SPECIFIC AND SENSITIVE DETERMINATION OF PREGNANCY-SPECIFIC β1-GLYCOPROTEIN BY RADIOIMMUNOASSAY

333 The most widely used biochemical indices of placental function are urinary or plasma steroids and hp.L. The major advantages of measuring a placental protein are firstly, the molecules circulate in their native state, secondly the day-to-day variation in plasma concentrations in any individual is at least threefold lower than for œstrogens,18 and thirdly the assays themselves are rapid and more precise. hp.L. has been shown to be a better predictor of fetal risk than plasma œestriol,19 but whether S.P.1 will prove superior to hp.L. remains to be determined. Theoretical considerations suggest that it might, inasmuch as the longer half-life (30-40 hours as against 15-20 minutes) would tend to yield more stable concentrations from hour to hour or day to day, and a single determination would therefore better reflect the absolute rate of placental synthesis. This idea has not as yet been tested. However, in one prospective survey of 850 pregnancies, 45% with I.U.G.R. had hp.L. concentrations below the tenth centile of the normal range, lU whereas in the present study S.P.1 concentrations were reduced in over 70% of l.U.G.R. pregnancies. The findings presented here are highly promising and provide a clearcut indication for further studies. We thank Miss K. Lo, Miss M. Leighton, and Dr J. D. Lewis for assistance with the collection of serum samples and statistical analysis 01 the data. Dr H. Bohn kindly provided the purified s.p., standard. J.G.G. was supported by the World Health Organisation.

Preliminary Communication SPECIFIC AND SENSITIVE DETERMINATION OF PREGNANCY-SPECIFIC &bgr;1-GLYCOPROTEIN BY RADIOIMMUNOASSAY A New Pregnancy Test Y. B. GORDON T. CHARD

J. G. GRUDZINSKAS DIANA JEFFREY

Departments of Reproductive Physiology and Obstetrics and Gynœcology, St. Bartholomew’s Hospital Medical College and the London Hospital Medical College, London A specific and highly sensitive radioimmunoassay (R.I.A.) for determination of pregnancy-specific &bgr;1-glycoprotein (S.P.1) in human plasma, urine, amniotic fluid, and breast milk has been developed. The minimum detection limit of S.P.1 was 8 µg/l of sample. The assay was applied to plasma and/or urine samples from 8 women in early pregnancy. S.P.1 was detected in the plasma of all three women obtained within 14 days of ovulation, the earliest positive result being at 7 days. In another woman plasma was negative at day 18 and positive by day 22. In the three remaining women plasma S.P.1 was detected when measured within 24 to 36 days of ovulation. S.P.1 was detected in four urine samples obtained between 20 and 28 days after ovulation. S.P.1 was also measured in breast milk, amniotic fluid, cord blood, and plasma of women with ectopic gestation and trophoblastic disease. It is suggested that the assay of S.P.1 may have important advantages over existing systems for detection and monitoring of early pregnancy.

Summary

INTRODUCTION

THE

measurement

of specific

proteins produced by the

Requests for reprints should be addressed to T.C., St. BarthoHospital Medical College, West Smithfield, London EC1A

lomew’s 7BE.

REFERENCES 1. Gruenwald, P. in The Placenta (edited P. Gruenwald), p. 1. Lancaster, 1975. 2. Beazley, J. M., Underhill, R. A. Br. med. J. 1970, iv, 404. 3. Campbell, S. J. J. Obstet. Gynœc. Br. Commonw. 1970, 77, 1057. 4. Gordon, Y. B., Landon, J. Rev. perinatal Med. (in the press). 5. Klopper, A. (editor) Plasma Hormone Assays in Evaluation of Fetal Well-

being p. 20. London, 1976. Diczfalusy, E. Physiology and Genetics of Reproduction (edited by E. M. Coutinho and F. Fuchs); p. 385. London, 1974. 7. Lin, T. M., Halbert, S. P., Kiefer, D., Spellacy, W. N., Gall, S. Am. J. Obstet. Gynec. 1974, 118, 223. 8. Bohn, H. Arch. Gynäk. 1971, 210, 440. 9. Towler, C. M., Horne, C. H. W., Jandial, V., Campbell, D. M., MacGillivray, T. Br. J. Obstet. Gynæc. 1976, 83, 775. 10. Thomson, A. M., Billewicz, W. Z., Hytten, F. E. J. Obstet. Gynæc. Br. Commonw. 1968, 75, 903. 11. Greenwood, F. C., Hunter, W. M., Glover, J. S. Biochem. J. 1963, 89, 114. 12. Leek, A. E., Ruoss, C. F., Kitau, M. J., Chard, T. Br. J. Obstet. Gynæc. 1975, 82, 669. 13. Gordon, Y. B., Ratky, S. M., Sola, C. M., Lewis, J., Baker, L. R. I., Chard, 6

