CA1307736C - Method of assay of inhibin - Google Patents

Abstract

Abstract A method of immunoassay for the estimation of inhibin in an inhibin-containing sample which comprises the step of using an antibody directed against inhibin. Preferably, the antibody is contained in an antiserum raised by injecting an animal with an antigen selected from the group consisting of naturally-occurring or recombinant inhibin, or sub-units, fragments or derivatives thereof. The assay may suitably be a radioimmunoassay, a fluorescence-based immunoassay, or an enzyme-linked immunosorbent assay using labelled 58kD or 31kD inhibin as tracer. Tracers and standards for use in the assay are described and claimed.

Description

i3~73~

METHOD OF ASSAY OF INHIBIN
This invention relates to methods for assay of inhibin, and in particular to methods or immunoassay o inhibin .

Two forms of inhibin from bovine follicular fluid 10 have been recently purified to homogeneity, with molecular masses of 58kD and 31kD ~International Patent Application PCT/AU/85/00119 and Robertson et al 1985, 1986). Under reducing conditions both forms consist of two subunit~ with molecular masses of 43kD and 15kD, and 20kD and 15kD
15 respectively. Their primary amino acid structures have been ~k q~

0~'736 elucidated following cloning and analysis of cDNA species derived from bovine granulosa cell mRNA (International Patent Application PCT/AU86/00097; Forage et al 1986). These studies indicate that 31kD inhibin is a processed form of the 58kD
molecule. 31-3lkD inhibin molecules with similar subunit structures to bFF inhibin have been isolated from porcine follicular fluid (Miyamoto et al 1985, Ling et al 1985) and sequenced (Mason et al 1985).
Currently inhibin activity is measured by a variety of in vivo and in vitro bioassay systems (Baker Ç~ 31 1981).
These systems are time consuming, expensive, have limited sensitivity and precision and are of limited practicability in their application to large sample numbers (Baker et al 1981, Lee et al 1982).
Summaxy_of the Invention The present invention relates to more convenient assays for the estimation of inhibin than have heretofore been possible. The preferred assays of the invention are radioimmunoassays and the following description, whilst being directed to the preferred assays, should not be construed as limiting the invention to radioimmunoassays. Other assays within the scope of the invention include ELISAs, immunoassay based on fluorescence detection, and related assays relying on polyclonal and monoclonal antibodies against inhibin.
Prefe~ed Embodiments of the Invention According to one aspect of the present invention there is provided a radioimmunoassay for determining inhibin suspected to be present in a biological sample, comprising the steps of:
(a) contacting the sample with an antibody directed against inhibin and incubating at 4- to 30-C for a period of 4 hours to 4 days, (b) adding 125I-inhibin and incubating for a period and at temperature selected from the group consisting of at overnight room temperature, 48 to 72 hours at 4-C or 16 hours at 30-C, (c) adding a second antibody and incubating for 30 minutes to 24 hours at 4-C, (d) separating precipitated material, and (e) measuring bound 125I-labelled inhibin in the precipitate.

13~)'77~

Preferably the antibody is contained in an antiserum raised by injecting an animal with an antigen selected from the group consisting of naturally-occurring or recombinant inhibin, or sub-units, fragments or derivatives thereof.
Particularly preferred antigens include preparation including inhibin, purified bovine 58kD inhibin, ,~

130'7736 purified bovine 31kD inhibin, human inhibin, or human or bovine inhibin or fragments thereof produced using recombinant DNA
technology.
Suitable animals include mammals such as mice, rabbits, horses, donkeys, dogs, sheep, and goats, and birds such as chickens.
Alternatively a monoclonal antibody or an IgG directed against any of the aforesaid inhibins may be used.
Most preferably the antibody is capable of neutraliz-ing inhibin bioactivity.
Preferably the immunoassay is further characterized by the step of using labelled 58kD or 31kD inhibin as tracer.
More preferably said tracer is labelled with 125iodine (125I) or with an enzyme, or with a fluorescent marker.
Preferably the assay is a radioimmunoassay or an enzyme-liked immunosorbent assay (ELISA), or a fluorescence-base immunoassay.
The invention provides a method for measuring inhibin in samples such as follicular fluid or serum from various species (including humans) wherein concentrations of inhibin in stan-dards are used to derive the concentration of inhibin in the follicular fluid or serum by competitive binding of 125I
labelled inhibin and inhibin from test samples with bovine 58kD
inhibin antiserum, followed by precipitation and counting of bound 125I labelled inhibin.
The preferred specific radioimmunoassay syste~ for inhibin of the invention is applicable to bovine and human follicular fluid and serum, and can employ an antiserum against (purified bovine) 58kD inhibin with iodinated 31kD or 58kD
inhibin as tracer.
According to a second aspect of the invention, there is provided a method for preparation and purification of 125I-labelled inhibin tracer which comprises the steps of iodination of inhibin using a Chloramine T procedure and purification of 5I-inhibin by an affinity fractionation step.
Preferably the affinity fractionation step uses Matrex Red A.
Preferably the purification procedure additionally comprises a gel filtration step.

~RR.ll '87 07:50 CLE~ENT H~CK ~UST +613 529 62 ~ 3 ~ ~ 7 ~ ~ P.07 According to a third aspect of the invention there is provided an assay standard selected from the group con~isting of naturally-occurring or recombinant inhibin, or fragments or derivatives thereof. Preferably the ~tandard S display~ parallelism in the assay with the samples under test.
Particularly preferred standards include bovine 31kD
inhibin, and partially purified or purified human inhibin.
The conditions of the assay, in particular incubation times, may be varied in order to attain desired 10 levels of sensitivity.
A form of the radioimmunoassay modified for increased sensitivity compri~es:
incubating s~mple and antiserum for 4 days at 4C, followed by the addition of 125I-31kD inhibin tracer, 15 incubating for 3 days at 4C and then adding second antibody, ~' precipitating, and counting bound 125I labelled inhibin.
According to one particularly preferred embodiment, suitable for measurement of inhibin in human serum samples, the tracer is 125I-31kD inhibin, and incubation with tracer i5 20 performed at elevated temperature (30C) in the presence of inhibin-free serum, in order to minimize non-~pecific effects.
Suitable sources of inhibin-free serum include steers or other castrated male animals, oophorectomized women, women with premature ovarian failure, and post-menopausal women.

25 Brief Description of the Drawings Figure 1 shows the fractionation of 125I-58kD and 125I-31kD inhibin on analytical SDS-PAGE under reducing conditions.
Figure 2 shows the time course of immunization of a 30 rabbit with 58kD inhibin.
Figure 3 ~hows the in vitro neutralization of bFF
inhibin by an antiserum raised to 58kD inhibin.
Figure 4 shows the radioimmunoassay dose response curves of bFF, hFF, purified 58kD and 31kD inhibin and bovine 35 granulosa cell culture medium (BGCM) using either 125I-31kD or 125I-58kD inhibin as tracers.

