Immunologic and Biologic Study of Chorionic Gonadotropindm5migu4zj3pb.cloudfront.net/manuscripts/105000/105000/... · 2014. 1. 30. · was observed by immunoelectrophoresis of HCG

Journal of Clinical InvestigationVol. 43, No. 6, 1964

Immunologic and Biologic Study of Human ChorionicGonadotropin *

SHINICHI HAMASHIGEANDEDWARDR. ARQUILLA

(From the Department of Pathology, University of California School of Medicinc,Los Angeles, Calif.)

Immunological methods have been used to meas-ure human chorionic gonadotropin (HCG) in nor-mal and pathological pregnancies (1-3). Immu-nohistological methods with the fluorescent anti-body technique have also been used to localizeHCGin placenta and tumors (4, 5). The anti-sera used in the above studies were produced withHCG preparations that contained contaminantsand, therefore, contained antibodies to a multi-plicity of extraneous antigens in addition to HCG.Antibodies to some contaminants can be re-moved from HCG antisera by adsorption withurine, serum, or their extracts (1, 3, 6). Previousstudies, however, have demonstrated that absorp-tion methods probably do not make such antiseraHCGspecific (7). This investigation is concernedwith the immunological characteristics of severalcommercial HCGpreparations in an attempt toevaluate the reliability and validity of the HCGantisera previously employed (1-5).

Seven antigens are demonstrable in commercialHCG preparations by immunoelectrophoresis.Five of these contaminating antigens are commonto HCGand nonpregnant urine extract and can bedemonstrated by immunoelectrophoresis and he-magglutination. HCGantisera after adsorptionwith normal male urine powder contain antibodiesto two antigens unique to pregnancy urine extractsand urinary HCGpreparations. These two anti-gens are dissimilar but share partial immunologi-cal identity by immunoelectrophoresis. Chro-matographic separation of HCGpreparations intobiologically active and inactive fractions has madeit possible to evaluate the relationship between im-munological and biological activities. The twoantigenic moieties unique to pregnancy are differ-

* Submitted for publication November 4, 1963; acceptedFebruary 6, 1964.

Supported by grant M6315 from the Population Coun-cil, California Institute of Cancer Research.

ent and distinct from the material possessing thebiological activity of HCG. It appears, therefore,that the biological activity of HCGcannot be at-tributed to the two antigenic moieties unique topregnancy urine that can be demonstrated by he-magglutination and immunoelectrophoresis. Thesepregnancy antigens, distinct from the biologicallyactive HCG, are probably the causative agents inpreparing antisera highly effective in testing forpregnancy. They probably do not possess biologi-cal activity, however, and antisera directed to oneor both of them are probably not specific for thelocalization of HCG (biologically active) histo-logically. Demonstration of precipitating or he-magglutinating antibodies against biologically ac-tive HCG was not possible in four differentantisera tested. All four of these antisera did,however, contain antibodies capable of neutraliz-ing the biological activity of HCG.

Methods

Human chorionic gonadotropin (HCG). HCG, lotno. 90231, 90259, and 90272, containing 3,400, 2,410, and2,710 U of HCG per mg, respectively, was obtainedcommercially.'

Cellulose acetate electrophoresis (7). Electrophoresisusing cellulose acetate strips 2 (8) was performed on theShandon Universal electrophoresis apparatus 2548 2 withbarbital buffer at pH 8.6, ionic strength 0.08, for 1.5hours at 0.4 ma per cm of strip width. A Heathkitmodel PS-3 3 was the current power supply. The elec-trophoretic strips were stained with light green S.F. orPonceau S and nigrosin.

Hemagglutination and hemagglutination inhibition meth-ods. The hemagglutination and hemagglutination inhibi-tion methods employed have been described by Arquillaand Stavitsky (9). Modifications in the method usedfor HCGhave been previously described (7).

Immunoelectrophoresis (7). Immunoelectrophoreticstudies by the Scheidegger modification (10) were per-

1Vitamerican Corp., Little Falls, N. J.- Consolidated Laboratories, Chicago, Ill.3 Heath Co., Benton Harbor, Mich.

1163

SHINICHI HAMASHIGEAND EDWARDR. ARQUILLA

formed on the LKB electrophoresis apparatus 6800A 4

with 1% Ionager no. 2 2 in barbital buffer at pH 8.6,ionic strength 0.08. Electrophoresis was performed at4.5 ma per slide for 1 to 3 hours. Incubation for de-velopment of precipitin lines was done at room tem-perature for 16 to 40 hours in a moist chamber. Thestaining of precipitin lines was accomplished with 0.5%phloxine B 5 and 0.5% eosin Y 5 in 95% ethanol.

Bioassay in the intact immature male rat. The ratprostate and accessory organ method of Diczfalusy (11)was employed for HCGbioassay with 21-day old Long-Evans rats. Formalin (10%) was used for tissue fixa-tion, and results were expressed in grams per 100 g bodyweight (7).