T. ibid. p. 958. Chard, T. in Plasma Hormone Assays in Evaluation of Fetal Wellbeing. (edited by A. Klopper), p. 1. London, 1976. 15. Tatra, G, Breitenecker, G , Gruber, W. Arch. Gynäk. 1974, 217, 383. 16 Horne, C. H. W., Jandial, V., Towler, C. M. International Symposium of Hypertensive Disorders in Pregnancy, 1976 (in the press). 17. Lin, T. M., Halbert, S. P., Spellacy, W. N. Am. J. Obstet. Gynec. 1976, 125,

14.

17. 18. 19.

Klopper, A.

Personal

communication.

Edwards, R. P., Diver, M. J., Davis, J. C., Hipkin, L. J. Br. J. Obstet. Gynœc. 1976, 83, 229. 20. Gordon, Y. B., Lewis, J. D. Unpublished.

placental trophoblast and secreted into the plasma and urine of pregnant women is widely used to detect early pregnancy and as an index of placental function in late pregnancy. The isolation of a new trophoblast-specific protein (S,P’l)l has promoted investigation of this material as a potential diagnostic marker.;/-4 S,P’1 is a &bgr;1-glycoprotein synthesised by the syncytiotrophoblast, with a molecular weight of approximately 90 000 daltons and a carbohydrate content of 29.3%. Most existing assays for s.p., have been based on immunoprecipitation systems, which are relatively insensitive and which therefore do

not allow detection before 8 weeks’ The gestation. development of a radioimmunoassay for leads to more precise measurement of (R.I.A.) S,P’l and a considerable increase in sensitivity. S.P.1 We describe here a highly sensitive R.I.A. for S.P.1 which can be applied to plasma, urine, amniotic fluid, and breast milk. We used this assay to determine s.p., at or around the time of implantation of the human zygote.

MATERIALS, METHODS,

AND

SUBJECTS

Highly purified S,P’l (lot 8VE), kindly supplied by Dr H. Bohn, Behringwerke AG., Marburg, was used for standardisation and iodination. Radioiodination was performed by the chloramine-I method with 125I-labelled sodium iodide (IMS 30, Radiochemical Centre, Amersham).6 The reaction mixture was purified by chromatography on ’Sephadex G100’, and mean specific activity was 80 fLCi per ug. The antiserum to S,P’l was also obtained from Behringwerke AG. (lot 2998C). The standards for the assay ranged from 8 to 500 µg/l and consisted of pooled "late-pregnancy" serum diluted in horse serum and calibrated against purified s.p.,. This range embraces the values of greatest physiological and clinical interest in early pregnancy. Samples of two control pools, representing high and low values, were prepared in a similar manner. In the R.I.A. procedure the final incubation volume was 500 µl. 50 µl of standard or sample, 350 µl of 10% horse serum

334 in 0.05 mol/l phosphate buffer (pH 7-5), and 50 µl antiserum t an initial dilution of 1 in 10 000 were incubated at 40C for 48 hours. 50 µl of 125I-labelled s.p., (1 ng) was then added. After incubation for a further 48 hours at 4°C, the antibodybound and free antigen fractions were separated by addition of 1 ml of 20% (w/v) polyethylene glycol 6000.7 In assays on urine samples, an additional 50 µl of horse serum was added immediately before separation. Over a period of 2 months the within-assay variation of the control pools was 4%, and the between-assay variation was 9%. Control blood-samples were obtained from fifty normal male volunteers, and paired daily blood and urine samples from a female volunteer during a normal menstrual cycle. Either blood (four women), or urine (one woman), or both (three women) were collected at frequent intervals from eight women wishing to conceive. The probable time of ovulation and fertilisation was determined by one or more of the following : menstrual history, basal-body-temperature records, serial plasma-progesterone levels, and the therapeutic use of ovulation stimulants. Blood was also obtained from a patient with a ruptured tubal gestation, from the umbilical cord at a premature delivery and a term delivery, and from a patient with a hydatidiform mole. Blood-samples were collected in heparinised tubes, the plasma separated by centrifugation and stored at -20°C until assay. Amniotic fluid obtained by amniocentesis, and breast milk obtained by manual expression were also assayed. RESULTS

