Radioimmunoassay of ACTH in plasma

Solomon A. Berson, Rosalyn S. Yalow

J Clin Invest. 1968;47(12):2725-2751. https://doi.org/10.1172/JCI105955.

Techniques are described in detail for a radioimmunoassay of plasma adrenocorticotropin(ACTH) that is capable of detecting hormone in unextracted normal human plasma at 1:5dilution under the conditions described. The sensitivity of the assay is at the level of 1µµg/ml (equivalent to 0.014 mU/100 ml).

In normal subjects ACTH concentrations averaged 22 µµg/ml (equivalent to 0.308 mU/100ml) plasma at 8-10 a.m. In a smaller group the concentrations averaged 9.6 µµg/ml(equivalent to 0.134 mU/100 ml) at 10-11 p.m. Although a circadian rhythm in normalsubjects was not always well marked throughout the daytime hours, plasma ACTH usuallyfell to its lowest value in the late evening. In hospital patients who were not acutely ill,concentrations were infrequently above 100 µµg/ml in the morning and usually fell tosignificantly lower levels in the late evening. Severely ill hospital patients occasionallyexhibited a.m. concentrations above 200 µµg/ml.

In a group of subjects showing frequent spiking of plasma 17-OHCS concentrationsthroughout the day parallel spiking of plasma ACTH as well was generally observed.

Metyrapone produced marked increases in plasma ACTH within 24 hr in all cases andgenerally within 3-6 hr except when started late in the day. Dexamethasone brought about apersistent reduction in plasma ACTH in a patient under continued treatment withmetyrapone.

Hypoglycemia, electroshock, surgery under general […]

Research Article

Find the latest version:

http://jci.me/105955/pdf

Pdf

Radioimmunoassay of ACTHin Plasma

SOLOMONA. BERSONand RosALYN S. YALoW

From the Radioisotope Service, Veterans Administration Hospital,Bronx, New York 10468 and the Department of Medicine, Mt. Sinai School ofMedicine, NewYork 10029

A B S T R A C T Techniques are described in detailfor a radioimmunoassay of plasma adrenocortico-tropin (ACTH) that is capable of detecting hor-mone in unextracted normal human plasma at 1: 5dilution under the conditions described. The sensi-tivity of the assay is at the level of 1 ,qLug/ml(equivalent to 0.014 mU/100 ml).

In normal subjects ACTH concentrations av-eraged 22 ,uug/ml (equivalent to 0.308 mU/100ml) plasma at 8-10 a.m. In a smaller group theconcentrations averaged 9.6 lqqg/ml (equivalentto 0.134 mU/100 ml) at 10-11 p.m. Although acircadian rhythm in normal subjects was not al-ways well marked throughout the daytime hours,plasma ACTHusually fell to its lowest value inthe late evening. In hospital patients who werenot acutely ill, concentrations were infrequentlyabove 100 /A/Ag/ml in the morning and usually fellto significantly lower levels in the late evening.Severely ill hospital patients occasionally exhibiteda.m. concentrations above 200 /Apg/ml.

In a group of subjects showing frequent spiking-of plasma 17-OHCS concentrations throughoutthe day parallel spiking of plasma ACTHas wellwas generally observed.

Metyrapone produced marked increases in plasmaACTH within 24 hr in all cases and generallywithin 3-6 hr except when started late in the day.Dexamethasone brought about a persistent re-duction in plasma ACTH in a patient under con-tinued treatment with metyrapone.

Hypoglycemia, electroshock, surgery under gen-eral anesthesia, histalog and vasopressin adminis-tration were usually followed by significant in-*creases in plasma ACTH concentration. Prior

Received for publication 1 April 1968 and in revisedJorm 26 August 1968.

administration of dexamethasone blocked the re-sponse to hypoglycemia.

Marked elevations in plasma ACTHwere ob-served in patients with adrenal insufficiency offsteroid therapy, in Cushing's disease after adrenal-ectomy even in the presence of persistent hyper-cortisolemia, and in some untreated patients withCushing's disease.

Umbilical cord blood contained higher plasmaACTHconcentrations than maternal blood at de-livery in seven of eight cases.

After suppression of ACTH secretion by dex-amethasone or cortisol, ACTHdisappeared fromplasma with half-times ranging from 22 min to30 min in three cases studied.

INTRODUCTION

The method of radioimmunoassay has been em-ployed for the measurement of several peptidehormones in plasma. Adrenocorticotropin (ACTH)poses an unusually 'difficult problem because of itslow concentration in the plasma of normal sub-jects. In a previous study (1) this difficulty waspartly circumvented by extracting and concen-trating ACTH from plasma samples containingtoo low a concentration for direct assay. In thispaper we describe methods with sensitivity suffi-cient to permit the quantitative determination ofendogenous ACTHin unextracted plasma of nor-mal human subjects.

Principles of radioimmunoassay with particu-lar reference to problems of sensitivity

The essential principle of radioimmunoassay in-volves the binding of an isotopically labeled peptideor protein to its specific antibody (2) and the

The Journal of Clinical Investigation Volume 47 1968 2725

0

70-

-i60-

Z 505 gLL_0

F-40-z

30-

20-

10-

200 500 ,upg/miHUMANACTH CONCENTRATION

FIGURE 1 Guinea pig anti-porcine ACTH sera were di-luted sufficiently to yield a B/F ratio of 0.6-1.2 with 'I-or 'I-labeled human ACTHin the absence of unlabeledhormone. The percentage change in B/F ratio is plottedas a function of the concentration of unlabeled hormone.

competitive inhibition of such binding by unla-beled hormone in standard solutions and in plasmaor other body fluids (3, 4). A "standard curve"may be constructed by plotting the ratio of bound(B) to free (F) labeled hormone, (B/F), as afunction of the total concentration of unlabeledhormone in known standard solutions. The con-centration of hormone in an unknown sample isthen determined from the B/F ratio observed inthe sample by reference to the standard curve. Itis a matter of experience that antisera obtainedfrom animals of the same species immunized onthe same schedule show considerable variabilitynot only in the quantity of antibody but also, andmore importantly for immunoassay, in the slopeof the standard curve ('Fig. 1). The antiserum thatyields the sharpest initial slope provides the great-est sensitivity for detection of low hormone con-centrations. The antiserum used in th present workwas selected from among approximately 230 anti-sera taken from about 50 guinea pigs immunizedwith porcine ACTH.

Since the slope of the standard curve is greatest

at the vertical axis and becomes progressively moreshallow with increase in hormone concentration,the initial slope approaches a maximum as theconcentration of labeled hormone ("tracer") tendsto zero. Ideally, the concentration of tracer em-ployed should be reduced to less than the lowestconcentration to be measured, but its practicallower limit will be determined by the statisticalaccuracy of counting and limitations in the count-ing time available. For a given concentration oftracer (weight of hormone per unit volume of in-cubation mixture) the counting rate will be pro-portional to (a) the efficiency of the counter, (b)the volume of mixture counted, and (c) the spe-cific activity of the tracer.

METHODSProduction of ACTHantibodies. Guinea pigs are im-

munized at irregular intervals by subcutaneous injectioninto the medial aspect of the thigh with porcine ACTH(Duracton, Nordic Biochemicals, Montreal, Canada, 40U ACTH/ml in carboxymethylcellulose) emulsified incomplete Freund's adjuvant. Approximately 75 mg ofmetyrapone is included in the emulsification mixture inthe hopes of preventing stimulation of cortisol secretion,which could theoretically inhibit antibody formation.ACTH antibodies are detectable in almost all guineapigs after two to three injections of 25 U each adminis-tered at intervals of approximately 2-4 wk. 8-10 daysafter imunization, blood is taken by cardiac puncture intoheparinized syringes. The separated plasma is stored at- 17'C but stock dilutions of guinea pig antiserum(1:500 or 1:1000 in normal saline solution containingmerthiolate 1: 5000 and control guinea pig plasma, 1: 50or 1: 100) can be maintained in the refrigerator manymonths without significant loss of antibody.

Specific activity requirements for ACTH-"'I andACTH-'I. Porcine ACTH and human ACTH aremolecules of approximately 4500 molecular weight andcontain two tyrosyl residues. For carrier-free 'I and 125I,the maximal specific activities theoretically attainable forone iodine atom/molecule ACTHand for completely io-dinated ACTH (four iodine atoms/molecule) are givenin Table I. Because of the longer half-life of '25I thetheoretically attainable specific activities for 'I-labeledpreparations are only 'A as high as for 'I-labeledpreparations.

Although 'I is obtainable virtually free of other iodineisotopes, the isotopic abundance of 'lI produced by tel-lurium irradiation has not, since January 1966, exceeded15-20% at the time of removal from the Union Carbidereactor at Sterling Forest (5). Although available onlyat monthly intervals, fission-produced 'I from Oak Ridgehas frequently shown an isotopic abundance of 30-35%at time of receipt in the laboratory (5). Table I alsoshows the concentrations of ACTH-1'I and ACTH-'31I,

2726 S. A. Berson and R. S. Yalow

TABLE I

Specific Activities and Tracer Concentration of ACTH-ulI and ACTH-"2UIfor Equal Counting Rates

Initial specific activity Concentration ofAverage I ACTH-I for

atoms/molecule Carrier-free Available* 1,000 cpm/ml1I I

ACTH-'311 1 3,600 mc/mg 1,100 Mc/mg 3.3 jstg/ml4 14,400 4,400 0.83

ACTH-nmL 1 500 500 1.44 2,000 2,000 0.35

* Available isotopic abundance: 11I 30%; I 100%.t In a sample counted 1 wk after receipt of isotope in well-scintillation counter with e-ray spectrometer set for photopeakof "II and 12-90 kv for 2HI. Counting efficiencies: 0.5 X 106 cpm/IAc 1311, 1.5 X 106 cpm/jUc 2Is1.

labeled with one iodine atom/molecule, required to yielda counting rate of 1000- cpm/ml in a sensitive automaticscintillation well-counter 1 wk after receipt of the radio-iodine, under optimal counting conditions. This table wascalculated on the basis that 'I is available with an iso-topic abundance as great as 30% and that the maximalefficiency for counting 'I in the usual well-type scintil-lation counter attached to a y-ray spectrometer is almostthree times that for 'I under conditions providing opti-mal counting rates at low background levels.

Preparation and purification of ACTH--I andACTH-'I. Immediately on receipt of 'I in the labora-tory its isotopic abundance is determined, as previouslydescribed (6), from a radioautograph of a starch gelelectrophoresis of crystalline insulin labeled at approxi-mately 800 and 1500 mc/mg. Only if the isotopic abund-ance exceeds 25-30% do we consider using the ACTH-mI,which is prepared at the same time at a specific activityof about 1000-1200 mc/mg, for an immunoassay.

We have previously described the preparation ofACTH-'I and its purification by adsorption to and elu-tion from cellulose columns or by addition to plasma andextraction therefrom with acetic acid-HCl mixtures(1). The following new procedure,' used for bothACTH-uI and ACTH-'I, is equally simple and re-covers a large fraction of the undamaged labeled ACTH.The method of iodination is modified from that used byHunter and Greenwood for growth hormone (7): to 20,ul 0.25 M phosphate buffer, pH 7.4, in a flint glass testtube are added, in turn, approximately 2-2.5 mc 'I or0.8-1.0 mc 'I, 2 Mg human ACTH (2 Ml of a 1 mg/mlsolution), 15 ul chloramine T (3.5 mg/ml in 0.25 M phos-phate buffer), 20 Al sodium metabisulfite (4.5 mg/ml in0.25 M phosphate buffer), and 1 ml normal human plasma.Thorough mixing is accomplished by bubbling gentlythrough the pipettes used to add the reagents and only afew seconds are allowed to elapse between steps afterthe addition of the hormone. The complete mixture isthen immediately added to 5 mg QUSOG32, micro fineprecipitated silica granules,2 for purification by the same

'Presented at the Annual Meeting of The EndocrineSociety, Miami, Fla., 17 June 1967 (Abstr. 163).