13~7~
iRf~. 12 '97 13:29 CLEI`~ENT HQCK ~LIST +613 529 6296 P.04 Pigure S ~ho~s the profi~e of inhi~in in vitro bioactivity and immunoactiv~ty follo~ing f~actionation of bFF
through the various ateps of the inhibi~ purification procedure of Robert~on et al tl986~.
S Figure 6 shows non-reduced SDS-PAGE profile~ of 125I-58kD and l25r-3lkD inhibin following incubation wi~h bPP
and serum under condition~ used in the RIA of ser~m inhib~n.
Figure 7 shows the effect of temperature on the binding of 125I-3lkD inhibin to the AntiSerum.
Figure 8 shows logit-log dose re~ponse lines of bovine and human serum, in the plasma RIA system employing 5~-31kD inhibin as tracer.
F~gure 9 ~hows the ovulation induction regime and ~erum level~ of FS~, ~H, inhi~in and oestradiol (E23 in 15 twenty-six women involved in an In Vitro Fertilisation ~IVF) programme and one normal ~oman (FL#27).
Figure 10 is a compari~on of plasma E2 and inhibin levels plo~ted for some of the data in Fig~re g.
Figure 11 shows the correlation between the number 20 of ova produced and ~2 or ;nhibin levels in ~erum.
Figure 12 shows the corxelation bet~een the number~
of ovariAn follicle~ detected ultrasonioally and peak inhibin level8 ~n ~erum.
Figure 13 shows inhibin, FS~, proge~terone and 25 oeatradiol concen~rations in the ser~ of non-pregnant subject~
~n ~ 16) on the days following oocyte r~trieval.
Figure 14 shows inhibin, FSH, progesterone ~nd oe~tr~diol concentra~ion~ in the ~era of pregnant subj~cts (n ~ 3) on ~he days ollowing oocyte retrieval.
Figure 15 shows the relationship between serum inhlbin and PSH duxing the lu~eAl phAse o~ non-pregnant subjoct~. For thi~ analysis, non-detectable inhibin values ~n 29) were asaigned the ll~it of a3say sensitivity.
Figure 16 show~ inhibin, FS~, LH, oestradiol and progesterone conoentrations in the ~era of normal women during the menstrual cyole, as~a~ed u~ing anti-31kD inhibln.

13~7'~6 ~R.11 '87 07:53 CLEI`1ENT Hf~CK ~UST +613 529 6Z96 P.09 Detailed Description of_the Invention Abbreviations bFF : bovine follicular fluid hFF : human follicular fluid oFF : ovine follicular fluid oRTF : ovine rete testis fluid HPLC : high performance liquid chromatography SDS-PAGE: sodium dodecyl sulphate polyacrylamide gel electrophoresis RIA : radioimmunoassay SS : steer serum ~S : bull serum CS : cow ~erum ~MS : human po~t-menopausal ~erum HFP : human female plasma Preparations (a) Purificition of bFF inhibin ~he purification of bFF 31kD and 58kD inhibin was based on the procedures described previously (Robert~on et al - 20 1985, 1986)(Figure 5):
(a) bFF was fractionated on a Sephacryl S200 (9 x 90 cm) gel filtration column in 0.05M ammonium acetate pH
7 Ø
~b) The void volume fraction from (a) was precipitated at pH 4.75 and ractionated on Sephadex G100 (9 x 90 cm) in 4M acetic acid.
(c) and ~d) : Peak I tS8kD inhibin) and Peak II ~31kD inhibin) frActions from ~b) were fractioned on an RPSC
Ultrapore column ~0.46 x 7.6 cm, Beckman) using a 0-50~ acetonitrile gradient in 0.1% trifluoracetic acid. In Figure 5 the continuous line indicates optical density at 280 nm (a) and (b) and 254 nm (c) and (d).
Hatched area denoteu inhibin bioactivity.
o---o RIA with 125I-58k~ inhibin as tracer.

13~7736 1~R. 11 ~87 07 5~ CLEMENT H~CK ~UST +613 529 6Z96 P. 10 o _ o RIA with 125I-31kD inhibin as tracer.
Vo = void volume.
BSA - bovine serum albumin (mol. wt 67,000).
OVA ~ ovalbumin (mol. wt 43,000).
The purified inhibin wa~ stored in SDS
electroelution buffer (approx. 3~ SDS in 10 mM NH4Hco3) prior to iodination. For bioassay, sample~ were methanol precipitated at -20C in order to remove SDS and solubilized by heating at 37C for 1 hour and sonication.
Similar profiles of both bio- and immunoactive inhibin were observed at each stage of the inhibin purification procedure. The biological to immunological activity ratios for a number of purified 31kD and 58kD inhibin preparations using both tracers in the radioimmunoassay ranged 15 from 0.30 - 0.43.

(b) Sample Preparation Human follicular fluid was obtained at oocyte collection in the in vitro fertilisation programme at the Queen Victoria Medical Centre/Epworth Hospital, Melbourne. It 20 was charcoal treated ~100 mg/ml dextran-coated charcoal for 1 hour at 4C), lyophillsed, stored at -20C and resolubilized prior to assay by æonication in assay buffer or culture medium. Ovine follicular fluid (oFF) was obtained by aspiration of ovaries collected at a local abattoir. Ovine 25 rete testis fluid (oRTF) is a lyophilised inhibin preparation (Baker et al 1985). Rat ovarian extract was a charcoal-treated rat ovarian cytosol preparation. Details of the biopotencies of these inhi~in preparations are outlined in Table 1.
Blood was collected in lithium heparin tubes from 40 women undergoing ovulation induction-therapy ~clomiphene citrate and human menopausal gonadotrophin treatment) with plasma estradiol level~ at time of plasma collection ranging from 40-2,900 pg/ml. Equal aliquots of serum from each 35 subject were combined to produce a plasma pool designated as 13~ '7~6 Ri;! . 11 ' 87 07: 55 CLEMENT H~CK RUST +613 5Z9 6296 P. 11 Human Female Plasma pool (HFP). Plasma from four post-menopausal women aged 52 and over were combined to give a pool designated Post-Menopausal Serum pool (PMS).
30vlne blood, ovaries and testes were collected on S ice from a local abattoir and processed within one hour. All samples were stored at -20C after snap freezing in solia C02/ethanol. Blood pools from adult intact (BS, n = 9) and castrate (SS, n = 1) male, and female (CS, n = 10) cattle were allowed to clot overnight at 4C prior to centrifugation and 10 storage. Bovine ovarian follicles were hemisected and granulosa cells were collected by aspiration and cultured for 40 hours at a concentration of 105 viable cells per well (Costar 48 well plate) in 400 ~1 DMEM/F12 complete medium.
Media were stored frozen at -20C prior to assay. Testes from 15 four bulls were decapsulated and homogenised in equal w/v Dulbecco's phosphate buffer using an Ultra-Turrax tissue disperser ~Janke and Xunkal XS, Staufen FRG) and centrifuged at lOO,OOOg x 1 hour at 4C and stored at -20C. Prior to assay the supernatants were charcoal treated with an equal 20 volume of 1% Norit A in Dulbecco's phosphate buffer and incubated at 4C for 30 minutes prior to centrifugation and bioassay ~Au et al 1983).

Immunizatlon Procedure Purified 58kD inhibin ~14 ,ug in 500 ~1 Dulbecco's 25 phosphate buffer) was emulsified in an equal volume of adjuvant ~Marcol 52 [Esso, Australia]: Montanide 888 [S.E.P.P.I.C., Paris] in the ratio 9:1) and injected into an intact male New Zealand white rabbit 4 intramuscular and one subcutaneous sites. Two booster injections of 14 ~g under the 30 same conditions were given at aix weeks and one year. Serum was collected throughout for assessment of its in vitro neutr~lization activity, its ability to bind iodinated inhibin and for plasma FSH estimations.
A similar procedure was used to raise antibody directed against purified 31kD inhibin.

l~R.ll '87 07:57 CLEMENT HPCK ~IUST +613 5Z9 6Z96 ~ P.12 Analytical SDS Polyacr~lamide Gel Electrophoresis Sera and bFF were incubated at various temperatures in an equal volume of 100 mM phosphate buffer pH 7.4 containing 0.15M NaCl, 0.1~ Tr~ton X-100 and either 0.5~ BSA
5 for studies with 125I-31kD inhibin or 0.5% Polypep for 125I-58~D inhibin. Equal volumes (5~1) of ~ample and 10~ SDS
and Dulbecco's Phosphate buffer pH 7.4 (30 ~ul) were placed in a boiling water bath for 2.0 minutes then in ice. Ten microlitres of bromophenol blue (0.006%) in glycerol (62.5~ in 10 H2O) was then added and the mixture centrifuged prior to electrophoresis on 12.5~ slab gels (3 hours, 20-30 mA).
Protein molecular weight markers were either co-electrophoresed with the iodinated sample in the absence of bPF and serum or on a separate track in their presence. The 15 gels were fixed and stained overnight in ethanol: H2O
formaldehyde (180: 420: 100) containing 0.1% Coomassie Brilliant Blue. Each track was divided into 50 2mm slice~ and counted in a gamma counter.