The assay had

a

minimum detection limit of 8

ug/1

1-Standard dose response curve for S.P., which was determined by means of 125I-S.P.1 and rabbit antiserum to S.P.,

Fig.

and a dynamic range of 8-500 ug/l (fig. 1). No inhibition of binding was observed in the presence of an excess of human chorionic gonadotrophin (hc.G.), human luteinising hormone (hL.H.), human follicle-stimulating hormone (hF.S.H.), human growth hormone (hG.H.), human prolactin (Pr), or alpha-fetoprotein (A.F.P.). S.P.1 was not detected in the plasma of the male volunteers, nor in the paired daily plasma and urine samples collected during a normal menstrual cycle. S,P’1 was detected in the plasma and/or urine of all the pregnant subjects studied (fig. 2). S.P’1 first appeared in plasma 7 days after fertilisation and in urine at 20 days. In all cases with serial samples there was a sharp increase in S,P’1 concentrations after the first appearance of the protein. s.p., was also detected in the plasma of patients with ruptured tubal pregnancy and hydatidiform mole, in umbilical-cord blood, amniotic fluid, and breast milk

(see accompanying table).

Fig.

2-Concentrations of S.P., in in early pregnancy.

plasma and

urine from I

women

DISCUSSION

We detected S,P’I in plasma and urine in early preg nancy, in abnormal early pregnancies, and in con blood, amniotic fluid, and breast milk. The measure ment is possible because of the development of a high ! sensitive R.I.A., the specificity of which has been closel defined in relation to a range of purified materials, and blood and urine samples from non-pregnant subjects. In this study S.P.1was detected in the plasma of thre women obtained within 14 days of ovulation, the earlies detection being at 7 days; however, in one subject negative result was obtained at 18 days followed by rise at 22 days (fig. 2). In the three remaining women plasma S,P’I was detected when tested for between day 24 and 36. These preliminary findings would sugges that S.P.1 can be detected in maternal blood immediatel after implantation in some but not all cases. S.P.1 wa first detected in urine 20 days after the likely date o conception. The early appearance of s.p., in the plasm of pregnant women establishes the measurement ofs.p. as a test of early pregnancy with potential advantage similar to those of the hc.G. 3-subunit assay. Thus it i likely that the determination of S.P.1 will provide valu able additional biochemical evidence in those situation in which the earliest possible detection is necessary, fo example infertility, ovulation induction, bromocriptin CONCENTRATIONS OF

S,P’l

IN VARIOUS BODY-FLUIDS DURING

NORMAL AND ABNORMAL PREGNANCIES

335

therapy, and menstrual regulation. The relatively late appearance of S,P’I in the urine of pregnant women suggests that the test will offer few advantages over existing urine tests based on determination of hc.cJ.. The very early appearance of s.p also suggests that it might be an appropriate target for the immunological destruction of an early pregnancy.1I This possibility has already been evaluated in the case of hc.G.’’ The great advantage of the use of s.p as a target is that it would be free from the potentially deleterious effects on the maternal pituitary gland which might result from crossreaction between hc.G. and hL.H. In addition, the use of a sensitive R.t.A. has permitted determination ofs.p in situations which have not been previously examined by means of a quantitative technique, for example, ectopic gestation, hydatidiform mole, amniotic fluid, breast milk, and the fetal circulation. In the latter, the levels are some 1000 times lower than in the maternal circulation, indicating that as with other placental proteins, the bulk of the secretion is into the intervillous space. Finally, there is a good prospect that measurement of s.p., may prove to be a valuable marker

Hypothesis IS BURKITT’S LYMPHOMA RELATED TO PERINATAL INFECTION BY EPSTEIN-BARR VIRUS?

GUY DE-THE

Unit of Biological Carcinogenesis, International Agency for Research

on

Cancer, Lyons, France

Burkitt’s lymphoma (B.L.) is very restricted geographically, whereas the Epstein-Barr virus (E.B.V.), a very likely causal factor, exists all over the world. It is proposed that perinatal infection (either transplacental, or, more probably, neonatal) is an important risk factor for the development of B.L. This hypothesis arose when the epidemiological characteristics of E.B.V. infection were compared in different parts of the world and is supported by the preliminary results of a seroepidemiological prospective study of B.L. E.B.V. would behave in a manner similar to that of animal tumour viruses whose oncogenic potential is greatly enhanced by neonatal infection.

