC studies we used an isocratic reversed-phase
system and a C8 or a C8/C18 column
run under different conditions. The mobile phase for the C8
column was 70 mL/L acetonitrile in water containing 0.5 mL/L TFA and
for the C8/C18 column was 200 mL/L methanol
containing 0.2 mL/L TFA. Flow rate for both columns was 1 mL/min. We
calibrated the columns with authentic ouabain. Plasma extracts were
obtained from four healthy subjects on C18 columns as
described earlier, resuspending the extract in 300 µL of acidified
water, mixing, and centrifuging before HPLC.
The test samples were loaded with a 50-µL loop. Eluate fractions (1 mL) were dried by vacuum centrifugation at 55 °C before resuspension in 300 µL of assay buffer for measurement of ouabain immunoreactivity. To ensure that no ouabain remained from any exogenous source, we flushed through at least five blank samples—0.5 mL/L TFA (C8 column) or 0.2 mL/L TFA (C8/C18 column)—between each run and assayed the fractions.
Stability studies.
To assess stability in whole blood, we
collected blood from two healthy subjects into EDTA tubes. Plasma was
separated by centrifugation (16 °C, 1400g, 10 min) and
frozen at -20 °C at time 0 and subsequently at 1, 2, 4, 6, and
24 h after collection. On the day of extraction, samples were
acidified and centrifuged as described in Sample
pretreatment.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
), or allowing for the 3.3-fold concentration effect of sample
extraction, 54–3000 pmol/L in the original plasma sample.
|
Minimum detectable concentration.
The lowest concentration
that can be distinguished from zero by statistical criteria was
calculated with Multicalc software program version 1.5 as the mean + 3
SD of the response at zero interpolated from the calibration curve.
Individual assay data were pooled (n = 13) to give a mean ±
SD minimum detectable concentration of 37 ± 11 pmol/L (range
21–57), equivalent to 11 ± 3.3 pmol/L in the original plasma sample.
Analytical recovery.
The mean recovery of 500 pmol/L ouabain
added to plasma from five healthy subjects was 100% (range 84–117%);
for plasma from six renal transplant recipients, it was 99% (range
80–119%). The mean recovery of 200 pmol/L ouabain added to plasma
from five healthy subjects was 90% (range 87–93%).
Parallelism.
Dilutions of plasma extracts from one healthy
volunteer and one renal transplant recipient with normal renal function
paralleled the calibration curve (Table 1
).
|
Specificity.
We determined the cross-reactivity of various
steroids and vasoactive compounds in the ouabain assay without prior
extraction. To calculate cross-reactivity, we divided the concentration
of cross-reacting material that inhibited 50% of the binding of the
zero calibrator by the amount of ouabain that gave the same inhibition,
expressed as a percentage (Table 2
). We detected no clinically critical cross-reactivity with any
of the endogenous mammalian steroids tested although, by 40 weeks of
gestation, serum progesterone concentrations could reach 800
nmol/L—which at 0.015% cross-reactivity equates to a ouabain
concentration of 120 pmol/L in this assay. However, recovery of
exogenous progesterone (670 nmol/L) added to duplicate plasma samples
before extraction was only 5% (35 nmol/L progesterone), illustrating
no substantial cross-reactivity.
|
HPLC.
HPLC of six extracted human plasma samples gave a single
immunoreactive peak that coeluted with authentic ouabain. Retention
times were 17 ± 1 min (Fig. 2
) for the C8 column and 14 ± 1 min (Fig. 3
) for the C8/C18 combination column.
Recoveries of plasma OLC after HPLC ranged from 70% to 90%. To
prevent carryover from one sample to the next, we extensively purged
the column between assays. Before loading each plasma extract, we
chromatographed a blank sample, which showed no evidence of containing
detectable ouabain immunoreactivity.
|
|
Stability studies.
OLC was stable in whole blood for 24 h
at room temperature (Fig. 4
). CVs for plasma OLC concentrations in two healthy volunteers
at the six storage time points were 9.4% and 8.8%, similar to the CV
for 10 extractions of a single plasma sample in one assay.
|
Reference values for plasma OLC.