Preparation, processing, and use of antisera. The im-munization of rabbits and processing and storage of HCGantiserum have been previously described (7). Identi-cal procedures were used to obtain antisera to normalmale urine extract and to an HCGfraction.

All lots of antisera used were individually pretestedimmunoelectrophoretically to determine the optimal anti-gen concentration required for maximal precipitin lineformation. Further tests were performed to define thezones of antibody and antigen excess and to detect op-timal proportions of other antigen-antibody systems.

Several lots of commercially available rabbit antihu-man serum8 were used for immunoelectrophoretic stud-ies. A pool of horse antiserum to Vitamerican HCG7was used exclusively for immunoelectrophoresis.

Preparation of human urine powders. Crude humanurine powders from normal males, normal females, andpregnant females were obtained by acetone precipitation(12). The crude acetone precipitate was suspended insaline and dialyzed against water at 40 C. The residueafter dialysis was separated by centrifugation and dis-carded. The dialyzed supernatant fluid was lyophilized,and about 1 mg of extract was obtained per 4 to 5 mlof freshly voided urine. Extracts used were freshlyprepared by dissolving weighed amounts of powder inbarbital buffered saline,8 and the insoluble residue was re-moved by centrifugation.

Various concentrations of the water insoluble resi-due of normal male urine acetone precipitate were dis-solved in barbital buffered saline and tested immunoelec-trophoretically with HCG antiserum. Adsorption withthe water insoluble residue failed to alter the character-istics of the antisera with respect to their ability to re-act with HCGimmunoelectrophoretically.

Bioassay studies performed with intact, immature male4 LKB Instruments, Inc., Washington, D. C.5 Allied Chemical Corp., National Aniline Division,

New York, N. Y.6 Hyland Laboratories, Los Angeles, Calif.7 Generously supplied by Dr. T. Asher of Hyland

Laboratories.8 Diluent for hemagglutination, immunoelectrophoresis,

and bioassay was barbital buffered saline (13), pH 7.4,containing 5 X 10' M MgCl2 and 1.5 X 10' M CaCl.per L.

rats showed no detectable gonadotropic activity in 5 mg(25 ml urine) of normal male or normal female urinepowders (p 0.0001). Similar amounts of identically pre-pared pregnant female urine powder contained HCG(p 0.92).

Adsorption of antiserumi. Weighed amounts of ex-tracts were added to antiserum, and the mixture was al-lowed to stand at room temperature for a minimum of3 hours. The adsorbed antiserum (supernatant fluid)was then obtained by centrifugation at 2,000 rpm for 10minutes. Twice the minimal amount of extract re-quired to delete the maximal number of precipitin linesor cause maximal inhibition of hemagglutination wasused. When an antiserum was adsorbed with more thanone material, each extract was added sequentially.

Preliminary studies indicated that 50 to 125 mg ofnormal male urine powder (250 to 650 ml of urine) wasnecessary to fulfill the criteria outlined for adsorptions.Adsorption with urine was not practical, since the amountrequired represented quantities greater than previouslyused (7) and antiserum would be diluted to a low orderof sensitivity.

Column chromatography with diethylaminoethyl celluf-losc. HCG, 250 mg, dissolved in 10 ml of buffer wasadded to a 2.5- X 30.0-cm column packed with diethylami-noethyl cellulose (DEAE). The DEAEwas equilibratedwith 0.03 M, pH 8.3, tris(hydroxymethyl)aminomethane(Tris) buffer, before packing the column. A gradientof sodium chloride (0.0 to 0.2 M) was obtained by al-lowing the eluant (0.4 M sodium chloride in 0.03 M Trisbuffer) to be diluted with an equal volume of 0.03 MTrisbuffer in a magnetically stirred mixing chamber beforeflowing into the column. The flow rate used in thesechromatographic studies was 2 to 5 ml of effluent perminute. Five-ml fractions were collected with a GMEvolumetric fractionator, model V152,9 continuouslyscanned for proteins at 280 m/A with a GMEultravioletadsorption meter, model UV-280IF,9 and recorded (Recti-riter) .10

The cores of the protein peaks were confirmed by ad-sorption at 280 m~uwith Beckman model DB spectro-photometer,1' dialyzed against water at 40 C, and lyophi-lized. The lyophilized powders were weighed and dis-solved in barbital-buffered saline just before use.'

Results

Characterization studies of HCGpreparations.A maximum of seven distinct precipitin lineswas observed by immunoelectrophoresis of HCGpreparations with concentrations ranging from 1to 5CO mg per ml. An example of the immuno-electrophoretic pattern with horse HCGantiserumis shown in Figure 1. The precipitin lines are

9 Gilson Medical Electronics, Middleton, Wis.10 Texas Instruments, Inc., Houston, Texas.11 Beckman Instruments, Inc., Fullerton, Calif.

