Radioimmunoassay of Serum Estrogen

Transcription

1 CLIN. CHEM. 19/7, (1973) Radioimmunoassay of Serum Estrogen Nai-Siang Jiang, Robert J. Ryan, and A. Albert Radioimmunoassay of serum estrogen (sum of estrone and estradiol-17f3) in 0.5 ml of human serum is described. After dilution of the serum, the estrogens are adsorbed on a Sephadex G-15 column in the presence of dilute HC1 and then eluted with benzene. After evaporation of the benzene eluate, the estrogens are dissolved in bovine serum albumin (5 g/. liter solution, at ph 8.0). An aliquot of this solution is used for the assay. Serum estrogen concentrations (mean ± SD) were 69 ± 23 pg/mi in 105 normal men and 199 ± 146 pg/mi in 26 fertile women. In 109 postmenopausal women, estrogen was undetectable in seven sera, but in the remaining 102 it was 52 ± 30 pg/mt of serum. Additional Keyphrases: estrone plus estradiol in serum #{149} of extensively purifying and separating them into two column chromatography on Sephadex #{149}normal values or three components. for men and women (pre- and postmenopausal) #{149} separation of free and antibody-bound hormone with use of dextran- and serum-coated charcoal #{149}urinary estrogen #{149}menstrual cyclic variations #{149}assessment of gonadal function The nature and amounts of estrogens in human body fluids have recently been reviewed (1). Estrogens in serum or plasma have been determined by gas chromatographic and double-isotope derivative methods (2-4), which require relatively large amounts of plasma or serum, are tedious to perform, and restricted in the number of samples that can be assayed in one run. Greater convenience (smaller volumes of sample, greater number of tests carried out per run, less tedium in performance) was achieved by use of radioimmunoassay (5-7), competitive binding assay (8-10), and receptor binding assay (11, 30, 31). However, preliminary extraction, purification, and separation of the estrogens are still required, so that the amounts of estrone (E1)1 and estradiol (E2) can be From the Mayo Clinic and Mayo Foundation, Departments of Laboratory Medicine and Endocrine Research, Rochester, Minn Nonstandard abbreviations used: E1, estrone; B2, estradiol-17fl; E3, estriol; TE2, [6,7-3H]estradiol-17$; DS charcoal, charcoal coated with dextran and serum; BSA, bovine serum albumin; E2- BSA, estradiol-17fl-succinyl bovine serum albumin; and Tris, tris- (hydroxymethyl)aminomethane buffer (50 mmol/liter, ph 8.0). Received Jan. 19, 1973; accepted May 1, determined separately. It is this preliminary purification and resolution of the estrogen fractions (E1 and E2) that make unmanageable the determination of estrogen in plasma or serum as a routine clinical determination. To achieve the distinctive features of a clinical assay, we approached the problem of serum estrogen determination from a different route. Under most ordinary circumstances the diagnostic value of estrogen assay is reportedly the same, whether one, two, or three of the three major estrogens are measured (13). If this indeed is the case, then meaningful clinical information can be obtained by a determination of collective plasma estrogens, thus avoiding the necessity This we did by preparing an extract of serum containing only E1 and E2, which can be done very quickly and quantitatively, and using radioimmunoassay to measure the sum of these two estrogens. Materials and Methods General procedures. All chemicals were reagent grade. Estrone, estradiol-17i9, and estriol were obtained from Sigma Chemical Co., St. Louis, Mo , and further purified by recrystallization. The purity of the final product was checked by paper chromatography. All other steroids were the best grade available from commercial sources and were used without further purification. Labeled E2 (6,7-3H estradiol-17f3 or TE2) with a specific activity of 42 Ci/ mmole was obtained from New England Nuclear, Boston, Mass The E2 antiserum was produced in our laboratory by injecting rabbits with estradiol- 17 succinyl bovine serum albumin (E2-BSA).2 The scheme for immunization was similar to that reported previously (12). Six rabbits were injected subcutaneously with 1 mg of E2-BSA per week. During the first two weeks, antigen was given in 1 ml of complete Freund s adjuvant. Subsequent injections were given in 1 ml of NaCl solution (9 g/liter). The animals 2 The antigen E2-BSA was a gift from the National Cancer Institute. 740 CLINICAL CHEMISTRY, Vol. 19, No.7,1973