2Available from the Philadelphia Quartz Co., Phila-delphia, Pa.

procedure as described earlier for the purification of la-beled parathyroid hormone (8). The QUSOgranulesadsorb the undamaged labeled hormone but leave the un-reacted iodide and the damaged fractions in solution. Thegranules pack well on centrifugation. After decantationof the supernatant plasma, followed by a single wash ofthe QUSOgranules with 3 ml distilled H20, the un-damaged labeled ACTH is almost quantitatively elutedfrom the QUSOby shaking briefly with 0.5 ml 40% ace-tone in 1% acetic acid; 1.5 ml distilled H20 is addedbefore centrifugation. The eluted labeled ACTH is testedfor suitability in immunoassay by addition of appropri-ately small quantities to nonimmune plasma and to a1: 50 dilution of antiserum, known to be capable of bind-ing essentially all of the tracer quantity of labeled ACTHwithin an hour or so. The mixtures are then examinedfor damaged fragments and for binding to antibody bypaper strip chromatoelectrophoresis (2) or by adsorptionto talcum powder (9) as described below. A good prepa-ration may contain less than 5% damaged fragments andless than 15% unreacted iodide even before purification(Fig. 2 a). The purified material should be essentiallyfree of damaged components even if there is significantdamage in the original preparation (Fig. 2 b).

Adsorption of labeled ACTH to glass and damage tolabeled ACTH; protection against adsorption and dam-age. Labeled ACTH has two important and trouble-some characteristics: a strong tendency for adsorptionto glassware (10) and a marked susceptibility to dam-age during incubation. We have tried various types ofplastic tubes to avoid adsorption but without signal suc-cess; however, we have found that soft ("flint") glass isless adsorptive than hard ("borosilicate") glass and thatsiliconization significantly reduces adsorption. Never-theless, it remains necessary to include a fairly highconcentration of carrier protein if losses by adsorptionare to be minimized. For this and other reasons, the in-cubations for assay are usually carried out in plasma di-luted 1: 5. As will be described below, the reaction ofACTH with the antibody may be partly inhibited athigh salt concentration, so that the standard curve maybe higher in diluted plasma than in whole plasma. How-ever, since sensitivity is quite satisfactory at a plasmadilution of 1: 5 and since the low concentrations of en-

Radloimmunoassay of ACTH in Plasma 2727

;K--i- t-- L-L2.-I:

LH LX--II-USOiOUS PEWATAT74._ F . _ .__.T_

9-v; __ -

a

COMPONENTS-/ -- ORIGINAL'..

ORIGIN

'-1-i-'1- - - ' :-X-- - .~t~r A DAMAGED~~~OMPO~NENTS

t --ACETONE-AETIC ACID - 1 I-- -u~;-XEXTRACTFROM QU0

<LIF------:,---l--+---'l--n--' -- '!+__ = QUSOEXTRACT

! . .... . .

[. 'I. t _ 4 _ / _i r _

b

FIGURE 2 Scans of radioactivity on paper strip chromatoelectrophoretograms of labeled ACTH before and afterpurification with QUSOgranules and of nonadsorbed QUSOsupernatant. Whatman 3 MCpaper was used in theseexperiments and approximately 100 1IA bromphenol blue-stained plasma was added to the preparations before appli-cation to the paper strips. The paper strip showing the stained serum albumin is placed under the scan of the origi-nal unpurified preparation in a. The unpurified preparation shown in b contained a larger fraction of damaged com-

ponents. Unextracted radioiodide and damaged components that migrate with the plasma proteins do not adsorb tothe paper strips at the site of application or to the QUSOgranules.

ADSORPTIONTO SILICONIZEDFLINT GLASS TEST TUBES (13 XIOOml)

(AFTER 15 MINUTES)

0 50

co0(na

0-

It

mc'J

0

LL0

zw

crw

VI

ADSORPTIONTO5-,ul MICROPIPETTES

PLASMA DILUTION(IN VERONALBUFFER, 0.02M)

A - 1: 20x -. I: 5* - UNDILUTED

A

1.5 ml/OLUME IN TEST TUBE

x. . .

60 sec

TIME HELD IN PIPETTEFIGURE 3 Adsorption of labeled human ACTH in plasma to glassware. Aftercontents were removed test tubes were rinsed with 2.5 ml water and pipettes were

rinsed with a single wash before counting.

2728 S. A. Berson and R. S. Yalow

ORIGIN

Y%

-X- - , -

N

dogenous plasma ACTHand the tendency for adsorptionto glass set limits to the plasma dilution that can betolerated, the conditions employed represent a convenientcompromise among the various factors governing sensi-tivity. Since adsorption to glass increases as the ratio ofsurface to volume increases, only slight percentage lossesare observed in a large full vessel standing quietly.However, a significant fraction of the ACTHpresent insmall volumes in test tubes or pipettes may be rapidlylost to the glassware (Fig. 3). Adsorption is slower at4°C than at room temperature and is significantly re-duced at acid pH.

It has been appreciated that damage to all labeledpeptide hormones occurs on prolonged incubation inplasma and that correction for damage must be made inthe radioimmunoassay (11). However, ACTHis prone togreater "incubation damage" than several of the otherhormones. Indeed, Richards and Sayers (12) have shownthat the biologic activity of ACTH incubated in plasmais lost within a few hours. Damage to ACTH-'I orACTH-15I is reflected in the release of small amountsof free radioiodide and in the appearance of componentsthat are no longer able to bind to antibody or of com-ponents that have otherwise lost such characteristics ofthe undamaged labeled ACTH as the ability to adsorbwell to cellulose, talcum, or QUSOgranules. Thus, insystems employing these media for separation of antibody-bound and free ACTH, the damaged components ap-pear with the nonadsorbed, antibody-bound fraction,whereas in double antibody systems they appear in thenonprecipitated free fraction.

It is known that the biologic activity of ACTH islost on exposure to H202 (13), a result of oxidation ofthe methionine residue in position 4 (14), and aftertreatment with plasmin (fibrinolysin) (15), a proteo-lytic enzyme in plasma. To prevent oxidation of themethionine of labeled ACTHwe had tried various thiolsand found mercaptoethanol to be most effective in pro-tecting against damage (1). However, since the additionof strong solutions of H202 to plasma does not signifi-cantly damage the labeled ACTHin plasma in the sameway that incubation in plasma does, it appears possiblethat mercaptoethanol may protect labeled ACTH fromincubation damage by inactivating certain disulfide-con-

TABLE I IEffect of Protective Agents on Stability of A CTH-1311

Added to 1 : 5Piasma and Incubated 16 hrat 250C

DamagedProtective agent components Iodide

A. None 53 3B. 6-Aminocaproic acid, 12.5 mg/mi 36 2C. Mercaptoethanol, 5 mg/ml 23 3D. Trasylol, 500 KIU/ml 19 2E. Mercaptoethanol, 5 mg/ml

+ Trasylol, 500 KIUlml 2 2

taining proteolytic enzymes. Whereas 50% or more oflabeled ACTH may be damaged within a day, or evena few hours, when kept at 4VC in plasma diluted 1: 5without mercaptoethanol, the labeled hormone is quitestable for weeks when stored in acetone-acetic acidwithout mercaptoethanol (see below).

At high concentrations of mercaptoethanol there ap-pears to be some inhibition of the antigen-antibody re-action but below a concentration of about 0.5% we havenot observed inhibition. Also, mercaptoethanol, 0.5%produces some coagulation of whole plasma but not ofa 1: 5 dilution of plasma in 0.02 MVeronal buffer. Incu-bation mixtures are therefore made up to contain about0.354.4% mercaptoethanol. Additional protection is pro-vided by certain inhibitors of proteolytic enzymes. Inour hands, Amicar (6-amino-caproic acid, Lederle Lab-oratories, Pearl River, N. Y.), an inhibitor of plasminogenactivation, has been only slightly protective even at fairlyhigh concentrations. However, both pancreatic trypsininhibitor and Trasylol (FBA Pharmaceuticals, Inc.), aninhibitor of trypsin and plasmin, derived from lung ex-tracts, have been useful (Table II).

On the basis of the foregoing observations, conditionsfor routine assay have been set as follows. Incubationmixtures of 2.5 ml are prepared in disposable flint glasstubes in which the plasma is diluted 1: 5 in 0.02 MVeronal buffer containing the following: mercapto-ethanol, 0.354.4%o; Trasylol, 400-500 KIU/ml; ACTH-31I or ACTH-5I, antiserum, and control (i.e. nonim-mune) guinea pig serum at a final concentration of 1-2%o;the latter is included to prevent adsorption of antibody tothe glass. Tubes are capped with parafilm to prevent lossof mercaptoethanol by evaporation. Mixtures are incu-bated at 4VC. Under the conditions described, adsorptionto the glass is negligible and incubation damage plus re-lease of radioiodide should not exceed about 8-10%o after7 days but some plasmas are more damaging than othersand the general level of damage varies slightly from oneassay to another depending on the quality of the la-beled ACTH. Although proteolytic enzymes appear tobe partly, if not completely, responsible for incubationdamage, we have frequently found that a higher con-centration of plasma is somewhat protective. Thus, inthe absence of any protective agents, labeled ACTHmaysuffer least damage in plasma diluted 1: 20 but slightlygreater damage in the same plasma diluted 1: 5 or 1: 10than at 1: 2.5 dilution. Whether this phenomenon is re-lated to differential effects of dilution on endogenousenzymes and enzyme inhibitors in plasma or to otherfactors is not known.

As already noted, free labeled iodide as well as dam-aged labeled peptide components may appear after pro-longed incubation. The release of radioiodide from thelabeled hormone appears to be caused by two factors.Firstly, it seems likely that part of the iodide that ap-pears is produced by dehalogenase activity in plasma sincethere is usually a greater amount of radioactive iodideafter storage of the labeled ACTH in plasma than in20% actone-0.25% acetic acid (Fig. 4). Secondly, wehave found that there is a more rapid appearance both

Radioimmunoassay of ACTH in Plasma 2729

of radioiodide and of altered labeled peptide componentsin stored solutions of ACTH-w'I than ACTH-mI whenlabeled with the same number of iodine atoms per mole-cule (Fig. 4). This phenomenon is probably related tothe fact that even when the average iodination level isonly one iodine atom/molecule, a significant fraction(over 50%o in the case of insulin-'I [6]) of the iodineis on molecules that contain more than one iodine atom.Thus, when all the iodine atoms are radioactive, as inACTH-'I, decay of a radioiodine atom on a moleculecontaining more than one iodine atom may result indisruption of the molecule and release of the other radio-active iodine atoms in the free form or attached to a pep-tide fragment. Wehave referred to this process as "de-cay catastrophe" (16). Since the isotopic abundance ofthe 'I used is frequently only 15%o or less, relativelyfewer of the iodinated molecules at the same iodinationlevel contain more than one radioactive iodine atom per

ACTH-131I

STOREDMN 20% ACETONE STOREDIN 1:10 PLASMA PROTECTED0.25% ACETIC ACID WITH 10mg/ml MERCAPTOETHANOL

~2/K0l i iit &,lI .,,f

- .: 3J/8;

4/7

4/26-

1"i'ft± ...E I 1-VI Iti

-1--- ' 1 1 *4-E.

.-i.F

molecule. Although the half-life of 1"I is 57 days andradioactive decay amounts to only about 8.5%o during the1st wk, 'I can be released in almost equal amounts fromdoubly labeled molecules and in almost twice this amountfrom triply labeled molecules during the same period.The progressive increase of damage with time tends tolimit the useful shelf-life of the "NI-labeled hormone.However, even after 7 wk storage in acetone-acid, the un-damaged fraction of ACTH-lI was still capable of bind-ing completely to ACTH antiserum (Fig. 5), but thedamaged peptide fraction migrating with the a-globulinsdid not bind to antibody.