164

IMMUNOLOGICAND BIOLOGIC STUDYOF HUMANCHORIONIC GONADOTROPIN

2 3 4 5

I \I / HCG25.%_

6 7

-.y\-

NMU5o

FIG. 1. IMMUNOELECTROPHORETICPATTERNS OF HCG AND NORMAL MALE URINE EXTRACTAGAINST HORSE HCGANTISERUM. HCG2= human chorionic gonadotropin preparation (25mg per ml); H-HCG= horse HCGantiserum; and NMU,,5o = normal male urine extract(150 mg per ml).

numbered 1 to 7 from anode to cathode. Thesenumbers will be used in referring to the homolo-gous antigen or antibody responsible for each pre-

cipitin line. Precipitin line 2 was faint and didnot reproduce well in the photograph. The HCGantisera from three rabbits used in these studiesshowed similar immunoelectrophoretic patternsand contained 10,000 to 50,000 hemagglutiinationunits (Table I, line 1).

Antigens common to HCGand normal maleurine extracts (50 to 800 mg) were demonstratedby immunoelectrophoresis with HCG antiserum(Figure 1). Five precipitin lines (no. 1, 2, 3, 4,and 7) were obtained with extracts of normalmale urine. The precipitin line similar to line 2

was faint and is not reproduced in the figureshown. Identical immunoelectrophoretic resultshave also been obtained with nonpregnant femaleurine extracts.

Immunoelectrophoresis of HCG preparationswith rabbit antiserum against normal male urineextracts failed to demonstrate the five (no. 1, 2,3, 4, and 7) precipitin lines. Hemagglutinationstudies with the same antiserum tested againstHCGconjugated sheep erythrocytes had a lowtiter (Table I, line 2). The poor hemagglutinatingtiter and paucity of precipitating antibodies inthese antisera were probably due to the low con-

centration of antigens in the extract used for im-munization.

TABLE I

Sunmmary of henmagglutination and hemnagglutination inhibition studies*

Line Rabbit antiserumt to: Inhibiting antigen:

Saline HCGio NMUio NFUio PFUio1 HCG 25,000 0 6,250 to 12,500 6,250 to 12,500 2002 NMU 1,280 0 40 40 20

F lo F Ilio3 HCG(NMU) 10,000 0 400 04 F II (NMU) 10,000 0 50 05 HCG(NMU, F I) 0

* Numbers indicate hemagglutiniation units that are reciprocals of titers. HCG= human chorionic gonadotropinpreparation, NNIU = normal male urine extract, NFU = normal female urine extract, PFU = pregnant female urineextract, F I = Fraction I, and F II = Fraction II.

t Adsorbed antiserum followed by letters in parentheses denoting adsorbent used.t Subscripts indicate amount of antigen in micrograms used for inhibition. Reaction with rabbit antiserum is indi-

cated by inhibition of hemagglutination of sheep erythrocytes to which HCGwas conjugated.

robot "lilwook"M

1165


1

b1FIG. 2. SERUMPROTEIN CONTAMINANTSDEMONSTRABLEIN NORMALMALE URINE EXTRACT.

R-HCG= rabbit HCGantiserum; NMUs,,O = normal male urine extract (150 mg per ml);and RAH= rabbit antihuman serum.

The presence of serum protein contaminants inHCG, tested immunoelectrophoretically with rab-bit antihuman serum demonstrated two precipitinlines. Two similar precipitin lines were also ob-served when various dilutions (1/1 to 1/100) ofhuman serum were tested with rabbit HCGanti-sera. These serum protein contaminants (precipi-tin lines b and c) were also identified in normalmale urine extracts (Figure 2). Precipitin linec, not visible in the photograph, was a short arcextending from just below the cathodic portion ofthe antigen well and appeared to merge with theanodic tip of precipitin line b. Precipitin line c(Figure 2) has been observed only under theseconditions and indicates the presence of an eighthantigen in HCGpreparations. Of the seven pre-cipitin lines (no. 1 to 7) previously demonstratedwith HCGand its homologous antiserum, onlyprecipitin line 2 appears to be due to a serum pro-tein antigen. Precipitin line a has not been ob-served in each of three lots of HCG. Furtherstudies show that antigen a possessed the electro-phoretic mobility of serum albumin and b and cmigrated as alpha globulins.

Each of three lots of a commercial HCGprepa-ration contains a minimum of seven antigenicmoieties capable of eliciting precipitating anti-bodies in rabbits. An eighth antigen (precipitinline c, Figure 2), not ordinarily detected, also ap-pears to be present in HCG preparations. Aminimum of five of these antigenic moieties is

common to HCGpreparations and nonpregnanthuman urine extracts. There appear to be twoantigens in each of three HCGpreparations testedthat demonstrate immunological cross reactionwith antigens in serum.

Studies of similarities and dissimilarities be-tween HCGand normal male, normal female, andpregnant urines by adsorption. Five (no. 1, 2, 3,4, and 7) of the seven antigens demonstrated werecommon to commercial HCG preparations andnonpregnant human urine extracts, whereas twoantigens (no. 5 and 6) appeared unique to HCGpreparations. To further study similarities anddissimilarities between HCGpreparations and anti-gens in both pregnant and nonpregnant urine, ad-sorptions of HCGantisera with urine powderswere performed.

