| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
1
National Diagnostics Centre—Bioresearch Ireland, University College, Galway, Ireland.
2
Department of Obstetrics & Gynecology, Asahikawa Medical
College, Japan.
3
Department of Biochemistry, University College, Galway,
Ireland.
4
Fertility Unit, Department of Obstetrics & Gynecology,
University College Hospital, Galway, Ireland.
a Author for correspondence. Fax 353 91 586570; e-mail: marian. kane{at}ucg.ie
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Key Words: indexing terms: menstrual cycle • in vitro fertilization
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Saliva has many practical advantages as a diagnostic medium. Sampling is noninvasive, it can be carried out without medical supervision, and frequent sampling is easy and cheap. However, because they closely reflect the free fraction of the steroid in plasma, the concentrations of steroids present in saliva are only 1–2% of their concentrations in blood. In recent years convenient, direct immunoassays have been established for some steroids in saliva, e.g., progesterone, cortisol, and testosterone (1)(2)(3)(4). However, the measurement of estradiol in saliva is particularly difficult because the concentrations are so low (1–36 pmol/L in nonpregnant women) (5).
Most published methods for measuring estradiol in saliva are radioimmunoassays and involve extraction with an organic solvent (5)(6)(7)(8). Direct, nonextraction immunoassays for salivary estradiol have been described, but these involve either radiochemical (9) or chemiluminescent (10) detection, both of which require specialized equipment.
Here we describe a direct, competitive, solid-phase, colorimetric enzyme immunoassay (EIA) with a monoclonal antibody and carried out on microtiter plates, as well as its validation and clinical application.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
reagents
Buffers and standards.
There were two general assay buffers:
phosphate-buffered saline containing pig collagen Prionex (Pentapharm
Ltd., Basel, Switzerland) (PBS/Prionex), which consisted of 10 mmol/L
sodium phosphate buffer, pH 7.4 (1.9 mmol/L
NaH2PO4, 8.1 mmol/L
NaH2PO4), 150 mmol/L NaCl, 1 g/L Prionex, and
0.1 g/L thimerosal, and PBS/Prionex/EDTA, which contained in addition
10 mmol/L EDTA. The coating buffer was 50 mmol/L
Na2CO3, 0.1 g/L thimerosal, pH 9.6. The enzyme
assay buffer contained 5.9 mmol/L H2O2, 5.5
mmol/L o-phenylenediamine, 103 mmol/L
Na2HPO4, 48.5 mmol/L citric acid, and 0.1 g/L
thimerosal, pH 5.0. Washing solution was 150 mmol/L NaCl, 0.5 mL/L
Tween 20. Enzyme conjugate stabilizing solution contained 10 mmol/L
sodium phosphate, 1.36 mol/L glycerol, 20 µmol/L cytochrome
c, 30 µmol/L bovine serum albumin, and 0.1 g/L thimerosal,
pH 7.4.
Stock solution of estradiol-17ß (1,3,5[10]-estratriene-3,17ß-diol, cat. no. E 8875; Sigma Chemical Co., St. Louis, MO), 734.2 µmol/L in ethanol, was stored at -20 °C. Working calibrators (3.67–367 pmol/L, 1–100 ng/L) were prepared in PBS/Prionex and stored at 4 °C.
Monoclonal antibody.
The monoclonal antibody against
17ß-estradiol–6-(O-carboxymethyloxime)–bovine serum
albumin (2F9) was purchased from Biodesign International (Kennebunk,
ME). 2F9 was an IgG2a isotype (Ka =
1 x 1010 L/mL) (12), supplied as an
affinity-purified immunoglobulin from mouse ascitic fluid in PBS with
10 mmol/L NaN3, and stored in aliquots at -20 °C.
Rabbit anti-mouse IgG was supplied by Dakopatts (High Wycombe, Bucks,
UK) as a purified immunoglobulin fraction in 100 mmol/L NaCl, 15 mmol/L
NaN3.
Steroid–enzyme conjugate.
The derivative,
17ß-estradiol–6-(O-carboxymethyloxime), was covalently
linked to horseradish peroxidase by the mixed-anhydride method
(13), as modified by Dawson et al. (14).
Unreacted material was separated by chromatography on Sephadex G-25
(Pharmacia, Uppsala, Sweden), and the conjugate was stored at 4 °C
in stabilizing buffer.
quality-control saliva
Male saliva was frozen and thawed, heat-treated (see below),
clarified by centrifugation, and assayed to confirm that estradiol was
<7.3 pmol/L (2 ng/L). The resulting pool was divided into four
portions, each supplemented with stock solution of 17ß-estradiol in
ethanol to give concentrations of 21, 32, 45, or 128 pmol/L, and stored
at -20 °C in small aliquots. A set of freshly thawed control saliva
samples were included in each assay run.
procedures
Heat treatment of saliva samples.
