Scandinavian Journal of Clinical and Laboratory Investigation
ISSN: 0036-5513 (Print) 1502-7686 (Online) Journal homepage: http://www.tandfonline.com/loi/iclb20
Removal of endogenous ligands from a highaffinity antiserum for radioimmunoassay V. Kruse To cite this article: V. Kruse (1979) Removal of endogenous ligands from a high-affinity antiserum for radioimmunoassay, Scandinavian Journal of Clinical and Laboratory Investigation, 39:6, 533-541, DOI: 10.1080/00365517909108831 To link to this article: http://dx.doi.org/10.1080/00365517909108831
Published online: 14 Feb 2011.
Submit your article to this journal
Article views: 3
View related articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=iclb20 Download by: [137.189.171.235]
Date: 16 June 2016, At: 11:38
Scand. J. din. Lab. Invest. 39, 533-541, 1979.
Removal of endogenous ligands from a highaffinity antiserum for radioimmunoassay V. KRUSE
National Institute of Animal Science, DK-1958 Copenhagen, Denmark
Downloaded by [] at 11:38 16 June 2016
Kruse, V. Removal of endogenous ligands from a high-affinity antiserum for radioimmunoassay. Scand. J . clin. Lab. Invest. 39, 533--541, 1979. A method for removal of endogenous ligands from high-affinity antisera (stripping) is described. A thyroxine antiserum of very high affinity was used to develop the method. The antiserum was incubated at 50°C in a glutamate buffer at pH 4.4 together with some ethanol and methyl cellulose and a large amount of activated charcoal. After incubation for up to 2 days, the stripped antibodies were separated from the ligand adsorbed to the charcoal by centrifugation. The estimated titre increased several fold by the stripping, although the stripping method caused a loss of about 13% of the total number of binding sites over a period of 2 days. When stripped antiserum was used instead of unstripped antiserum in an assay system, the sensitivity was up to 3 times better, and the concentration, which could be measured with the best relative precision, was 3 times lower. The stripped antiserum showed a poorer specificity than the unstripped antiserum when a short incubation period was used. However, when a long incubation period was used, the specificity was nearly the same. Keywords: antiserum; charcoal ; endogenous ligand ; radioimmunoassay ; removal; stripping of antibodies; stripping of antiserum; thyroxine
V. Kruse, Ph.D., Novo Research Institute, Novo Alle, DK-2880 Bagsvaerd, Denmark
It has been known for some years that antibodies produced against a ligand will be partly saturated if the ligand is a normal component of the blood of the immunized animal [2, 7, 8, 13, 151. Ligand concentrations in immunized animals can be up to 1000 times higher than in non-immunized animals [2, 151. It has recently been reported that ligands derived from the ligand-protein conjugates used for immunization may be a significant source of saturating ligand [16]. 0036-55 13/79/ 1000-0533$02.00 0 1979 Medisinsk Fysiologisk Forenings Forlag
Several attempts have been made at removing endogenous Iigands from antiserum [3, 6, 14, 15, 19, 201. However, only a few investigators have reported some degree of success with the treatment [3, 6, 151. Our own attempts at improving a triiodothyronine (T3) antiserum by the method of Fyhrquist & Wallenius [6] were unsuccessful (Avivi & Kruse, unpublished). Other investigators have also found the method to be unsatisfactory [20]. It is necessary for an antiserum to contain at least some antibodies of high affinity when it is to be used for radioimmunoassay purposes. If
533
534
V. Krirse
not, the standard curves will be of no use. The high affinity of the antibodies makes it difficult, however, to remove endogenous ligands from radioimmunoassay antisera. This paper reports a simple and efficient method for the removal of endogenous ligands from highattinity antibodies [9].