Ten healthy human volunteers
(five men, five women), mean age 28 years (range 26–32), had a mean
OLC concentration of 92 pmol/L (range 55–168), assuming 100%
cross-reactivity of OLC in the ouabain assay.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This assay is the most completely described immunoassay since the work of Harris et al. (6). Our results confirm their findings in human plasma of an OLC that coelutes with authentic ouabain. This is important because other workers (23)(24) have not been able to reproduce the results of Hamlyn's group and therefore cast some doubt on their original findings. Our antiserum, like others described, recognized similar cardiac glycosides but cross-reacted little with endogenous steroids (6)(23)(24). Cardiac glycosides have dimensions similar to the antibody binding site, and it is possible to examine the most important antigenic determinants for a given anti-ouabain antibody (32). The poor cross-reactivity of dihydroouabain indicates that the unsaturated lactone ring is an essential recognition epitope. The rhamnose moiety, which is absent from ouabagenin, is less important, leading to only a partial reduction in immunoreactivity. Digoxin and digitoxin display marked differences in antibody binding, despite being practically identical (digitoxin lacks a hydroxyl group at C-12 on the steroid backbone). Ouabain and strophanthidin also do not contain a hydroxyl group at C-12, and it can be assumed that this site is important for antibody recognition. Collectively these data indicate that our assay detects a digitalis-like compound, i.e., a cardenolide whose steroid moiety has the cis–trans–cis configuration between its ring junctions. The antiserum did not cross-react substantially with any of the endogenous mammalian steroids tested. The weak cross-reactivity of progesterone was clinically negligible after sample extraction because of the selectivity of the extraction technique.
Reported OLC concentrations in healthy subjects are 300–1000 pmol/L (19)(26), 40 pmol/L (9), <75 pmol/L (24), and <5 pmol/L (23). The OLC concentrations detected by our EIA (~100 pmol/L) are between these extremes. The differences could be partially due to assay- and laboratory-dependent differences in calibration, extraction recovery, and antibody specificity (see below).
These HPLC results show that immunoreactive human plasma OLC has elution characteristics identical with ouabain under two different elution and separation conditions—further evidence that OLC is probably very similar structurally to authentic ouabain. This is in agreement with recent structural analyses of OLC from human plasma and bovine hypothalamus, which found that ouabain and OLC are structurally alike but not identical (14)(25). Some groups (24)(30) have not been able to reproduce the original HPLC results (5) that showed ouabain coeluting with immunoreactive human OLC in an acetonitrile gradient. The difficulties associated with running reproducible gradient separations may explain this. Because our work with acetonitrile gradients gave erratic results (data not shown), we opted for an isocratic elution. Differences in antibody specificity are likely to be a major factor contributing to both the ability to detect OLC and its apparent concentration. If, as recent evidence suggests, OLC is not ouabain but a stereoisomer of ouabain, some anti-ouabain antiserum may cross-react with OLC far less than other antisera. Thus an OLC peak coeluting with ouabain might not be detected by anti-ouabain antibodies in some assay systems. Therefore, an antiserum with a very high specificity for ouabain, e.g., that described by Gomez-Sanchez et al. (30), which showed minimal cross-reactivity for related cardiac glycosides, might prove less useful as an immunoassay reagent for detecting OLC than is the antiserum described here. These differences in specificity may result from the different procedures used for immunogen synthesis—although all methods use ouabain conjugated to a carrier protein through the rhamnose moiety. The majority of groups who, like ourselves, have used (9)(11)(26) or partially used (6) a BSA–ouabain conjugate to raise an antiserum have achieved a higher degree of OLC detection than those using other conjugates of ouabain, e.g., linked to ovalbumin (23) or jack bean urease and porcine thyroglobulin (30). However, Doris et al. (24) and Worgall et al. (29), using an assay with an antiserum raised with a BSA conjugate, have both reported low or undetectable plasma OLC concentrations. The use of a spacer between ouabain and the carrier protein does not appear to confer a greater detection of OLC. Investigators who have not used a spacer include Masugi et al. (26), who reported relatively high concentrations of OLC, and Gomez-Sanchez et al. (30), who could barely detect any OLC immunoreactivity. The reasons for the differences in ability to detect OLC are unclear.
In conclusion, this new, rapid, and easy-to-perform immunoassay for OLC has an obvious advantage over some of the other assay methods: its ability to detect OLC in measurable quantities. However, the degree of OLC cross-reactivity in our assay is unknown, and therefore its absolute accuracy is uncertain. Because human OLC dilutes in parallel with the ouabain calibration curve and because added ouabain can be recovered with great accuracy, we think it likely that OLC is highly cross-reactive in our assay. However, until this and the other immunoassays have been calibrated with human OLC, all assigned OLC concentrations should be treated as relative measurements only and not necessarily be considered comparable between assays. The HPLC data indicate that the chromatographic characteristics of plasma OLC are identical with those of authentic ouabain, providing further evidence for structural similarity. Elucidation of the origin, bioactivity, and pathophysiological role of plasma OLC is the subject of ongoing studies.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  O. Vakkuri, S. S. Arnason, P. Joensuu, J. Jalonen, O. Vuolteenaho, and J. Leppaluoto Radioiodinated Tyrosyl-Ouabain and Measurement of a Circulating Ouabain-like Compound Clin. Chem., January 1, 2001; 47(1): 95 - 101. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. M. A. M. Qazzaz, M. A. El-Masri, and R. Valdes Jr. Secretion of a Lactone-Hydrogenated Ouabain-Like Effector of Sodium, Potassium-Adenosine Triphosphatase Activity by Adrenal Cells Endocrinology, September 1, 2000; 141(9): 3200 - 3209. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