Hormones. Several highly purified preparations ofhuman ACTH, obtained through the generosity of Dr.C. H. Li and of Doctors A. Lerner and V. Upton, wereused for labeling and as standard. The Li preparationwas sent to us in September 1964 and was reportedto contain 140 U/mg. Preparation 8B of Upton and

ACTH-125 I

STOREDIN 20% ACETONE STOREDIN 1:10 PLASMA PROTECTED025% ACETIC ACID WITH 10mg/ml MERCAPTOETHANOL

'' 2/l 4&A l

; :

5/u

14tl- !g1' 1 -4-

FIGuRE 4 Chromatoelectrophoretograms of labeled ACTHpreparations showing release of radioiodide and appear-ance of damaged components from ACTH-'I and ACTH-'I (both labeled with an average of about 1 atom totaliodine/molecule) after storage in solutions indicated. Both preparations were made and purified on 2/10. Portionswere removed at intervals thereafter and added to 100 I&d plasma before application to paper strips for chromatoelec-trophoresis. Note the greater damage in plasma than in acetone-acetic acid and the generally greater release of radio-iodide from ACTH-'I in both storage media.

2730 S. A. Berson and R. S. Yalow

.

I

.

FIGURE 5 Radioautograph of paper electrophoretogramof labeled ACTH prepared 3/23/67 and added to im-mune plasma on 5/11/67. Binding of ACTH-mI to anti-body after 7 wk storage of the labeled hormone in ace-tone-acetic acid. Note light radioactive band of damagedhormone in inter a-region that does not bind to antibody.In nonimmune plasma undamaged ACTH-'I remainedat origin.

Lerner was reported to contain 139 U/mg. These can-not be distinguished immunochemically (Fig. 6 a) butseveral other preparations, which are of lower biologicpotency, are weaker immunochemically as well and werenot used as standards for radioimmunoassay. Hormonewas dissolved in 5 X 10' M HC1 at 1 mg/ml and analiquot was further diluted successively in whole humanplasma (previously tested and found to contain negligbleamounts of ACTH) to concentrations of 10 Aug/ml and200 mgg/ml for use as standards. The latter solutionswere divided among 10-20 vials, each containing 0.5 mland frozen immediately. Previous experience with lossesby adsorption from solutions containing lesser concentra-tions of carrier protein led us to the storage of standardsin whole plasma. Original stock solutions at 1 mg/mlwere also divided among multiple vials which werefrozen and melted when needed for iodination.

Preparation of standard solutions. For assay, a stockstandard solution is prepared at 2.5 m~tg/ml in plasmafrom hypopituitary subjects or from dexamethasone-treated normal subjects if tests revealed the completeabsence of hormone in the latter. Plasma containingstandard is then mixed with the hormone-free plasma inappropriate proportions to provide final concentrationsranging usually from 1 uag ACTH/ml to 200-500 AqgACTH/ml in the final incubation mixtures.

Preparation of incubation mixtures. When final incu-bation volumes are 2.5 ml, 500 1.l of each unknown plasmaare pipetted into individual tubes. To all tubes containingstandards or unknowns 50./Al labeled ACTHfollowed by1.95 ml diluted antiserum is then added using a semiauto-matic, foot-operated dilutor 8 that can process over 350tubes per hour. The antiserum is usually prepared at1:100,000 dilution -in 0.02 M Veronal buffer containingcontrol guinea pig plasma and protective agents as de-scribed above. Labeled ACTH is prepared in the samediluent fortified by 20% hormone-free plasma to preventadsorptive losses. Thus, the final concentration of anti-serum is 1: 125,000 and all tubes, standards and unknowns,contain plasma in the same final dilution (approximately1: 5) and all other reagents, except unlabeled ACTH,at identical concentrations. Control tubes for the deter-

3 York Instrument Corp., New York.

mination of free radioiodide and damaged componentsare set up for each patient. Plasmas assayed at differentdilutions are diluted in the hormone-free plasma. Mix-tures are incubated at 4VC for 5-7 days. Under the con-ditions employed, equilibrium is not complete (i.e. maxi-mumB/F ratio is not achieved in tubes containing onlythe labeled ACTH) before 4-5 days, as determined bydaily analysis. Since mixtures containing lower ACTHconcentrations require a longer time to reach maximalB/F ratios than those with higher concentrations, theinitial slope of the standard curve is sharpest when in-cubation is carried to equilibrium.

Separation of antibody-bound and free hormone.Separation is effected either by paper chromatoelectro-phoresis (2) or by adsorption of the free hormone totalc (9). All tubes are kept at 4VC until separation iseffected. When ACTH-I is employed, only the talcseparation is used since the counting efficiency for 'I inautomatic paper strip Geiger counters is very poor.Since it is inconvenient to apply more than 300400 /Al topaper strips, the chromatoelectrophoretic technique isused only when ACTH-lI can be prepared at a specificactivity of 1000 mc/mg or greater. Otherwise it is diffi-cult to obtain satisfactory counting rates with a suitablylow tracer concentration.

Conditions for separation are as follows:Paper strips: Whatman 3 MCor Toyo (No. 514) pa-

per strips, width 11 inches; chromatoelectrophoresis in0.05 M phosphate buffer, constant voltage 600 v. Suffi-

4 Veronal buffer, used for other hormones (2), in-hibits, slightly, the adsorption of free labeled ACTH tothe paper and tends to broaden the peak of free hormoneat the origin.

0.8-

N--

I('a()

z

mii.

ANTISERUM GP 330-5-271:125 X 103 DILUTION

|x Li U* LERNER-UPTON8B

200 500 pupug/mlHUMAN ACTH CONCENTRATION

FIGURE 6a Standard curve with human ACTH prepa-rations of C. H. Li and of A. Lerner and V. Upton. Mix-tures prepared in hypopituitary plasma diluted 1: 5 in0.02 M Veronal buffer.

Radioimmunoassay of ACTH in Plasma 2731

.:,A,

0.8-

I

-)< 0.4-z

m

co I LL-

ANTISERUMGP 330-5-271:125 X 103 DILUTION

STANDARDIN PLASMAOF* HYPOPITUITARY PATIENTX DEXAMETHASONE-

SUPPRESSEDNORMAL

500 ,uug/mlHUMANACTH CONCENTRATION

FIGURE 6b Standard curve with Lerner-Upton humanACTH 8B prepared in 1: 5 dilutions of plasmas fromhypopituitary patient and dexamethasone-treated normalsubject.

cient human plasma is added to incubation mixtures im-mediately before application to paper strips to bring thetotal volume of applied plasma to approximately 100-125,u per strip.

All strips are counted twice in an automatic stripcounter at a rate of 12 inches/hr. Generally, at least4000-5000 (usually more) counts are accumulated foreach peak (B and F) of radioactivity.

Talc: One 50 mg talc tablet is added to each incubationtube containing 2.5 ml volume. After dissolution of thetablet, which requires about 1-2 min, the contents arebriefly agitated by touching the bottom of the tube to aVortex mixer; the tubes are then centrifuged for 25-30min at 3000 rpm. Supernatant solutions containing anti-body-bound hormone are decanted from the well-packedtalc, to which the free hormone adsorbs. Supernatantsolutions and talc precipitates are counted separately inan automatic well-counter attached to a y-ray spectrom-eter; appropriate corrections are made for differences ingeometry, where necessary. Tubes are counted at leasttwice for a sufficient time to collect at least 10,000 countsfor each tube.

Miscellaneous. Data are processed with the aid of aWang Loci -2 Computer, which is programmed to correctfor damage in control mixtures and to correct for dif-ferences in geometry of talc and supernatant solutionscounted in the well-counters.

Plasma concentrations of 17-OHCS were determinedby Dr. Dorothy Krieger by a modification of the Sweatfluorescent method (17). Plasma glucose concentrationswere determined in the AutoAnalyzer by the Somogyi-Nelson method (18); blood for these determinations wascollected in tubes containing sodium fluoride.

Blood samples for ACTH assay were taken intoheparinized syringes and centrifuged immediately in arefrigerated centrifuge. The separated plasma was storedat - 15'C.

RESULTS

Standard curves and sensitivity of assay. Al-though administration of dexamethasone did notalways completely obliterate plasma ACTH, thestandard curves obtained in plasmas with unde-tectable ACTH were essentially identical withthose obtained in plasmas from cortisol-dependentpatients with classical hypopituitarism (Figs. 6 b,c). With one exception, B/F ratios in all hypo-pituitary subjects were essentially the same whenthe plasmas were assayed at the same dilution.When assays were carried out in hypopituitaryplasmas at different plasma dilutions, the B/Fratios were always higher in the more dilutedplasma, at least between dilutions of about 1: 2 to1: 10 (Fig. 6 c). This difference appears to beprincipally, if not entirely, the result of a salteffect. If NaCl is added to the buffer or if thebuffer strength is increased significantly, thecurve obtained with plasma at high dilution ap-proaches that obtained with more concentratedplasma. Standards prepared in hypopituitaryplasma at 1: 5 dilution and in 1% albumin in thesame buffer fortified with 0.18% NaCi give similarcurves (Fig. 6 d).

It is well known that antigen-antibody com-plexes are dissociated at extremes of pH. This isalso true in the present system but we have not

ANTISERUM GP 330-5-27 125 X 103 DILUTIONASSAY 2/14/67

1.0

H--C-)I

z

Im

TRACER CONCENTRATION3,u~ug/mlTALC SEPARATION

Y, ----I

100 200,upug/mlHUMAN ACTH CONCENTRATION

FIGuRE 6 c Standard curves in 1: 5 and 1: 10 dilutionsof plasma.

2732 S. A. Berson and R. S. Yalow

0.8

F-

(-)I

< 0.4-z

D -

ImnL

ANTISERUM GP 330-5-27 1:50 X 103 DILUTIONASSAY 3/10/66

TRACER CONCENTRATION12 ,u,ug/mlPAPER SEPARATION

1: 5 DILUTION OF PLASMA FROMHYPOPITUITARY PATIENT

. HUMANSERUMALBUMINg/IOOml IN 0.18% NaCI

I

250HUMANACTH CONCENTRATION

FIGURE 6 d Standard curve in 1: 5 dilution of plasma in 0.02 M V1% serum albumin plus 0.18%o NaCi in 0.02 MVeronal.

500,uug/mi

Ieronal and in

noted any definite differences in plasma diluted about 50%o from the initial ratio at a concentration1: 5 in phosphate buffer 0.02 M, pH 7.4 and in of 10-20 ,uwg unlabeled ACTH/ml and a readilyVeronal buffer 0.02 M, pH 8.6. measurable decrease (220%o) in B/F ratio at

When tracer concentrations did not exceed 1.5- 2 /A/Ag/ml (Fig. 7 a). In assays in which a tracer2 /,ug/ml, the B/F ratio usually showed a fall of concentration of about 1-1.5 juqLg/ml was used, a

ANTISERUM GP 330-5-27 1:125 X 103 DILUTION

0.81

-r

I

z

MILL

10 20 30 40 50 60 70 80 200 400HUMANACTH CONCENTRATION-ppug/ml

FIGURE 7 a Standard curve for human ACTH in 1:5 dilution of plasma.Tracer ACTH-'I concentration 1.5 jug/m1.

Radioimmunoassay of ACTH in Plasma 2733

concentration of 1 pjuqg/ml could be detected. Inplasmas assayed at 1: 5 dilution, 5-1 0 p~g ACTH/ml whole plasma is therefore measurable, the lowervalue obtaining in the more recent assays.

