319

Clinica Chimica Acta, 66 (1976) 319-330 @ Elsevier Scientific Publishing Company,

Amsterdam

- Printed in The Netherlands

CCA 1437

DIRECT RADIOIMMUNOASSAY

E. ROLLERI,

M. ZANNINO,

OF PLASMA CORTISOL

S. ORLANDINI

and R. MALVANO

*

Laboratory of Clinical’Physiology C.N.R., Pisa, S.O.R.I.N., Biomedical Researches, Saluggia, and Faculty ofMedicine, Znstitute of Chemistry, E. U.L.O., Brescia (Italy) (Received

July 14, 1975)

Summary The simplification of the measurement of circulating cortisol by direct radioimmunoassay of plasma samples sets the problem of inhibiting the carrier, proteins competing with antibodies. This was accomplished by exploiting the much higher effectiveness of pH and temperature variations on steroid binding to carrier proteins than to antibody sites. A solid-phase system was set up, using antisera to cortisol-21-BSA conjugates coupled to CNBr-activated cellulose. The standardized procedure consisted of an incubation at pH 3.5 and room temperature, directly assaying 10 ~1 of plasma. A methodological and clinical validation of the measurement was carried out through a series of tests aimed at assessing the reliability of results (assay of steroid-deprived plasma, recovery test and serial dilution of samples, comparison between different antisera and with different methods including extraction, responsiveness to well-established physiological situations). The results obtained are reported and the validity of the method discussed in terms of more general applicability to steroid assay.

Introduction In its more general structure, steroid radioimmunoassay (RIA) consists of a three-step analytical sequence including extraction, chromatography and the assay itself. Starting from this situation, which entails time-consuming and delicate operations, simplified methodologies are increasingly attempted. From this viewpoint the availability of highly specific antisera or, in particular cases like for plasma cortisol, the predominance of the steroid under analysis

* Correspondence to: Dr. R. Malvano. Laboratory of Clinical Physiology C.N.R.. 56100 Pisa. Italy.

Via Savi 8.

320

in biological fluids, favour a measurement which does not include chromatographic purifications. Further improvements of the assay practicabilit/, through the elimination of the preliminary extraction step, have to deal with the problems set by the presence in samples of carrier proteins competing with the antibodies. Direct assays in unextracted plasma have been reported, in which different techniques were adopted to circumvent protein effects, such as heat denaturation [1,2], incubation in ethanolic medium [3,4], compensation for the protein content of samples [ 5,6], and protein inhibition by competition with massive amounts of steroids weakly cross-reacting with antibodies [ 7,5]. A different approach was followed at our laboratories exploiting, for a selective hindrance of the interferences, the much higher effectiveness of pH and temperature changes on steroid binding to plasma carrier proteins than to antibody sites. A cortisol-anticortisol system was used as a model, being the assay of plasma samples largely simplified by the limited requirement of sensitivity and specificity. On the other hand, the feasibility of direct assay of plasma cortisol can be of practical interest by itself, owing to its wide applications in the clinical practice as a screening test of the adrenal function. Experimental Materials and reagents

[1,2-3H] Cortisol, with a specific activity of approx. 50 mCi/mmole, was supplied by CEA-IRE-SORIN. Cortisol antisera were elicited in rabbits against a cortisol-21-conjugate to bovine serum albumin (prepared from the 21-hemisuccinate derivative [9] ) by repeated subcutaneous injections of 1 mg of immunogen in complete Freund’s adjuvant (1 : 1) as booster, performed every 2 weeks during 4 months. Coupling of two whole immune sera to CNBr-activated cellulose was carried out according to the procedure reported by Wide [lo], using an initial serum-matrix ratio of approx. 10% (w/w). The resulting immunoadsorbents, characterized by multiple titration with competing steroids (see the example of Fig. l), were kept in buffer at 4°C. 0.03 M phosphate/citrate buffer containing 0.25% lysozyme was used as diluent of the RIA reagent. Steroid-free plasma was prepared by treating pooled plasma from normal subjects with grossly-ground charcoal (approx. 0.25 g per ml plasma, overnight agitation at 4” C, centrifugation and filtration on Millipore membrane); tritiated cortisol was used to monitor the steroid removal (less than 1% residual radioactivity). Methods Solid-phase

procedure

For the standardized procedure, 0.1 ml of the immunoadsorbent suspension (corresponding to 2-3 mg) kept under agitation were added to glass

321

812 UNLASELLED

STEROID

Fig. 1. Inhibition curves relative to cortisol and some competing steroids, using the solid-phase RIA system (see Experimental). The data refer to one of the two immunoadsorbents prepared; the specificity of the other was essentially similar. 1, Cortisol; 2, cortexolone; 3. corticosterone; 4, progesterone; 6. cortisone: 6, testosterone; 7. estradiol.