Saliva samples were
heat-treated before assay to reduce interference by the saliva matrix
(15) and to inactivate infectious contaminants. Thawed
samples (overnight at 4 °C) were placed in an oven at 56 °C for
2 h and centrifuged at 3000g for 20 min. Supernatants
were either assayed immediately or stored at -20 °C until assay.
Coating of microtiter plates.
Each well was coated with a
second antibody by addition of 200 µL of rabbit anti-mouse IgG
diluted 2000-fold in coating buffer, incubated at 37 °C for 90 min,
emptied, washed four times with 300 µL of wash buffer, and tap-dried.
Then 100 µL of 2F9 monoclonal antibody diluted 12 000-fold with
PBS/Prionex was added, incubated at 4 °C overnight, and washed as
above.
Estradiol EIA.
To each well we added 50 µL of
17ß-estradiol calibrator, control saliva, or unknown saliva, covered,
and incubated at room temperature (16–20 °C) for 60 min. Then 150
µL of estradiol–peroxidase conjugate (5.3 µg/L) was added, mixed,
covered, and incubated at room temperature for 30 min. After the
conjugate was washed four times as above, 150 µL of enzyme assay
buffer was added and incubated for 30 min at room temperature in the
dark. The enzyme reaction was stopped by the addition of 50 µL of 4
mol/L H2SO4 to the wells in the same order and
at the same rate as the substrate solution was added. The solutions
were mixed in the wells and then the absorbance was measured at 492 nm
with a microtiter plate reader (SLT EAR 400 AT). The absorbance reading
for each well (B) was divided by the average absorbance for
the zero calibrator wells (B0) (n = 4) to
give a series of B/B0 ratios. A
calibration curve of the logarithms of the estradiol concentrations vs
the B/B0 ratios for the calibrators
was drawn, and the control and unknown concentrations were
interpolated.
Other assay kits.
We measured salivary progesterone
concentration with the kit from BioResearch Ireland (2).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
represents the mean of 12 curves, obtained separately, each
with duplicate concentrations of calibrators. The mean absorbance of
the zero calibrator was 0.968 (CV = 5.3%) and the mean
ED50, defined as the calibrator concentration
equivalent to a B/B0 value of 0.5,
was 66.1 pmol/L (18 ng/L; CV = 8.2%). The lower detection limits
of the assay, i.e., the concentrations equivalent to
B0 - 2 SD and B0 - 3
SD, were 260 and 365 amol (72 and 100 fg) per well or 5.2 and 7.3
pmol/L saliva, respectively. The between-replicates CV calculated from
duplicate data was 
15% over the whole analytical range (Fig. 1
).
|
precision
Repeated assay in one run (n = 6) of four quality-control
saliva samples yielded the following concentration-dependent intraassay
CVs: at 22 pmol/L, 8.9%; at 33 pmol/L, 11.5%; at 46 pmol/L, 8.0%;
and at 127 pmol/L, 6.8%. Interassay CVs for duplicate determinations
of four quality-control saliva samples in 14 separate assays gave
15.3% at 21 pmol/L, 12.3% at 31 pmol/L, 9.2% at 46 pmol/L, and
10.3% at 132 pmol/L.
independence of volume
We examined the ability of the assay to measure estradiol in
saliva independently of the volume of sample used. Saliva samples from
five different subjects were diluted to various degrees in calibrator
buffer, and 50-µL aliquots of each dilution were assayed. Sixty- and
70-µL aliquots of undiluted saliva were also assayed for the
lower-concentration samples. The lack of consistent deviations from the
horizontal in the dilution plots for all the samples tested indicates
minimal interference by normal saliva components up to and including
the concentrations obtained by adding 70 µL of sample (Fig. 2
); the volume normally used is 50 µL.
|
analytical recovery
The ability of the assay to accurately quantify estradiol in
saliva samples was tested. Low, medium, and high concentrations of
estradiol (14, 36, and 114 pmol/L, respectively) were added to portions
of each of four saliva samples containing different endogenous
concentrations of estradiol (0–165 pmol/L). We then determined the
total amounts of estradiol present in each sample and calculated the
recovery in each case. The range of analytical recovery was 83–110%,
and the mean was 94.8% (±7.5% SD, n = 12).
specificity
The cross-reactivities of various steroids, either closely related
structurally to estradiol or of physiological importance, were assessed
as described by O'Rorke et al. (2) and Abraham
(16) (Table 1
). As recommended by Krouwer (18), the maximum
concentration of each potential cross-reactant that might be
encountered in patients' samples is listed (when this could be located
in the literature), and the potential for interference in the assay at
that concentration is estimated.
|
clinical applications
Natural menstrual cycle.