Reversed-phase HPLC

2 1-Inhibin was applied to an Ultrapore RPSC column (0.46 x 7.5 cm, Beckman, Berkeley, Ca., USA) and fractionated using a 30 min linear gradient o 0.50~ acetonitrile in 0.1 trifluoroacetic acid at 1 ml/min and 0.5 ml fractions using Waters HP~C apparatus (model 6000A pumps and a model 660 25 Programmer, Nilford, Mass., USA).

In Vitro Bioassay Inhibin activity was determined using an in vitro bioassay based upon the dose-dependent suppression of ~SH cell content in rat pituitary cell cultures utilizing a parallel 30 line bioassay design (Scott et al 1980~. The charcoal-treated bovine follicular fluid preparation employed a lymph reference preparation with an arbitrary unitage of 1 unit/mg (Scott et al 1980) Hormone Assays 1;~0~36 1QR,11 '97 07:5a CLEI`1ENT Hl:~CK RUST +613 5Z9 6296 P.13 (i) Rabbit FSH Radioimmunoassay Rabbit FSH was determined using an RIA kit kindly provided by Dr. A.F. Parlow (Torrance, Ca, USA) employing 15~ polyethylene glycol to separate bound and free hormone. The sensitivity of the a~say was O.9 ng/ml using rabbit FSH AFP.538.C as standard.
The within assay coefficient of variation was 8.1%
and all samples were assayed in the one assay.
(ii~ Rat FSH Radioimmunoassay Rat FSH generated by the pituitary cells in culture was mea6ured by a specific radioimmunoassay using reagents supplied by the NIAMDD. 125I-rat FSH (I5) was used as tracer and FSH RP-2 used as standard.
The within-assay coefficient of variation was 7%.
(iii) RIA of hum~n hormones Serum PSH was measured by RIA ~Amerlex-M, Amersham, USA) using 2nd IRP FSH as standard with an interassay CV of 7.0% from 31 assays. LH was measured by 2IA ~LH RIA, Diagnostic Products Corp., ~.A., USA) using the 2nd IRP LH as standard with an interassay CV of 10.1~ from 31 aCsays. Both oestradiol and progesterone were measured using RIA
(Coat-a-Count, Diagnostic Products Corp., ~.A.) with interassay CVs of 8.7~ and 8.1~ respectively from 150 assays. Serum beta subunit of hC~ was measured by RIA (B-hCG RIA-Quant, Mallinckrodt Inc. St.
Louis, USA) using the hCG 2nd IS as standard with an intera~say CV of 10.4% from 30 assays.

Calculations The RIA dose-response curves were lineari~ed using a logit-log dose transformation. Parallelism was asse~sed from a comparison of slope values of dose-response curves using the multiple range test for groups of unequal size (Kramer, l9S6) or by paired t-test. Potency estimates were determined using 35 standard parallel line bioa~ay statistics. In situations where non-parallelism was observed between dose response lines 1hF~. 12 !~37 1~:30 rLEMENT HI~CK hlJST +613 5Z9 ~2g~ P.05 of unknown and standard preparation,s, potency estimates were determined ro~ the rat~o of their ~D50 values. The sensitivity (~Dlo) was deined as the mass of hormone required to give lO~ diçplacement in the as~ay whil6t ED50 corresponded 5 to the mass required for 50~ displacement.
The index of precision (Gaddum (l933~; F~nney (19641) w~s used ~o describe assay precision. The between assay variati~n was calculated from the coef~icient of variation o~ the repeated meaaurement of a partia~ly purified lO inhibin pr~paration. The dissociation cons~ant (KdiS) was determined by scatchard anal~sis using 125I-tlvrmor~e and incre~sin~ ~mounts of unlabelled ho~monc. ~ e mass of 5I-hormone l~sed in the analy~is was determined from its ~pecific activity (~ g).

The invention will now be illustrAted by reerence to thc following non-limitin~ examples.

~xample 1 Antiserum Char~cterizat~on Antisera against 58kD and 31kD inhibin were character~
~zed by showing that following immunization, parall-l changes in plasma FS~ and inhib~n an~ibody titre were ob~erved, indicating inhibin neutralization in vivo. ~he antisera neutralized bFF, hFF and purified 31kD and 58kD inhibin activity in an in vitro bioassay. The results des~ribed below refer to anti-58kD
inhlbin, but similar result~ wers obtained using anti-31kD
inhibin.

(a) Response In Vitro to ~mmunization Following the fir~t booster (Fig. 2a) the antibody titre was assessed by the ability of the antibody to neutralize inhibin bioactivity in vitro and its capacity to bind l25I-58kD inhi~in.
A sharp pa~allel elevation in these activitieS was ob-~erved between 1-8 weeks ~ost-boo~ter in~ection. During ~his period significant (P 0.05) elevation in serum PSH ~6.~3 + 0.95 ng/ml, n ~ 5 vs 4.97 ~ 0.87 ng~ml, n ~ 6 (mean + l S~)) was noted.
Following the second boostor ~Fig. 2b), an immediate and sustained elevation of serum 125-31kD inhibin-binding capacity and serum PS~ was observed.

1RR.11 '87 013:01 CLEI`1ENT H~CK ~UST +613 5~9 6296 7~6 P.15 Basal levels of serum FSH were assessed from the mean + 2 S.D.
of 14 observations over the preceeding five months ~hatched area Figure 2b).
These results indicate that purified inhibin from 5 bovine follicular fluid c~n be u~ed to immunize rabbits, producing an antiserum which has the capability of neutralizing inhibin bioactivity. The elevated levels of plasma FSH in the rabbit observed during the period of pea~
antiserum titre ~as assessed by iodinated inhibin binding and 10 in vitro neutralizing capacity) indicate that the antibodies produced are capable of neutralizing endogenous inhibin. The combination of the neutralization of inhibin both in vivo and in vitro and the clo~e relationship of the neutralizing activity and iodinated inhibin binding capacity of the 15 antiserum provides convincing evidence o its specificity.

tb) Inhibin Neutralisation In Vitro Bovine follicular fluid inhibin (2 units) was quantitatively neutralised by 1 ~1 antiserum per culture well while 75% inhibition of bioactivity was achieved with 0.35 ~1 20 per well (Fig. 3, Table 1). In Flgure 3, the vertical dotted line indicates the volume of antiserum required to achieve 75%
neutralization of inhibin activity, an arbitrary parameter of antiserum neutralizing titre. Purified 58 and 31kD inhibin gave corresponding 75~ inhibition values of 0.33 and 0.38 ~1 25 respectively. In comparison, neutralisation of hFF inhibin bioactivity required 1.32 ~1 antiserum (n = 2, Table 1) corresponding to 27% cross-reactivity in comparison to bFF
inhibin. Inhibin from ovine sources (follicular fluid, rete testis fluid) showed ~ and 6% cross-reactivity respectively.
This antiserum, at a maximum non-toxic level of 4 ul per well, did not neutralise 2 units of inhibin activity in rat ovarian cytosol extracts.

~3~','736 1~R. 11 '87 08:0ZCLEMENT HRCK fWST +613 529 6296 P, 16 i ~ V
O _ _ O
a O a ::~ o ,~, A
O ~i ~: O a~ ~ O v o.~ 0 ~( V
o C ~ o ~
a~

V X O N 0~ 0 1-~ ~1 O +
~1 .C ~.
O ~ O O O
~U O V V V
'J

_ O~ "

~ ~:

e C ~ ~ Oo O o ` c ~¦ a --o e 2 2 i 2 ~ ; .