There are some doubts concerning the immuno-chemical integrity of the labeled ACTH. In earlierstudies with insulin-131I (19) and human growthhormone (HGH)-131I (20), we found no differ-ences in immunoreactivity between labeled and un-labeled hormone; i.e. a tracer quantity of labeledhormone fortified by increasing amounts of un-labeled or labeled hormone at the same hormoneconcentrations gave similar curves. However, insimilar experiments with labeled ACTHwe havefrequently observed that the B/F ratio is lowerwhen the same increment in ACTHconcentrationis provided by the labeled hormone than by theunlabeled hormone. The possibility that overio-dination may be at fault is suggested by observa-tions that selection of heavily iodinated insulin(6) or HGH(21) components results in dimin-ished immunoreactivity compared to componentsof lower iodine numbers in the same preparations.Nevertheless, until proven otherwise with ACTHlabeled at a low iodination level, the possiblity can-not be excluded that the preferential site for iodina-tion is near or within the antigenic site for theantiserum used in these studies and that the io-dine substitution reduces the imunoreactivity. Itshould be emphasized that whatever influence thelessened reactivity of ACTH may have on thesensitivity of the assay, it is, per se, irrelevant tothe validity of the assay, which depends on theidentity of reactivities, not of labeled and unla-beled ACTH, but only of standard pituitaryACTHand endogenous plasma ACTH.

Specificity of assay. As indicated above, theimmunochemical reaction is sensitive to the ionicenvironment. However, since both standards andunknowns are incubated in plasma diluted simi-larly in the same buffer, there need be concernonly for the possible effects of variations in thenonionic constituents of plasma. Of those plasmasubstances that are likely to vary widely, neitherglucose nor urea, in concentrations found clini-cally, appears to influence the reaction.

A necessary, if not sufficient, condition forspecificity is that the measured ACTHconcentra-tion in plasma decrease linearly with dilution ofthe plasma. This condition is met when plasma

ANTISERUM GP 330-5-27 1: 50 X 103 DILUTIONASSAY 12/16/65

1.2

H0-r

rna0.6-z

Imilt

DILUTION IN HYPOPITUITARY PLASMAX STANDARDHUMANACTH* PLASMA OF DOA PLASMA OF RE

HUMANACTH200 puug/ml

CONCENTRATION

FIGURE 7 b Plasmas of two patients (Do. and Re.) withhigh concentrations of endogenous plasma ACTH wereassayed at multiple dilutions by mixing with appropriatequantities of hypopituitary plasma. All tubes includingstandards contained plasma at 1: 2.5 dilution in 0.02 MVeronal buffer. The concentration in whole plasma wasdetermined from the assay at 1: 2.5 dilution (see TableIII). The concentrations at other dilutions were thencalculated from the dilution factors and plotted againstthe observed B/F ratios.

containing a high concentration of hormone isprogressively diluted in hypopituitary plasma andthe points fall along the standard curve (Fig. 7 b,Table III). This indicates that in this antiserumthere is no discernible difference in immunochemi-cal behavior of endogenous plasma hormone andpurified extracted pituitary hormone added toplasma.

We have earlier (22) compared the cross-re-activities of certain fragments of ACTH and ofthe intact hormone in the antiserum used for as-say. The hormonally inactive 11-24 and the Cterminal 25-39 fragments reacted very poorly(0.2-1%); the hormonally active 1-23 (tricosa-peptide) and 1-24 (tetracosapeptide) fragmentsreacted more strongly but still no more than5-10% as well as the intact hormone when cross-reacting with the labeled intact hormone. Thus, ifmiddle or C terminal fragments should be presentin the blood stream they are not likely to be de-tected. If the tricosapeptide or tetracosapeptide arepresent they would be measured like ACTH bybioassay but would be grossly underestimated byimmunoassay. 8-MSH was previously found to beessentially inactive in another antiserum (1) and

2734 S. A. Berson and R. S. Yalow

$

TABLE I I IEndogenous ACTHin Unknown Plasmas Diluted in

Hypopituitary Plasma

ACTHconcentration,j&g/ml

Plasma Undiluteddilution In sample plasma

1:100 Patient D. O. 5 5001:40 9 3601:20 18 3601:10 47 4701:5 110 5501:25 200 500

Mean 460

1:100 Patient R. E. 2 2001:40 4 1601:20 8 1601:10 22 2201:5 40 2001:2.5 70 175

Mean 186

was not tested in the antiserum used in the pres-ent study.

Perhaps the most important criteria for speci-ficity of any assay are related to appropriate bio-logic behavior of the hormone under conditionsknown to be associated with stimulation or.sup-pression of hormone secretion. Observations ofthis nature are reported subsequently.

Reproducibility

Reproducibility of B/F ratios in same incuba-tion sample when bound and free hormone areseparated by paper chromatoelectrophoresis andby talc. Wehave previously shown good agree-ment between the two separation techniques (9).This is confirmed for most samples in anotherlarger series here (Fig. 8 a). However, in theplasma samples of one subject, M, which werequite damaging to the hormone, agreement wasless satisfactory. When damage is excessive it isbest to discard results or to judge them only semi-quantitatively. Fortunately, such highly damagingplasmas as M's are rare under the protective con-ditions employed presently and have not been en-countered among the last thousand samplesassayed.

Reproducibility in duplicate separate incubation

tubes prepared from the same plasma samples inthe same assay. When duplicates of the sameplasma samples were run as the first and lastgroups to be set up in an assay, results were gen-erally in excellent agreement, except for the nearzero values (Fig. 8 b). Wehave found it desirablenot only to run a number of such samples in dupli-cate at the beginning and end of each assay butalso to divide the standards into two sets of al-ternating numbers, one to be set up at the begin-ning and the other at the end of the assay. Thepurpose of this procedure is to confirm that anti-body concentration remained the same throughoutthe assay. In some earlier experiments we foundthat, because of the relatively large volume (1000-2000 ml) of the antiserum-diluent mixture em-ployed for an assay, the marked dilution of anti-serum, and the dead space volume of the auto-matic dilutor, it was advisable not only to mix theantiserum-diluent mixture more thoroughly thanwas originally thought necessary, but also to cir-culate it through the dilutor for at least an hour,in order to avoid inhomogeneities in antibodyconcentration.

Reproducibility in different assays. When thesame plasma sample was reassayed, the plasmabeing stored frozen between assays, the reproduci-

0.8

0.7

0.6-

0.5-J

mL0.3-

0.2

wD

,00.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

T (PAPER STRIP CHROMATOELECTROPHORESIS)

FIGuRE 8a Separation of bound and free labeled hor-mone by talc and paper strip chromatoelectrophoresis.B/F ratios (corrected for damage in control tubes) ob-served in the same samples separated by chromatoelec-trophoresis and by talc. Samples of Mwere particularlydamaging to the labeled hormone.

Radioimmunoassay of ACTH in Plasma 2735

m

0I *

A

A

. A

a 0

AMAYA 12/4 Xl

0* 13/24 1It~~~~~~~~~/

A I~~~~~~0/24 A000 0 , z /12 0

I I I I 10 I,20 40 60 80

PLASMA B

FIGURE 8 b Reproducibility of plasma ACTHconcentrations in same assay. Hormone concen-trations in duplicates of the same plasma sam-ples determined in the same assay.

100 200 300 ,Pg/ml

bility (Fig. 8 c) was less good than that observedbetween duplicates in the same assay (Fig. 8 b).This was particularly true of the samples contain-ing either very high or fairly low concentrations inthe comparison assays of 4/24 and 6/1. This is at-

ppgJ gnlaIdnn*tvv

300-

200-

loo0

.M

(f)

a.

80

60

40-

20

11~~~~~~~1

/0

Oy

1.1

tributable to the fact that the standard curve on theassay of 4/24 was somewhat flatter initially andsharper at the end of the curve than was usuallyobserved, which could be accounted for by ran-

dom errors in one or two points of each of these

0tX*0

DATES OF ASSAYS2/14, 3/173/17, 3/244/24 6/1 *

20 40 60 8O 100 200 300 400 500 600 ,upg/mlPLASMA B

FIGURE 8 c Reproducibility of plasma ACTH concentrations in different assays. Hor-mone concentrations in duplicates of the same plasma samples determined in differentassays.

2736 S. A. Berson and R. S. Yalow

'I,'-

200'

100o

80'

60

40-

20

owl

portions of the curves. Except for some near zerovalues, the largest percentage differences wereobserved with a group of plasma samples that as-sayed at about 8-12 pug/ml on the assay of 6/1and about 25 pjug/ml on the assay of 4/24. Toavoid such errors we now set up two standardcurves for each assay (12 points each), either twoseparate hormone preparations (known to beequally potent) in the same hypopituitary plasmaor one hormone preparation in two different hypo-pituitary plasmas.

Plasma ACTHin subjects without pituitary oradrenal disease

In hospital personnel and in clinic patients with-out serious illness, plasma concentrations at 8-10a.m. only infrequently exceeded 50 u~ug/ml andaveraged 22 Nsg/ml (Fig. 9). However, consid-erably higher concentrations were often observed

in hospital patients with minor disorders as wellas in more seriously ill patients (Fig. 9; see alsoFig. 10). If the five values greater than 250pu.&g/ml are excluded, the average for this groupof hospital patients is 45 pqig/ml and the medianvalue is 25 uug/ml.

In a small group of nine normal subjects (Fig.10) the mean plasma ACTH concentration at10-11 p.m. was 9.6 ,u~g/ml.

Diurnal variations (Figs. 10-12)

Although a well-marked circadian rhythm, simi-lar to that reported for plasma cortisol, was notalways evident in healthy personnel, the lowestvalues of the day were usually found between 6p.m. and 11 p.m. (Fig. 10). Because of the incon-venience associated with blood sampling throughthe night, an early morning peak between 4 a.m.and 6 a.m. may have been missed in some of the

NDR ESWIELL ADRENIAL EXTUATUfTARY OEUSNGSDISEASETEM IN HOSP ATENS AVROMEGALY TNORWT POST BILATERAL

OfF STEROID OFF STEROID OFFSTEROIDONSROID ON STEROID ON STEROID

roo'

> ?50-z0

5200

H

Ca:

Z 250LiJ0zo 200-0

H- 150-0

U)

1i50-0-

PERSISTANT

8emDEXA/DAY

DEX/W-.-

FIGURE 9 Plasma ACTH concentrations, 8-10 a.m. Cases shown here do not includethose shown in the other figures except for the laboratory personnel in Fig. 10. Thefive in-hospital patients with the highest ACTH levels were all severely and acutely illand carried the following diagnoses: Graves' disease, pneumonia and congestive heartfailure, bronchogenic carcinoma, severe nonmalignant chest disease, and nephritis withuremia.

Radioimmunoassay of ACTH in Plasma 2737

* 0 0

00 :::*

0 0

4'* :

4. : :: ::

: ": :":: :. . : .. . 0

0: "!.O.O.,- 0-

LABORATORYPERSONNEL

150-, I' .,

I150sO SIL-CONVALESCENTPOST EYE

We.- SEIZURES- SURGERYDILANTIN

lI,_

I1

150 iZo-NO DISEASE

8A 8P

NOSLEEP100- 50- v

o k ~~ ~~~~50-E. We

50

SLP

50- 0

SLEEP

8A 8P 8A 8A 8P

TIMEFIGURE 10 Plasma concentrations of ACTH throughout the day in healthy hospital personnel andin hospital patients without severe acute illness. The data of the four hospital personnel in the leftcolumn of that group have been reported previously (22).

subjects. Although concentrations were frequentlyrather high at 6 a.m.-8 a.m. in hospital patients,including a group of subjects with a variety ofobscure chronic neurologic disorders but withoutsevere acute disability, a fall to less than 40-501tyug/ml later in the day was usually observed(Fig. 10; see also Table V B, C). Two of theneurologic patients (We. and Au.), who had per-

sistently high ACTH concentrations during theday, had been receiving dilantin. In two other pa-

tients, not reported here, plasma ACTH concen-

trations were considerably higher while on dilantintherapy than prior to therapy.