Summary

Association between Epstein-Barr virus and Burkitt’s

lymphoma EVIDENCE for the association between the herpesvirus discovered by Epstein and Barr (E.B.V.)’ and the childhood lymphoma described by Burkitt1 in equatorial Africa rests on :(a) the presence of viral markers (E.B.V./D.N.A. and E.B.v. nuclear antigen [E.B.N.A.]) in tumour cells;3 4 (b) the high serological reactivities observed in Burkitt’s lymphoma (B.L.) patients to a series of virally determined antigens-viral capsid antigen (v.c.A.), early antigen (E.A.), E.B.N.A., and soluble complement-

fixing antigen (c.F./S.);5 (c) the capacity of the E.B.v. to "immortalise" B lymphocytes;67 (d) the in-vivo oncogenic potential of E.B.v. in some New World primates.8-12

in the treatment of trophoblastic disease since, in conhc.G., there would be no ambiguity with results at low concentrations because of cross-reaction with materials present in the normal non-pregnant state. trast to

We wish to thank Mrs M. Welch, Prof. A. Klopper, Dr D. M. Saunders, I)r C. H. Mortimer, and Dr M. ’I’horner tor assistance with sample collection. Dr H. Bohn kindly provided the purified .I"1 standard. was supported by the World Health Organisation. Requests tor reprints should be addressed to T. C., St. Barthotomcw’s Hospital Medical College, West Smithfield, London EC1A 7H)’

REFERENCES S 1. Bohn, H. Arch. Gynäk. 1971, 210, 440. 2. Bohn, H., Protides biol. Fluids, 1976, 24, 117. 3 Towler, C. M., Horne, C. H. W., Jandial, V., Campbell, D. M., McGillivray, T. Br. J. Obstet. Gynœc. 1976, 83, 775. 4. Gordon, Y. B., Grudzinskas, J. G., Jeffrey, D., Chard, T. Lancet, 1977, i,

331. 5. Towler, C. M., Jandial, V., Horne, C. H.

W., Bohn, H. Br. J. Obstet. Gynœc. 1976, 83, 368. 6. Greenwood, F. C., Hunter, W. M., Glover, J. S. Biochem. J. 1963, 89, 114. 7. Leek, A. E., Ruoss, C. F., Kitau, M. J., Chard, T. Br. J. Obstet. Gynœc. 1975, 82, 669. 8. Bohn, H., Weinmann, E. Arch. Gynäk. 1974, 217, 209. 9. Stevens, V. C. in Immunization with Hormones in Reproduction Research, (edited by E. Nieschlag); p. 217. Amsterdam, 1975.

The determination of the nature of the association between E.B.v. and B.L. is not easy, since E.B.v. infection is ubiquitous, while B.L. represents a rare event in very restricted geographical areas.’3 After the discovery that infectious mononucleosis (i.M.) was caused by a late primary E.B.v. infection in adolescents of high socioeconomic class, 14 the question was raised as to whether B.L. represented a malignant i.M. and was the consequence of a late (with regard to the general population) E.B.v. infection. To answer this question, a prospective seroepidemiological study on B.L. was carried out by the International Agency for Research on Cancer (I.A.R.C.) in the West Nile district of Uganda, where a cohort of 45 000 children (aged 1-8 years) were visited, registered, bled once, and carefully followed up to detect all tumour cases.t5 Active B.L. case detection in this population and comparison of the serological profiles of "pre-B.L." sera-i.e., sera taken from children who subsequently developed the disease-with the sera from children representative of the general population allowed the testing of the following hypotheses: (i) E.B.v. is not causally related to B.L.-according to this "null" hypothesis the serological profile of the pre-B.L. sera should not differ from that of the controls. (ii) B.L. results from a relatively recent primary infection by E.B.v. This hypothesis, based on the infectious mononucleosis model, predicts that pre-B.L. sera, when collected before the "incubating period", should be E.B.v.-negative and when collected during the incubating period should reflect a recent infection. (iii) B.L. occurs in a child who has had a long and heavy exposure to E.B.v. In such a case, the pre-B.L. sera should exhibit much higher E.B.v. antibodies than control sera.

Comparative epidemiology of B.L. and E.B. V. infection The main epidemiological characteristic of B.L. is its relation to the environment, as demonstrated by its limited geographical distribution within the tropical belt,13 the role of temperature, humidity, and altitude,’6 the time-space clustering," and seasonal variation of