J. Endocrinol. Invest. 4: 275, 1981

Sandwich-type enzyme immunoassay for human chorionic gonadotropin

T. Sekiya*, Y. Furuhashi*, S. Goto*, S. Kaseki*, Y. Tomoda*, and K. Kato** *Department of Obstetrics and Gynecology, Nagoya University, School of Medicine, Showa Ku, Nagoya 466, Japan and **Department of Biochemistry, Institute for Developmental Research, Aichi Prefectural Colony, Kasugai, Japan

ABSTRACT. We have developed a sandwich-type enzyme immunoassay for human chorionic gonadotropin (hCG), in which antibody Fab'- fJ -D-galactosidase complex and an antibody­immobilized silicone rubber solid phase were used. Despite the fact that this assay system cross­reacted about 40% with human luteinizing hormone (LH) that contains an immunologically very similar subunit to that of hCG, it proved'to be highly sensitive with hCG measurable at levels as low as 0.3 mlU per assay tube. Using 25 tJl of serum sample or 100 tJl of urine sample, hCG levels in serum (10-1000 mlU/ml) or in urine (3-300 mlU/ml) could be determined with the same degree of precision as in radioimmunoassay without sample interference with the assay. The coefficients of variation in within-run, and between-run were 9.2-13.3%, and 4:2-18.8%, respectively. Values obtained with enzyme immunoassay correlate well with those of radioimmunoassay (r = 0.961, slope = 1.129, y-intercept = 3.7 mlU/ml for 35 serum samples) and hemagglutination assay (r = 0.954, slope = 0.951, y-intercept = 1.8 mlU/ml for 88 urine samples).

INTRODUCTION Human chorionic gonadotropin (hCG) is a glycopro­tein hormone which is increased in serum and urine of the pregnant women and patients with tropho­blastic disease. Radioimmunoassay and hemagglutination assay methods are most frequently used for the determina­tion of hCG in the current clinical medicine (1, 2). The radioimmunoassay method, however, has several disadvantages resulting from the usage of ra.dioactive isotope and the hemagglutination assay method is less sensitive than the radioimmunoassay and cannot be applied for the assay of serum hCG levels. Recently, enzyme immunoassay of hCG has been reported as another non-isotopic method (3-5). The competitive enzyme immunoassay for hCG reported by Kikutani et al. (6) was as sensitive as the radioim­munoassay, but the assay has encountered the problem of sample interferences, limiting the applica­tion of that method for clinical chemistry. In this report, we present a sandwich-type enzyme immunoassay for hCG, in which the sample interfer­ences with the assay were improved by separating

Key-words: Human chorionic gonadotropin (hCG). enzyme immunoassay. trophoblastic disease. immunoassay. tl-D-galactosidase.

Correspondence: Dr. Tatsuhiro Sekiya. Department of Obstetrics and Gyne­cology. Nagoya University. School of Medicine. Showa-ku. Nagoya 466. Japan.

Received June 6. 1980; accepted January 12. 1981.

275

the immunoreaction with biological samples from that with the enzyme-labelled antibody.

MATERIALS AND METHODS

Anti-hCG antibody

Purified hCG (12,000 IU/mg) was prepared by the method of Van Hell et al. (7). One mg of hCG in 0.5 ml of physiological saline with equal volume of Freund's complet~ adjuvant was injected intrader­mally to New Zealand white rabbit, with a similar booster injection done on the 30th day. A week later, antisera were bled from the rabbit. The antisera were treated with child -urinary proteins and serum proteins to adsorb unwanted antibodies before use. Immu­noglobulin G (lgG) fractions were prepared from the antisera by (NH4):?804 (50% saturation) precipitation and chromatography on DEAE-cellulose. F (ab')2 fragments of the IgG fractions were prepared by digesting the IgG fractions with pepsin (8).

Silicone rubber-immobilized anti-hCG

The F(ab')2 fragments of the IgG fractions of anti­hCG serum were immobilized non-covalently on silicone rubber pieces (3 mm in diameter and 4 mm in length, from Sanko Plastic Co, Osaka) (9) and stored in buffer A (0.01 M sodium phosphate buffer pH 7.0 containing 0.1 M NaCI, 1 mM MgCI2, 0.1 % bovine 'serum albumin from Armour Pharmaceutical Co, Chicago, and 0.1 % NaN3 ) at 4 C for at least one week pefore use.

