American Journal of
Obstetrics and Gynecology volume 134
AUWST 1, 1979
Human chorionic gonadotropin-like Presence in normal human tissues YOSHIO ADA
With the use of radioreceptor assay for gonadotropin and a beta-chain radioimmunoassay for human chorionic gonadotropin (hCG), we have been able to demonstrate the presence of hCG-like material in all normal human tissues tested. This activity was completely absorbed by a hCG-antibody &inity column and &as demonstrated to have altered carbohydrate content by studies showing this material did not bind to concanavalin A (con A) -Sepharose affinity columps. In identical studies greater than 90% of placental hCG was bound to con A. However, if placental hCG is rendered carbohydraie free it also does not bind to con A. It is known that carbohydrate-free hCG is cleared rapidly from the circulation and thus possesses liffle or no biological potency in vivo. Accordingly, we hypothesize: (1) that this hCG-like material in normal tissues has the protein structurs of hCG but d&x not possess the carbohydrate moieties of placental hCG and probably has little or no bioactivity in vivo and (2) that the trophoblastic cell is not unique in its ability to synthesize hCG but has developed the ability to glycosylate hCG, transforming a ubiquitous cellular protein into a hormone. hCG might better be called humtin cellular gonadotropin. (AM. J. OEISTET. GYNECOL. 134:729, 1979.)
IT COMMONLY has been believed that human chorionic gonadotropin I(hCG) is secreted by normal or neoplastic trophoblastic cells, by teratocarcinomas,’ and by 10% to 20% of a variety of neoplasms, e.g., carcinoma From the Division of Endocrinology, Department of Medicine, and Department of Pathology, University of Calfork at Los Angeles Medical Center-Harbor, Received,,‘or Accepted
Reprint requests: Dr. W. Odell, Chairman, Medicine, Harbor-UCLA Medical Center, Carson St., Torrance, Calzfornia 90509. 0002-9378/79/150729+05$00.50/O
Department 1000 W.
@ 1979 The C. V. Mosbv
of the lung, pancreas, and colon.2. ” In 1975, Braunstein and colleagues” reported Jhat extracts of normal human testes contain a material indistinguishable from hCG as determined by hCG-P radioimmunoassay (RIA) and concanavalin A (con A) chromatography. in 1976, Chen and associates5 showed that such an hCG material was present in pituitary and urine of a patient with Klinefelter’s syndrdme. In 1977, we reported that extracts of liver and colon from patients dying without cancer contained hCG-like material as identified by hCG-P RIA and gonadotropin radioreceptor assay (RRA) but that this hCG-like material was devoid of 729
2 5 IO hCG and LH (ngl
hCG RADIORECEPTOR ASSA,’
Fig. 2. Correlation (abcissa) 02
IO hCG - RIA (rig/g-Iissue)
tissue 11(X; det~ minctl by RRA (ordinate).
05 I 2 5 IO 20 50 100 hCG end P sub tin,! of hCG (ng)
Fig. 1. A: Homologous
hCG-RIA. Dose-response lines are shown for the hCG reference preparation, for lung and kidney extracts, and for purified human LH. Results are shown as log dose of hCG standard or volume of lung or kidney extracts added per tube versus the logit transformation of the lz51 counts bound with labeled hCG. response. B, = Maximum B = Counts bound in the presence of labeled and unlabeled ligand. Purified human pituitary LH (LER 960) cross reacted to 10% at the point of 80% of B/B,, but did not show parallelism with the hCG reference standard. B: RRA for hCGLH. Dose-response lines are shown for the hCG reference preparation, for lung and kidney extracts, and for the beta subunit bf hCG. Results are shown as log dose of hCG standard or volume of lung or kidney extracts added per tube versus the logit transformation of the response. LI, and B are as defined for A. The purified P-subunit of hCG did not react in the RRA.
carbohydrate.” We report herein more extensive studies on this material from other human tissues which show that in all probability most or all human tissues elaborate hCG with altered carbohydrate content.
of glacial acetic acid. ‘I-tie homoger~ates were agitated in a water bath at X5” C for 30 minutes and centrifuged al 10,000 X g for 50 miriules at 4’ C;. Tllp supernatants were lyophilized and reconstituted in 0.0 I M phosphate-buffered saline (PBS), pH 7.6, at the cqui\~alent concentration of 3 gm of’unextracted tissue per milliliter and centrifuged at 15,000 x g for 30 tninutes. Aliquots of the supernatants were suejected to hCG-p deRIA7 and LH-hCC RRA,’ which were previously scribed.’ (Purified hCG [CR- 1191, the /J-subunit ot hCG [P-subunit of hCG CR-l 131 for iodination. and reference standard and antiserum to tile /3-subunit of hC:G-SBCi
Material and methods Extraction of tissues. Normal human tissues were obtained at autopsy, performed within 12 hours ot death, and rapidly frozen at -70” C until extractiou. The tissues were weighed, minced, and homogenized with a Polyrron at a concentration of 1 gm tissue/.? ml
Arthritis, Metabolism, and Digestive Diseases.) Recol-ered amounts of hCC; (Antituitrin. Parke, Davis Xr Co.) added to the liver, and colon tissues were 6ti.S’i; and 94.3% , respectively. Osmolality and pH were carefullv adjusted and fully tested to ensure that these paramcters did not affect assat results. Con A chrornatogra&y. (;on A-Scpharttse atli nit! chromatography was done utlder conditiorls described .” 14 total
2 To 8 Inl
applied to the column, Ivhich IV;IS i*luted with 20 ml 01’ PBS, follo~ved by PBS conraining 0.2X1 Inethyl-cu-IIalucOPYr;I11osi(lC (M&P) (Calbiochem). IIIW 01‘ the most potent inhibitors of co11 :I-substrate interac-tivn. Fractions conraitling
50 ~1 of
in 2 1111 aliquots I % bo\
x** con, all)iurlitl---I’RS.
