459
A RAPID SPECIFIC RADIOIMMUNOASSAYFOR UNCONJUGATED ESTRIOL IN PLASMA Thomas K. Brown (l), Margaret A. Brammall and Harvey S. Schiller Departments of Obstetrics and Gynecology and Laboratory Medicine, University of Washington, Seattle, Washington 98195 Received: l2/08/75
ABSTRACT
A rapid, non-chromatographicradioimmunoassay for unconjugated estrio1 in pregnancy plasma has been developed which utilizes a commonly available antiestrogen antisera. Estradiol-178 and estrone demonstrate 135% relative cross-reactivitywith our antiserum, as compared with 100% for estriol. Specificity is achieved by purification of estrio1 with solvent partitioning using benzene: petroleum ether (1:l). The results obtained using this method are similar to a radioimmunoassay utilizing a highly specific, but commercially unavailable, antiestriol antiserum. The method is precise, with coefficients of variation ranging from 3.0 to 8.2%. INTRODUCTION Monitoring of 24 hr urinary estriol (2) during the last trimester of human pregnancy is a well proven method for detecting fetal distress (3-5). Fluorometric (6) and gas chromatographic (7) methods were formerly employed to measure either urinary or, more recently, plasma estriol. Many investigators are now reporting radioligand assays for plasma estriol utilizing either uterine cytosol (8, 9) or antiestrogen antibodies. This report describes a rapid nonchromatographicRIA for plasma, unconjugated estriol during pregnancy. The method is specific for estriol and uses commercially available, non-specific antiestrogen antisera. MATERIALS Antiserum: A lyophilized antiserum (S-310 No. 5), prepared in a ewe against the estriol-3,16,17-trisuccinate-HSA conjugate, was obtained commercially (ProfessionalStaff Association, Harbor General Hospital, Los Angeles, California). It was reconstituted in distilled water, and small aliquots were frozen at -40 OC until used. The antiserum was diluted with the buffer to achieve an initial binding of 60-70% of 13H) estriol in the absence of standard estriol (usually 750-fold dilution). Steroids: The following radioactive steroids were obtained from New England Nuclear Cor ., Boston, Mass.: {3H}2,4,6,7-estradi+-178 (S-A. { 5H)6,7-estrone (S.A. 40 Ci/mmole)and { H)6,7-estriol 105 Ci/mmole) , Volume
27, Nwnber
4
S
TIIEOXDI
April,
1976
(S.A. 53.1 Ci/mmole) all radioactive steroids were shown to have greater than 95% purity. Radioinert steroids were purchased from Sigma Chem. Co. and Steraloids Inc. The purity of estriol, estradiol-178 and estrone was tested by their melting points. The stock estriol solution was prepared twice a year in benzene:ethanol (9:l). From this, a working standard solution of 0.5 ng/ml was prepared monthly by diluting the stock solution with the buffer. Reagents and solutions: Benzene (J.T. Baker Chem. Co.) was purified by passage through a column of silica gel (28-200 mesh, grade 12, Fisher). Diethyl ether (Mallinckrodt Chem. Works) and absolute ethanol (U.S. Industrial Chemicals) were used as purchased. The bovine y-globulin fraction II was obtained from Miles Laboratories and prepared as a 1 mgl 0.05 ml solution in the buffer. The polyethylene glycol reagent, used as a precipitant, was prepared as a 300 g/kg solution of polyethylene glycol ("Carbowax 6000"; Union Carbide) in the buffer. The buffer (pH 7.0) was prepared as a sodium barbital (1.47 g/liter)-sodium acetate (0.97g/liter) sodium chloride (8.5g/liter) solution. METHOD Extraction, solvent partitioning and recovery: Venous blood was collected in heparinized tubes and the plasma stored at -20 'C until the assays were performed. Duplicate 0.1 ml plasma aliquots were pipetted into 15 ml Kimax screw top tubes, diluted to 1 ml with water and extracted twice with 6 ml of benzene: petroleum ether (1:l) by mixing rapidly on an automatic vortex-type stirrer for 30 sec. The organic phase was discarded. The aqueous phase was extracted twice in similar fashion with 6 ml of diethyl ether. The ether extracts were dried under nitrogen and redissolved on 1.0 ml buffer. Then 0.1 and 0.4 ml aliquots, along with triplicate standards containing 0 to 160 pg of estriol, were pipetted into 15 x 75 mm plastic disposable test tubes cgntaining the appropriate amount of buffer. 