Example 2 Iodination of inhibin Iodination of either 58kD or 31kD inhibin has been achieved using a conventional Chloramine T iodination procedure and was a~ociated with considerable iodination 5 damage. Purification of the tracer was therefore nece~ary and it was not achieved following gel filtration chromatography on SephadeX G25. Specific binding of either iodinated hormone to Matrex Red A achieved satisfactory purification although recoveries were low. Either iodinated 10 inhibin form thus purified had the physico-chemical properties of its non-iodinated form.
~ lternative iodination procedures using Iodogen, Iodobeads, lactoperoxidase, or Bolton-Hunter reagent were found to cause les~ damage to the inhibin molecule, but 15 re~ulted in poorer incorporation of radioactivity.
Consequently iodination using Chloramine T was preferred.
Purified 58kD or 31kD inhibin (1-2 pg in 25 ~1 electroelution buffer was added to 25 ul 0.5 M phosphate -buffer, pH 7,2. Nal25I ~0.5 mCi, 5 ,ul; Amersham, Bucks" UK) 20 was added. Chloramine T (40 ~1) was added at a ratio of 8:1 Chloramine T to hormone. The reaction proceeded for 60 second3 at room temperature with stirring and was terminated with 20 Jul sodium metabisulphite (3 mg~ml). The reaction mixture was made up to 50 yl in 20mM phosphate buffer 0.~ BSA
25 or 0.5~ Polypep (Sigma, St. Louis, Mo., USA) pH 6.0 and gel filtered on a Sephadex G25 column tPD10, Pharmacia, Uppsala, Sweden) to remove 125I. The void volume fractions were pooled, made up to 20 ml and applied to a column of 200 pl Matrex Red A (Amicon, Danvers, Ma~s., USA) and then washed 30 with phosphate buffer containing 400 mM KCl, the eluted counts being di3carded. 125I--inhibin was eluted with lM XCl/4M urea in phosphate buffer. The iodinated inhibin was further gel filtered on a Sephadex G25 column (PD10) with the appropriate RIA buffer t~ee below~ to remove the KCl/urea.
Following iodination of 58kD and 31~ inhibin, 60 ~Ci and 25 ,uCi reipectively were recovered in the void volume fractions following gel chromatography on Sephadex G25.

* TRADE MARK
.~
,J

13~7~36 '1f~R. 11 '87 03:04 CLE~`1ENT HI~CK QUST +613 5Z9 6296 P. la Approximately 30~ was eluted with the lM KCl/4M urea buffer.
125I-inhibin, as assessed by its molecular weight on SDS-PA~E, was found in this fraction.
The specific activity of the iodinated preparations 5 was asse~sed in the radioimmunoassay using a self-displacement procedure (Marana et al 1979) with the hormone used for iodination as standard. Specific activities of 50-60 ~Ci/~g for 58kD inhibin and 24 ~uCi/~g for 31kD inhibin were obtained, with recoveries ranging from 5-25%.

10 Example 3 Characterisation of iodinated inhibin The physico-chemical char~cteristics of 125I-inhibin were assessed using RP-HPLC and SDS-PAGE. A close correspondence was observed between the radioactive and bioactive profiles on RP-HPLC for both the 58kD and 31kD
15 preparations (data not shown). The molecular mass of 125I-58kD inhibin following fractionation on SDS-PAGE was similar to purified non-iodinated inhibin under both non-reducing (58kD) and reducing (43kD and 15kD) conditions except that a 58kD material of unknown identity was observed 20 in relatively low proportions (18%) under reducing conditions (Fig. 1). The molecular weight markers employed were BSA
(bovine serum ~lbumin) 67,000; OV tovalbumin) 43,000: CA
(carbonic anhydrase) 29,000; GL (goose egg lyso2yme) 20,300;
and CL (chick egg lysozyme) 14,300. The arrow, in figure 1, 25 refers to the point of sample application. Radioactivity found in fractions beyond fraction 47 represents free iodine in the solvent front. The purity of the inhibin used for iodination as assessed by silver staining on SDS-PAGE suggests that the 5I-58kD material is not an iodinated contaminant.
30 In support, 125I-58kD inhibin was ~ractionated by microelectrofocusing procedure on the pH range 3.5-10 and 4-8 (Foulds and Robertson 1983), and 3 peaks of radioactivity with pI values of 7.4, 6.2 and 5.2 were observed. Upon reduction on SDS-PAGF each of these peaks showed persistence of 35 125I-58kD material. The results suggest that the presence of 130~736 ~1RR.11 '87 08:05 CLEI`1ENT H~CK l:lUST +513 529 6296 P.l9 the 125I-58kD material is attributable to difficulties in reduction of the iodinated hormone rather than to the iodination of a contaminating protein.
Practionation of the 125I-31kD inhibin on SDS-PAGE
5 revealed molecular weights of 30,200 under non-reducing conditions and 20,000 and 15,000 subunits following reduction;
these values are si~ilar to those for the non-iodinated hormone. A second antibody RIA system using either tracer yielded a parallel displacement between purified 31kD and 58kD
10 inhibin.

Example 4 Radioimmunoassay Procedure The assay buffer used was 10 mM phosphate, 0.15 M
NaCl, O.5~ BSA, pH 7.2. A delayed tracer addition, second antibody assay system was employed. The sample and antiserum 15 were incubated in a volume of 300 ~1 for 16 hours at room temperature followinq which 125I-inhibin (10,000 cpm, 1OOJU1) was added and the incubation continued either overnight at room temperature or for 48 hours at 4C. second antibody (goat antiserum to rabbit IgG, 100 ~1) was added and incubated 20 for 1 hour at 4C following which 1 ml 6~ polyethylene glycol was added. The tubes were vortexed and incubated for a further 30 min, spun at 2000 g for 30 min at 4C, decanted and counted. The inclusion of Triton X-100 (final concentration 0.025%) in the assay buffer reduced non-specific binding from 25 4 to 0.5~.
Radioimmunoassay procedures were established using both 31 and 58kD inhibin tracers. FollowLng a logit-log dose transformation of the response curves, linear displacement of each tracer was observed for a range of inhibin preparations, 30 with the exception of 31kD inhibin when using 125I-58kD
inhibin as tracer, in which a deviation from linearity below logit -0.5 (38~ B/~o) was seen (Fig. 4). In figure 4, each value represents the mean ~ SD of triplicates. The characteristics of each assay are outlined in Table 2.
35 Scatchard analysis revealed similar affinities for the antiserum of either inhibin form. Non-parallel dose response 13V'7736 ~RR.ll '~7 03:07 CLE~ENT HRCK RUST +613 529 6296 P.20 lines were observed between bFF and either 31kD inhibin with 125I-31kD inhibin as tracer of 58kD inhibin with 125I-58kD
inhibin as inhibin tracer. ~he sen~itivity (EDlo) and ED50 values were comparable in each assay with either hormone.

13~7~36 11:1R.11 '87 013:08 CLE11ENT H~CK FIUST +613 529 6296 P.21 Character~stics of the two radioimmunoassaY sv6tems w{th L ~ ~ ~2 __ I-31kD ~nhibin and r-58kD inhibin as tracers 5I-31kD inh~bin25I-58kD inhibin Antibody Dilution 1: 8000 1: 4000 Tracer binding (8Ol 30~ 18%
Affinity (Kdis) x 10 lOM 0.66 0.72 10 EDlo (ng, fml 3 lkD inhibin 0.10, 3 . 0 0 . 13, 4 . 4 58kD inhibin 0.07, 1.2 0.13, 2.2 ED50 ~ng ) 31kD inhibin 0.30, 10.1 O.Sl, 17.1 58kD inhibin 0.26, 4.3 0.43, 7.0 Slope*
b~F 1-37 1 O.O9a(8)1.47 ~ 0.09C(8) 31kD inhibin 1.53 ~ o.09b(5)1.68 + 0.08 (3) 58kD inhibin 1.50 ~ 0.07 (3)1.73 ~ 0.14d(5) 20 Precision**, 0.036 (5) 0.038 (5) Between Assay Variation** 14 (5) 8. 5~ (5) Bio/Imm Ratio 31kD inhibin 0.34 * 0.09 (16)0.43 ~ 0.13 (4) 58kD inhibin 0.30 ~ 0.12 (7)0.37 t 0.12 (5) BGCM 0.25 a vs b P~ 0.05) ) as assessed by paired t-test a vs c P ~0.01) 30 Mean + SD
~ Number in brackets: number of preparations *~ Number in bracXets: number of assays BGCM - bovine granulosa cell culture medium ~or details see text.