2 were bled six weeks after the first injection and every two weeks thereafter. The serum was separated from the clot, sodium azide added in a final concentration of 1 g/liter, and stored in the deep freeze until used. Purification of solvents. Benzene was purified by repeated washing with tenth volumes of concentrated H2S04 until the acid layer was colorless. The organic layer was then washed with a tenth volume of 0.1 molar NaOH and with several portions of water until the aqueous phase was ph 5.5, then dehydrated over anhydrous Na2SO4 and distilled twice. Ethyl acetate was purified by distillation over CaO and stored at - 20#{176}C until used. Preparation of charcoal coated with dextran and serum (DS charcoal). Two hundred milliliters of a 5 g/liter suspension of Norit A (Fisher Scientific Co.) and 200 ml of Dextran T70 (0.5 g/liter) were mixed and centrifuged, and the supernatant fluid was discarded. The precipitate was resuspended in 160 ml of a mixture of pooled human serum (35 mi/liter) and Tris buffer (0.05 mol/liter, ph 8.0) and stirred slowly (magnetic stirrer) in the cold room overnight. The mixture was centrifuged and the supernatant fluid was discarded. The sediment was resuspended in 20 ml of cold 0.05 molar Tris buffer, ph 8.0. Tritium counting. Radioactivity was determined by mixing 1 ml of aqueous sample with 15 ml of Bray s solution (15) in a glass liquid scintillation counting vial, which was counted in a Packard 3375 liquid scintillation spectrometer. Counting efficiency was 23%. Each sample was counted for 5 mm; the counting error was less than 2.5%. Separation of bound and free estrogens. E2-antiserum (0,5 ml of a 500-fold dilution) was incubated with 40,000 cpm of TE2 in a final volume of 5 ml of Tris buffer (0.05 mol/liter, ph 8.0). Simultaneously, a control tube was prepared under identical conditions except 0.5 ml of normal rabbit serum (500-fold dilution) was used instead of E2 antiserum. The mixtures were incubated at room temperature overnight. After incubation, the following experiments were performed on the mixtures: a. Sephadex G-75 column chromatography. One milliliter of each of the incubation mixtures was added to the top of a Sephadex G-75 column (1 X 7.5 cm). After the 1 ml solution percolated through, the column was eluted with 0.05 molar Tris buffer, ph 8.0. The column was operated at 4#{176}C. Thirty-six fractions (1 ml each) were collected, and radioactivity was counted. b. Treatment with Dowex-l-Cl resin. One milliliter of the incubation mixture was treated with 20 mg of Dowex-1-Cl (Ag 1-x2, Bio-Rad, mesh) for #{149} 30 mm in an ice-water bath, with occasional mixing, then centrifuged at 1,000.X g at 4#{176}C for 5 mm and the supernatant solution was subjected to Sephadex G- 75 column chromatography as described above. The radioactivity of the fractions was measured. c. Treatment with DS charcoal. One milliliter of the incubation mixture was added to 0.25 ml of DS charcoal (containing 12.5 mg of Norit A) and mixed with a vortex-type mixer, incubated in an ice-water bath for 30 mm, and centrifuged at 1,000 x g at 4#{176}C. The supernatant solution was chromatographed on Sephadex G-75 column and the radioactivity of the fractions was measured as described above. Second-stage antibody precipitation (14). To 1 ml of the incubation mixture was added 0.1 ml of normal rabbit serum-a threefold dilution in 0.05 molar Tris buffer, ph 8.0-and 0.1 ml of goat antiserum against normal rabbit serum. The mixture was vortex-mixed and incubated at 4#{176}C overnight. At the end of the incubation the mixture was centrifuged at 10,000 x g at 4#{176}C for 20 mm. The supernatant solution was decanted, the precipitate was redissolved in 1 ml of the Tris buffer, and the radioactivity was measured. Specificity study. We studied the specificity of the E2-antiserum by incubating, in duplicate tubes, 0.4 ml of E2-antiserum, 210-fold dilution; 10,000 cpm TE2; and various amounts of the