T. Sekiya, Y. Furuhashi, S. Goto, et al.

(Anti-hCG) Fab'-f3-D-galactosidase complex

The F(ab')2 fragments prepared from the antiserum, as described above, were reduced with 2-mercap­toethylamine and coupled to/3-D-galactosidase from Escherichia coli (Boehringer Mannheim, Mannheim) by use of N, N'-o-phenylenedirilaleimide (Aldrich Chemical Co., Milwaukee) as reported previously (8). Amounts of the complexes were expressed as units of f3-D-galactosidase activity, and one unit of the activity is defined as that which hydrolyzes 1 JJmol of substrate/min under the assay condition used (9).

Reagents for immunoassay

Buffer G: 0.Q1 M sodium phosphate buffer, pH 7.0 containing 0.3 M NaCI, 1 mM MgCI2, 0.1 % bovine serum albumin, 0.5% gelatin (Difco Laboratories, Detroit) and 0.1 % NaN3, which was used for the assay of serum samples. Buffer U: 0.06 M sodium phosphate buffer, pH 7.0 containing 1 mM MgCI2, 0.1% bovine serum albumin, 0.5% gelatin and 0.1 % NaN3, which was used for the urine samples. Substrate solution for enzyme assay: 0.3 mM 4-methylumbelliferyl-/3-D-galactoside (Sigma Chem­ical Co., SI. Louis). Enzyme-reaction stopping solution: 0.1 M glycine­NaOH buffer pH 10.3. Standard hCG and hLH: The Second International Standard hCG and the Second International Refer­ence Preparation of hMG, which were kindly supplied from the National Institute for Medical Research (Mill Hill, London), were used as Standards (10).

Assay procedure

A piece of the silicone rubber-immobilized antibodies was incubated in duplicate with various amounts of standard hCG, or serum sample or urine sample in 0.5 ml final volume with buffer G or buffer U. After incubation at 30 C for 3 h with shaking, the reaction medium was discared by aspiration and each piece was washed twice with 1 ml of buffer A. The piece was then incubated at 4 C overnight with 3 mU of the (anti-hCG) Fab'-/3-D-galactosidase complex in 0.2 III I of buffer A. After the reaction, each piece was washed twice with buffer A and transferred to another test tube which contained 0.1 ml of buffer A. The piece was preincubated for 5 min at 30 C and than the enzyme reaction was started by adding 50 pi of substrate solution to each tube. After incubation at 30 C for 20 min with shaking, 2.5 ml of the stopping solution was added to each tube. Fluorescence intensity of 4-methylumbelliferone produce.d by the enzyme reaction was measured by. a spectrofluo­'rometer (MPF-3, Hitachi, Tokyo) at 360 nm for exitation and 450 nm for emission analysis.

Serum and urine samples

Most of the serum and urine samples used for assay

276

were freshly frozen samples obtained from the pa­tients with trophoblastic disease. Some samples were those from the pregnant women.

Radioimmunoassay of hCG

Double-antiboJy radioimmunoassay of hCG was performed with the HCG+ 125 kit obtained from CIS (CEA-IRE-SORIN), France.

Hemagglutination assay of hCG

It was performed with HIGONAVIS kit from Mochida Pharmaceutical Co., Tokyo.

RESULTS A standard curve of the sandwich-type enzyme immunoassay for hCG is shown in Figure 1. The assay was highly sensitive with measurable levels of hCG as small as 0.3 mlU/assay tube.

200

100 80

~ 60 ., _---a

c: .! 40 .: Q) 0 c:

20 Q) 0 ., ~ 0 :::l 10 iL

8 6

Fig. 1 - Standard curve for sandwich-type enzyme immu­noassay of hCG and cross-reaction with hLH. Fluorescence intensity, 100 = 0.1 JJmol of 4-methylumbellif­erone.

The assay system cross-reacted with hLH as ex­pected, since the antiserum prepared by injecting the whole hCG molecules cross-react with hLH. Simi­larly, the assay system cross-reacted with the a subunit and /3 subunit of hCG about 5% on the basis Of mole equivalent (data not shown). Specific immuno-, reactions at 2 sites on the antigens (with the antibody-solid phase, and with the antibody-enzyme complex) in the s~ndwich immunoassay may result in a relatively low cross-reaction of the assay with the a and /3 subunits. The effects of serum and urine samples on the immunoassay system were examined by adding various volumes of each sample, and the values of hCG estimated from the standard curve were plotted as a function of sample volumes. As shown in Figure 2, a linear increase in the amounts of hCG was

Fig. 2 - Influence of sample volumes on the enzyme immunoassay of hCG. Indicated volumes of 3 serum samples (A) or 3 urine samples (B), were subjected to the immu­noassay as described in the text.

30

~ .a 20 m Ul os ......

::::> E C!) 10 () .t:

o

3.16212.5 25

Serum sample added (J.LI)

Enzyme immunoassay of hCG

30

Q) .0

.a 20 >-os Ul Ul os ......