0.1 ml (25,000 dpm) of ( H) estriol was added to the samples and standards. After the addition of 0.1 ml of antiserum to all tubes, except those necessary to determine the total radioactivity, the tubes were gently shaken and allowed to incubate in a final volume of 0.7 ml at 0 'C for 2 hr. Antibody-bound and free 13H) estriol were separated by precipitating antibody with 0.75 ml of the polyethylene glycol solution after the addition of 1 mg (0.05 ml) of bovine y-globulin (10,ll). The tubes were immediately mixed on a vortex-type stirrer and centrifuged at 2800 rpm (1750 g) for 30 min. at 4 OC. The supernate was decanted, 10 ml of a modified Bray's solution (12), lacking ethylene glycol and methanol, was added and the radioactivity was counted in a Packard liquid scintillation counter with an efficiency of 54%. The percent recovery of estriol was calculated from triplicate 0.1 ml plasma aliquots incub ted for at least 30 min with 0.8 ml buffer and 4 0.1 ml (40,000 dpm) of ( H) estriol in buffer and extracted simultaneous ly with the plasma samples. The pe cent bound of the sample and stan5 dards was calculated from the free ( H) estriol in the supernatant as follows: percent bound = (total cpm - supernatant cpm) (lOO)/total cpm.
S
TREOTDrn
Standard curves were constructed from the percent E3H} estriol antibodybound, and plasma estriol concentrations, in pg/liter, were calculated as follows: estriol (pg/liter) = (0.1 x X x V)/(P x R x B), where X=pg from standard curve, V = total volume of buffer used for redissolving residue (ml), P = plasma volume assayed (ml), R = % recovery, and B = volume of buffer aliquot (ml). RESULTS I.
Antiserum Characteristics A.
Affinity constant: The equilibrium constant of association,
(Ka), of the antiserum was evaluated by the method of Scatchard (13). At 4 'C and under the conditions employed for our routine RIA, the antiserum had high affinity for estriol (Ka-1 x 101oM-l) and was therefore potentially capable of being used for a sensitive RIA. B.
Specificity: The cross reactivity of steroids with this anti-
serum was determined by the method of Abraham (14).
Estradiol-178
and estrone demonstrated 135% relative cross-reactivity,as compared with 100% for estriol (Table 1).
The relative cross-reactivities
obtained are typical for antisera prepared against estriol derivatives conjugated to the functional groups. II. Partial Purification of Estriol High assay specificity for unconjugated estriol is achieved by preliminary solvent extraction of interfering estrogens present in significant concentrations in plasma which cross-react with the antiserum. Only 2.5-3% of estradiol-178 and Z-2.52 of estrone are recovered in the final ether extract.
The percent of estriol recovered in this extract
is very consistent within, 86.721.1 (IS.D)% (n=lO), as well as between 86.922.3 (IS.D)% (n=ll), assays.
S
462
TPROID=
TABLE 1 Relatively Cross-Reactivity
of Sreroids with Antiserum
s 310 No. 5 Steroid
Cross-Reactivity
Estriol
100
Estradiol-178
135
Estrone
135
16-Epiestriol
66
17-Epiestriol
46
6-Oxoestradiol
16
16,17-Epiestriol
13
6-Oxoestriol
(X)
9
III. Sensitivity, Blank, Precision and Accuracy The sensitivity of the standard curve, defined as the least amount of estriol standard which could significantly estriol from the antiserum, was 8 pg.