Fig. 3
shows the concentrations of salivary estradiol and progesterone
in daily samples collected from a healthy woman who had regular
menstrual cycles. As expected, there were two estradiol peaks, located
at the preovulatory and midluteal phase of the cycle. Among the five
healthy women from whom daily samples over full cycles were obtained,
the estradiol peak ranged from 15 to 31 pmol/L (mean ± SD:
22 ± 7 pmol/L) in the preovulatory phase and from 9 to 33 pmol/L
(21 ± 11 pmol/L) in the midluteal phase. Also the concentrations
of salivary progesterone increased ~6 days after the preovulatory
estradiol peak in each case. These results are comparable with those
previously published (5).
|
Stimulated cycle.
We also measured the concentrations of
estradiol in saliva samples collected during eight stimulated cycles
from patients participating in the IVF and embryo transfer program in
our hospital. Again we found two estradiol peaks in each cycle,
resembling those in the natural cycle. However, the peak maxima were
often much greater and ranged from 29 to 224 pmol/L (105 ± 65
pmol/L) in the preovulatory phase and from 24 to 108 pmol/L (82 ±
60 pmol/L) in midluteal phase. When the profiles of cycles resulting in
pregnancy were compared with those of nonpregnant cycles, there was
generally a distinct difference between the mid–late luteal phase
profiles (Fig. 4
).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To our knowledge, this is the first report of the validation of a direct, colorimetric EIA for estradiol in saliva. Other direct assays have been described, but these involve either radioactive (9) or chemiluminescent (10) tracers, which require expensive and nonstandard equipment to determine the end point. The basis of the current assay is a stable, horseradish peroxidase-labeled estradiol conjugate that can be quantified by a simple colorimetric assay. The conjugate solution is stable for >3 years at -20 °C and at least 1 year at 4 °C.
The performance of the current assay compares well with that of the chemiluminescent assay described by DeBoever et al. (10), which uses the same monoclonal antibody. The detection limit (at B0 - 2 SD) of 260 amol/test is lower than the equivalent 756 amol/test obtained with the chemiluminescent assay and much lower than the limits obtained with a competitive colorimetric assay (3.67 fmol/test) (21) and a noncompetitive idiometric EIA (12.8 fmol/test) (22), all with the same monoclonal antibody. Jackson and Ekins (23) predicted a theoretical minimum detection limit of 400 amol/test for a competitive assay involving an antibody of similar affinity. The assay described here has achieved this detection limit. Accuracy and precision are also comparable with the chemiluminescent assay despite the lower sample volume (50 as compared with 100 µL) used in the current assay (10). Cross-reactivity studies have shown that the assay is not susceptible to interference from other steroids at the concentrations likely to be found in saliva, except for estriol in the third trimester of pregnancy.
Most steroids in blood are protein-bound, either to sex-hormone-binding globulin or to albumin (24), whereas the concentrations of salivary steroids such as estradiol are believed to reflect more closely the unbound fractions in plasma (25). The low concentration of the sex- hormone-binding globulin that has been demonstrated in saliva probably arises from contamination with gingival fluid, which may constitute up to 0.5% of the volume of saliva in healthy subjects (26). However, DeBoever et al. (10) reported no substantial difference between salivary estradiol concentrations obtained in the absence and presence of displacing agents. Similar results were also found with the current assay format (results not shown).
The suitability of this assay for the measurement of the low concentrations of estradiol in saliva during the normal menstrual cycle was demonstrated. In all five unstimulated cycles examined, we found the expected two peaks in each cycle at the preovulatory and midluteal phase with concentrations in the reported ranges (4). That ovulation and luteinization occurred in each cycle was supported by the demonstration of a typical rise and fall in progesterone concentration during the luteal phase of the cycle. In some other cycles examined more recently (data not shown), no estradiol peaks were identified, but the luteal phase progesterone peak was also absent, strongly suggesting that these cycles were anovulatory.
From an analytical viewpoint, the measurement of estradiol in the saliva of patients undergoing cycle stimulation as part of an IVF and embryo transfer program presents much less of a challenge because of the higher estradiol concentrations encountered during these treatments. However, a wide variation in the peak values of estradiol attained during the preovulatory and luteal phases of individual patients was noted, although both peaks were clearly seen in all patients. We also noted a tendency towards a higher estradiol concentration in the late luteal phase of cycles that resulted in pregnancy than in unsuccessful treatments. This tendency was also noted by Wong et al. (27), suggesting the possible use of salivary estradiol measurements for prediction of outcome of an IVF and embryo transfer program. The variations of the salivary estradiol profiles obtained during the follicular phase of unsuccessful IVF and embryo transfer attempts may also prove useful when deciding on the treatment variation to be given during subsequent attempts, as suggested by Ronnberg et al. (28).
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  G Wycherley, D Downey, M T Kane, and A C Hynes A novel follicle culture system markedly increases follicle volume, cell number and oestradiol secretion Reproduction, June 1, 2004; 127(6): 669 - 677. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