tubes with 0.9 ml of buffer at pH 3.5 containing 0.15 ng of tritiated cortisol and either variable amounts of unlabelled cortisol (standard curve) or 10 ~1 of plasma (unknowns). After 1 h incubation at room temperature and a 5-min centrifugation at 1500 X g, 0.5-ml aliquots of the supernatants were transferred

B

ng

UNLAEELLED

CORTISOL

h

INCUBATION

Fig. 2. Effects of agitation during incubation on the response curve (A) and on the time to reach esuilibrium (B) in the solid-phase RIA system (see Experimental). CJ-, Without agitation: a-. under agitation.

322

for @counting. Different incubation conditions were occasionally used, as specified hereafter, in the experiments aimed at standardizing the method. In particular, a rotatory agitation at 10 rev./min was employed to keep the immunoadsorb~nt suspended, whenever binding parameters had to be derived which resulted in the modification of the response curve shown in Fig. 2. procedure [11,123 2 ml of buffered solution containing 0.3 ng of tritiated cortisol, variable amounts of unlabelled cortisol, antiserum at the proper dilution, and 1 ~1 of plasma (when specified), were equilibrated at 4°C for 2 h with 300 mg of Sephadex G-25 coarse (previoudy swelled in buffer for 24 h) under rotary agitation (10 rev./min). After a 5 min cent~fugation at 1500 X g, 0.5 ml of the supernatant were transferred and counted. The bound and free fractions were evaluated from the counts related to the total radioactivity in the external volume, by applying the relationships Gel-equilibration

L,=

Li+a+ L,;

Fe=

-Li+a K P

GO = -...

B=L,-FF,=Lt

L,=

Fe+3

(1)

-&IL,) RP

i

2t

(2)

-

1 - LJLt K P

(3)

where L, = total ligand concentration; Li+a = ligand concentration in the internal volume and ligand non-specifically adsorbed; L, = ligand concentration in the external volume (bound and free forms); F, = concentration of unbound ligand in the external volume; K, = apparent partition factor of ligand between external and internal volumes in the absence of antiserum. In this case, owing to reversible adsorption processes, K, was found to depart from the theoretical value of the true partition factor K,, as determined with blue dextran [ll], which was used to normalize concentrations in the external volume. The same procedure was followed to evaluate the extent of cortisol binding to proteins, using protein solution instead of antiserum in the incubation medium. Results Effects of uninhibited

~l~rna proteins

The effects of plasma proteins competing with antibodies for the common reactant cortisol were evaluated for both solid-phase and gel-equilibration method by adding 1% steroid-deprived plasma to the incubation mixture at pH 7.4. The resulting response curves are compared in Fig. 3 with those obtained in the absence of plasma: modifications of the response, which diverge in the two cases, are apparent. Study of the c~rtiso~-protein

in teruc tions

The extent of cortisol binding to carrier proteins was evaluated with the gel-equilibration method, incubating tritiated cortisol with plasma or al-

323

B

01 0

4

SOLID-PHASE

8

UNLABELLED CORTISOL

(ng /ml I

Fig. 3. Modification of the response curve obtained with gel-equilibration (A) and solid-phase system (B) in the presence of steroid-deprived plasma (4’C, rotatory agitation). o-, No plasma; e-, 1% deprived plasma.

bumin to equilibrium, in different conditions of pH and temperature. The results shown in Fig. 4 indicate a complete inhibition of the interaction occurring at pH values around 4 in all the cases investigated (1.1 f 1.5% binding, plasma concentrations up to 0.26%, albumin concentration up to O.l%, temperature ranging from 15 to 37 “C).

4

6

6

PH Fig. 4. pH-dependence of cortisol binding to plasma PYOM~S at 15’C (o), 24OC (gel-equilibration). one, 1% plasma; m, 2.5% plasma; A, 0.05% albumin.

(A,

A

and A) and 37’C (0)

324

pti

0 pH

7.4

0 4°C

pH

4.7

A 20°C

fJ pH

3.5

0 30°C

q

1

0

I 2

I ng

Fig.

5. Effect

of pH and

/

4

temperature

on the

RIA

I

0

2

UNLABELLED response

3.5

4

CORTISOL

curves

(rotary

agitation).