Two merging precipitin arcs with a single spur,which correspond to precipitin lines 5 and 6, wereconsistently observed by immunoelectrophoresisof HCGpreparations with each of three rabbitHCGantisera adsorbed with normal male urineextract. Horse HCGantiserum was used in theexample shown (Figure 3, top half of montage)and is identical to results obtained with rabbitHCGantisera. These antisera showed a decreasefrom 25,000 to a minimum of 6,250 hemagglutina-tion units. As much as 800 mg of normal maleurine powder per ml of HCGantisera was usedin adsorptions with no detectable change in theimmunoelectrophoretic pattern or further decrease

R H C Gpop"

Nmu 5 0

--.---R. A H.-OF

1166

2 3 4 7


+ H C G2 5

P F U00

FIG. 3. ILLUSTRATION OF ANTIGENS UNIQUE TO PREGNANCYBY ADSORPTION OF HCGANTI-

SERUMWITH NORMALMALE URINE EXTRACT. HCGW= human chorionic gonadotropin prepara-tion (25 mg per ml); H-HCG NMU05=horse HCGantiserum (two different lots) adsorbedwith normal male urine extract (100 mg per ml of antiserum); and PFU555 = pregnant femaleurine extract (500 mg per ml).

in hemagglutination titer. Nor was the HCGneutralizing ability of these three antisera ad-sorbed in this fashion altered when tested bio-logically (Table II, line 3). Similar immunoelec-trophoretic, hemagglutination, and bioassay re-

sults were also obtained when these HCGantiserawere adsorbed with urine extracts from nonpreg-

nant females.It has been possible to demonstrate that preg-

nant female urine extracts contain all of the anti-genic moieties demonstrable in HCGpreparationswith the methods employed. A total of seven anti-gens common to HCGand pregnant female urinecan be demonstrated by immunoelectrophoresiswith HCGantisera. By varying the concentrationof the pregnant female urine extract, the five anti-genic moieties common to HCGpreparations andpregnant and nonpregnant urine extracts were

TABLE II

Summary of bioassay studies*

Mean wt,prostate andaccessory sex

Line Antigen (biological units) Rabbit antiserumt to: organs$ p

mg ml g/100 gbody wt

1 HCG, 0.004 (10 U) 0 0.252 HCG, 0.004 (10 U) HCG 0.02 0.15 0.00013 HCG, 0.004 (10 U) HCG(NMU) 0.02 0.17 0.00014 HCG, 0.004 (10 U) HCG(NMU, F I) 0.02 0.15 0.00015 HCG, 0.004 (10 U) F II 0.05 0.17 0.00016 F II, 0.002 (10 U) HCG 0.05 0.16 0.0001

* HCG= human chorionic gonadotropin preparation, NMU= normal male urine extract, F I = Fraction I, andF II = Fraction II.

t Adsorbed antiserum followed by letters in parentheses denoting adsorbent used.t Average SD per group = 0.02 g per 100 g body weight. Range of SD per group = 0.01 to 0.03 g per 100 g body

weight.

W-1- v..-c--

I..In - .* -Wm

1167

H.


demonstrable. Precipitin lines 5 and 6 were ob-served with higher concentrations (500 mg per

ml) of pregnancy urine extracts (Figure 3, lowerhalf of montage). The horse HCGantiserumused in this illustration was incompletely adsorbed,as evidenced by a faint third precipitin line near

the anode. Further studies indicated that thisthird line was precipitin line 1 and could be re-

moved by adsorption without affecting the two

merging precipitin lines (no. 5 and 6). The anti-genic similarity existing between HCGprepara-

tions and pregnant female urine extracts was

further demonstrated by complete adsorption ofdemonstrable antibodies (precipitating and he-maglutinating) in rabbit HCGantisera.

These studies further confirm the presence offive contaminant antigens (biologically inactive)common to HCG preparations and all humanurines. Furthermore, two antibody species pres-

ent in HCGantisera have been shown to be di-rected to antigens unique to HCGpreparationsand pregnancy urine. These antigens share an

immunological identity but are distinct and sepa-

rate as indicated by the single spur (Figure 3).The hemagglutination of HCGconjugated cells

(Table I, line 2) by antisera against extracts ofnormal male urine was significantly less (1,280)than seen with HCGantisera (25,000). Inhibi-tion of hemagglutination of HCGconjugated cells

by antisera against extracts of normal nmale urinewas very similar with extracts of pregnant andnonpregnant urines (Table I, line 2). The inhi-bition of hemagglutination of HCGantisera withnormal urine extracts was relatively minor (6,250)compared with the marked inhibition (200) ob-tained with pregnancy urine extracts (Table I,

line 1). Therefore, the components that are re-

sponsible for most of the immunological activityof HCGantisera as measured by hemagglutinationand hemagglutination inhibition are probably con-

tributed by the two "pregnancy antigens."These studies also demonstrate that HCGanti-

sera can be made reliable for determining preg-

nancy by proper adsorption methods. The per-

sistence of two different antibody species, how-ever, implies that adsorbed antisera against HCGpreparations are not HCGspecific.