In another group of healthy subjects, in whomplasma 17-OHCS concentrations every 4 hr hadbeen determined by Dr. Dorothy Krieger, a peakin plasma ACTHconcentration was not always de-tected preceding the rise in plasma steroid (Fig.11, note particularly all studies on St.). Becauseof the short half-life of ACTH it is possible tomiss the rise and fall in plasma ACTHconcentra-tion associated with a transient spurt in secretionwhen the interval between sampling is as long as

4 hr. To evaluate the temporal relationship be-tween plasma ACTHand steroid levels, we tookplasma samples at intervals of 1/2-1 hr over one or

two 24-hr periods from the subjects of Fig. 11 as

well as from another subject. Comparison of

plasma ACTHand steroid levels in these subjectsreveals a good degree of parallelism in most cases

(Fig. 12), the spikes in ACTH concentrationspreceding the rises in cortisol concentration. Theparallelism is most remarkable in Fr. 1/25 andFu. 2/1, but is singularly poor in St. in whom an

anomalous relationship between the tropic hor-mone and the steroid had been observed on threeprevious occasions (Fig. 11). The range of ACTHlevels varies more greatly than the range of plasma17-OHCS concentrations among individuals.

Effects of metyrapone (SU 4885) (Table IV)

With the exception of C. Iv., who is a normalcontrol, studies in this group were performed on

hospital patients who were not severely ill; pa-

tient Lap., who was suffering from a postoperativewound infection, is an exception. When a singledose of metyrapone, 750 mg, by mouth, was givenat 9 a.m. after two control samples taken 1 hrapart, plasma ACTHconcentrations changed littleor continued to drop for 2-3 hr longer, after whichthere was usually at least a slight transient rise inconcentration by the 3rd to 6th hr (Table IV A).Repeated doses of metyrapone, 750 mg every 4 hr,always resulted in very high ACTH concentra-tions by the following day.

When metyrapone was started at 6 p.m. at a

2738 S. A. Berson and R. S. Yalow

PATIENTS

150

A 150' us- uEL..,..I .. Au. -SE IZURES-

Mc. - SEIZURES--LANTIN

NO THERAPY

8A BP 8A 8P

150 Fu.1 J1/3-I1/4. AZ.~~~~~~~~~I

C"I50 Mc.

I I

F-

< 200 - Ke.1/5-1/6

U)

0-

IjN.M.

St.11/3-11/4

O

-.,. .04*s ,

St.10/12-10/13

i ~

40-4bfla-.% ]

St.9/20-9/21

0%I %4%

1 "1

*---q.

4P 4A 4PTIME

FIGURE 11 Plasma concentrations of ACTH and 17 OHCSthroughout theday in normal subjects. Plasma samples in this and the following figure were

supplied by Dr. D. Krieger who also performed the plasma 17-OHCSdeterminations.

dosage of 750 mg every 2 hr for three doses fol-lowed by 750 mg every 4 hr till the next morning,there were observed by 11 p.m. a distinct rise inplasma ACTHin only one subject and less defi-nite rises in two other subjects of the eight studiedon this regimen (Table IV B). By the followingmorning, however, plasma ACTHwas elevated inall cases.

When metyrapone was given in doses of 750mg, by mouth, at 2-hr intervals starting at 9 a.m.,

plasma concentrations usually began to rise within2-3 hr reaching quite high levels during the earlyafternoon but tending, in three of the four cases,

to decline by the evening (Table IV C). Simi-larly, when metyrapone was started at 4 p.m.

plasma ACTH concentrations tended to fall by11 p.m. even when an increase had been produceda few hours earlier. Again, in all cases, very highconcentrations were observed on the followingmorning (Table IV D).

When metyrapone was given by continuous in-travenous infusion over a 4 hr period in the morn-

ing, four of the six patients showed a distinct risein plasma ACTH by the end of the infusion or

within 3 hr. The mean increase was 300 ±140%o.

Effects of dexamethasone

After institution of dexamethasone treatment,0.5 mg every 6 hr, plasma ACTHconcentrations

at 8 a.m. were frequently undetectable and almostalways less than 10-20 tqpg/ml, even when mety-rapone stimulation had been carried out the previ-ous day (Table IV B) . Patient Lap. (Table IV B),referred to above, is an exception.

In a patient (J. F.) whose plasma ACTHhadrisen from 20 to 325 Juqug/ml after 1 day's treat-ment with metyrapone (750 mg every 6 hr), a

single oral dose of 2 mgof dexamethasone produceda sharp fall in plasma ACTHto 19 ,ug/ml in 3%hr. Under continued dexamethasone treatment(0.5 mg every 4 hr) plasma ACTHremained atless than 12 ju/g/ml for the next 20 hr in spite ofcontinued treatment with metyrapone (750 mg, bymouth, every 4 hr).

Effects of "stressful" stimuli (Table V A, Fig.13)

Surgery (Table VA). Plasma ACTH con-

centrations were frequently high even preopera-

tively but were generally much higher postopera-tively in patients given general anesthesia even

though the latter samples were obtaind later in theday. No rise in plasma ACTHwas observed post-operatively in two patients who had undergonesurgery under spinal anesthesia.

Electric shock therapy (EST) (Table V B).In a group of patients undergoing EST for variouspsychodepressive disorders, a sharp transient rise

Radioimmunoassay of ACTH in Plasma 2739

-30

00

3o I-30 Ck

en

or0,--

30 <

* U)0~AL-

4AI

TABLE IVEffects of Metyrapone and Dexamethasone on Plasma A CTHConcentrations (jA1g/ml)

A. Metyrapone 750 mg p.o. at 9 a~m. and q 4 hr from 3 p.m.-7 a~m.Day 1 Day 2

Subject 8 a.m. 9 a.m. 10 a~m. 11 a~m. Noon 1 P.M. 2 p.m. 3 p.m. 8 a.m. 3 p.m.

Ha. >1000 175 25 50 70 115 140 140 50 300Re. 40 90 55 55 100 80 70 60 > 1000 200Mim. 25 15 20 15 20 50 30 70 >1000 >1000Go. 35 50 40 40 70 30 25 30 180 70Gel. 40 40 50 50 70 65 60 70 200 60SC. 20 20 20 20 15 90 15 15 125 >1000Mis. 10 35 15 15 20 55 60 30 750 220De. 70 40 25 40 60 50 20 40 1000 500Lab. 60 40 45 95 105 - - 1000 250 250Bi. 65 95 90 130. 110 125 150 95 1000 300

B. Metyrapone 750 mg p.o. q 4 hr from 6 p.m. day 1 to 6 a~m. day 2 followed by dexamethasone1 mg p.o. q 6 hr 9 a.m. day 2 to 3 a~m. day 3

Day 1 Day 2 Day 3

Subject 6 a.m. 8 a.m. Noon 6 p.m. 8 P.M. 11 P.M. 8 a.m. 8 a~m.

Lin. 230 50 30 20 20 45 500 10Lap. 160 130 120 140 -140 300 90Bas. 10 15 45 10 15 35 250 10Arn. 50 40 30 30 -15 100 0Mul. 100 180 70 55 50 130 100 10DeG. 55 25 30 10 10 10 300 10New. 70 55 50 30 25 20 200 15Hat. 60 50 40 60 30 20 130 5

C. Metyrapone 750 mg p.o. q 2 hr 9 a'm.-il p.m.; 3 a~m., 7 a.m.

Day 1 Day 2

Subject 8 a.m. 9 a.m. 10 a.m. 11 a.m. Noon 1 P.M. 2 p.n. 3 p.m. 6 p.m. 8 p.m. 11 P.M. 8 a.m. 9 a.m.

K. T. 230 175 45 90 110 100 115 175 250 135 10 170 470G. S. 110 57 40 125 170 175 160 100 180 58 225 700 700E. M. 48 5 10 15 27 50 30 25 150 135 95 - 425C. Iv. 0 23 43 70 43 80 150 175 80 15 25 185 280

D. Metyrapone 750 mg p.o. q 2 hr 4 p-m.-10 p.m.; 2 a.m.. 6 a.m.

DaylI Day 2

Subject 3 p.m. 4 p.m. 5 P.M. 6 p.m. 7 p.m. 8 p.m. 9 P.M. 10 P.M. 11 P.M. 8 a.m. 9 a.m.

L. P. 40 20 18 35 35 30 35 90 20 200 500W. M. 45 55 35 28 80 125 115 110 55 430 750J. E. 100 70 55 115 200 115 160 180 175 440 -C. D. 20 15 15 35 0 40 30 25 30 250 200

in plasma ACTHcharacteristically occurred within pital patients, a result perhaps attributable to the5-10 min of the shock, after which concentrations apprehension associated with anticipation of thefell precipitously and remained below preshock treatment.levels. It appeared that ACTHsecretion was sup- Hypoglycemia. When a fall in plasma glucosepressed after the initial stimulation. to less than 50 mg/lOG mfl was induced by intra-

It may be noted that the control values averaged venous administration of insulin, 0.1 U/kg body75 IA/g/ml, which was somewhat higher than the weight, plasma ACTHuniformly rose sharply andaverage concentration of 45 p.qpg/mil for other hos- then fell rapidly again as blood glucose rose (Fig.

2740 S. A. Berson and R. S. Yalow

TABLE V

Effects of Certain Stresses in Plasma ACTHConcentratzon (js1sg/mt)

A. SurgeryPreoperation Immediate postoperation 24 hr after preoperation

Concen- Concen- Concen-Patient Operation Time tration Time tration Time tration

Aiqg/ml jqsg/ml jqg/mlGeneral anesthesia

R. T. Arthrotomy 8:15 a.m. 115 11:15 a.m. 320 8:55 a.m. 20V. A. Bile duct exploration 8:15 10 3:40 p.m. 40 8:25 15F. G. Cholecystectomy 8:00 115 2:00 280 9:50 0H. G. Herniorrhaphy 8:15 75 3:30 150 9:20 70C. McG. 1. nephrectomy 8:20 10 3:00 175 9:00 25F. A. Subtotal thyroidectomy 8:35 250 12:10 500 9:30 135W. McD. Right lower lobectomy 9:15 85 12:15 45 8:15 45A. A. Craniotomy 8:25 180 1:45 130 8:40 160D. A. r. adrenalectomy in

patient with Cushing'ssyndrome previouslysubjected to 1. adrenal-ectomy, 75 mgcorti-sone q 6 hr 9:00 65 11:15 a.m. 275 9:00 65

Spinal anesthesiaM. B. Prosthesis 1. hip 8:20 250 12:25 p.m. 65 9:10 20H. T. Herniorrhaphy 10:00 60 1:15 90 8:45 80

B. Electroconvulsive shock therapy (sedated, 30 sec shock)Patient Date 0 5-10 15 30 60 120 min

Co. 3/1 48 75 37 35 27 22Co. 3/8 125 200 100 60 30 10Co. 3/20 75 180 75 32 17 10'De. W. 3/3 100 250 125 220 100De. W. 3/20 100 400 166 102 55 50Gu. 8/28 95 450 400 95 45 30Gu. 9/6 60 500 270 175 52 10Go. 8/18 55 500 230 125 55Go. 8/28 80 300 - 115 45 20Cg. 9/13 25 280 230 165 115 0Cg. 9/22 45 90 50 35 0Pa. 9/18- 160 800 650 450 180 70Mu. 9/13 60 320 200 125 80 40Li. 2/1 60 80 15 12 22Le. 2/1 27 40 25 26 17 12

C.Histalog, 100 mgPatient Injection -10 0 10 20 30 40 50 60 min

1.C.*~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I. C.* i.m. 45 30 35 50 80 175 300 370E. B. s.c. 0 0 0 15 30 0 0 0J. C. s.c. 0 0 0 0 10 0 0 0

D. Lysine vasopressin, 10 U, i.m.Patient -10 0 10 20 30 60 90 120 min

J. R. 60 60 65 130 130 80 65 50A. P. 30 30 30 42 60 55 35 45J. T. 30 20 20 20 28 20 20 15

* Complained of headache (15 min) and abdominal pain (30 min).

Radioimmunoassay of ACTH in Plasma 2741

WC 12/13

I AtJt~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Wa 1/11 a-

or

./\,01~~~! I~A

Fu. AI /4

lI~4

I !ThCWA

N 4P 8P M 4A 8A N

-24

00

C-,)C-

(fr)

I32 0

c:EC!)