Study of the cortisol-an tibody interaction Using the insolubilized antiserum, response curves were obtained for pH values ranging from 3.5 to 7.4 and for temperatures included between 4 and 30°C. Evidence for a limited responsiveness of the cortisol-anticortisol system to variations of incubation pH and temperature can be obtained from the comparison of the curves of Fig. 5 and from the binding parameters and the approximate thermodynamic data listed in Tables I and II, as derived from the curves themselves. Assessment

of the RIA method

On the basis of the above results, incubation at pH 3.5 and room temperature was chosen for solid-phase RIA. In order to check to what extent these

TABLE

I

EFFECT

OF

pH

ON

THE

BINDING

PARAMETERS

OF

THE

CORTISOL-ANTICOCORTISOL

(20°C).

PH

Equilibrium K x 109

3.5

1.79

4.7

2.52

7.4

3.39 * Calculated

from

constant

*

(M-1)

the response

curves

applying

the

Sips’

relationship

[ 131.

SYSTEM

325 TABLE II APPROXIMATE (PH 3.5)

THERMODYNAMIC

DATA

Equilibrium constant *

K X lo+

Free energy change

AF (kcallmole)

FOR

(M-* )

Enthalpy change

AH (kc&d/mole)

Entropy change

AS (e.u.)

THE CORTISOL-ANTIBODY

4% 20°C 30% 4% 20% 30°C

INTERACTION

2.41 1.79 1.39 -11.90 -12.46 -12.68 -

3.79

+29.30

* Calculated from the response curves applying the Sips’ relationship [131.

condition8 actually prevent protein interferences, response curves in the presence and in the absence of 1% deprived plasma were compared. The coincidence of the response is indicated in Fig. 6, and further confirmed by the correlation in Fig. 7 referring to six successive experiments with the two insolubilized antisera. The extent of modifications of the assay response induced by uninhibited plasma proteins was better put in evidence by the comparison of the results obtained at pH 3.5 and 7.4 for some random samples: decidedly higher estimates are apparent in the latter condition from Fig. 8. Validation of the RIA The consistency increasing doses and usual test of recovery

method of the analytical response of the standardized method for plasma protein content was assessed by means of the and serial dilution. The results reported in Fig. 9 indicate

pH 7.4

I 0

4

*

nS

UNLASELLED

CORTISOL

Fig. 6. Effect of the presence of 1% deprived plasma on the RIA response curves at pH 7.4 and 3.5 10 ~1 deprived plasma. (room temperature). l -0, No plasma. o----o,

326

Y = 15 l1,49 x N=ll

i

2 = Y = 0.04 +0.98

X

r = 0.948 300

f

.

N=42

I

c

r = 0.995 c

200

l 0

l 0

100

!

0

0 0

0.2

B/T,

BUFFER

0

0.6

0.4

_

100

pH 3.5,

ALONE

300

200

n&l/ml

Fig. 7. Correlation of the RIA response values obtained in the presence and in the absence of 1% deprived plasma. There is no significant difference between equivalence and calculated regression @ > 0.5). Fig. 8. Correlation of the results obtained with random plasma samples by RIA at pH 3.5 and 7.4.

a quantitative recovery of the exogeneous cortisol added throughout the dose range investigated, and a parallel response up to at least three times the plasma volume fixed for routine measurements. Evidence for the absence of protein-induced modifications of the assay response for the protein concentrations expectedly involved in the RIA system, can also be derived from the

r

RECOVERY TEST

0

ng CORTlSOl. Fig. 9. Results of recovery

10

AWED

and dilution te9ts using the standardized RIA method.

20

30

)I’

PLASMA

327 TABLE III EFFECT OF INHIBITED PLASMA OF SOLID-PHASE RIA (pH 3.6)

PROTEINS

ON INITIAL BINDING AND NON-SPECIFIC

% Plasma

@/T)o (%)

NIT (%) *

0 1.25 2.6 6 10

42.0 41.2 41.3 42.6 44.6

1.6 0.1 0.6 0.2 2.5

+ 1.6 f 0.6 ?: 1.0 f 1.1 f 0.9

+ + + f f

COUNTS

1.0 0.9 0.6 0.2 0.7

* Data referring to celIuIose_coupled y%Iobulins.

. Y.-9.4+1.02x

Yr 96+69lx

.

N=48

N.

.

r.a96ll

. .

. .

k g,

.

.*

36

r= a993

i!

DIRECTRIAOF PLASMA Fig. 10. CorreIations of the results obtained with the standardized RIA method and with extract assay by competitive protein binding and fluorimetry. No significant difference results between ectuivaIence and calculated regressions @ > 0.75).

TABLE IV COMPARISON OF RESULTS UBILIZED ANTISERA * Sample

OBTAINED

IN STANDARDIZED

CONDITIONS

Cortisol (rig/ml) Immunoadsorbent 2 6 24 73 98 110 163 370

A

Immunoadsorbent

B

3.5 8 21 80 99 102 141 395

* Affinity constant of 1.79 X lo9 and 0.61 X lo9 M’,respectively.