different steroids tested in a total assay volume of 1.2 ml, using the Tris buffer to adjust all tubes to the same volume. Two tubes without steroid served as 100% controls and two tubes without E2-antiserum served as blank controls. The tubes were incubated at room temperature overnight; each tube was then treated with 0.25 ml of DS charcoal, as described above, to remove free TE2. The radioactivity of 1 ml of the supernatant solution was measured. The average counts per minute of the duplicates, with blank subtracted, was expressed as a percentage of that of the 100% tube. A displacement curve was constructed for each steroid, with logarithm of the steroid added (in picograms) as abscissa and the percent TE2 bound as ordinate. Purification of E1 and E2 from human serum. The procedure developed by Van Baelen et al. (16) for purification of estrogens from urine was modified for processing serum samples. Sephadex G-15, 800 mg, was packed in a column (1 cm i.d.) having a porous glass plate at the bottom. The whole column was immersed in water overnight, or longer, in a 1.5 x 15 cm test tube. The column was washed with 10 ml of water immediately before use. The serum to be assayed was diluted eightfold by mixing 1.1 ml of serum, 4.7 ml of water, and 3.0 ml of 0.16 molar HC1. Four milliliters of the diluted serum was transferred to the top of the Sephadex G-15 column. This procedure was done in duplicate for each sample. For column blanks, two columns, to each of which 4 ml of dilute acid was applied instead of the serum sample, were developed for each assay. After the 4-ml sample had percolated through, the column was washed with 7.5 ml of water to remove serum proteins and other interfering substances, and residual water on the column was removed by briefly applied suction. The estrogens, still retained on the column, were eluted with 10 ml of either ethyl acetate or benzene, which was forced through the column by compressed air. In the final procedure, benzene was selected for the elution because it eluted only E1 and E2. The organic solvent CLINICALCHEMISTRY, Vol. 19, No.7,

3 eluate was evaporated in a 65#{176}C water bath under a gentle stream of nitrogen. The residue was redissolved in 1 ml of BSA solution (5 g/liter of Tris buffer) by shaking on a rotary shaker for 40 mm. The resulting solution was used for assay. Radioimmunoassay of estrogen. The assay mixture contained 2,500 cpm of TE2 (33 pg), standard E2, or sample extract, and 0.1 ml of 210-fold diluted E2- antiserum and sufficient Tris buffer to make a total volume of 1.2 ml. Duplicate standard tubes contained 0, 10, 20, 40, 80, or 160 pg of E2, which was transferred to the tubes in a volume of 0.7 ml of the BSA-Tris solution. Two tubes without antiserum served as blank. The amount of benzene extract used in the assay depended on the expected E1 and E2 concentration in the sample. When the expected concentration was less than 400 pg/ml of serum, 0.7 ml of the final extract was used for the assay. When the concentration was expected to exceed 400 pg/ml serum, 0.2 ml of the extract was used. In the latter case, 0.5 ml of BSA-Tris solution was added to adjust the final incubation volume to 1.2 ml. The incubation mixture was left overnight (15 h) at room temperature. The next morning, the tubes were chilled in an ice-bath for 10 mm; then 0.25 ml of DS charcoal (containing 12.5 mg of charcoal) was added to each tube and vortex-mixed. The mixture was set in an ice-water bath for an additional 30 mm with occasional mixing. After centrifugation at 4#{176}C and 1,000 X g for 5 mm, the supernatant fluid was poured into a liquid scintillation counting vial without disturbing the precipitate. A standard displacement curve was constructed as described above and, from the percent bound TE2 found in the tubes containing unknowns, the quantity of E1 and E2 could be read from the standard curve. The final calculation was simple: 1 ml of extract was the equivalent of 0.5 ml of original serum, so 0.7 ml of extract represents 0.35 ml of serum, and 1 ml of serum = 1/.35 or 2.86 volumes of the volume assayed. Thus, the picograms of estrogen corresponding to the percent bound counts x 2.86 = pg estrogen/ml