(k2S.D.) displace j3H)
The solvent blank was invaria-
bly less than the sensitivity of the method except for a few instances when the benzene and petroleum ether had not been repurified for over 6 weeks.
Repurification
of these solvents eliminated the blank.
The
smallest amount of estriol which can be detected by our routine assay using 0.1 ml of plasma is 250 pg/ml.
Estriol could not be detected in
plasma from men or non-pregnant women. The precision of the method was studied by repetitive measurement of estriol in pooled plasma, in replicate and from day to day. intra-assay estriol mean for two plasma pools was 13.7kO.46
The
(1S.D.)
ng/liter (n=18) and 11.9kO.98 (1S.D.) ug/liter (n=24), with respective coefficients of variation of 3.4% and 8.2%. The inter-assay estriol means for two plasma pools was 5.720.35 (1S.D.) ng/liter (n=9) and 12.5kO.35 (1S.D.) ngfliter (n=6), with respective coefficients of varisation of 6.1% and 3.0%. Accuracy of the method was evaluated by several methods. First, various quantities of estriol were added to plasma and the samples were assayed. 93-108% of the added estriol was recovered (Table 2).
Second,
various plasma volumes were assayed. Linear regression analysis calculated by the least squares method resulted in a straight line with a correlation coefficient of 0.992 and y-intercept of -1.91 pg (Fig. 1). Table 2 Recovery of Estriol Added to Plasma Estriol Added (pg)
Estriol Measured (pg)fIS.D. N
C.V. (W) Recovery (X)
400
432k65
9
15
108
800
823242
9
5
103
1000
1047+23
4
2.1
105
1200
1178+115
9
9.8
98
1600
1488+81
9
5.4
93
In addition, fifty samples were assayed by this method and by a specific RIA for estriol employing a highly specific antiserum, which was prepared in a ewe against estriol-6-(0-carboxymethyl)oximederivative (15). Relative cross-reactivitywith other estrogens was low: estradiol-178, 0.8%; estrone,
A RAPID SPECIFIC RADIOIMMUNOASSAYFOR UNCONJUGATED ESTRIOL IN PLASMA Thomas K. Brown (l), Margaret A. Brammall and Harvey S. Schiller Departments of Obstetrics and Gynecology and Laboratory Medicine, University of Washington, Seattle, Washington 98195 Received: l2/08/75
ABSTRACT
A rapid, non-chromatographicradioimmunoassay for unconjugated estrio1 in pregnancy plasma has been developed which utilizes a commonly available antiestrogen antisera. Estradiol-178 and estrone demonstrate 135% relative cross-reactivitywith our antiserum, as compared with 100% for estriol. Specificity is achieved by purification of estrio1 with solvent partitioning using benzene: petroleum ether (1:l). The results obtained using this method are similar to a radioimmunoassay utilizing a highly specific, but commercially unavailable, antiestriol antiserum. The method is precise, with coefficients of variation ranging from 3.0 to 8.2%. INTRODUCTION Monitoring of 24 hr urinary estriol (2) during the last trimester of human pregnancy is a well proven method for detecting fetal distress (3-5). Fluorometric (6) and gas chromatographic (7) methods were formerly employed to measure either urinary or, more recently, plasma estriol. Many investigators are now reporting radioligand assays for plasma estriol utilizing either uterine cytosol (8, 9) or antiestrogen antibodies. This report describes a rapid nonchromatographicRIA for plasma, unconjugated estriol during pregnancy. The method is specific for estriol and uses commercially available, non-specific antiestrogen antisera. MATERIALS Antiserum: A lyophilized antiserum (S-310 No. 5), prepared in a ewe against the estriol-3,16,17-trisuccinate-HSA conjugate, was obtained commercially (ProfessionalStaff Association, Harbor General Hospital, Los Angeles, California). It was reconstituted in distilled water, and small aliquots were frozen at -40 OC until used. The antiserum was diluted with the buffer to achieve an initial binding of 60-70% of 13H) estriol in the absence of standard estriol (usually 750-fold dilution). Steroids: The following radioactive steroids were obtained from New England Nuclear Cor ., Boston, Mass.: {3H}2,4,6,7-estradi+-178 (S-A. { 5H)6,7-estrone (S.A. 40 Ci/mmole)and { H)6,7-estriol 105 Ci/mmole) , Volume