MATERIAL AND METHODS
Downloaded by [] at 11:38 16 June 2016
Chemicals. Activated charcoal, Med. Norit, Ph. Nordica, was obtained from Norit, Amersfoort, Holland; methyl cellulose, type low substitution from BDH Chemicals, Poole, England; and L-glutamic acid from Riedel-de H a h , SeelzeHannover, W. Germany. Reagents. Glutamate buffer 0.04 molil, pH 3.4; glutamate-EDTA buffer: glutamate buffer with 3 mmol/l EDTA and 0.12 mmol/l merthiolate; acetate buffer 0.14 mol/l with 0.15 mol/l NaCl and 20 mmol/l EDTA, pH 5.0; phosphate buffer 0.2 mol/l with 10 mmol/l EDTA, pH 7.0; tris buffer 0.12 mol/l with 20 mmol/l NaN,, pH 8; and trichloroacetic acid (TCA) 0.6 mol/l. Antiserum. Thyroxine (T4) antiserum from three bleedings of rabbit number 8 (Ref. 8) was used. The bleeding from 22 July 1975 contained 350 nmol/l endogenous T4. It was taken after 10 months of T3 suppression [8]. The bleedings from 12 and 23 March 1976 were obtained after a period without Ts suppression and contained 1600 and 1500 nmol/l endogenous T4, respectively. The last bleeding contained T4 antibodies with dissociation rate constants of 0.550, 0.045 and 0.006 h-' at 23°C [lo]. These values correspond to T4-antibody complex half-lives of 1 h, 15 h and 115 h (4.8 days). The raw antisera were stored at about -20°C until use. Removal of endogenous ligands from arttisera (stripping) Principle. The stripping method is based on three principles: (I) a high dissociation rate is brought about by physico-chemical means, (2) the dissociated ligand is adsorbed to particles, and (3) the stripped antiserum is separated from the particles by centrifugation.
Protocol. Over a number of experiments, the following procedure has proved the most efficient one: mix 300 mg methyl cellulose with 3 ml ethanol, add 26 ml glutamate buffer, add 1 ml of the antiserum to be stripped, and adjust to about pH 4.4. Heat to 50"C, and add 500 mg of activated charcoal. Keep the teniperature at 50°C and mix gently. Stop the stripping after about 6 h by removing the charcoal by high-speed centrifugation ( 12,000 g ) . Treated antiserum was stored at 4-10°C for a few days without further dilution. In case of longer storage, the antiserum was diluted with 6 volumes of tris buffer. Evalmtion of stripping eflicienry Stripping curves. In order to ensure complete mixture of labelled T4 with the whole pool of endogenous Tq,the antiserum was incubated with lo6 cpm/ml of [1251]T4for 1 day at 50°C plus 4 days at room temperature before stripping. Sodium azide 15 mmol/l was used as a preservative (Caution: poisonous gases are liberated at low pH!). The equilibration procedure did not cause any detectable loss of binding sites. The stripping process was followed by analysis of samples of the stripping suspension taken after I min to 2 days of stripping. The charcoal was removed from 2 ml of stripping suspension by high-speed centrifugation and precipitable radioactivity in the supernatant was determined. The precipitation was carried out on 150 ,uI supernatant plus 400 pl normal bovine serum with 4 ml TCA. The precipitate contained predominantly labelled T4-antibody complexes and the supernatant labelled iodide. A control tube with normal rabbit serum, instead of antiserum, was stripped according to the protocol. Less than 1:4 precipitable radioactivity was found after 10 min of stripping. Assay for thyroxine in thyroxine antisera. The antisera were diluted 25 to 100 times, and T,-free serum was added to make the unknown samples similar to the standards in respect of serum proteins. Samples and standards were boiled for 15 min [4] in glutamate-EDTA buffer. The T4 concentration was measured by a sequential saturation assay [ 121. Non-specific binding was measured for all dilutions of
Stripping of antisera
boiled antiserum. No antibody binding activity was detected after boiling.