serum. When only 0.2 ml of extract was used, the calculation factor was 10. Other methods. Serum luteinizing hormone was measured by the radioimmunoassay of Faiman and Ryan (17), with LER-9073 as standard. Total urinary estrogen was determined fluorometrically (18). The method of Mikhail et al. (7) was used to measure E1 and E2 separately. Results Separation of bound and free estrogen. Figure 1 shows the elution patterns of free and antibody-bound TE2 from a Sephadex G-75 column. In the presence of rabbit E2-antiserum, an appreciable amount of the labeled hormone emerged at the void volume (fractions 6 through 9), indicating that E2 was bound LER-907 is a standard human pituitary LH preparation that is distributed by the Endocrine Study Section of NIH. to the antibody. The free hormone was eluted as a second peak at fractions 17 through 27. In the presence of normal rabbit serum at the same dilution as the E2-antiserum, most of the TE2 was eluted from the column as free hormone and only an insignificant amount was eluted at the void volume. When the incubation mixture was first treated with either Dowex-1-Cl or DS charcoal and then chromatographed, the peak representing bound hormone was still present. The peak representing free hormone was completely eliminated by DS charcoal treatment but was only partially removed by the Dowex-1-Cl treat- #{149} ment. Control experiments showed that, in the absence of antiserum, the free hormone remaining after Dowex-1-Cl treatment was 9% of the total counts added, whereas after DS charcoal treatment it was only 2%. Recovery of bound hormone by the doubleantibody technique was 94% of that obtained by the DS charcoal method. If the dextran-coated charcoal was not pretreated with serum, both free and bound hormone were adsorbed by the charcoal. Thus, all methods explored for the separation of free and bound hormone gave generally similar results. The DS charcoal method was chosen for the separation of free and bound estrogen because of its simplicity. Specificity of the antiserum. In addition to the compounds shown in Figure 2, cholesterol, ethinyl estradiol-17f., diethylstilbestrol, 5a-androstan- 17- ol-3-one, z5-androstene-3,17-dione, Ei-3-sulphate, E2-3-sulphate,4 and E2-3,17-disulphate4 were also tested. All the steroid hormones tested showed some interference with a sample size of 10 pg. Cholesterol, diethylstilbestrol, and BSA did not show cross reaction at 106 pg, the largest amount tested. Ethinyl estradiol-173 interfered with the assay at 300 pg. E1-3-sulfate, E2-3-sulfate, and E2-3,17-disulfate began to show cross-reaction at 10 pg. Estrone and estradiol- 17i3 gave nearly identical displacement curves, whereas E3 was only a tenth as effective as E2. The original antigen, E2-BSA, displaced TE2 very effectively. Sensitivity of the antiserum. When 2,500 cpm of TE2 and 0.1 ml of 210-fold diluted antiserum were used in the incubation mixture, the least amount of E1 and E2 detectable was 10 pg. Figure 3 illustrates 12 standard curves obtained during a period of three months. Purification of estrogens from serum. Preliminary attempts to measure estrogens in whole serum were unsuccessful, because of the presence of a substance (or substances) in serum interfering with the binding of TE2 to the antibody. Various purification procedures were explored, and the method we found most satisfactory was the Sephadex G-15 column procedure. Recovery was best obtained when serum was diluted eightfold with water in the presence of dilute E2-3-sulfate and E2-3,17-disulfate were kindly supplied by Dr. Hannah E. Rosenthal, Roswell Park Memorial Institute, Buffalo, N.Y. 742 CLINICALCHEMISTRY, Vol. 19, No.7, 1973