0-

TIMEFIGURE 12 Plasma concentrations of ACTH and 17-OHCS throughout the day in normal subjects.

13). However, on the subsequent decline in plasmaACTHthe concentration rarely fell below controllevels. In two nondiabetic subjects in whom thelowest plasma glucose levels achieved were 51mg/100 ml (F.E.) and 55 mg/100 ml (S.E.)and in one diabetic subject (R.O.) whose plasmaglucose concentration fell 100 mg/100 ml to 72mg/100 ml in 30 min, only slight transient in-creases were observed in plasma ACTH. How-ever, in two other diabetic patients a significant

rise in ACTHconcentration occurred despite thefact that the lowest plasma glucose concentrationsrecorded were 70 mg/100 ml (A.R., Fig. 13) and80 mg/100 ml (G.O., not shown); the latter pa-

tient exhibited typical symptoms and signs of a

hypoglycemic reaction at a time when the bloodglucose was 80 mg/100 ml (preinsulin controllevel 125 mg/100 ml), after which plasma ACTHrose to an extremely high concentration (7001'g/ml).

2742 S. A. Berson and R. S. Yalow

3751200-

30*St. I11/4

II160 -

I

LH-

CD-

AC!

u

F-I0Z

Ke 1/6

itISG ,2

-40 220

40

200 Mc 1/6

Any- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~in-24

F24 60 241501

F24

The effectiveness of hypoglycemia in stimulatingACTHwas not significantly different at 9:30 a.m.and 4:30 p.m. but hypoglycemia was ineffective inthe morning or afternoon after 1 day of suppres-sive treatment with dexamethasone (Fig. 14).5

Histalog administration (Table V C). One pa-tient (I. C.) given 100 mg Histalog intramuscu-larly complained of headache and considerableabdominal pain and showed a marked sustainedrise in plasma ACTHconcentration. Two patientsgiven the same dose subcutaneously without dis-comfort showed smaller, transient, increases inplasma ACTHwithin a half hour.

5In most patients treated for 1 day with dexametha-sone, 1 mg every 6 hr, a higher fasting blood glucoseconcentration was observed than on the preceding dayand (or) the usual insulin dose of 0.1 U/kg failed toproduce a sufficiently marked hypoglycemia to provide anadequate test of the ACTH response, which was usuallycompletely negative. In two such patients, tests per-formed 24 hr later, after discontinuation of dexametha-sone, showed hypoglycemia and ACTHstimulation equalagain to that in the control studies.

0.1 U INSULIN/KgBODYWT IV

Effect of vasopressin (Table V D)

After intramuscular injection of 10 U vasopres-sin, two of three patients exhibited a doubling ofthe plasma ACTH concentration within 20-30min.

Plasma ACTH in patients with pituitary-adrenal disease

Hypopituitarism. In plasma samples from eightpatients known, on clinical grounds, to be deficientin ACTHproduction, B/F ratios in the same as-say were equal to each other within experimentalerror, and equal or higher than ratios obtained indexamethasone-suppressed normal subjects. How-ever, in one patient (Sm.) receiving steroid ther-apy, who shows good clinical evidence of hypo-pituitarism and is thought to be steroid-dependent,the B/F ratio determined twice on the same a.m.plasma sample was lower than that observed inother hypopituitary patients and corresponded to aplasma ACTH concentration of 15 1,/g/ml. We

|FACTH |lG COSE

-

00

I E

Io

w0C-)

IU)

-Ja.

60 120 60 120 60 1I2Q 60 120

TIME IN MINUTESFIGURE 13 Effects of insulin-induced hypoglycemia on plasma ACTHconcentrations.

Radioimmunoassay of ACTH in Plasma 2743

BEFORE DEXAMETHASONEAFTER DEXAMETHASONEmg q 6hrx4

20 U INSULIN 30 U INSULIN400 .AV. M IV.9 30A A930A

; I\ ~~C.W.300

I 200

E 100

:21

400 10 U INSULIN 20 U INSULIN i,,40 i.v. iv.V

< 430P 4 30P (I ~~~~~~~~J.F.D

:Ea:; l'\ J X 150s~~~~10~300 -J5

~200 .--

100--' /C

0 60 120 0 60 120TIME IN MINUTES

FIGURE 14 Effects of hypoglycemia on plasma ACTHconcentrations in morning and afternoon and the influ-ence of prior treatment with dexamethasone, 1 mg every6 hr for 24 hr.

were unable to study ACTH reserve in thispatient.

Hypoadrenalism. In patients with hypoadrenal-ism due to primary Addison's disease or secondaryto bilateral adrenalectomy for severe diabetes mel-litus or pheochromocytoma, plasma ACTH con-centrations were markedly elevated when steroidtherapy was withdrawn (Fig. 9) and normal inone patient receiving maintenance doses of steroids(Fig. 9).

Cushing's disease (Fig. 15). We have hadthe opportunity to study only a few untreated pa-tients with this disease. In one patient (Car.)plasma ACTHconcentrations preoperatively wereabout 200 uf~g/ml throughout the day but wereconsiderably higher after bilateral adrenalectomyeven though the patient was receiving 50 mg cor-tisone every 6 hr. In another patient (La.) withpersistent disease after removal of the right adrenaland % of the left adrenal, both of which weremarkedly hyperplastic, plasma ACTH level re-mained about 800 ,u~g/ml throughout the day.In a third patient (D'A') with persistent hyper-cortisolemia after unilateral adrenalectomy, plasmaACTH levels remained above 120 f,~ug/ml

throughout most of the day. After removal of thesecond adrenal, plasma ACTH levels remainedabove 350 tu~g/ml even when the patient was onmaintenance cortisone (37.5 mg/day). In twoother cases the patterns were not sufficiently ab-normal under the conditions of the studies to con-clude that ACTHsecretion was not suppressed byhigh plasma cortisol levels.

After bilateral adrenalectomy plasma ACTHconcentrations were very markedly elevated whensteroid therapy was withdrawn for 48 hr or longer(Fig. 9). In some cases, the B/F ratios were lowerthan that corresponding to the last point on thestandard curve and assay at a higher dilution ofplasma would be needed to provide a precise quan-titative determination of ACTHconcentration; incases in which this has been done, values as highas 10 m/ig/ml have been obtained. Plasma ACTHremained high postadrenalectomy in patients onreplacement steroid therapy.

Congenital virilizing adrenogenital syndrome ofthe 11i-hydroxylase deficiency type. One patientwith this syndrome had an ACTH level of 4001Wg/ml.

Plasma ACTH concentrations in mother andnew born at time of delivery. In seven of eightcases, plasma ACTHconcentrations were signifi-cantly higher in the cord blood (mean 161 ±29[SD] ,jqtg/ml) than in maternal venous blood(mean 56 ±23 [SD] IA/Ag/ml). Concentrationsabove 60 puLg/ml were observed in only two of themothers but in all of the cord blood specimens.

Turnover of plasma ACTH

The disappearance of ACTH from plasma af-ter acute stimulation by EST or hypoglycemiacannot be taken as a measure of the turnover ofthe hormone since the initial decline in concentra-tion after a short-lived peak must be related prin-cipally to distribution into extravascular spacesrather than to metabolic turnover alone. However,the fall in concentration to very low levels insuch studies and others to be described, does serveto indicate that metabolic products of ACTHca-pable of reacting in the immunochemical systemdo not return in significant quantities to theplasma.6

6 Return to plasma of peptide fragments released byproteolytic degradation of ACTH in tissues would notbe detected if such fragments were only poorly immuno-

2744 S. A. Berson and R. S. Yalow

k4LA

Old'

)o

0

C-)R4 O_) p<a IJ. Ia, ICKJ I

a.I II- I

J II4 I<: IILI

LJm %

I

0 1C) I

I

Qia-

Ix

J E

4a:L1

WZL E

-J ..

I I0< (-

0.r-I

00

wo-JzOw

oO

I 0

L:

cr

0LL

/)0

a.MO

I-LU

Za:

aL

I-a:

/)W

Ct) Li

a_

L')

)'-J

zLiJ0

cn _C,)4I

a- U) >-x 0oa. w>- (1)1I e

amz:I -Ct)

Li

Z F-

>0w IVCLw

dJ.

V

-jv

41

a0z

0

E'A

N'CYN

CA)ux0

a:4Z

ia0

I IT* N

o 0o 00 dJSV-ld

CUC

~cd0d

- 0cdU

I-

OC)o

en

C'sa e 0

CUd

00C)

C..

' 0

ob E

CUCLC)0

LU Y

co=

H

00

CP

A

00CUd

C..0_)

-0c

Cdt

* 0U..

C C

a, C

0

u

.5d cn

'-,n'ab S.

o V

corZ 04

0') E,a

z )

cU)

Radloimmunoassay of ACTH in Plasma

cn

-Ja.

I

crJm

LLI-

4

-J

z

Li0ar-0-I.Cti

z

Ntw

Cob

0)

EE E

oQ z

ON 0

Ec c

"

w

K8

ON <4F:0

NU~

o 0o 0

IW/i- HiOV

E3iCP

ol

0)

CEEf)CZ

VA

2745

RECURRENTPHEDCIOOCYTOMAPOST BILATERAL ADRENALECTOMY

fOFFMIMNTENANCE-1 y, CORTISONE72 hr

830 830 930 0

TIME IN HOURS

FIGURE 16 Effect of i.v. hydrocortisoneACTH concentrations. We are indebted toBrown, UCLA, for providing the samples ftient with pheochromocytoma.

In subjects who have maintaine(plasma ACTH levels for some hours ocan be assumed that the hormone hasits volume of distribution. The rate ofplasma ACTHafter suppression of se(then provide some estimate of the uppcturnover half-life.- Since suppression iscomplete the fractional rate of turnoverbe underestimated.

In a bilaterally adrenalectomized sul400 mg cortisone by constant intravesion there was observed on one occasioiincrease in plasma ACTH concentratlater repeat study, however, plasma Yfrom the outset with an initial half-life (

min (Fig. 16). In the normal subjectearlier in whomplasma ACTHhad rislevels under the influence of metyrapotfall in concentration was produced by t]ministration of dexamethasone (patieSince samples were not taken at suffi(quent intervals it is not possible to detewhen suppression took effect. Howeverriod from about 2%2 to 3% hr after in'dexamethasone treatment the plasma Areactive and can therefore not be excluded asparticularly in the case of the hormonallyterminal fragments.

centration fell from 125 to 19 ,utg/ml in 80 minindicating a half-life of not more than 30 min. Ina similar experiment in a patient given 10 mg hy-drocortisone intravenously, plasma ACTH fellfrom 165 to 82 ppg/ml between 8:58 and 9:20a.m., indicating a half-life of not more than 22 min(Fig. 16).