Downloaded by [] at 11:38 16 June 2016
Eflcct of c tripping on atitiseriim characteristics Titre. Curves showing the titre as a function of stripping time were constructed on the basis of the samples taken of the stripping suspension. The titre was taken as the factor with which the antiserum must be diluted to yield a 50% binding of the tracer under the given titration conditions. The incubates had a volume of 1 ml and were made in barbita;l-buffer with 40 pl T,-free serum, 1.25 pmol ammonium 8-anilino1 -naphthalene-sulphonate (displaces T, from serum proteins), 110 fmol labelled T4 and diluted antiserum. Incubation was carried out at 23°C for 4 h, and separation of bound and free T4 was performed with charcoal at room temperature [8]. Sensitivity and precision. Standard curves were made on the basis of equilibrium-type assays.
0
I
2
535
The incubates had a volume of 1.25 ml and were made in barbital buffer, with reagents added in the following order: 50 rcl serum standard, 1.6 pmol ammonium 8-anilino-lnaphthalene-sulphonate, 100 fmol labelled T, and diluted antiserum. The incubation time was 6 days at 23°C. Bound and free T, were separated as for the titration. The assay sensitivity was estimated as 2 x SD at zero concentration. The sensitivity was compared for unstripped and stripped antiserum on the basis of curves showing the same initial binding (Bo). All assays were run simultaneously, with the same batch of tracer. A precision profile was constructed for each of the standard curves. A precision profile shows the coefficient of variation as a function of the measured T, concentration. The concentrations measurable with the lowest coefficient of variation (the best relative precision) were read off from these curves. An example of a precision profile has been published by Rodbard (Fig. 6G in Ref. 18). The calculations
3
4
5
6
Time ( h )
FIG.1 . Stripping curves. The effect of 50 (. . . . .), 150 (- - -),
500 (----) and 1500 (.--.-.) nig charcoal per ml antiserum on the removal of labelled thyroxine from a thyroxine antiserum is shown as a function of stripping time. Acetate buffer (pH 5.0) without methyl cellulose was used. Other reagents and conditions were as described in the protocol. Antiserum from rabbit number 8, bled 12 March 1976.
536
V. Kruse Effect of buffer and p H . Preliminary studies showed that it was essential to perform the stripping under acid conditions. Stripping in acetate buffer at pH 5 was very effective (Fig. 3). However, about 15% of the T 4 antibodies were lost by precipitation when this buffer was used at 50°C. Precipitation could be avoided to a large extent if use was made of glutamate buffer. Figure 3 shows that a high stripping rate combined with a small loss of titre units was achieved when glutamate buffer was used at pH 4.4.
of sensitivity and precision were made according to the principles of Ekins [5].
Cross-reaction studies. Calculations were made according to Abraham [ I ] on a molar basis.
Downloaded by [] at 11:38 16 June 2016
RESULTS E j e c t of charcoal concentration on stripping rate and titre. Figure 1 shows that an increase in charcoal concentration causes an increase in the disappearance rate of labelled Ta-antibody complexes. However, the titre curves in Fig. 2 clearly show that there was a great loss of antibodies when 1500 mg charcoal was used per ml antiserum. When only 50 or 150 mgiml was used, the titre increased by a factor of about 5 . The loss of binding sites found at high charcoal concentrations was most likely caused by adsorption of antibodies to the charcoal. When no charcoal was used (Fig. 2) there was a 14%, loss of binding sites per 2 days of stripping.
Efect of ethanol. It was observed that the
stripping rate could be enhanced considerably by including some ethanol in the stripping medium (Fig. 4). Use of 10% ethanol caused a loss of about 13% of the original titre when the antiserum was incubated for 2 days with all the reagents except charcoal (Fig. 5 ) . The loss was unacceptable when higher concentrations of ethanol were used (Fig. 5 ) . The destruction of binding sites appeared to take place
2oooc
............ .............?,- - -.-.-
.-.-
- ................... - - - -- - - ...... .................
~ r r ;
I
1500C
cE
I0000
5000 I
--.-_
-.--
\.