4 Fig. 1. Elution pattern of antibody-bound and free [6,7-3H}estradioI-17 from Sephadex G-75 column Left, #{149} E2-antiserum. 1:500; X normal rabbit serum, 1:500. Right, 0, treated with DS charcoal; #{149} treated with 20 mg of Dowex-1-Cl in ice bath for 30 mm TUBE NUMBER TUBE NUMBER Fig. 2. Specificity of E2-antiserum 1, BSA; 2, cortisol; 3, deoxycorticosterone; 4, progesterone; 5, testosterone; 6, 5-androstene-3a- 1 7/3-diol PICOURAMS OF UNLABELED STEROIDS Table 1. Recovery Experimentsa Amount Amount recovered, pg added, Estrogen pg Mean % Estrone Estrone Estradiol Estradiol Estradiol a Estrogen was added to human male serum and the mixture analyzed for estrogen (2-3 runs per experiment). HC1. Experiments with tritium-labeled E1, E2, and E3 showed that benzene eluted E1 and E2 almost quantitatively from the column, but only 2% of E3, while ethyl acetate eluted the three estrogens equally effectively. Because benzene was used in the final procedure and the E2 antiserum cannot differentiate E1 and E2 quantitatively, the final result obtained was the sum of E1 and E2 concentration of the sample. When E1 and E2 were added to male serum and benzene was used to elute the column, recovery was good (Table 1). Table 2. Comparison of Results Obtained for Eight Sera by Method of Mikhail et al. (7) and Present Method E1 Method of Mikhail et al. E2 pg/mi Sum of Present method EtandE2 E1+E Estrogen concentration in human serum. Table 2 compares the results obtained by the method of Mikhail et al. (7), in which E1 and E2 are separated by Sephadex LH-20 column chromatography and measured individually by radioimmunoassay, and by the present method, which measures E1 and E2 simultaneously. Evidently, the values obtained by the CLINICAL CHEMISTRY, Vol. 19, No.7,

5 foe z 0 I- z w U w a O L VV V?.#{149}...! : % 2 1 T Iii 300 e 280 -I 4,,( z fbi C, ISO sc I.. u 60 fbi d NORMAL HYPO NORMAL HYPO PM 40 _ MALES FEMALES CHILDREN Fig 4. Serum estrogen concentration of normal and hypogondal men and women, postmenopausal women, and children #{149} #{149} #{149} Line at 29 pg/mi represents extinction points of assay Fig. 3. Mean for estradiol PICOORAMS ESTRADIOL and individual data for 12 standard curves present method are the same as the sum of E1 and E2 values obtained by the method of Mikhail et al., within experimental error. Serum estrogen concentrations for normal men, fertile women, and postmenopausal women are given in Table 3. Figure 4 shows the histogram of serum estrogen concentrations in normal subjects, postmenopausal women, hypogonadal patients, and children. In most of the hypogonadal patients and all of the children examined, estrogen was undetectable in serum, whereas none of the normal subjects examined had serum estrogen values of less than 30 pg/ml. Figure 5 illustrates the patterns of urinary total estrogen, serum estrogen (measured by the present method), and serum luteinizing hormone during a normal menstrual cycle. The pattern of serum estrogen parallels reported patterns of estrogen in both blood and urine during the menstrual cycle. Table 4 shows the inter- and intra-assay precision of the present method. The radioimmunoassay of serum estrogen can be outlined as follows: 1. Apply 4 ml of dilute serum (1.1 ml of serum ml of water ml of 0.16 mol/liter HC1) to each of two Sephadex G-15 columns. Wash the columns with 7.5 ml of water, elute with 10 ml of benzene, and flush with compressed air. Evaporate the eluate at 65#{176}C under nitrogen. Dissolve the residue in 1 ml of albumin-containing (5 g/liter) buffer, ph Use 0.7 or 0.2 ml of the extract in radioimmunoassay, with concomitant standard solutions of E2, 100% tube, and blanks, in a total reaction volume of 1.2 ml of buffer. Incubate for 15 h (overnight) at room temperature. Mix with 0.25 ml of DS charcoal, centrifuge, pour the supernatant fluid into a counting vial, and measure radioactivity. 3. From the percent counts bound of unknown, read from the standard curve the corresponding value in picograms of estrogen. If 0.7 ml of extract was used: Table 3. Estrogen Concentration in Serum of Normal Men and Fertile and Postmenopausal Women Sublects n Mean ± SD Actual range MedIan 95% range pg/mi Men ± Fertile women ± Postmenopausal women ± 300 O-22O a Seven samples that had undetectable amounts were omitted from calculation. 4Zero Is used to denote any value that is undetectable in the assay (i.e.. <30 pg/mi serum). 744 CLINICAL CHEMISTRY, Vol. 19, No. 7, 1973