DISCUSSION

The improvement in sensitivity of the radioim-munoassay for plasma ACTHdescribed here hasbeen made possible principally by selection of asensitive antiserum and by the use of conditionsthat permit reduction of the "tracer" concentra-tion of labeled ACTH to about 1 futg/ml. Thelatter measure would not alone suffice with other

2 3 antisera that yield significantly less sharp slopesin the standard curve. In several experiments we

on plasma have observed that a still greater sensitivity is ob-Dr. Josiah tainable with this antiserum by using a "tracer"

From the pa- concentration of 0.5 jupLg/ml. This can be effectedeither by reducing the amount of "tracer" in the

d elevated same volume or by doubling the volume of incuba-)r longer it tion mixture. However, the present sensitivity of

s mixed in 1 ujug/ml is adequate for all practical purposes anddecline in the needs do not justify the further inconvenience

cretion can of prolongation of counting time or the use ofer limits of larger volumes. Under present conditions it isnot always possible to set up a 700 tube assay without diffi-is likely to culty. Once the plasma samples are pipetted, the

labeled hormone, antiserum, and diluent are addedbject given to 350 tubes/hr. With the talc tablets, bound andnous infu- free hormone are separated at the rate of more

n an initial than 150 tubes/hr.tion. Qn a Since a biologic unitage cannot be applied to im-kCTH fell munochemical reactivity, all plasma ACTH con-

of about 22 centrations have been given in tuyg/ml. However,t discussed for the purpose of comparison with the results ofsen to high bioassay it is useful to make the conversion fromie, a sharp ju~g/ml to mU/100 ml, in which units the resultshe oral ad- by bioassay have always been presented. The Lint j F.). human ACTH and the Lerner-Upton human:sently fre- ACTH8B did not differ significantly in immuno-Lrmine just chemical potency and we may take the biologic po-

in the pe- tency of the standards used as 140 U/mg. Astitution of plasma concentration of 20 ,pg/ml is thereforeLCTH con- equivalent in immunochemical potency to that of

a possibility, a standard with a biologic potency of 0.28 mU/r inactive C 100 ml. There is a paucity of bioassay data on

normal concentrations in human subjects because

2746 S. A. Berson and R. S. Yalow.

NORMALSUBJECTMETYRAPONESTIMULATION

2500 - METYRAE750 mgq 4hrx6

XAAA n1000

-.,50O-250-

<t 100-21

cnAJ 50-0L

0HYDR0c0RI.0WE0rmgOm..

..,A

,of the lack of sufficient sensitivity of the bioassaysto measure the low plasma concentrations. On thebasis of inability to detect hormone in oxycelluloseextracts of 40 ml blood by the adrenal ascorbicacid depletion test, Sayers (23) concluded thatnormal human plasma contains less than 0.5 mU/100 ml whole blood. Since then, other workersextracting larger volumes (200-1000 ml) wereable to detect the presence of hormone by thismethod and estimated its concentration in humanplasma to be about 0.7 mU/100 ml whole blood(24) and 1 mU/100 ml whole blood (25). UnlessACTH is present in blood cells, the values inplasma would be almost twice as great. Thegreater sensitivity of the Lipscomb-Nelson method(26, 27), which measures the increase in corti-costerone secretion into the adrenal vein of thehypophysectomized rat, enabled Vance, Reddy,Nelson, and Thorn (28) to measure the hormonein 10 ml plasma after concentration by lyophiliza-tion to dryness and redissolution in 5 ml 0.01 NHC1; at 8-9 a.m., the concentrations were foundto be in the range 0.4-1.0 mU/100 ml plasma.Smaller volumes of plasma (3-5 ml) do not gen-erally reveal hormone by this method (28-30).However, by extraction and concentration from250 ml plasma, Ney, Shimizu, Nicholson, Island,and Liddle (31 ) estimated concentrations at 6 a.m.and 6 p.m. to average 0.25 mU/100 ml plasma and0.11 mU/100 ml plasma, respectively. These esti-mates assumed that the hormone was completelyextracted from the large volumes of normal plasmaon the basis of earlier (30) recovery experimentson smaller volumes of plasma containing high hor-mone concentrations. In a previous study on radio-immunoassay of extracted normal plasma (cor-recting for incomplete recovery), we found a.m.concentrations to average about 0.6 mU/100 mlplasma (range 0.34-0.88) in nine normal sub-jects (1). By the same extraction procedure andradioimmunoassay method, Demura, West, Nu-gent, Nakagawa, and Tyler (32) found a meanvalue of 0.5 mU/100 ml (range 0.3-0.7 mU/100ml) in 10 normal subjects at 8 a.m.

In the present larger series, the concentrationsdetermined by radioimmunoassay in normal sub-jects (equivalent to 0.31 mU/100 ml plasma ina.m. and 0.13 mU/100 ml plasma in p.m.) cor-respond closely to the lowest values reported bybioassay. However, from the results of the present

study it would appear that the range of normalconcentrations may be quite great. The very rapiddisappearance of hormone from the circulation inthe stimulation experiments with insulin and ESTand the equally rapid spontaneous decreases fre-quently observed in control and hospital patientsindicate that peak concentrations of plasma ACTHmay be very short-lived. Samples taken a halfhour or an hour apart from the same individualmay show very great differences in concentration.The disagreements in values obtained from smallnumbers of subjects by different investigators maytherefore be more apparent than real. However,the values obtained by Ney et al. (31) at 6 a.m.average significantly lower than those obtainedeven later in the morning by other investigatorsusing the same assay system (28); it is possiblethat the extraction of the large volumes of plasmaby Ney et al. (31) was not as efficient as thatobserved with smaller volumes (30).

From the diurnal variations earlier observedby us in four cases (22) and Demura et al. in twocases (32) and from the additional observationsreported here, it would appear that an early morn-ing peak may frequently be missed in samplestaken at 8 a.m. or later, especially in persons ac-customed to early rising. Control subject C. Iv.,who showed an essentially flat curve from 6 a.m.to 11 p.m., usually leaves for work about 5 :30 a.m.and could be expected to have his peak ACTHconcentration several hours earlier. In hospitalpatients, however, high levels were frequently ob-tained at 6 a.m. although concentrations usuallyfell off later in the day.

Perhaps the essential point to be made regard-ing these plasma ACTHconcentrations is the lackof information that can be derived from isolatedobservations, except that the demonstration of ahigh level excludes the existence of ACTHdefi-ciency and a near zero value in association withhypocortisolemia provides evidence of pituitary in-sufficiency. Transient spurts in ACTHsecretion,the rapid turnover of plasma ACTH, anomalousdeviations from the "typical" circadian rhythm innormal subjects and the characteristic response ofACTH secretion to the acute disturbances inphysiologic equilibrium that are classified as"stressful" stimuli all contribute to fluctuations inplasma ACTHand complicate what might other-wise be a simple feedback regulation of ACTH

Radioimmunoassay of ACTH in Plasma 2747

secretion by steroid levels in one or more con-troller sites. It has elsewhere been emphasized(22) that, unlike the serum levels of calcium,uric acid, thyroxine, glucose, etc., concentrationsof several of the peptide hormones vary ratherwidely among normal subjects and tend to un-dergo marked fluctuations during the day in indi-vidual subjects. These characteristics apply toplasma ACTH, which even when closely parallel-ing the plasma cortisol concentration, shows amuch greater spread between high and low points.So, too, the 10-20-fold rises in ACTHconcentra-tion after hypoglycemia or other stimulation arefar more impressive than are usually seen in theplasma cortisol level in the general circulation.Even were the adrenal cortex to respond propor-tionately to the logarithm of the plasma ACTHconcentration, the longer half-life of plasma corti-sol would tend to blunt the changes in steroid con-centration but, in addition, there is reported to be adiurnal variation in the responsiveness of the cor-tex to ACTH (33). The present study indicatesalso that there are significant variations in plasmaACTH levels among normal individuals.

In view of the sporadic fluctuations in plasmaACTH, a full appreciation of the dynamics ofACTHsecretion requires multiple determinationsat frequent intervals and an evaluation of the re-sponses to stimuli and suppressants.

The results of the metyrapone experiments indi-cate that a single dose at 9 a.m. was not as ef-fective as repeated doses during the morning instimulating ACTH release. However, even whenan effective level of metyrapone was maintainedby repeated doses, plasma ACTH declined to-wards the evening in most cases. Also, the later inthe day that metyrapone was started, the moredifficult it was to elicit an increase in plasmaACTHover the next few hours. Similar conclu-sions were reached by Martin and Hellman (34)who found a smaller increment in urinary 17-OHCSexcretion in response to i.v. metyraponein the late afternoon and evening than in theearly morning, a difference that could not be ex-plained by the decreased responsiveness of theadrenal cortex to ACTH late in the day. Theseresults are consistent with the suggestion of Liddle(35) that this agent increases ACTH secretionprincipally through its inhibitory action on the11,8-hydroxylase activity and therefore on corti-

sol synthesis in the adrenal cortex. Others havesuggested that another action may be necessaryto explain its effect (28, 36). Indeed, Vance et al.(28) were unable to demonstrate a rise in plasmaACTH in normal subjects given metryapone in-travenously or by mouth or in Addisonian patientsgiven it intravenously; only slight increases wereobserved by Liddle, Island, and Meador (30)within 24 hr when metyrapone was given orally,although by 72 hr plasma ACTH concentrationshad risen to 1-2 mU/100 ml in all of the six sub-jects reported. In the present study of a muchlarger number of patients, concentrations hadreached 100-500 /qug/ml (equivalent to 1.4-7.0mU/100 ml) by the morning after institution oftherapy in every case even when metyrapone hadbeen started late in the preceding afternoon. Theinability of metyrapone to provoke a release ofACTH in the evening of the 1st day is not sur-prising if its action is solely through its inhibitionof cortisol synthesis, since the normally low plasmacortisol concentrations at that time are likewiseineffective. The diurnal rhythm of plasma ACTHand plasma cortisol are a consequence of the in-sensitivity of the hypothalamico-hypophyseal sys-tem to feedback stimulation at night although thepituitary-adrenal system may be even more sen-sitive to stimulation by vasopressin (37) and pyr-ogen (38) at this time. The ability of dexametha-sone (patient J. R.) and hydrocortisone (Fig. 16)to bring about an abrupt and sustained decline inplasma ACTH concentration in patients main-tained on metyrapone further supports the con-clusion of Liddle (35) regarding the mechanismof action of this agent.

The results obtained on the stimulation ofACTHsecretion by hypoglycemia, electric shock,Histalog, surgical procedures and vasopressin andthe inhibitory effect of dexamethasone on hypogly-cemia-induced secretion do not require extensivediscussion since they are simply confirmatory ofearlier work in animals and man in which bioassayof ACTH or plasma steroid determinations hadbeen performed. In particular, the classic studiesof Sayers and Sayers (39) established that ap-plication of a variety of stressful stimuli was fol-lowed by the appearance in blood of adrenal as-corbic acid-depleting activity which behaved likepituitary adrenocorticotropin and that adminis-tration of cortisol and other steroids could block

2748 S. A. Berson and R. S. Yalow

the response. The ability of vasopressin to provokeACTHsecretion (40) was also first demonstratedby bioassay of plasma ACTH.

Likewise, high plasma ACTH concentrationsin Addison's disease have been repeatedly con-firmed by bioassay methods (23, 39, 41, 42) andby radioimmunoassay as well (1, 32), after theinitial report of Taylor, Albert, and Sprague (43).Extraordinarily high levels were found by Nelson,Meakin, and Thorn (44) in patients subjected tobilateral adrenalectomy for Cushing's syndromeassociated with bilateral adrenal hyperplasia. Ele-vation of blood ACTHlevels was also reported inchildren with congenital adrenal hyperplasia (45)after the prediction of Bartter, Albright, Forbes,Leaf, Dempsey, and Carroll (46) that this wouldbe the case and this finding was confirmed by De-mura et al. (32) using radioimmunoassay.

Results on plasma ACTHconcentrations in un-treated Cushing's syndrome associated with bi-lateral adrenal hyperplasia have not been con-sistent. Davies (24) measured plasma ACTHbythe adrenal ascorbic acid depletion assay and re-ported concentrations of 2.16 mU/100 ml and1.80 mU/100 ml respectively in two cases com-pared to 0.7 mU/100 ml in two large pools (500-1000 ml) of normal whole blood extracted byoxycellulose. Using the Lipscomb-Nelson assay(26, 27), Miyake, Imura, and Muranaka (29)did not detect plasma hormone in one case andVance et al. (28) found normal values in six pa-tients; a seventh patient in the latter study hadhigh values but had previously undergone bilateraladrenalectomy. With the same assay system butusing extracts of plasma hormone, Ney et al. (31)found the 6 a.m. concentrations in nine patientswith untreated Cushing's disease to average 2.5times the normal concentration although only intwo patients were the concentrations distinctlyabove normal; these authors suggested that sincethese patients usually lacked diurnal variation inplasma 17-OHCS concentration (47), they prob-ably also lack diurnal variation in ACTH secre-tion. The present study bears out this suggestion.Although hypercortisolemic patients with Cush-ing's disease did not always have a.m. levels ofplasma ACTH distinctly elevated above thoseoccasionally encountered in other hospital pa-tients, concentrations usually remained elevatedthroughout the day.