::::> E C!) () .t:

50 ~~~--~------~ 50 100

Urine sample added (f..II)

observed up to 25 J.l1 with each serum sample, and up to 100 J.l1 with the urine sample. Similar results were obtained with the other samples (5-8 samples) tested. These results suggested that the maximal sample volume applicable on the assay system were 25 J.l1 for serum and 100 J.l1 for urine, and indicated that the measurable levels of hCG ranged from 10 to 1000 mlU/ml in serum and 3-300 mlU/ml in urine or in duplicate in 6 consecutive assays (between-run). As shown in Tables.1 and 2, the coefficients of variation in the assay of serum and urine were less than 19%. To evaluate the accuracy of the present method, we com­pared the vlues of 35 serum samples obtained by the enzyme immunoassay, with those obtained by a double-antibody radioimmunoassay, or the values of 88 urine samples estimated by the enzyme immu­noassay with those estimated by a hemagglutination assay. As shown in Figure 3, there are good correla­tions between the enzyme immunoassay and the radioimmunoassay of serum sample with the regres­sion equation y = 1 .13x + 3.7, and the correlation coefficient r = 0.96 (Fig. 3A), and between the en­zyme immunoassay and the hemagglutination assay of urine sample (Fig. 38, Y = 0.95x + 1.8, r = 0.95).

Table 1 - Precision of hCG assay in serum.

Fig. 3 - A comparison of 35 serum ·hCG levels (A) obtained by this enzyme immu­noassay (EIA) and by a radioimmunoassay (RIA). r = 0.96, Y = 1.13x + 3.7 and that of 88 urine hCG levels (B) by this enzyme immunoassay (EIA) and by a hemaggluti­nation assay (HAR). r = 0.95, Y = 0.95x + 1.8

Within-run Sample 1 Sample 2 Sample 3

Between-run Sample 4 Sample 5 Sample 6

1 Mean ± SO

No. of assays

8 8 8

6 6 6

hCG, mlU/ml

18.0± 2.4' 80.0 ± 10.2

262.0 ± 24.0

31.3 ± 5.9 81.0 ± 10.5

134.2 ± 23.1

CV,%

13.3. 12.9

9.2

18.8 12.9 17.2

Table 2 - Precision of hCG assay in urine.

E 100 ...... ::::> E C!)

~ 10

."

Within·-run Sample 1 Sample 2 Sample 3

Between-run Sample 4 Sample 5 Sample 6

1 Mean ± SO

.. ..

. . .. . ..

.. . ... .. ,.. ....

J-4~1 1111!lId I 11111111

o 10 100 1000

hCG (mIU/ml) by RIA

277

No. of assays

8 8 8

6 6 6

250

~ 200 w 11 W 150 ...... ::::> E 100

C!)

~ 50 : . i: I'

hCG, mlU/ml

14.6 ± 1.7' 47.3 ± 5.1 73.5 ± 7.5

11.8 ± 2.0 22.8 ± 2.8 58.8 ± 2.4

CV, %

11.6 10.8 10.2

17.1 12.2

4.2

o 50 100 150 200 250

hCG (mIU/ml) by HAR

T. Sekiya, Y. Furuhash/~ S. Golo, et al.

DISCUSSION The enzyme immunoassay for hCG that we just described is as sensitive as the radioimmunoassay capable of estimating a low concentration of hCG (10 mlU/ml serum, or 3 mlU/ml urine) without complications and limitations associated with radioi­sotopes. A competitive enzyme immunoassay for hCG re­ported by Kikutani et al. (6), in which ,B-D-galactosi­dase from E. Coli was used for labelling hCG is as sensitive as the present method. However, their assay system is interfered, when the serum samples are present in the assay medium. This interference lowers the practical sensitivity of the assay, since the volume of sample applicable to the assay is limited. To improve the problems mentioned above, we have developed a sandwich-type enzyme immunoassay system for hCG. In this method, biological sample present in the first immunoreaction is washed out before the addition of the enzyme-labelled anti­bodies, which prevents a direct effect of biological fluid on the immunoreaction of the antibody-enzyme complex. The interference still observed with the sandwich method is removed by using the gelatin-containing buffer (11). The antibody-,B -D-galactosidase complex is stable at least for 6 months at 4 C in buffer A. The complex prepared from 20 mg IgG fractions of the antiserum is sufficient for 10,000 assays. The antibody­immobilized silicone rubber solid phase is stable for 1 month at 4 C in buffer A. The cross-reaction of hLH with the assay system was observed to be about 40%, when the reactivity was compared on the basis of international units. This is reasonable, since the antiserum to hCG also reacts with LH. The specific immunoassay system without the cross-reaction with LH could be prepared by the above described manner with an antibody specific to ,B subunit of hCG (12, 13) at least for preparing the