Chromatographic fractionation of HCG( prepa-rations. Figure 4 is a chromatogram of HCGlotno. 90272 in which two major peaks were obtainedand labeled Fractions I and II, respectively. Frac-tion I was consistently obtained as a sharp spikewith a sodium chloride concentration (14) greaterthan 0.01 Mand less than 0.05 M. Rechromatog-raphy of both fractions, under identical condi-tions, showed no significant changes in the pat-terns. Similar chromatographic results were ob-tained with HCGlot no. 90231.

HCGpreparation: Vitomericon Corporation Lot No. 90272Column: DEAE Cellulose, 2.5x 45 cm

Buffer: Tris, 0.03m, pH 8.3Eluant: NaCI, 0.4m, in Tris buffer~~~~~~~~~~~~~~~~~~., _.

0° '-TU BE NOSH

J:A, tt0'H

I e90 _ Ievil ti-F_-_- _

FIG. 4. CHROMATOGRAPHYOF HCG.

1168


HCG25+! ' __I_~~~~~~~~~~~~~~~~~~~~~.....

F 15 + F 1110

FIG. 5. MIXTURES OF FRACTIONS I AND II EXAMINED IMMUNOELECTROPHORETICALLYWITH

HCG ANTISERUM. HCG., = human chorionic gonadotropin preparation (25 mg per ml);H-HCG=horse HCGantiserum; and F 15+F II10=mixture of Fraction I (5 mg per ml)and Fraction II (10 mg per ml).

Chromatograms of HCGlot no. 90259 were simi-lar to the other two lots of HCGexcept for a bifidfirst peak (chromatogram not shown). The bifidnature of Fraction I, HCGlot no. 90272, wasdemonstrated by further chromatographic studies.Cellulose acetate electrophoresis, hemagglutina-tion, hemagglutination inhibition, immunoelectro-phoresis, and bioassay studies repeatedly failedto demonstrate qualitative differences of com-ponents in the three lots of HCGused. The bifidfirst peak in HCGlot no. 90259 apparently repre-sents a significant quantitative variation peculiarto this lot. The chromatographic results wererepeated several times with each HCGpreparation.

The specific biological activity of the effluentfractions was not determined. Random qualita-tive determinations were made and detectable bio-logical activity was present with the ascending por-tion and descending portion of Fraction II andin the long trailing tail that follows Fraction II.

The amounts recovered comprised about 14%oand 36%o of the starting HCGpreparation forFraction I and Fraction II, respectively. Wheneach of these fractions was rechromatographed,70 to 75%o of the sample added was recovered.About 45%o of the HCG(biological activity) wasrecovered after Fraction II was chromatographed.

Cellulose acetate electrophoresis of HCGprep-arations and a mixture of Fraction I and Frac-tion II resulted in identical patterns. Fraction Irevealed two protein zones migrating toward thecathode that corresponded to two similar zones

with HCG. Fraction II presented a dumbbell-shaped pattern that corresponded to an oblong andmore diffuse electrophoretic pattern observed withHCGpreparations.

Significant antigenic alteration was not demon-strated when mixtures of Fraction I and FractionII were examined immunoelectrophoretically withHCGantisera (Figure 5). A single precipitinline (no. 1) nearest the anode is not seen in therecombined mixture. The missing antigen repre-sents the only element eliminated by the chromato-graphic procedure and probably adhered to thecolumn. Immunoelectrophoresis of recombinedmixtures of Fractions I and II tested against HCGantisera adsorbed with normal male urine powdersreproduced the two merging precipitin arcs (no.5 and 6) with a single spur previously demon-strated (Figure 3) and shown to be unique topregnancy.

Chromatography apparently did not alter the im-munological or biological activities of any of thethree HCGpreparations. Hemagglutination andhemagglutination inhibition studies demonstratedFraction II to contain specific activity (immuno-logical) comparable to HCGpreparations (TableI, lines 3 and 4). Rabbit antisera to Fraction IIdemonstrated hemagglutination activity similar toHCGantisera (Table I, line 4). Biological neu-tralization of gonadotropic activity in HCG orFraction II was readily accomplished by antiseraagainst either HCG or Fraction II (Table II,lines 5 and 6).

1169


Fl5

HiHCG

F 1110

FIG. 6. IMMUNOELECTROPHORETICSTUDY OF FRACTIONS I AND II WITH HCGANTISERUM.F I = Fraction I (5 mg per ml); H-HCG= horse HCGantiserum; and F II,, = Fraction II

(10 mg per ml).