-
-_
L .\
L 0
I
3
2
4
5
6
Time f h )
FIG. 2. Titre of antiserum stripped with 50 (. ' .. .), IS0 (--), 500 (-) and 1500 (.--.--.) mg charcoal per ml antiserum as shown in Fig. 1 . Antiserum incubated under stripping conditions without charcoal is included for comparison (. --- ). +
Downloaded by [] at 11:38 16 June 2016
Stripping of antisera
I 0
1
I
I
l
2
3
I 4
I
I
5
6
537
Time ( h )
FIG.3. Stripping curves. Effect of buffer and pH on the removal of labelled thyroxine from a thyroxine antiserum. Phosphate buffer, pH 7.2 (.. ".), glutamate buffer adjusted to p H 5.0 (- - -), pH 3.8 (-) or pH 4.4 (-) and acetate buffer, pH 5.0 (.--.--.) were used. The titre of samples taken after 3.5 h of stripping was not determined (. . . .), 18000 (-- -), 15000 (--), 18000 (--) and 13000 (.-.-.). No methyl cellulose was used. The antiserum was from rabbit number 8, bled 12 March 1976.
especially during the first few hours of incubation (Fig. 5 ) . Effect of methyl cellulose. The problem of adsorption of antibodies to charcoal at high charcoal concentrations (Figs 1 and 2) could be overcome by adding 300 mg of methyl cellulose to the stripping suspension. A suspension was made as described in the stripping protocol and compared with one made without methyl cellulose. The methyl cellulose had practically no effect on stripping and denaturation curves. The titre, however, was 17% higher when methyl cellulose was used. Eflect of serum proteins. Normal rabbit serum was added to stripping suspensions to see if this would increase the titre by preventing the adsorption of antibodies to charcoal. The suspensions were made according to the protocol, except that 0, 1 or 4 ml serum was substituted for glutamate buffer. No effect was seen on stripping and titre curves.
Endogenous thyroxine in thyroxine antiserum. The estimated concentration of Tgin unstripped antiserum is given with the description of the antisera (see Material and methods). The concentration of T4 in the stripped antiserum was too low to be measured by the described method. Sensitivity and precision. Figure 6 shows standard curves for unstripped and stripped antiserum. The assay sensitivity varied from 0.09 to 0.57 nmol/l, and was up to 3 times better for stripped antiserum than for unstripped antiserum. The largest difference in sensitivity was seen when curves with high initial binding (Bo) were compared. The lowest coefficients of variation were achieved when concentrations of about 1.5 and 5 nmol/l were measured with stripped and unstripped antiserum, respectively. Cross-reactivity. In general, stripped antiserum gave higher cross-reactions than unstripped
538
V. Kruse
Downloaded by [] at 11:38 16 June 2016
I00
0
I
I
I
I
I
I
2
3
4
5h'
&
I 2d
Time
FIG. 4. Stripping curves. Effect of 0 (.....), 5 (- - -), 10 (-), 15 (.--.--.) and 20 (-) % ethanol on the removal of labelled thyroxine from a thyroxine antiserum. The antiserum was from rabbit number 8, bled 22 July 1975. 100000
EOOCIO
---___ \
GOO00 '.
-. -. -. - -
'.
-,
- . -.'. '.
?
\
40000
.-
L
k
!
\
I
L
I
I
I
I
0
I
2
3
4
h:
'%d
Time
(...**), 5 (- - -), without charcoal.
FIG.5. Titre of antiserum incubated under stripping conditions with 0 10 (-),
I5 (--.--.) or 20 (--
--)
% ethanol as in Fig. 4-but
Stripping of antisera
Downloaded by [] at 11:38 16 June 2016
80
0
539
-
L/
I
I
I
I
0.5
I
2
4
I 8
I
16
Thyroxine (nrnol /I)
FIG.6. Standard curves made with unstripped o r stripped thyroxine antiserum in different dilutions. The unstripped antiserum was used in final dilutions between I :2,500 a n d 1 :12,600 (- - -). Antiserum from the same bleeding was stripped for I day and used in final dilutions between 1 :13,000 and I :40,000 (--). Antiserum from rabbit numher 8, bled 23 March 1976.