Finally, some "order of magnitude" estimate ofACTH turnover may be derived on the basis ofthe data presented. Accepting an average plasmaconcentration of 22 ,uug/ml at 8-10 a.m. and 10)utxg/ml at 10 p.m. and short-lived spikes in con-centration at 4-6 a.m., a crude estimate for theintegrated mean normal concentration throughoutthe day would appear to be about 18 xuug/ml. Ifthe space of ACTH distribution corresponds toextracellular space (or 14 liters in a 70 kg man)and the mean turnover tj is 25 min, the turnoverin normal subjects is calculated to be 7.24 ug or1 U/day. Since the turnover tj for endogenousACTHis probably less than 25 min, as has previ-ously been observed in one case (1), the actualturnover is probably somewhat higher. The re-sults are not incompatible with the observationsof Liddle et al. (30) who found that slight eleva-tions of plasma and urinary steroids were producedby a constant intravenous infusion of ACTH atthe rate of 1.2 U/day.

ACKNOWLEDGMENTSWeare indebted to Doctors Choh Hao Li, Aaron Lerner,and Virginia Upton for supplies of human ACTH usedin these studies; to Doctors Dorothy Krieger, RichardDoe, Josiah Brown, Maria New, Leonard Kleinman,W. J. Blankenship, and numerous other physicians whoprovided samples for our studies; to Doctors NoraVarsano and Elsa Echemendia of our laboratory for aidin obtaining samples in some of the special studies re-ported; to Charles Iversen, Walter Ormond, and ManuelVillazon for technical assistance and to Violet Mallory,Frieda Steiner, and Carol Montemurro for secretarialassistance; to Gloria Spevacek, Elizabeth Walsh, GlenHarahan, and Bernard Harris of the Medical Illustra-tion Department for the preparation of the figures.

REFERENCES

1. Yalow, R. S., S. M. Glick, J. Roth, and S. A. Berson.1964. Radioimmunoassay of human plasma ACTH.J. Clin. Endocrinol. Metab. 24: 1219.

2. Berson, S. A., R. S. Yalow, A. Bauman, M. A. Roth-schild, and K. Newerly. 1956. Insulin-I' metabolismin human subjects: demonstration of insulin bindingglobulin in the circulation of insulin-treated subjects.J. Clin. Invest. 35: 170.

3. Berson, S. A., and R. S. Yalos. 1958. Isotopic tracersin the study of diabetes. In Advances in Biologicaland Medical Physics. Academic Press Inc., NewYork. 6: 349.

4. Yalow, R. S., and S. A. Berson. 1959. Assay ofplasma insulin in human subjects by immunologicalmethods. Nature. (London). 184: 1648.

Radioimmunoassay of ACTH in Plasma 2749

5. Yalow, R. S., and S. A. Berson. 1966. Labeling ofproteins-problems and practices. Trans. N. Y. Acad.Sci. 28: 1033.

6. Berson, S. A., and R. S. Yallow. 1966. Iodoinsulinused to determine specific activity of Iodine-131.Science. 152: 205.

7. Hunter, W. M., and F. C. Greenwood. 1962. Prepa-ration of Iodine-131 labeled human growth hormonehigh specific activity. Nature. (London). 194: 495.

8. Yalow, R. S., and S. A. Berson. 1966. Purification of"9-I-parathyroid hormone with microfine granules ofprecipitated silica. Nature. (London). 212: 357.

9. Rosselin, G., R. Assan, R. S. Yalow, and S. A. Berson.1966. Separation of antibody-bound and unbound pep-tide hormones labeled with iodine-131 by talcumpowder and precipitated silica. Nature. (London). 212:355.

10. Ferrebee, J. W., B. B. Johnson, J. C. Mithoefer, andJ. W. Gardella. 1951. Insulin and adrenocorticotropinlabeled with radio-iodine. Endocrinology. 48: 277.

11. Yalow, R. S., and S. A. Berson. 1960. Immunoassayof endogenous plasma insulin in man. J. Clin. Invest.39: 1157.

12. Richards, J. B., and G. Sayers. 1951. Fate and ex-cretion of adrenocorticotrophic hormone. Proc. Soc.Exptl. Biol. Med. 77: 87.

13. Dedman, M. L., T. H. Farmer, and C. J. 0. R. Mor-ris. 1955. Oxidation-reduction properties of adreno-corticotrophic hormone. Biochem. J. 59: xii.

14. Dedman, M. L., T. H. Farmer, and C. J. 0. R. Mor-ris. 1961. Studies on pituitary adrenocorticotrophin.Biochem. J. 78: 348.

15. Mirsky, I. A., G. Perisutti, and N. C. Davis. 1959.The destruction of glucagon, adrenocorticotropin andsomatropin by human blood plasma. J. Clin. Invest.38: 14.

16. Yalow, R. S., and S. A. Berson. 1968. Topics onradioimmunoassay of peptide hormones. I. Labelingof hormones. II. Separation of free labeled hormonebound to antibodies. International Symposium onProtein and Polypeptide Hormones, 19-25 May 1968,Liege, Belgium. Excerpta Medica Foundation. 3644.

17. Sweat, M. L. 1954. Sulfuric acid-induced fluorescenceof corticosteroids. Anal. Chem. 26: 773.

18. Nelson, N. 1944. A photometric adaptation of theSomogyi method for determination of glucose. J. Biol.Chem. 153: 375.

19. Berson, S. A., and R. S. Yalow. 1959. Quantitativeaspects of reaction between insulin and insulin-bind-ing antibody. J. Clin. Invest. 38: 1996.

20. Berson, S. A., R. S. Yalow, S. M. Glick, and J. Roth.1964. Immunoassay of protein and peptide hormones.Metab. Clin. Exptl. 13: 1135.

21. Berson, S. A., and R. S. Yalow. Recent advances inimmunoassay of peptide hormones in plasma. In Pro-ceedings of International Diabetes Federation 6thCongress, Stockholm, Sweden, July 30-August 4,1967. Excerpta Medication Foundation. In press.

22. Berson, S. A., and R. S. Yalow. Peptide hormonesin plasma. The Harvey Society, N. Y. Academy of

Medicine, Dec. 15, 1966. In The Harvey Lectures,The Maple Press Co., Inc., York. 107.

23. Sayers, G. 1955. Editorial. Blood ACTH. J. Clin.Endocrinol. Metab. 15: 754.

24. Davies, B. M. A. 1960. Blood corticotrophin in Cush-ing's syndrome. Nature. (London). 188: 1203.

25. Fujita, T. 1957. Determination of corticotropin(ACTH) in human blood and urine by a modifiedoxycellulose method. J. Clin. Endocrinol. Metab. 17:512.

26. Lipscomb, H., and D. H. Nelson. 1959. Measurementof corticosterone in rat adrenal venous plasma as abioassay for ACTH. Federation Proc. 18: 95.(Abstr. 373)

27. Lipscomb, H. S., and D. H. Nelson. 1960. Dynamicchanges in ascorbic acid and corticosteroids in adrenalvein blood after ACTH. Endocrinology. 66: 144.

28. Vance, V. K., W. J. Reddy, D. H. Nelson, andG. W. Thorn. 1962. Adrenocorticotropic hormone inhuman plasma. J. Clin. Invest. 41: 20.

29. Miyake, T., H. Imura, and H. Muranaka. 1962.Plasma adrenocorticotropic hormone in normal sub-jects and patients with endocrine disorders; withobservations on the effect of adrenal steroids andadrenal llp-hydroxylase inhibitors.Endocrinol.Japan.9: 223.

30. Liddle, G. W., D. Island, and C. K. Meador. 1962.Normal and abnormal regulation of corticotropin se-cretion in man. Recent Progr. Hormone Res. 18: 125.

31. Ney, R. L., N. Shimizu, W. E. Nicholson, D. P.Island, and G. W. Liddle. 1963. Correlation of plasmaACTH concentration with adrenocortical responsein normal human subjects, surgical patients, and pa-tients with Cushing's disease. J. Clin. Invest. 42:1669.

32. Demura, H., C. D. West, C. A. Nugent, K. Naka-gawa, and F. H. Tyler. 1966. A sensitive radioim-munoassay for plasma ACTH levels. J. Clin. Endo-crinol. Metab. 26: 1297.

33. Perkoff, G. T., K. Eik-Nes, C. A. Nugent, H. L.Fred, R. A. Nimer, L. Rush, L. T. Samuels, andF. H. Tyler. 1959. Studies of the diurnal variationof plasma 17-hydroxycorticosteroids in man. J. Clin.Endocrinol. Metab. 19: 432.

34. Martin, M. M., and D. E. Hellman. 1964. Temporalvariation in SU-4885 responsiveness in man: evi-dence in support of circadian variation in ACTH se-cretion. J. Clin. Endocrinol. Metab. 24: 253.

35. Liddle, G. W. 1960. Tests of pituitary-adrenal sup-pressibility in the diagnosis of Cushing's syndrome.J. Clin. Endocrinol. Metab. 20: 1539.

36. Vernikos-Danellis, J. 1965. The regulation of the syn-thesis and release of ACTH. Vitamins Hormones. 23:97.

37. Clayton, G. W., L. Librik, R. L. Gardner, and R.Guillemin. 1963. Studies on the circadian rhythm ofpituitary adrenocorticotropic release in man. J. Clin.Endocrinol. Metab. 23: 975.

38. Takebe, K, C. Setaishi, M. Hirama, M. Yamamoto,and Y. Horivchi. 1966. Effects of a bacterial pyrogen

2750 S. A. Berson and R. S. Yalow

on the pituitary-adrenal axis at various times in the24 hours. J. Clin. Endocrinol. Metab. 26: 437.

39. Sayers, G., and M. A. Sayers. 1947. Regulation ofpituitary adrenocorticotrophic activity during responseof rat to acute stress. Endocrinology. 40: 265.

40. Nagareda, C. S., and R. Gaunt. 1951. Functional re-lationship between the adrenal cortex and posteriorpituitary. Endocrinology. 48: 560.

41. Parrott, D. M. V. 1955. Assay of adrenocortico-trophic activity in plasma extracts. J. Endocrinol. 12:120.

42. Bethune, J. E., D. H. Nelson, and G. W. Thorn.1957. Plasma adrenocorticotrophic hormone in Ad-dison's disease and its modification by the administra-tion of adrenal steroids. J. Clin. Invest. 36: 1701.

43. Taylor, A. B., A. Albert, and R. G. Sprague. 1949.Adrenotrophic activity of human blood. Endocrinol-ogy. 45: 335.

44. Nelson, D. H., J. W. Meakin, and G. W. Thorn. 1960.ACTH-producing pituitary tumors following adrenal-

ectomy for Cushing's syndrome. Ann. Intern. Med.52: 560.

45. Sydnor, K. L., V. C. Kelly, R. B. Raile, R. S. Ely,and G. Sayers. 1953. Blood adrenocorticotrophin inchildren with congenital adrenal hyperplasia. Proc.Soc. Exptl. Biol. Med. 82: 695.

46. Bartter, F. C., F. Albright, A. P. Forbes, A. Leaf,E. Dempsey, and E. Carroll. 1951. The effects ofadrenocorticotropic hormone and cortisone in theadrenogenital syndrome associated with congenitaladrenal hyperplasia: an attempt to explain and cor-rect its disordered hormonal pattern. J. Clin. Invest.30: 237.

47. Doe, R. P., J. A. Vennes, and E. B. Flink. 1960.Diurnal variation of 17-hydroxycorticosteroids, so-dium, potassium, magnesium and creatinine in normalsubjects and in cases of treated adrenal insufficiencyand Cushing's syndrome. J. Clin. Endocrinol. Metab.20: 253.

Radioimmunoassay of ACTH in Plasma 2751