27, Nwnber
4
S
TIIEOXDI
April,
1976
(S.A. 53.1 Ci/mmole) all radioactive steroids were shown to have greater than 95% purity. Radioinert steroids were purchased from Sigma Chem. Co. and Steraloids Inc. The purity of estriol, estradiol-178 and estrone was tested by their melting points. The stock estriol solution was prepared twice a year in benzene:ethanol (9:l). From this, a working standard solution of 0.5 ng/ml was prepared monthly by diluting the stock solution with the buffer. Reagents and solutions: Benzene (J.T. Baker Chem. Co.) was purified by passage through a column of silica gel (28-200 mesh, grade 12, Fisher). Diethyl ether (Mallinckrodt Chem. Works) and absolute ethanol (U.S. Industrial Chemicals) were used as purchased. The bovine y-globulin fraction II was obtained from Miles Laboratories and prepared as a 1 mgl 0.05 ml solution in the buffer. The polyethylene glycol reagent, used as a precipitant, was prepared as a 300 g/kg solution of polyethylene glycol ("Carbowax 6000"; Union Carbide) in the buffer. The buffer (pH 7.0) was prepared as a sodium barbital (1.47 g/liter)-sodium acetate (0.97g/liter) sodium chloride (8.5g/liter) solution. METHOD Extraction, solvent partitioning and recovery: Venous blood was collected in heparinized tubes and the plasma stored at -20 'C until the assays were performed. Duplicate 0.1 ml plasma aliquots were pipetted into 15 ml Kimax screw top tubes, diluted to 1 ml with water and extracted twice with 6 ml of benzene: petroleum ether (1:l) by mixing rapidly on an automatic vortex-type stirrer for 30 sec. The organic phase was discarded. The aqueous phase was extracted twice in similar fashion with 6 ml of diethyl ether. The ether extracts were dried under nitrogen and redissolved on 1.0 ml buffer. Then 0.1 and 0.4 ml aliquots, along with triplicate standards containing 0 to 160 pg of estriol, were pipetted into 15 x 75 mm plastic disposable test tubes cgntaining the appropriate amount of buffer. 0.1 ml (25,000 dpm) of ( H) estriol was added to the samples and standards. After the addition of 0.1 ml of antiserum to all tubes, except those necessary to determine the total radioactivity, the tubes were gently shaken and allowed to incubate in a final volume of 0.7 ml at 0 'C for 2 hr. Antibody-bound and free 13H) estriol were separated by precipitating antibody with 0.75 ml of the polyethylene glycol solution after the addition of 1 mg (0.05 ml) of bovine y-globulin (10,ll). The tubes were immediately mixed on a vortex-type stirrer and centrifuged at 2800 rpm (1750 g) for 30 min. at 4 OC. The supernate was decanted, 10 ml of a modified Bray's solution (12), lacking ethylene glycol and methanol, was added and the radioactivity was counted in a Packard liquid scintillation counter with an efficiency of 54%. The percent recovery of estriol was calculated from triplicate 0.1 ml plasma aliquots incub ted for at least 30 min with 0.8 ml buffer and 4 0.1 ml (40,000 dpm) of ( H) estriol in buffer and extracted simultaneous ly with the plasma samples. The pe cent bound of the sample and stan5 dards was calculated from the free ( H) estriol in the supernatant as follows: percent bound = (total cpm - supernatant cpm) (lOO)/total cpm.