6 Table 4. Inter- and Intra-Assay Precision Between assays Within assays Statistic Male pool Female pool Male pool Female pool n * ± SD 68 ± ± ± ± 8.1 CV, % F.R.a a FR., Fischer s ratio. DAYS Fig. 5. Urinary total estrogen, serum estrogen, and serum Iuteinizing hormone during normal menstrual cycle; LH is expressed as nanograms of LER-907/ml pg estrogen x 2.86 = pg estrogen/mi serum. If 0.2 ml of extract was used: pg estrogen X 10 = pg estrogen/ ml serum. Discussion Some additional information obtained during the development of the present method will be given here in order that others who may wish to introduce some modifications will benefit from our failures. Although the assay is generally simple, meticulous attention to technical details is required if results are to be consistent and reproducible. The maximum sample volume (after dilution) for the Sephadex G-15 column is 4 ml; use of larger volumes results in poor recovery. For good recovery, the serum sample must be diluted eightfold; hence, the maximum serum volume for extraction is 0.5 ml. The benzene must be evaporated at or below 65#{176}C; use of higher temperatures results in poor and variable recovery (50 to 70%). Benzene must be completely removed, ensured by keeping the container at 65#{176}C for 90 mm under a constant, gentle stream of nitrogen. Even traces of benzene interfere with the binding of estrogen to the antibody. Use of the BSA solution facilitates solution of the dried estrogen residue. All the radioimmunoassays and competitive binding assays for estrogens described in the literature have been applied after chemical separation of E1 and E2. We agree with the general rule that it is best to assay a single substance whose chemical identity is known. However, lengthy extraction and chromatographic procedures are required to separate the two estrogens. Because recoveries are generally low in these procedures, relatively large quantities of tntium-labeled hormones have to be added to the sample to correct for losses, and the extracts obtained generally give relatively high blank values. The complexity of the separation methods and the resulting difficulties imposed by separation constitute grounds for not following the general rule when the assay is performed for routine clinical purposes. In the present method, the sum of E1 and E2 is measured, because the antiserum does not differentiate significantly between them and there are the advantages that the purification steps are considerably simplified and no correction factor is needed because the recovery is consistently excellent. One technician can easily use 46 columns (23 duplicate samples) and complete the assays within one and one-half working days, exclusive of counting time. The blank value also is low in the present method. The mean ± SD for 42 column blanks (data collected over a period of several months) was 96.8 ± 4.3% of control tubes. Table 5 compares mean normal values as determined by the present method with the values reported by previous investigators using other methods. For normal men, our value agrees well with that reported Method Table 5. Compari son of Various Reported Mean Normal Values (pg/mi) Men Fertile women Postmenop ausal women Age n E1 E2 E1 + E2 Age n ET E2 E1 + E2 Age n E1 E2 Double-isotope deny. (4) Receptor binding (30) Receptorbinding (11, 31) Present radioimmunoassay E1 + E2 CLINICAL CHEMISTRY, Vol. 19, No.7,

7 by Nagai and Longcope (30) and Korenman et al. (11, 31), but is somewhat lower than that reported by Baird and Guevara (4). For postmenopausal women, our value agrees well with that of Korenman et al. (11, 31) but is higher than that reported by Nagai and Longcope (30). The difference nay be attributed to the fact that the subjects studied by Nagai and Longcope were considerably older. Although urinary total estrogen (E1, E2, and E3) determination (18, 19) has been used as an index of gonadal function for some time, the present paper represents the first attempt to apply the same principle to serum. We believe that the combined E1 and E2 value can be used as a valid index of gonadal function because these two estrogens are interconvertible (20-22) in the body, both have biological activity, and are both secreted by the gonads (23-25). The cyiticism could be raised that the combined serum E1 and E2 value may not be a valid index of gonadal function since a considerable amount of E1 in the blood derives from extra-gonadal sources (26-28). However, the