The immunoelectrophoretic study of FractionsI and II with HCGantiserum is shown in Fig-ure 6. Each fraction was characterized by fourprecipitin lines. Only three of the precipitin linesin Fraction I photographed well. Further im-munoelectrophoretic studies show that Fraction Icontained antigens 4, 5, 6, and 7 and Fraction II

+

contained antigens 2, 3, 4, and 5. Antigens 4 and5 appear to be common to both fractions. Frac-tion I contains the pregnancy antigen 6 (Figure7). Pregnancy antigen 5, however, appears inboth Fraction I (Figures 6 and 8) and Fraction II

(Figure 7).Bioassay of the chromatographic fractions by

F I

H -HCG@NMUIO

I F1I I

FIG. 7. IMMUNOELECTROPHORETICSTUDY OF FRACTIONS I AND II WITH HCGANTISERUMAD-

SORBED WITH NORMALMALE URINE EXTRACT. F I,= Fraction I (5 mg per ml); H-HCG-

NMU1, = horse HCGantiserum adsorbed with normal male urine extract (100 mg per ml);and F IIo Fraction II (10 mg per ml).

170


the rat prostate and accessory organ methodshowed Fraction I (0.1 mg) to contain no detecta-ble gonadotropic activity (p 0.0001). FractionII, obtained from each of three lots of HCGandtested separately, contained 4,000 to 6,000 U per

mg.12 The chromatographic procedure has doubledthe specific biological activity of Fraction II as

compared to the activity in the HCGpreparations.In these series of experiments it has been pos-

sible to detect variations in commercial HCGpreparations that appear to be quantitative ratherthan qualitative differences in the antigens pres-

ent. DEAEcolumn chromatography resulted inpartial separation of antigens and complete ex-

clusion of biologically active HCGfrom FractionI. The studies also indicate that Fraction II con-

tains the major portion of the hemagglutinatingactivity (Table I, lines 3 and 4) and all of thebiological activity recovered.

Immunologic and biologic study of HCG. Theseparation of antigens into two fractions and con-

centration of the immunological and biological ac-

tivity into a single fraction was achieved by col-umn chromatography. However, there was sig-

12 Fraction II (HCG lot no. 90259) was bioassayed at

4,773 U per mg by South Mountain Laboratories, Inc.,Maplewood, N. J., using the immature rat uterine weightmethod employing a 3-dose assay with 15 rats per doselevel.

+.X

nificant contamination of Fraction I with moietiesfrom Fraction II. Immunoelectrophoretic stud-ies testing Fraction I with rabbit antiserumagainst Fraction II demonstrated four precipitinlines identical to those observed with the homolo-gous antigen (Fraction II). Only two of the ma-

jor precipitin arcs are easily visible in the photo-graph shown (Figure 8). Consequently, it isprobable that Fraction I contains antigens (2 and3) in addition to antigens 4, 5, 6, and 7. WVhenFraction II antisera, adsorbed with normal maleurine powder, were tested against Fractions Iand II, a single precipitin arc possessing the char-acteristics of pregnancy antigen 5 was obtainedwith each fraction. The two rabbit antiseraagainst Fraction II contain demonstrable pre-

cipitating antibodies to only one of the two preg-

nancy antigens previously described.Fraction I contains the pregnancy antigens 5

and 6. It is not surprising, therefore, that HCGor Fraction II antisera adsorbed with both normalmale urine powder (50 to 125 mg per ml of anti-serum) and Fraction I powder (2 to 5 mg per mlof antiserum) fail to give precipitin lines whentested against either HCGor Fraction II. Fur-thermore, HCGor Fraction II antisera so ad-sorbed contained no demonstrable hemaggluti-nating antibodies (Table I, line 5). Similarresults were obtained with antisera from three

F I

R Fi

R|F11V 1

10

FIG. 8. IMMUNOELECTROPHORETICSTUDYOF FRACTIONS I AND II WITHRABBIT ANTISERUMTO FRACTION II. F I, = Fraction I (5 mg per ml);R-F II = rabbit antiserum to Fraction II; and F II,, = Fraction II (10mg per ml).

1171


rabbits immunized with HCGpreparations andalso antisera from two rabbits immunized withFraction II.

Four HCGantisera from which all detectableprecipitating and hemagglutinating antibodieswere removed by adsorption were capable of com-pletely neutralizing the biological activity of 10U of HCGtested in the male rat (Table II, line4). Consequently, we feel that HCG, biologicallyactive or inactive, failed to elicit the productionof precipitating antibodies in the five rabbits im-munized. It is possible that biologically inactiveHCGelicits precipitating antibodies that do notcross-react with biologically active HCGand maybe responsible for one or both of the pregnancyantigens described.

The antibodies in HCGantisera that neutralizethe biological activity of HCGwere not associ-ated with antigens responsible for precipitatingor hemagglutinating antibodies present in thesame antisera. The precipitating and hemag-glutinating antibodies in antisera directed againstHCGpreparations and specific for pregnancy areelicited by the two pregnancy antigens previouslydescribed and not by biologically active HCG.