S
TREOTDrn
Standard curves were constructed from the percent E3H} estriol antibodybound, and plasma estriol concentrations, in pg/liter, were calculated as follows: estriol (pg/liter) = (0.1 x X x V)/(P x R x B), where X=pg from standard curve, V = total volume of buffer used for redissolving residue (ml), P = plasma volume assayed (ml), R = % recovery, and B = volume of buffer aliquot (ml). RESULTS I.
Antiserum Characteristics A.
Affinity constant: The equilibrium constant of association,
(Ka), of the antiserum was evaluated by the method of Scatchard (13). At 4 'C and under the conditions employed for our routine RIA, the antiserum had high affinity for estriol (Ka-1 x 101oM-l) and was therefore potentially capable of being used for a sensitive RIA. B.
Specificity: The cross reactivity of steroids with this anti-
serum was determined by the method of Abraham (14).
Estradiol-178
and estrone demonstrated 135% relative cross-reactivity,as compared with 100% for estriol (Table 1).
The relative cross-reactivities
obtained are typical for antisera prepared against estriol derivatives conjugated to the functional groups. II. Partial Purification of Estriol High assay specificity for unconjugated estriol is achieved by preliminary solvent extraction of interfering estrogens present in significant concentrations in plasma which cross-react with the antiserum. Only 2.5-3% of estradiol-178 and Z-2.52 of estrone are recovered in the final ether extract.
The percent of estriol recovered in this extract
is very consistent within, 86.721.1 (IS.D)% (n=lO), as well as between 86.922.3 (IS.D)% (n=ll), assays.
S
462
TPROID=
TABLE 1 Relatively Cross-Reactivity
of Sreroids with Antiserum
s 310 No. 5 Steroid
Cross-Reactivity
Estriol
100
Estradiol-178
135
Estrone
135
16-Epiestriol
66
17-Epiestriol
46
6-Oxoestradiol
16
16,17-Epiestriol
13
6-Oxoestriol
(X)
9
III. Sensitivity, Blank, Precision and Accuracy The sensitivity of the standard curve, defined as the least amount of estriol standard which could significantly estriol from the antiserum, was 8 pg.
(k2S.D.) displace j3H)
The solvent blank was invaria-
bly less than the sensitivity of the method except for a few instances when the benzene and petroleum ether had not been repurified for over 6 weeks.
Repurification
of these solvents eliminated the blank.
The
smallest amount of estriol which can be detected by our routine assay using 0.1 ml of plasma is 250 pg/ml.
Estriol could not be detected in
plasma from men or non-pregnant women. The precision of the method was studied by repetitive measurement of estriol in pooled plasma, in replicate and from day to day. intra-assay estriol mean for two plasma pools was 13.7kO.46
The
(1S.D.)
ng/liter (n=18) and 11.9kO.98 (1S.D.) ug/liter (n=24), with respective coefficients of variation of 3.4% and 8.2%. The inter-assay estriol means for two plasma pools was 5.720.35 (1S.D.) ng/liter (n=9) and 12.5kO.35 (1S.D.) ngfliter (n=6), with respective coefficients of varisation of 6.1% and 3.0%. Accuracy of the method was evaluated by several methods. First, various quantities of estriol were added to plasma and the samples were assayed. 93-108% of the added estriol was recovered (Table 2).
Second,
various plasma volumes were assayed. Linear regression analysis calculated by the least squares method resulted in a straight line with a correlation coefficient of 0.992 and y-intercept of -1.91 pg (Fig. 1). Table 2 Recovery of Estriol Added to Plasma Estriol Added (pg)
Estriol Measured (pg)fIS.D. N
C.V. (W) Recovery (X)
400
432k65
9
15
108
800
823242
9
5
103
1000
1047+23
4
2.1
105
1200
1178+115
9
9.8
98
1600
1488+81
9
5.4
93
In addition, fifty samples were assayed by this method and by a specific RIA for estriol employing a highly specific antiserum, which was prepared in a ewe against estriol-6-(0-carboxymethyl)oximederivative (15). Relative cross-reactivitywith other estrogens was low: estradiol-178, 0.8%; estrone,