extra-gonadal contribution of E1 is relatively constant. The major source of extra-gonadal E1 is the peripheral conversion of M-androstenedione to E1. MacDonald and his associates (27) have shown that the fraction of radioactivity appearing in the urinary metabolites of estrogens after administration of radioactive i.4-androstenedione is similar in normal and castrated individuals. Longcope (29) found that there is little if any difference in the rate at which androstenedione is converted to E1 by fertile and postmenopausal women. Longcope et al. (24) have also shown that the amount of androstenedione converted to E1 in men and women varied from zg/day. Therefore, a relatively constant concentration of estrogen represents the extra-gonadal contribution of E1, whereas a fluctuating proportion represents the gonadal contribution. The former probably is below detectability by the present assay, because most hypogonadal individuals and some postmenopausal women showed undetectable amounts (see Figure 4). All castrated individuals that we have examined to date yield undetectable values. The part that is measurable in the present assay probably represents estrogens secreted by the gonads. We believe that, owing to its simplicity and speed, the present method will be useful as a routine diagnostic procedure because the estrogen values are in general accord with the clinical condition of the individuals, and second, the serum estrogen values in one complete menstrual cycle showed a typical biphasic pattern related to the surge in luteinizing hormone, a pattern rather fully documented by a variety of chemical methods for assay of estrogen. Further evidence that the combined measurement of E1 and E2 serves as an index to estrogen production by the gonad can be derived from Table 5. Disregarding the variability of estrogen determination by various authors using different methods in very small populations, a simple mean for E1, E2, and E1 + E2 Table 6. Comparison of Mean Values for Serum Estrogena Postmenopausal Fertile women Men women Estrogens, pg/mi of serum E E E1 +E a Derived from Table 5. in Table 5 was made for men, fertile women, and postmenopausal women. These three indices of estrogen production were arranged in Table 6 in ascending order from left to right. E1 is about the same for postmenopausal women and men and about twice this amount in fertile women. E2 is lowest in postmenopausal women, twice this value in men, and 10-fold this postmenopausal concentration in fertile women. The sum of E1 and E2 is about 42% greater in men than in postmenopausal women, and four times greater in fertile women. Thus, while E2 measurements may show a more active response as a function of gonadal activity, it is true, as earlier work on urinary estrogens showed, that the sum of E7 and E2 likewise reflect gonadal activity. Skillful technical assistance was provided by Mrs. Marie E. Bezdicek. This investigation was supported in part by Research Grants AM-01738, HD-03726, and FR-5530, and Contract from the MH, USPHS. References 1. Paulsen, C. A., Ed., Estrogen Assays in Clinical Medicine, University of Washington Press, Seattle, Wash., 1965, 396 pp. 2. Wotiz, H. H., Charransol, G., and Smith, I. N., Gas chromatographic measurement of plasma estrogens using an electron capture detector. Steroids 10, 127 (1967). 3. Svendsen, R., and Sorensen, B., The plasma concentration of unconjugated oestrone and 17-oestradione durink the normal menstrual cycle. Acta Endocrinol. (Copenhagen) 47, 245(1964). 4. Baird, D. T., and Guevara, A., Concentration of unconjugated estrone and estradiol in peripheral plasma in nonpregnant women throughout the menstrua, cycle, castrate and postmenopausal women and in man. J. Clin. Endocrinol. Metab. 29, 149(1969). 5. Abraham, G. E., Solid-phase radioimmunoassay of estradiol- 17fl. J. Clin. Endocrinol. Metab. 29,866(1969). 6. Midgley, A. R., Jr., Niswender, G. D., and Sri Ram, J., Hapten-radioimmunoassay: A general procedure for the estimation of steroids! and other haptenic substances. Steroids 13, 731 (1969). 7. Mikhail, G., Wu, C. H., Ferin, M., and VandeWiele, R. L., Radioimmunoassay of plasma estrone and estradiol. Steroids 15, 33(1970). 