1307~
~R. 11 '87 08:09 CLE~ENT H~CK ~UST +613 529 6Z96 P.22 Example S Specificity of the assay The specificity was asqessed on the following grounds. First, a similar hierarchy of cross-reaction of inhibin from various species in the RIA using either tracer 5 and in vitro neutralisation studies was observed ~Table 1).
The cross-reaction in the radioimmunoassay of inhibin from different species when expressed in terms of their bioactivity was bFF 100%, hF~ 30% ovine FF 1% and rat ovarian extract non-detectable. With respect to this antiserum it is apparent 10 that both male and female bovine and human inhibin share common antigenic determinants not found in inhibin from the other two species. This implies close structural similarity between inhibin from both sexes and species. Secondly, no cross-reaction tO.5%~ occurred for a range of purified 15 glycoproteins and polypeptides. Rat LH and FSH, ovine LH and FSH, hCG, bovine TSH, LHRH, ovalbumin and bovine serum albumin showed less than 0.5% cross reactivity using either tracer.
Alternatively, medium from the bovine granulosa cell culture (Fig. 4), and bovine testis extract (data not shown), both 20 containing inhibin bioactivity, gave parallel displacement curves to bFF inhibin in the RIA. The parallel dilution of inhibin bio- and immunoactivity of medium from bovine granulosa cell culture with the inhibin standard provides evidence for these cells being the site of inhibin production, 25 as has been previously suggested (Erickson and Hsueh 1978;
Henderson and Franchimont 1981). Thirdly, similarities were observed in the profiles of both biological and immunological activities following fractionation of bFF on gel filtration chromatography and RP ~PLC. However, in the 40-60k~ molecular 30 mass region of the Sephacryl S200 column (Figure 5a), an 8-40 fold excess of immunoactivity over bioactivity was present, accounting for 12-18% of the recovered immunoactivity.
A large variation in the ratio of biological/immunological activities with charcoal-treated bFF
35 as standard was observed following fractionation of bFP
inhibin on gel filtration and RP-HPLC ~Fig. S) and between purified inhibin preparations ~Table 2J. The ratios ranged 13~'7736 R. 12 !87 13:31 CLEMENT HhCI< I~UST +61~ 529 62~ P.06 -- ZO --from 0.02-2.09 in fractions obtained during the purification procedure and ~rom 0.30-0.43 with the purified inhibin preparations.
It is concluded that the ~IA procedures are not 5 detecting molecular entities devoid of biological activity and VlCe ~ers~ except in the lower ~olecular weight region (40-60kD) of the Sephacryl S200 chromatogram. Whether this lower molecular weight material represents a protein distinct from inhibin ~hich cros~-reacts in the RIA or inhi~in devoid 10 of bioloqical activity ~as not been established.
The cross-reactivities of inhibin-r~lated prot~ins in the RlA relative to 31 kDa bFF inhibin were as follows:
porcine transforming growth factor-beta ~R & D Systems, Minn.
~SA) ~0.9~, bovine Mullerian inhibitory substance (kindly 15 provided by Dr. J. Hutson, Royal Children's Ho~pital, Melbourne) ~0.3~ purified bovine inhi~in B subunit dimer ~2 and the subunits of 31 kDa bFF inhibin obtained ~ollowing reductio~ and alkylation ~0.1~. A range of glycoproteins and growth factors hAve been previously tested (MeLachlan et al, 20 1986) and showed cros8 reactivities against anti-58kD inhibin of less than 1.0~. The specii~i~y of anti-31kD i~hibin wan similar, with Cros--reactiviti-- Of less than 1. 0~ .

Examp~e 6 Application of the Radioimmunoassay to Serum The RIA in its appllca~ion to serum required substantlal modification. Firstly lOOmM phogphate ~uffer pl~
7.4 containing 0.15M NaCl, 0.5~ BSA was used, and, b~cause of 25 its stability in gerum, 125I-31~P inhibin was preferred to 5I-58kP inhibin as RIA tracer. Sccondly, a tempcrature~dependent interference of steer serum with 125I-31kD in~ibin binding to the antiserum was observed, with an increase in binding (B/Bo) from 57~ at 4C to 94~ at 37C
30 ~Fig. 7). This figure demonstrates the temperature dependence following a 16 hour incubation in the presence of various serum an~ inhibin preparations ~bF~, 31k~ inhlbin, steer serum ~SS), cow serum ~CS), bull serum (BS), human post menopausal serum ~PMS), human female serum pool (~FP)). Tracer binding 35 ~/T) was maxim~l at 30~ in the presence of steer serum, being 87.1 ~ 3.4% ~n ~ 5). Displacement of 125I-31kD inhibin 13~)~7736 R. 12 ~7 13: 31 CLEMENT H~CK hU~T +~.13 5~ ~Z96 P. 07 by bFF or 31~D inhibin was largely unaffected by temperature Human PMS showed no temperature related interference upon binding although the bindin~ w~ elevated (B/~o 110-120~).
Ba~ed on these data, conditions for the as~ay of 5 elther bov~ne or human serum inhibin were established. T~ese involved using 125I-31kD inhibin as tracer in an overnight 30C incubation ~nd, in order to compen~ate for the low level of interference ~y SS or PMS (presumed to contain no inhibin1 standards and samples were diluted in SS or PMS accordingly.
No detectable activity was determined in ~teer or in human post-menopaus~l serum, whil~t bull and human female serum ~howed parallel dose-response curvefi to their re3pective follicular fluid standards, with circ~lating levels of 0.9 and 1.1 ng re~pectively.
Inhibin preparation~ or ~erum samples were diluted in SS or PMS to a sample volume of 200 ~1. Anti~erum (100 ~1, final dilution 1:8000) and samples t200 ~1) were incubated for 4 hour~ at 30C, followed by a further incubation of 16 hours at 30C in the presence of trace~. Second antibody was added 20 and the tubes were $ncubated for 24 hours at ~~, following which 2 ml 0.15M NaCl was added and the tubes were aentrifug-d.
Thirdly, with respect to a choice of stAndard in the RIA of bovine serum inhibin, purified 31kD inhibin i~ faYoured 25 in view of its stability in serum. In ~he absence of a p~rified human inhibin preparation, 8lkD bovine inhibin may be used as the ~tandard in the ~IA of human serum inhibin.
However, the partially purified hFF inhibin preparation described above i8 preferred, and purified h~P inhibin, when 30 available, would be the most pre~erred standard. The detectable levels of inhibin $mmunoactivity in serum from women under going ovarian stimulation with exogenOus gonadotrophin is analogous to the findings o Lee et al (lg82~, where circulatinq levels of inhibin activity were 35 detected in PMSG treated immat~re female rats, particularly directed against 58kD and 31k~ inhlbin, 13~ 36 lf:ll~'. 1~ '87 13: 3Z CLE~`lEl`lT HRCI< fl~lST +61_: 5~g 6zg6 P.
- 21a -Individual anti3era may behave differently in the ~s6~y, and assay parameters may have to be determined for each casc. Considerable variation~ in 6ensitivity between antisera have been ob~erved, particularly between antisera directed against 31kD and 5akD lnhibin. Anti-31kD inhibin appeared to glve greater ~ensitivity than anti-58kD inhibin ~n the samples tested 80 far, 13~7736 1FIR. 11 '87 08:12 CLE11ENT Hf~CK ~UST +613 529 6296 P.Z5 Exampls 7 Improved Sen~itivit~c RIA for human ~erum In order to improve as~ay sensitivity, the assay procedure above wa~ modified as follows: the total volume of the assay was reduced from 400 to 300 ~1 (comprising 200 ~1 5 sample, 50~ul tracer and 50 ~1 anti~erum). The assay buffer was 150 mM phosphate, 0.2~ ~SA pH 7.4, and the incubation of sample and antiserum was 4 days at 4C followed by the addition of tracer and a further 3 days at 9C prior to the addition of second antibody. Using this method, a 2.5 fold 10 increase in sensitivity was achieved. ThiS modified assay procedure has been applied to the measurement of human plasma inhibin. The modified assay allows the quantification of plasma inhibin in normal male plasma and in plasma throughout the normal menstrual cycle.
.