DiscussionThe majority of immunological studies of HCG

have been performed with antiserum obtained byimmunizing rabbits with commercial preparations(1-3, 6, 7). The presence of contaminants inthese preparations has been demonstrated byelectrophoresis, immunodiffusion, or immunoelec-trophoresis (6, 7, 15, 16). The number ofcontaminants shown to be present has been vari-able, depending on the preparations and char-acterization techniques used. Adsorption methodsusing serum or urine have been employed to in-crease the HCGspecificity of the antiserum (1, 3,6). A previous study from this laboratory ques-tioned the validity of adsorption and the specificityof the antiserum after adsorption (7).

Immunoelectrophoresis of commercial HCGpreparations with homologous antisera (Figure 1)consistently demonstrates seven precipitin lines.This number and the pattern of precipitin lineshave been reproduced with antisera from threedifferent rabbits. An eighth antigen probably ispresent in the HCGpreparations employed. An-tigens 1, 2, 3, 4, and 7 are contaminants, lack

gonadotropic activity, and are common to all hu-man urines. Only two of these contaminant anti-gens demonstrate precipitin lines (Figure 2, b andc) with rabbit antihuman serum and appear to beof serum protein origin. These two contaminantsmay represent the "long alpha2" and the secondalpha2 globulins described in urine by Berggaird(17). Bergga'rd was unable to demonstrate theseantigens in human serum. Antisera against thecommercial HCGpreparations employed in thisstudy can cause two precipitin lines (Figure 2,precipitin lines b and c) similar to those observedby Bergga'rd when tested against whole human se-rum. The remaining contaminating antigens inHCG preparations probably represent urinarymacromolecules, not originating in whole serum.This is not surprising, since at least two-thirds ofurinary macromolecules have no serum counter-part (18).

The adsorption of HCGantiserum with urineis not practical, since the amount required repre-sents quantities greater than commonly used andwould dilute the antiserum to a very low order ofsensitivity. After adsorption with normal, non-

pregnant human urine powders, two precipitatingantibody species remain in HCGantisera demon-strable by immunoelectrophoresis (Figure 3).These precipitin lines (no. 5 and 6) showed no

change in character or intensity, despite adsorptionwith increasing increments of extracts up to 800mg (4 L of urine) per ml of antisera. Theseantigens unique to pregnancy urine are distinctbut share partial immunological identity. It ispossible, therefore, to prepare antisera againstHCGpreparations, by adsorptions with extractsof nonpregnant urine, which can be used as re-liable indexes of pregnancy. Such antisera, how-ever, contain at least two antibody species and,therefore, cannot be considered specific for anysingle moiety.

It was decided that these materials had to bepurified and characterized before reliable and validantisera for measuring and localizing HCGcouldbe produced. Two major fractions were repeat-edly obtained by chromatography. One antigen(no. 1) was no longer discernible after DEAEchromatography. All of the remaining antigens(no. 2, 3, 4, 5, 6, and 7) were detectable in Frac-tion I. Fraction II, however, contained antigens2, 3, 4, and 5.

172


The possibility exists that antigens 4 and 5 may

be macromolecules that dissociate to give antigenicmoieties responsible for precipitin lines 2 and 3.In this way the presence of all of the antigens inFraction I that come off the column first may beexplained. Both pregnancy antigens 5 and 6 were

present in Fraction I, whereas Fraction II con-

taimed only pregnancy antigen 5. No detectablegonadotropic activity (p 0.0001) was present inFraction I (0.1 mg) when repeatedly tested inmale rats. Therefore, the possibility that preg-

nancy antigens 5 and 6 are not associated withbiologically active HCGwas considered.

The contamination of Fraction I suggests thepossibility of adsorbing all precipitating antibodiesfrom HCGantisera. Several antigens, however,were present only in trace amounts; consequently,large quantities of Fraction I would be required.Therefore, a combination of normal male urineand Fraction I powders was used to remove allprecipitating and hemagglutinating antibodiesfrom HCG antisera. Neither Fraction I nor

normal male urine extracts contained any meas-

urable amounts of biologically active HCG. Con-sequently, the possibility of preparing antiseradevoid of demonstrable precipitating and hemag-glutinating antibodies but capable of neutralizingthe biological activity of HCGwas tested. Thistest was performed with antisera from four rab-bits (Table II, line 4). Complete neutralizationof biologic activity (10 U of HCG) was achieved.A previous study (7) in which the biologic neu-

tralizing ability of HCG antisera was titratedshowed that neutralizing activity is negated be-fore loss of hemagglutinating activity. Hemag-glutination is markedly more sensitive for meas-

uring antibodies than is the biological neutraliza-tion of HCG. Therefore, it was considered un-

likely that the antibodies responsible for neutrali-zation were the same as those causing hemaggluti-nation. The presence of biological neutralizingantibodies in adsorbed HCGantisera devoid ofprecipitating and hemagglutinating antibodies is,therefore, best explained by assuming that the an-

tibodies which neutralize HCGare directed to an-

tigen (or antigens) different and distinct fromthe antigens eliciting the production of precipi-tating and hemagglutinating antibodies. Immuno-logical methods probably measure moieties notresponsible for the activity of HCGpreparations

measured with biological methods. This thesiswas previously proposed, and supporting evidencewas presented by Wide (1).