WITH TWO INSOL-

328

TABLE V RESULTS

OBTAINED

WITH THE STANDARDlZED No. of cases

Condition

_._____ 8 12 4 8 12

METHOD IN NORMAL

HUMANS

Cortisol (nglml)

.____~~

Mi?alI SD. ______ __._____-l_l__-~~---.

--.--

a.m. noon p.m. p.m. midnight

30 9 9 9 9

126 91 65 50 34

41 35 11 24 22

ACTH stimulus (80 Ui.m., 7 a.m.) 8 a.m.

5

358

78

Dexamethasone inhibition (1 mg per OS, 11 p.m.) 8 a.m.

5 -._------

5

3

Rallga .__~ 55-200 42-135 28-110 20-103 20-75 276-

.----

468

2-7

effects on the values of initial binding B/T (i.e. fraction of tracer bound in the absence of unlabelled cortisol) and of nonspecific counts N/T (i.e. fraction of tracer misclassified as bound in the absence of ~tise~m) reported in Table III. No indication for systematic differences of the analytical information depending on the individual antiserum was given by parallel measurements with the two different immunoadsorbents prepared (see Table IV). As a further criterion to assess the assay reliability, the RIA method was compared with two techniques, both including extraction, i.e. competitive protein binding [14,15] and fluorimetry [16] ; the correlation of results displayed in Fig. 10 does not show any consistent inter-method difference. The assay of samples related to well-defined physiological situations was assumed as an additional check of the validity of the method. The results collected in Table V are in agreement with the data of the literature [l&3, 14,15,17,18]. Discussion Competition of plasma proteins (transcortin and albumin, in this case) with antibodies disfavours the formation of immunocomplexes, while there results a higher overall concentration of bound ligand. Therefore, depending on the method, the classification of the bound and free fractions may eventually differ. Methods such as gel equilibration, which do not distinguish between ligand complexes with antibodies and ligand complexes with carrier proteins, result in larger proportions of bound radioactivity when plasma is added to the incubation medium. The opposite situation is encountered with immunoadsorbents, as only the immunocomplex is classified as bound form. Underestimation and overestimation errors are thus entailed by the two separation procedures, respectively, when competition of plasma proteins is not hindered and the response obtained for samples is referred to the calibration solutions. The information furnished by a direct assay without any precaution to circumvent protein effects could retain some validity as an

329

index related to particular pathophysiological situations [19] . However these data are in no way generalizable, being any extrapolation and comparison prevented by the dependence on method, antiserum characteristics and possibly on the individual protein content of samples. Even a relatively low plasma concentration as that used introduces a large bias in the analytical system. The selectivity of the technique proposed to inhibit plasma interferences relies upon the different features of the cortisol bonds with antibodies and with carrier proteins. This study has given some experimental evidence that hydrophobic components predominate in the cortisol-antibody interaction, exactly as they do in steroid binding to plasma proteins [ 201. A much stricter interdependence of conformational structure and binding function can be hypothesized for carrier proteins and particularly for transcortin, than for immunoglobulins, for which unfolding at low pH should not cause severe modifications of interacting sites. The analogous behaviours of two cortisol antisera with different characteristics could be an argument to show that this situation is not peculiar to a given individual antiserum. Furthermore, indication that an entropy-driven interaction with antibodies is a general feature of steroidal molecules emerges from experimental data previously reported in the literature [21,22] and from our own recent results concerning estradiol, testosterone and progesterone [12,23,24,25] , thus supporting the view that this approach to direct RIA could be extended to other cases. Some advantages appear to be associated with the present method, with respect to those reported for cortisol [1,2,3] , corticosterone [ 41, estradiol [ 71, testosterone [8] and digitalis glycosides [ 5,6] . The losses of sensitivity involved are apparently limited, as compared to the ones possibly arising from a weakening of interaction, when incubation is performed in ethanolic solution [3,4], or from an increase of reactive sites, when a mere compensation for protein content is made through the addition of plasma not containing the substance under analysis to the calibration solutions [ 5,6] . The singlestep procedure implied adds an element of methodological practicability with respect to the techniques using a preliminary step of heat-denaturation of proteins [ 1,2]. As far as selectivity of inhibition of protein interferences relies on the nature of ligand interactions rather than on the specificity of antibody sites to close molecular structures, the method proposed could be considered of more general validity than that based on competition for carrier proteins of related substances weakly interacted by antibodies [ 7,8]. The specificity characteristics, which define the actual performances of the latter technique in terms of selectivity and sensitivity, appear to be much more subject to the individual variability of antisera than the mode of steroid-antibody interaction itself. For both procedures based on steroid competition [7,8] a correction for blank was required, probably as a consequence of the separation method used, i.e. adsorption of free radioactivity on charcoal-dextran: methodological artefacts, stemming from immunocomplex dissociation to variable degrees (depending on the protein content), were in fact put in evidence for different RIA systems employing irreversible adsorbents [26,27,28]. Conversely, no significant plasma effect was observed with the method studied. Though this is partly attributable to the low plasma concentrations involved in routine