15 Example 8 Stability of 125I-31k~ and 125I-58kD Inhibin in Serum SDS-PAGE profiles of 125I-58kD inhibin following overnight incubation with serum (SS and PMS) showed an increased formation of an 125I-30k~ component (12~ of 20 recovered activity at 4C; 17~ at 30C) in comparison with either buffer or bFF (6% 4C and 30C; (Fig. 6). ~he tracers were incubated overnight at either 4 or 30C with either bFF, steer serum (SS) or human post-menopau~al serum ~PMS).
Incubation of either tracer in RIA buffer alone gave ~imilar 25 profiles to the bFF incubation shown. Molecular weight markers are described in Figure 1. No radioactivity was observed between the position of the marker carbonic anhydrase and the solvent front with either tracer. Results are presented as the mean I SD of three replicate experiments. In 30 contrast, SDS-PA~E profiles of 125I-31kD inhibin under the same incubation conditions showed no significant changes.
Recoveries of radioactivity with either tracer were not affected by either temperature or by the presence of bFF or serum.

1RR, 11 '87 08:14 CLE11ENT H~CK ~UST +613 525 62961 P.26 Example 9 Radioimmunoassay of inhlbin in bovine serum The application of the inhibin RIA to serum from cattle resulted in parallel logit-log dose response lines of BS with either b~F or 31kD inhibin as standards (Fig. 8). The 5 response shown in figure 8 is for bovine and human ~erum, diluted in steer serum or post-menopausal serum respectively, in the plasma RIA system employing 12sI3lkD inhibin as tracer.
Potential RIA standards (bFF, 31kD inhibin, hFF) were assayed in the presence of either phosphate buffer ~200 ~ul, o) or 10 steer serum (200~ ) or post-menopausal serum (200~ ) and their logit plots were calculated using their respective non-specific binding and Bo values. Cow serum shows a minimal detectable immunological response. When the immunoactivity was expressed in terms of 31k2 inhibin standard the level of 15 inhibin in BS was 0.91 ~ 0.27 ~n = 3) mg/ml and CS 0.1 ng/ml.
A parallel response line of HFP was observed in the RIA with 31kD inhibin and hFF as standards corresponding to levels of 1.05 = 0.07 (n = 3) ng/ml. The levels in normal plasma (n =
8) were equal to or less than the sensitivity of the assay 20 (0.1 ng/ml).

Example 10 Radioimmunoassay of inhibin in infertile human The radioimmunoassay previously described was applied to pla~ma and serum from post-menopausal subjects 25 (presumed to be inhibin free, n = 8) and young women with ovsrian failure (premature ovarian failure n - 2, Turner's ~yndrome n - 1, ovariectomized n ~ 1). There was no difference in 125I-31k2 inhibin binding in the assay between these two groups and therefore post-menopausal serum was used 30 as a diluent in the assay.

Example 11 Radioimmunoassav of inhibin in human serum The method was applied to:
(a) Plasma from normal women during ~pontaneous menstrual cycles from the early follicular phase until the 3S time of ovulation (n = 2, example 'FL' #27, Fig. 9). Inhibin ~3~'736 1FIR. 11 '~37 0~: lS CLEMENT H~CK flUST +613 529 6296 P.27 -- 2g --immunoactivity was below the limit of assay detection prior to day 13 and its increase correlated with the increa3e in circulating level~ of oestradiol (E2), LH and FS~ on day 13 and 14.
(b) Inhibin immunoactivity in serum from normal men (n e 7) was at the lower limit of assay detection.
(c) Plasma was obtained from 26 unselected women undergoing their fourth cycle of ovulation induction in the in vitro fertilization programme at the Epworth Hospital, 10 Richmond, Victoria. Briefly this involved the administration of clomiphene citrate 100-150 mg da~ly on day 5-9 of the menstrual cycle, followed by human menopausal gonadotrophin ~hMG) 75-225U daily for the next 5-7 days. Adequate follicular development was assessed by the progressive 15 increase in plasma oestradiol and by ovarian ultrasound.
Ovulation occurred spontaneously if an endogenous LH spike was observed, or in the absence of the latter, ovulation was induced by administration of 5000 IU of human chorionic gonadotrophin (hCG). For an example of a spontaneous 20 ovulation see 'BU' # 11, Fig. 9k and for an hCG-induced ovulation sample see '~O' 1, Pig. 9a.
Inhibin immunoactivity in the plasma samples, as defined in terms of the biological activity of a purified 31kD
inhibin standard, showed a highly significant correlation with 25 plasma oestradiol levels (Fig. 10). The correlation coefficient values have been calculated from the total data in figure 9. An example of the correspondence of plasma oestradiol and inhibin during an ovulation induction cycle is seen in example 'BE' # 9, Fig. 9b. There was also sign~ficant 30 correlation between peak plasma oestradiol concentrations and the number of oocytes recovered, and peak plasma inhibin concentrations and the number of oocytes recovered (Fig. 11), as well as a strong correlation between the peak plasma inhibin concentration and the number of ovarian follicles 35 detected ultrasonically prior to oocyte aspiration (Fig. 12).
Tt is therefore apparent that both plasma inhibin and plasma 13~7736 ~l~R. 11 '87 0S: 16 CLEMENT Hf:~CK ~UST +613 529 6296 P.26 oestradiol are parameters of follicular development and health, and in the majority of cases these show a close correspondence.
In certain examples ('BY' # 6, 'BR' # 10, Fig. 9h 5 and j), a dissociation between plasma inhibin and plasma oestradiol concentrations was observed, suggesting different regulation of these two parameters of follicular development.
As inhibin is a peptide produced by ovarian granulosa cells and plasma oestradiol in the human is predominantly an ovarian 10 theca cell product, the assay of plasma inhibin is the first direct plasma parameter of granulosa cell/oocyte health and maturation. The di~sociation of plasma inhibin and E2 may therefore be of therapeutic importance in that the plasma inhibin is a direct measure of follicular development, and its 15 assessment may affect the timing of ovulation induction and oocyte collection.