LUNG O-4 u
O.ZM W&P -cc
Fig. 3. Affinity chromatography on con A-Sepharose of extracts of kidney (A), lung (B), colon (C), liver(D), stomach (I?), and placenta(F). None to 14.6% of hCG activities of extracts of normal human organs, measured by RIA and RRA, eluted with 0.2M MaGP: and no hCG activity was found after adding 1.5M MaGP or borate buffer. In extract of placenta, a small fraction eluted before McuGP; the majority (>90%) eluted after 0.2M MolGP was added. The void volume of this column by blue dextran was 0.75 ml. RIA, o----o: RRA, o-o.
Affinity chromatography with hCG antiserum. Antibodies to hCG were produced in rabbits immunized with commercial hCG.g The gamma-globulin fraction of the antiserum was coupled to CNBr-Sepharose 4B (Pharmacia) according to the method of Crossley.‘* Affinity columns measuring approximately 2.5 by 1.0 cm were prepared. Adsorption capacity of these columns was assessed by applying trace amounts of
radioiodinated, highly purified hCG alone and with unlabeled hCG in increasing doses ranging from 0.01 to 100 IU. Each dose was followed by 20 ml of PBS. At doses of unlabeled hCG less than 10 IU >90% of unlabeled hCG was adsorbed to the affinity column. At doses of unlabeled hCG of 10 IU and greater, radioiodinated hCG appeared in increasing amounts in the void volume, indicating column saturation. For
Yoshimoto et al.
I. hCG binding
Normal tissue Cancer tissue Placenta Pregnant serum Cancer serum
to con A
6.1 t 1.6 31.2 +- 9.1
(0.0 - 14.6) (4.0 - 86.0) (90.1
hCG-/3 RIA or gonadotropin RRA. 1n contra\t IO !t(X; from placenta or pregnant! bt-rutn. 0’; to i i.li’, oi total hCG activity from extracts of t~or~ttal tissue hutt(1 to con A and eluted with O.2hl M~ltr(;f’: no II(:(; ;t(:i\it\ was found aftct I .SM Ma(;P or borate br~f’lcr u.ti ;~tldcd. That is, X4:1 10 100’4 of the II(:(; from 11orma1 tissues
studies of tissue hCG, atnounts (by RIA) smaller than the dose of hCG which displaced radioiodinated 11% as determined above were used. hCG was eluted from the affinity column using O.lM acetic acid.
Results All extracts of normal kidney, lung, stomach. colon, liver, and heart contained hCG-like material in detectable amounts.* Dose-response lines parallel to hC(; reference standard were produced in both RIA and RRA (representative data are shown in Fig. 1, .,f and B). Since human purified luteinizing hormone (LH) (LER 960) gives a dose-response line which is not pat-ailel to hCG in the RIA and the P-subunit of hC(; doe< not cross react in the RRA, the material extracted front normal human kidney, lung, stomach, colon, liver, and heart resembled hCG, not human LH or the beta chain of hCG. Fig. 2 shows correlation between concentration of tissue hCG quantified by RIA and concentration quantified by RRA in various tissues. The ratio ot RRAiRIA ranged from 2.2 to 5.9. These tissue extracts were applied to the hCG antibody affinity column. All RIA and receptor-active material was adsorbed onto the column. To characterize further the hCG present in normal tissues, con A -Sepharose colutnn studies were done. All of the hCG derived from sera of pregnant women and >90% of hCG from extracts of normal placenta were bound to con A. All this bound hCG was displaced or eluted by 0.2M MaGP. hCG in normal tissues could be fractioned by this affinity chromatography into a con A-nonbinding component and a con A-binding component, both of which were indistinguishable b) *The amount of’ hCG present in tissue extracts is stated in mass units (nanograms) in terms of our highly purified hCG reference preparation. Stated in terms of International Units, this reference preparation contains 10,000 to 15,000 units per microgram or 10 to 15 mlU per nanogram. It is important to note, we believe, that because of its altered carbohydrare structure tissue hCG possesses little or no bioactivity in viva and can be compared to an International Reference Preparation only by in vitro assays in which metabolism of the hCG does not occur. Based on these translations, normal human tissue contains approximately 10 to 120 mIU hCG/gm.
to co11 .4.
in Fig. S, .-I to F aiiti summarized