330

measurements, the ineffectiveness of the assay response shown in the case of much higher protein amounts, suggests that a limited susceptibility to nonspecific interferences from samples is an intrinsic feature of the solid-phase system. The use of immunoadsorbents also implies a simplification of the assay sequence, as the binding reagent coincides with the separation agent. Once the losses of immunological properties have been minimized, through the optimization of immunoadsorbent preparation [ 12,251, solid-phase RIA is therefore to be regarded as a method of choice in improving the practicality of measurement.

References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

Foster, L.B. and Dunn, R.T. (1974) Clin. Chem. 20. 365-368 Donohue, J. and Sgoutas. D. (1975) Clin. Chem. 21.770-773 Farmer, R.W. and Pierce, C.E. (1974) Clin. Chem. 20,411-414 GomezSanchez, C., Murry, B.A., Kern, D.C. and Kaplan, N.M. (1975) Endoerinologv 96, 796-798 Smith, T.W., Butler, V.P. and Haber. E. (1969) New Engl. J. Med. 281.1213-1217 Smith, T.W. and Haber. E. (1970) J. Clin. Invest. 49.2377-2386 Jurjens, H., Pratt, J.J. and Woldring, M.G. (1975) J. Clin. Endocrinol. Metab. 40,19-25 Pratt, J.J., Wiegman, T.. Lappiihn, R.E. and Woldring, M.G. (1976) Chn. Chim. Acta 59, 337-346 Erhmger. F., Boxek, F., Beiser, SM. and Lieberman, S. (1969) J. Biol. Chem. 234,1090-1094 Wide, L. (1969) Acta Endocrinol.. suppl. 142.207-218 Pearlman, W.H. and Crepy. 0. (1967) J. Biol. Chem. 242,182-189 Comoglio, S., Massaglia, A., Rolleri, E. and Rosa, U. (1976) in preparation Nisonoff. A. and Pressman, D. (1958) J. Immunol. 80.417-428 Kolanowski, J. and Pirarro, M.A. (1969) Ann. Endocrinol. 30. Suppl. 1 bis, 177-182 Leclereq, R., Copinschi, C. and Franckson, J.R.M. (1969) Rev. Franq. Etud. Chn. Biol. 16,816~819 Mattingly, D. (1962) J. Clin. Pathol. 15,374-382 Ruder, H.J.. Guy, R.L. and Lipsett, M.B. (1972) J. Clin. Endocrinol. 35, 219-224 Vecsei. P., Penke, B.. Katzi, R. and Back, L. (1972) Experientia 8.1104-1105 Verdonck, L. and Vermeulen, A, (1974) J. Steroid. Biochem. 5,471-479 Westphal. U. (1971) Steroid-protein interaction. Springer Verlag, Berlin. Heidelberg, New York Abraham, G.E. and Odell, W.D. (1970) in Immunological methods in steroid determination (Peron. F.G. and Caldwell, B.V., eds.), pp. 87-10’7, Appleton-Century Croft,New York Kley, R. and Hansen, W. (1974) Arrtl. Lab. 20.202-208 Massa&ia. A., Rolleri, E.. Barbie& U. and Rosa, U. (1974) J. Clin. Endocrinol. Metab. 38,820--826 Malvano. R., Rolleri. E. and Rosa, U. (1974) in Radioimmunoassay and related procedures in Medicine, Vol. II, pp. 97-122. IAEA, Vienna Malvano, R. and Rolleri, E. (1975) in Radioimmunoassay of steroid hormones (Gupta, D., ed.), Verlag Chemie, Veinheim Rolleri, E., Malvano, R., Gandolfi, C. and Rosa, U. (1974) Horm. Metab. Res. 6,57-61 Malvano, R., Zucchelli. G.C., Gasser, D. and Bartolini, V. (1974) Chn. Chim. Acta 50, 161-171 Malvano, R.. &uesada. T.. Rolleri, E.. Gandolfi. C. and Zucchelli, G.C. (1974) Clin. Chim. Acta 51. 127-139