Example 12 hFF inhibin as standard for radioimmunoassay Human follicular fluid (hFF) obtained at oocyte collection in the IVF programme was prepared for use a~ the 20 radiummunoassay (RIA) standard by two gel chromatographic steps and reversed phase HP~C as described for bFF inhibin (Robert~on et al, 1985). This material yielded parallel dose reponse lines to human female serum inhibin obtained from women undergoing ovarian hyperstimulation for in vitro 25 fertilisation. This partially purified human follicular fluid (hFF) inhibin standard preparation was defined in term~ of its in vitro inhibin bioactivity using an inhibin bioassay ba~ed on the dose-related suppresion of FSH cell content (Scott et al, 1980). This material was used as the RIA standard and 30 gave dose response lines parallel to serum inhibin obtained from women undergoing ovarian hyperstimulation for in vitro fertilization and also to that of pregnant serum.
The unitage of the hFF inhibin standard was calibrated in terms of an ovine testicular lymph standard 35 preparation of defined unitage 1 U/mg using the inhibin 'lQR. ll '87 139 le CLE~`lENT HflCK flU5T +613 529 6296 13~77~6 P.29 bioassay. The RIA has an interassay coefficient of variation (CV) of 6.4% (n = 5 assay) and the sensitivity (logit + 2~ was 0.37 U/ml.
The RIA was specific to bovine and human inhibin and 5 cross-reacted less than 0.3S with a range of glycoproteins and growth factors. In addition, inhibin-related peptides cross-reacted as follows: porcine transforming growth factorp c 0.9%, bovine Mullerian Inhibitory Substance 0.3~, purified bovine inhibin 8 subunit dimer ~1~ and the subunits of 31kDa 10 bF~ inhi~in following reduction and alkylation ~0.1~. No immunoactivity was detectable in the sera of castrate subjects, post-menopausal women, nor in a subject with Turner's syndrome. The RIA had an interassay coefficient of variation of 8.3% (n = S assays), and a sensitivity of 0.37 15 U/ml.
Inhibin levels were determined at 1 dilution level against the partially purified hFF standard preparation using an iterative curve-fitting procedure (Burger et al., 1973).
In the calculation of results, a lognormal distribution of 20 individual observations (Gaddum et al 1933) was assummed, i.e. all calculations were performed using logarithmically transformed va~ues to give geometric means and 67% confidence intervals. Statistical compar~son between pregnant and non-pregnant groups was performed using the unpaired t-test.

25 Example 13 Inhibin levels during luteal phase and early pregnancy Nineteen women pre~enting consecutively for treatment in the Monash university IVF program were studied.
Clinically their infertility resulted from tubal disease (n =
30 7), endometriosis ~n = 6), unkown causes (n 5 5) or poor semem quality (n = 4). The protocol of ovulation induction has been described elsewhere (Wood and Trounson, 1984). Briefly, all subjects received clomiphene citrate (Clomid, Merell Dow, Sydney) 100-150 mg daily between days 5 and 9 of the cycle and 35 HMG (Pergonal, Serono, Rome) 75-225 units daily from day 6.
The dosage and duration of HMG therapy were optimized - 13(~7736'~R . 11 ' 57 08: 19 CLEMENT Hl:~CK RUST +613 529 6296 P. 30 according to daily plasma oestradio1 concentrations and follicular size as a~sessed by ovarian ultrasound. HCG
(Pregnyl, Organon, OS8) 5, OOO IU intramuscularly was administered to induce ovulation, and oocyte retrieval was 5 undertaken 36 hours later. Embryo transfer was performed as described by Wood and Trounson (1984). Blood was taken on day 1 post laparoscopy and every second day from day 2 to day 14 and sera stored for measurement of FSH, LH,~ subunit hCG, oestradiol, progesterone and inhibin. Three of the 19 women 10 became pregnant.
At various stages of gestation, a single serum sample was obtained from each of 24 normal pregnant women.
Samples were assayed for inhibin using the hFF
inhibin standard described in Example 11. In the 16 patients 15 who did not conceive, luteal phase inhibin levels rose to a peaX level of 2.5 U/ml on day 6 and then fell to undetectable levels by day 14. These results are shown in Figure 13.
The number of subject serum samples per day was 13-16 except at day one when only eight were available.
20 ~esults are expressed as the geometric mean + 67~ confidence intervals. The broken line indicates the limit of sensitivity of the inhibin radioimmunoassay. The number of sub~ects showing non-detectable inhibin values is shown in parentheQes.
Non-detectable values are not included in the mean 25 confidence intervals.
In three subjects who conceived, inhibin levels were similar to non-conception cycles between days 2 and 8, increaslng thereafter and becoming significantly higher (p 0.001) than in the non-pregnant group by day 12 30 po8t-laparoscopy. Figure 14 shows these results, expressed as the geometric mean l 67~ confidence intervals. The broken line indicates the sensitivity of the inhibin radioimmunoassay. ~pc0.05, **p~0.01, ~**p~0.001 comparing hormone values for the pregnant and non-pregnant groups on the 35 same day. Significance values in the second panel refer to serum FSH. The late luteal phase rise in serum inhibin in the 7~6 R. 11 '~37 03:20 CLE11E~T H~CK RUST +613 529 6296 P.31 pregnant patients coincided with both the rise in 5erum ~hCG
and with the decline in serum FSH to values below thoQe seen in the non-pregnant group.
Serum FSH showed a significant inverse correlation 5 with inhibin in the luteal phase of the non-conceptlon cycles (r = 0.51, n = 113, p ~0.001) (Fig. 15). Similar Qignificant inverse relationships were observed between FS~ and inhibin when the data were analysed wcording to whether the progesterone concentrations were in the normal ovulatory 10 luteal phase range (25-100) nm) or greater, (r = 0.38, n =
76, p ~0.001 vs r = 0.37, n = 37, p c0.05, slopes not statistically different). Significant inverse correlations also existed between luteal phase serum FSH and both progesterone (r = 0.64, n - 113, p ~0.001) and oestradiol (r -15 0.52, n ~ 114, p ~0.001).
Plasma inhibin and progesterone concentrations weresignificantly correlated in the luteal phase of cycles in which pregnancy did not occur (r = 0.81, n - 85, pcO.001), as were plasma inhibin and oestradiol concentrations (r - 0.65, n 20 ~ 85, p ~0.001). In pregnant subjects, luteal phase inhibin levels did not show significant correlations with either progesterone or oestradiol. Serum LH levels (not shown) fell sharply from day 1 (21.0 (17.0-26.1] mIU/ml) to a nadir (3.5 ~1.2-9.8~ mIU/ml) on day 8.
ln a separate study of serum inhibin, levels were determined during gestation in 24 normal pregnant women. The mean level prior to 20 weeks gestation (1.31 (0.95-1.80) U/ml, n - 13) was significantly lower (p ~0.02) than levels after this time (2.02 U/ml (1.32-3.10) U/ml, n = 11).
Thus there is a rise in circulating inhibin concentrations during the luteal phase of stimulated menstrùal cycles and during pregnancy.

130~;'736 SRO. 1~ ~, 13 ~3 6LE~I1ENT H~ UST +613 SSS 6ZS5 P.0 -- 2~ --EXAm~1e 4 Inhibin Levels in the Nor~al Men5trUal ~ele In a further study~3erum inhibin was determined daily in 6 normal wo~en throughout the menstrual cycle, using a rad~ OimmunoasSAy smploylng an antiser~m directed to 31kD inhibin. ~he normalcy of the menstrual oycle was a~oessed from the serum profiles of ~S~, ~H, progesterOne, and oestradlol.
The increAse in ~en~it~vity of the as6ay using this antiserum permitted the detectlon of inh~bln in over 97~ of sAmples~ The results are shown in Flg. 16.

13~ 36 . lc '~7 1:~: 33 CLEMENT H~U< ~UST +~13 529 6296 P. 1 10 Advant~es and applications of the assaY accordihq to the invention.
1 The assay may be u~ed for determining inhibin concentration in a wide range of biological samples, such as serum, plasma, urine, follicular fluid, tissue homogenates, 15 and culture fluids.
2. The as~ay may be used to monitor the purification of inhibin from tissue, biological fluids, or culture medium, or to moni~or transfection studies.
3. Inhibin levels may be u~ed as a marker of parameters of reproductive function, ~uch as granulosa cell function, fol~ieular development, number of ovarian follicles following ovarian hyper~timulation, and foetal well-being during early pregnancy, and Sertoli cell function.

lt will be clearly understood that the invention in 25 its general a~pects i5 not limited to the specific details referred to hereinabove.
The following terms referred to hereinbefore are trade marks: Amerlex-M, Clomid, Coat-a-Count, Marcol 5 2, Matrex Red A, Montanide 888, Norit A, Pergonal, Polyp~p, 30 Pregnyl, RIA-Quant, Sephacryl, Sephadex, Triton X-100, Ultrapore, and Ultra-Turrax.
Refqrences citcd herein are listed on the following paqe~, 1RR 11 '87 08 24 CLE~ENT HPCK RUST +613 529 6296 1307736 P 34 REFERENCES

Au, C.L., Robertson, D.M. and de Kretser, D.M. ~1983) Endocrinology 112, 239-244.