. antibody-immobilized solid phase (and, if possible, with an antibody specific to a subunit for preparing the antibody-enzyme complex). The antibodies coupled with ,B -D-galactosidase were not the Fab' fragments of purified (mono­specific)' IgG, but those of the IgG fractions of antiserum. To evaluate the concentration of' the effective complex in the preparation, aliquots of the complex preparation were passed through a small column of goat (anti-rabbit IgG) IgG-coupled Se­pharose 4B, or a column of hCG-coupled Sepharose 4B as described (8). More than 90% of the ,B­D-galactosidase activity applied were trapped in the column of the anti-rabbit IgG-Sepharose, indicating the fact that most of the enzyme was coupled with rabbit Fab' fragments. But the enzyme activity trapped in the column of the hCG-Sepharose was found about 10% of the activity applied. These

278

results indicate that only 10% of the rabbit Fab'-,B -D-galactosidase complex was effeCtive for immu­noassay. From these observations, it is suggested that the sensitivity of the present immunoassay is probably increased by using the purified antibodies for pre­paring the complex and the solid phase. Develop­ment of such an improved immunoassay system, using a purified ,B subunit-specific antibody prepara­tion, is now in progress in our laboratories.

ACKNOWLEDGMENTS This study was supported in part by a Grant to Tomoda (54-30) for Cancer Research from the Ministry of Health and Welfare and by a Grant to Tomoda (348299) from the Ministry of Education, SCience and Culture, Japan.

REFERENCES 1. Midgley A.A.

Radioimmunoassay; A method for human chOrioniC gonadotropin and human luteinizing hormone. Endocrinology 79: 1 0, 1966,

2. Tomoda Y., Hreschghyshyn M.M. Radioimmunoassay for human chorionic gonadotropin. Am. J. Obstet. Gyneco!. 100: 118, 1968.

3. Van Weemen B.K., Schuurs A.HW.N. Immunoassay using antigen-enzyme conjugates. FEBS Lett. 15: 232, 1971.

4. Van Weemen BK, Schuurs A.HW.N. Immunoassay using antibody-enzyme conjugates. FEBS Lett. 43: 215, 1974.

5, Yorde D.E., Sasse EA, Wang T.Y., Hussa A.O., Gar­ancis J.C. Competitive enzyme-linked immunoassay with use of soluble enzyme/antibody immune complexes for la­belling. 1. Measurement of human choriogonadotropin. Clin. Chem. 22: 1372, 1976.

6. Kikutani M., Ishiguro M., Imamura S., Miura S . Enzyme immunoassay of human chorionic gonado­tropin employing ,ll-D-galactosidase as labe!. J. Clin. Endocrino!. Metab. 47: 980, 1978.

7. Van Hell H., Mattijsen R, Homan JD.H. Studies on human chorionic gonadotropin: 1. PUrifica­tion and some physico°ochemical properties. Acta Endocrino!. 59: 89, 1968.

8, Kato K., Fukui H., Hamaguchi Y., Ishikawa E. Enzyme-linked immunoassay: Conjugation of the Fab' fragment of rabbit IgG with ,ll-D-galactosidase from E. Coli and its use for immunoassay. J. Immuno!. 2116: 1554, 1976.

9. Kato K" Hamaguchi y" Okawa S., Ishikawa E., Kobaya­shi K" Katunuma N, Use of rabbit antibody IgG-loaded silicone pieces for the sandwich enzymoimmunoassay of macromolecular antigens. J. Biochem. 81: 1557, 1977.

10. Bangham D,R., Grab B.

The second International Standard for Chorionic Go­nadotropin. Bull. WHO. 31: 111, 1964.

11. Kato K., Umeda Y., Suzuki F., Hayashi D., Kosaka A. Evaluation of a solid-phase enzyme immunoassay for insulin in human serum .. Clin. Chem. 25: 1306, 1979.

12. Matuura S., Chen H.C., Hodgen G. Antibodies to the carboxyl-terminal fragment of human

279

Enzyme immunoassay of hCG

chorionic gonadotropin .a-subunit: Characterization of antibody recognition sites using synthetic peptide ana­logues. Biochemistry 17: 575, 1978.

13. Vaitukaitis J.L., Braunstein G.D., Ross G.T. A radioimmunoassay which specifically measures human chorionic gonadotropin in the presence of human luteinizing hormone. Am. J. Obstet. Gynecol. 113: 751, 1972.