antiserumwith the T4 analogues tetraiodothyropropionic acid, tetraiodothyroacetic acid, 3,3’, 5’-triiodo-L-thyronine (reverse T3), the X-ray contrast medium a-ethyl-3-hydroxy-2,4,6triiodohydrocinnamic acid (Teridax) and 3,5,3’triiodo-L-thyronine (T3). However, the crossreactions estimated for reverse T3, Teridax and T, were highly dependent on incubation time, especially when stripped antiserum was used, The cross-reactions were in the range of 0 to 19% after a 2 h incubation period. When incubation was prolonged to 24 h, the crossreactions were only 0 to 4% and almost the same for unstripped and stripped antiserum. The effect of incubation time was not seen for the tetraiodo analogues. These analogues showed cross-reactions in the range of 22-81 Further details will be published elsewhere [I I].
z.
DISCUSSION In principle, the only criterion of successful stripping is that the T4 has been removed from
the antiserum (without harm to the antibodies). The effectiveness of stripping can be measured direct by determining the concentration of iodine, T4 or labelled T4 in the antiserum before and after stripping. This has only been done in rare cases [9, 191. If the concentration of labelled T4 is to be taken as a measure of the efficiency of stripping, it is of course essential that the labelled T4 is in equilibrium with the whole pool of endogenous T4. There has been a tendency to evaluate the effectiveness of a stripping method on the basis of estimates of titre [3, 6, 15, 201, apparent equilibrium constant [6, 151, assay sensitivity [6, 15, 201 or assay specificity [6]. However, such indirect estimates of stripping efficiency should be regarded as unreliable indicators of stripping, as already pointed out by Skrabanek et al. [20]. The estimates depend to a large extent on the conditions under which they are measured, as shown for specificity in Results and as discussed for titre and sensitivity below. Some of the T4 antibodies in the antiserum used for development of the stripping method
Downloaded by [] at 11:38 16 June 2016
540
V. Kruse
have a dissociation rate constant of 0.006 h - ’ at 23°C (see Material and methods). This value corresponds to a half-life of 4.8 days for the T,-antibody complexes. It is evident that with such stable complexes in the antiserum, it was necessary to take quite drastic measures in order to obtain a fast and complete stripping. However, these measures all tend to cause a loss of antibody activity. As the aim is a rapid and complete stripping, and as any loss of binding sites should be avoided, the stripping protocol must necessarily represent a compromise solution to the problem of removing ligands from high-affinity antisera. The method described does give a rather fast and practically complete stripping, albeit at the expense of about 13 ”/, of the total binding sites. A thorough study of the loss has been done for one antiserum only (Fig. 5). The question naturally arises as to what types of antisera can be stripped by the method. It was found possible to strip antisera against glucagon, ACTH, angiotensin 1, angiotensin IT, thyrotropin releasing hormone, Ts, progesterone, testosterone, estradiol, cortisol and corticosterone. The initial part of the stripping curves was steeper for all these antisera than for the T4 antiserum. This indicates that the antibodies dissociated relatively fast. About 5 % of the labelled glucagon could not be removed from the antiserum. The removal of all the other tracers was practically complete. It seems likely that the more stable the complexes of endogenous ligand and antibody, the greater the advantages gained by stripping. If it takes 4 5 days for half of the ligand-antibody complexes to dissociate, and if such stable complexes constitute a major part of the antibodies, it is clear that a titration involving incubation for 2, or even 24 h, does not give a measure of all the binding sites. The endogenous ligand would simply block a large fraction of the binding sites for days, and the tracer would not be distributed over all binding sites during the incubation period. The T4 antiserum was found to have more available binding sites after stripping in spite of the 13% loss of total binding sites caused by the stripping. Some bleedings from the rabbit showed as much as an 8-fold increase in titre after stripping. Only a few of the other hormone antisera, which were stripped in the same way as the T4