Baker, H.W.G., Eddie, L.W., Higginson, R~Eo~ Hudson, B., 5 Keogh, E.J. and Niall, H.D. (1981) Assays of Inhibin. In P.
Franchimont and C.P. Channing (Eds), Intragonadal Regulation of Reproduction, Academic Press, pp. 193-228.

Baker, H.W.G., Eddie, L.W., Hudson, B. and Niall, H.D. (1985) Australian J. Biological Sciences (in Press).

10 Burger, H.G., Lee, V.W.K., and Rennie, L.C. ~1983) J. Lab.
Clin. Med. 80 302-308.

Erickson, G.F., and Hsueh, A.J.W. (1978) Endocrinology 103, 1960-1963.

Finney, D.J. (1971) Statistical Method in Biological Assay, 15 2nd ed. C. Griffin & Co. Ltd., London.

Forage R.G., Ring, J.M., Brown, R.W., McInerney, ~.V., Cobon, G.S., Gregson, R.P., Robertson, D.M., Morgan, F.J., Hearn, M.~.W., Findlay, J.X., Wettenhall, R.E.H., Burger, H.G., and de Kretser, D.M. (1986) Proc. Natl. Acad. Sci. V.S.A. 83, 20 3091-3095.

Foulds, L.M. and Roberston, D.M. (1983~ Mol. Cell. Endocr.
3~ 7-130.

Gaddum, J.H. (1933) Medical Research Council special Report Series, 183.

25 Henderson, K.M. and Franci~ont, P (1981) J. Reprod. Fert. 63, 431-442.

130~736 ~R.ll ~a7 08:Z5 CLE~1ENT Hf~CK f:lUST +613 5Z9 6Z96 P.35 Kramer, C.Y. (1956) Biometrics 12, 307-310.

Lee, V.W.K., McMaster, J., Qulgg, H. and Leversha, L. (1982) Endocrinology 111, 1849-1854.

Ling, N., Ying, S-Y, Ueno, N., Esch, F., Denoroy, L. and 5 Guillemin, R. (1985) Proc. Natl. Acad. Sci. U.S.A. 82, 7217-7221.

Marana, R., Suginami, H., Robertson, D.M. and Diczfalusy, E.
(1979) Acta Endocrinol. (Xbh.) 92, 585-598.

Mason, A.J., Hayflick. J.S., Ling, N., Esch, F., Ueno, N., 10 Ying, S-Y, Guillemin, ~., Niall, H., and Seburg, P.H. (1985) Nature _ , 659-663.

Miyamato, X., Hasegawa, Y., Fu~uda, M., Nomura, M., Igara~hi, M., Kangawa, K., and Matsuo, H. (1985) Biochem 3iopys Res.
Comm. 129, 396-403.

15 Robertson, D.M., ~oulds, L.M. LeverQha, L., Morgan, F.J., Hearn, M.T.W., Burger, H.G., Wettenhall, R.E.H. and de Xretser, D.M. (1985) ~iochem. Biophys Res. Comm. 126, 220-226.

Robert90n, D.M., de V05, F.L., Foulds, L.M., McLachlan, R.l.
20 Burger, H.G., Morgan, F.J., Hearn, M.T.W. and de Xretser, D.M.
(1986) Mol. Cell. Endocrinol. 44, 271-277 Scott, R.S., ~urger, H.G. and Quigg, H. (1980) Endocrinology 1 , 1536-1542 Wood, E.C., and Trounson, A.O. (eds). In vitro fertilization 25 and embryo transfer ~1984) Churchill Livingstone

Claims (13)

1. A radioimmunoassay for determining inhibin suspected to be present in a biological sample, comprising the steps of:
(a) contacting the sample with an antibody directed against inhibin and incubating at 4° to 30-C for a period of 4 hours to 4 days, (b) adding 125I-inhibin and incubating for a period and at temperature selected from the group consisting of at overnight room temperature, 48 to 72 hours at 4-C or 16 hours at 30-C, (c) adding a second antibody and incubating for 30 minutes to 24 hours at 4°C, (d) separating precipitated material, and (e) measuring bound 125I-labelled inhibin in the precipitate.

2. A radioimmunoassay according to Claim 1, comprising the steps of:
(a) incubating sample and antiserum for 4 days at 4°C, (b) adding 125I-inhibin and incubating for 72 hours at 4°C, (c) adding a second antibody and incubating for 30 minutes to 24 hours at 4-C, (d) separating precipitated material, and (e) measuring bound 125I-labelled inhibin in the precipitate.

3. A radioimmunoassay according to Claim 1, in which the antibody is contained in an antiserum raised by injecting an animal with an antigen selected from the group consisting of inhibin, purified bovine 58kD inhibin, purified bovine 31kD
inhibin, human inhibin, human or bovine inhibin produced using recombinant DNA technology, or immunologically cross-reactive sub units, fragments or derivatives thereof.

4. A radioimmunoassay according to Claim 1, 2 or 3, in which the 125I-inhibin is 58kD or 31kD inhibin.

5. A radioimmunoassay according to Claim 1, 2 or 3, in which samples to be assayed are diluted in inhibin-free serum.

6. A radioimmunoassay according to Claim 1, 2 or 3, in which polyethylene glycol is added following incubation with the second antibody and incubation is continued for a further 30 minutes.

7. A radioimmunoassay according to Claim 1, 2 or 3, in which Triton X-100 is incorporated into samples to be assayed.

8. A radioimmunoassay according to Claim 1, 2 or 3, in which an assay standard sample is used, said standard sample being selected from the group consisting of naturally-occurring or recombinant bovine 31kD inhibin and naturally-occurring or recombinant human inhibin, or immunologically cross-reactive sub-units, fragments or derivatives thereof.

9. A radioimmunoassay according to Claim 1, 2 or 3, in which the 125I-inhibin is prepared by the steps of:
(a) iodinating inhibin with Chloramine-T, (b) isolating 125I-inhibin by affinity chromatography on Matrex Red A, and optionally (c) further purifying the 125I-inhibin by gel filtration.

10. A test kit for the estimation of inhibin in a sample, comprising at least one of:
(a) 125I-labelled inhibin, (b) an antibody directed against inhibin, (c) an assay standard selected from the group consisting of naturally-occurring or recombinant bovine 31kD inhibin and naturally-occurring or recombinant human inhibin or immunologically cross-reactive sub-units, fragments or derivatives thereof.

11. A method of monitoring a parameter of reproductive function in a mammal selected from the group consisting of ovarian granulosa cell function, ovarian follicular development, number of ovarian follicles following ovarian hyperstimulation, foetal well-being during early pregnancy and Sertoli cell function, comprising the step of measuring inhibin content in a sample of plasma from that mammal using a radioimmunoassay method according to Claim 1, 2 or 3.

12. A method of assessing health and maturation of oocytes or ovarian granulosa cells of a mammal comprising the step of measuring inhibin content, in a sample of plasma from that mammal using a radioimmunoassay method according to Claim 1, 2 or 3.

13. A method of selecting a preferred time for ovulation induction or oocyte aspiration for in vitro fertilization of ova from a female mammal comprising the step of measuring inhibin content in a sample of plasma from that mammal using a radioimmunoassay method according to Claim 1, 2 or 3.