8. Corker, C. S., and Exley, D., The determination of plasma estradiol-17$ by competitive protein binding radioassay. Steroids 15, 469(1970). 9. Mayes, D., and Nugent, C. A., Plasma estradiol determined with a competitive protein binding method. Steroids 15, 389 (1970). 10. Dufau, M. L., Dulmanis, A., Catt, K. J., and Hudson, B., Measurement of plasma estradiol-17 by competitive binding assay employing pregnancy plasma. J. Clin. Endocrinol. Metab. 30, 351 (1970). 11. Korenman, S. G., Perrin, L. E., and McCallum, T. P., A radioligand binding assay system for estradiol measurement in human plasma. J. Clin. Endocri no!. Metab. 29,879(1969). 746 CLINICAL CHEMISTRY, Vol. 19, No.7,1973

8 12. Jiang, N. S., and Ryan, R. J., Radioimmunoassay for estrogens: A preliminary communication. Mayo Clin. Proc. 44,461 (1969). 13. Albert, A., Significance of clinical estrogen assays. In Estrogen Assays in Clinical Medicine, C. A. Paulson, Ed. University of Washington Press, Seattle, Wash., 1965, pp Feinberg, R., Detection of non-precipitating antibodies coexisting with precipitating antibodies using P3 labeled antigen (abstract). Fed. Proc. 13, 493 (1954). 15. Bray, G. A., A simple efficient liquid scintillator for counting aqueous solutions in a liquid scintillation counter. Anal. Biochem. 1,279(1960). 16. Van Baelen, H., Heyns, W., and De Moor, P., Measurement of urinary estrogens using adsorption on Sephadex. J. Clin. Endocrino!. Metab. 27, 1056(1967). 17. Faiman, C., and Ryan, R. J., Radioimmunoassay for human luteinizing hormone. Proc. Soc. Exp. Biol. Med. 125, 1130(1967). 18. Hobkirk, R., and Metcalfe-Gibson, A., Urinary estrogens. Stand. Methods Clin. Chem. 4, 65(1963). 19. Ittrich, G., Untersuchungen uber die Extraktion des voten Kober-Farsbstoffes durch organische Losungsmitted zur Oestrogenbestimmung im Ham. Acta Endocrinol. (Copenhagen) 35, 34 (1960). 20. Beer, C. T., and Gallagher, T. F., Excretion of estrogen metabolites by humans. I. The fate of small doses of estrone and estradiol- 17fl.J. Biol. Chem. 214,335(1955). 21. Fishman, J., Bradlow, H. L., and Gallagher, T. F., Oxidative metabolism of estradiol. J. Biol. Chem. 235, 3104 (1960). 22. Longcope, C., Layne, D. S., and Tait, J. F., Metabolic clearance rates and interconversions of estrone and 17ft-estradiol in normal males and females. J. Clin. Invest. 47, 93 (1968). 23. Lloyd, C. W., Lobotsky, J., Baird, D. T., McCracken, J. A., Weisz, J., Pupkin, M., Zanartu, J. and Puga, J., Concentration of unconjugated estrogens, androgens and estrogens in ovarian and peripheral venous plasma of women: The normal menstrual cycle. J. Clin. Endocri no!. Metab. 32, 155 (1971): 24. Longcope, C., Kato, T., and Horton, R., Conversion of blood androgens to estrogens in normal adult men and women. J. Clin. Invest. 48, 2191 (1969). 25. Kelch, R. P., Jenner, M. R., Weinstein, R., Kaplan, S. L., and Grumbach, M. M., Estradiol and testosterone secretion by human, simian and canine testes, in male with hypogonadism and in male pseudoherniaphrodites with the ferwinizing testes syndrome. J. Clin. Invest. 51, 824 (1972). 26. Baird, D. T., Horton, R., Longcope, C., and Tajt, F., Steroid prehormones. Perspect. Med. Biol. 11, 384(1968). 27. MacDonald, P. C., Rombaut, R. P., and Siiteri, P. K., Plasma precursors of estrogens. I. Extent of conversion of plasma i- androstenedione to estrone in normal males and nonpregnapt normal, castrate and adrenalectomized females. J. Clin. Endocrinol. Metab. 27, 1103 (1967). 28. Longcope, C., Kato, J., and Horton, R., The contribution of androgens to the blood production rates of estrogens. Program of the ifi mt. Congr. Endocrinol., Mexico City; Excerpta Med., Abstr. 452(1968). 29. Longcope, C., Metabolic clearance and blood production rates of estrogens in postmenopausal women. Amer. J. Obstet. Gynecol. 111,778(1971). 30. Nagai, N., and Longcope, C., Estradiol-17 and estrone: Studies on their binding to rabbit uterine cytosol and their concentration in plasma. Steroids 17,631 (1971). 31. Tulchinsky, D., and Korenman, S. G., A radio-ligand assay for plasma estrone; normal values and variations during the menstrual cycle. J. Clin. Endocrinol. Metab. 31,76(1970). CLINICALCHEMISTRY, Vol. 19, No. 7,