Summary

1) Seven antigens are demonstrable in comminer-cial human chorionic gonadotropin (HCG) prepa-rations by immunoelectrophoresis. Five of theseven antigens present are contaminants, lackgonadotrophic activity, and are common to HCGpreparations and acetone precipitated urine ex-tracts.

2) HCGantisera after adsorption with non-pregnant urine extracts contain precipitating anti-bodies to two antigens unique to pregnancy urinesand HCG preparations. Such antisera appearhighly specific for pregnancy.

3) The two antigens unique to pregnancy aredissimilar but share partial immunological identityby immunoelectrophoresis. The precipitating andhemagglutinating antibodies in HCGantisera aredirected against these antigens.

4) The antibodies that neutralize the gonado-trophic activity of HCGas tested in the rat areindependent of precipitating and hemagglutinatingactivity in antisera prepared against three dif-ferent commercial HCGpreparations.

5) Antibodies directed against biologically ac-tive HCGare detected with certainty only by neu-tralization experiments. Current immunologicalmethods appear to be of questionable value in thedetection, assay, and localization of biologicallyactive HCG.

References1. Wide, L. An immunological method for the assay of

human chorionic gonadotrophin. Acta endocr.(Kbh.) 1962, 41, suppl. 70.

2. Mishell, D. R., Jr., L. Wide, and C. A. Gemzell.Immunologic determination of human chorionicgonadotropin in serum. J. clin. Endocr. 1963, 23,125.

3. Brody, S., and G. Carlstrom. Immuno-assay of hu-man chorionic gonadotropin in normal and patho-logic pregnancy. J. clin. Endocr. 1962, 22, 564.

4. Midgley, A. R., Jr., and G. B. Pierce, Jr. Immuno-histochemnical localization of human chorionic go-nadotropin. J. exp. Med. 1962, 115, 289.

5. Thiede, H. A., and J. W. Choate. Chorionic gonado-tropin localization in the human placenta by im-munoflourescent staining. II. Demonstration ofHCGin the trophoblast and amnion epithelium of

1173


immature and mature placentas. Obstet. and Gy-nec. 1963, 22, 433.

6. Midgley, A. R., Jr., G. B. Pierce, Jr., and W. 0.Weigle. Immunobiological identification of humanchorionic gonadotropin. Proc. Soc. exp. Biol.(N. Y.) 1961, 108, 85.

7. Hamashige, S., and E. R. Arquilla. Immunologicalstudies with a commercial preparation of humanchorionic gonadotropin. J. clin. Invest. 1963, 42,546.

8. Chromatographic and Electrophoretic Techniques,Zone Electrophoresis, Ivor Smith, Ed. New York,Interscience, 1960, vol. 2.

9. Arquilla, E. R., and A. B. Stavitsky. The productionand identification of antibodies to insulin and theiruse in assaying insulin. J. clin. Invest. 1956, 35,458.

10. Scheidegger, J. J. Une micro-methode de l'immuno-electrophorese. Int. Arch. Allergy 1955, 7, 103.

11. Diczfalusy, E. An improved method for the bio-assay of chorionic gonadotrophin. Acta endocr.(Kbh.) 1954, 17, 58.

12. Frank, R. T., U. J. Salmon, and R. Friedman. De-termination of luteinizing and follicle-stimulatingprinciple in castrate and menopause urine. Proc.Soc. exp. Biol. (N. Y.) 1935, 32, 1666.

13. Kabat, E. A., and M. M. Mayer. Experimental Immunochemistry, 2nd ed. Springfield, Ill., CharlesC Thomas, 1961, p. 149.

14. Schales, O., and S. S. Schales. A simple and ac-curate method for determination of chloride in bio-logical fluids. J. biol. Chem. 1941, 140, 879.

15. Brody, S., and G. Carlstrdm. The problem of speci-ficity in serologic determination of human chorionicgonadotropin. Acta endocr. (Kbh.) 1963, 42, 485.

16. Got, R., G. Levy, and R. Bourrillon. Analyse im-munoelectrophoretique de la gonadotropine chori-ale humaine. Experientia (Basel) 1959, 15, 480.

17. Bergga'rd, I. Studies on the plasma proteins in nor-mal human urine. Clin. chim. Acta 1961, 6, 413.

18. King, J. S., Jr., and W. H. Boyce. High MolecularWeight Substances in Human Urine. Springfield,Ill., Charles C Thomas, 1963, p. 56.

174

Documents

Immunologic and Biologic Study of Chorionic Gonadotropindm5migu4zj3pb.cloudfront.net/manuscripts/105000/105000/... · 2014. 1. 30. · was observed by immunoelectrophoresis of HCG