antiserum, showed a gain in titre (unpublished). The titre of antisera stripped by other methods is reported to have either increased considerably [3], remained the same [19], or have decreased [6, 15, 201. It has been shown [lo] that the good antibodies in an antiserum, i.e. the slow-dissociating antibodies, are utilized much better as binders in the stripped form than in the unstripped form under radioimmunoassay-like conditions. This improvement in utilization of good binders is considerable for high-affinity antisera, but slight for antisera of lower affinity [lo]. The assay sensitivity was up to three times better with the stripped T, antiserum. Other investigators have also found an increase in sensitivity after stripping [6, 151. The reason for the improved sensitivity is probably that endogenous ligand does not dilute the labelled ligand in the assay tube when stripped antiserum is used, in contrast to the use of unstripped antiserum. This may be illustrated by the following example. The standard curves for unstripped antiserum shown in Fig. 6 were set up with 100 fmol labelled T, and antiserum in final dilutions between 1 :2,500 and 1 :12,600. The amount of endogenous T4 added to each assay tube together with the unstripped antiserum was between 150 and 750 fmol. The specific radioactivity of the labelled T, was thus reduced to between 40 and 12% of that in the tubes with stripped antiserum. A similar reduction to 23 % of the original specific radioactivity of the tracer has been reported in the description of a T 3 assay [2]. The specificity of the T, antiserum did not appear to be improved by the stripping. Theoretically, the specificity is highest when the incubation i s carried out to equilibrium [17]. Unfortunately, the specificity was not estimated at equilibrium in the present case. However, the finding that the cross-reactions with reverse TB,Teridax and T 3 lessened the longer the incubation period, and that the decrease in cross-reactivity was most pronounced for the stripped antiserum, indicates that the specificity at equilibrium is at least as good for the stripped as for the unstripped antiserum. Incubation to equilibrium would have little effect on the cross-reactions with the tetraiodo analogues because no effect of incubation time on cross-reactivity was detected for these substances.
Stripping of antisera No problems have been encountered with the stability of stripped antiserum when neutralized with 1 volume of tris buffer and stored at 410°C. Freezing in 10 ml tubes up to six times did not cause any detectable loss of titre.
ACKNOWLEDGMENTS
Downloaded by [] at 11:38 16 June 2016
I wish to thank Karen Saxtorph Jensen for excellent technical assistance. Antisera were generously provided by G.D. Niswender, L.G. Heding, A.M. Kappelgaard, A. Veje, H. Worsaae and L. Hummer. This work was supported by The Scandinavian Contact Agency for Agricultural Research as Project 34.
REFERENCES 1 Abraham, G.E. Solid-phase radioimmunoassay of
estradiol-17p. J. din. Endocr. 29, 866, 1969. 2 Alexander, N.M. & Jennings, J.F. Radioimmunoassay of serum triiodothyronine on small, reusable Sephadex columns. Clin. Chem. 20, 1353, 1974. 3 Auletta, F.J., Caldwell, B.V. & Hamilton, G.L. Androgens: Testosterone and dihydrotestosterone. p. 359 in Jaffe, B.M. & Behrman, H.R. (eds) Methods of’ Hormone Radioimmunoassay. Academic Press, New York and London, 1974. 4 Dunn, R.T. & Foster, L.B. Radioimmunoassay of thyroxine in unextracted serum, by a single-antibody technique. Clin. Chem. 19, 1063, 1973. 5 Ekins, R.P. Basic principles and theory. Br. med. Bull. 30, 3, 1974. 6 Fyhrquist, F. & Wallenius, M. Improvement of antisera against polypeptide hormones. Nature (Lond.)254,82, 1975. 7 Hillier, S.G., Cole, E.N., Groom, G.V., Boyns, A.R. & Cameron, E.H.D. Effect of active immunisation against testosterone-3-BSA on circulating levels of testosterone, LH, prolactin and testosterone antibody titre in the male rat. Steroids 22, 227, 1973.
541
8 Kruse, V. Production and evaluation of highquality thyroxine antisera for use in radioimmunoassay. Scand. J. din. Lab. Invest. 36, 95, 1976. 9 Kruse, V. A method for removal of endogenous ligands from high-affinity antibodies to be used in radioimmunoassay. p. 57 in Abstracts 2nd European Congress on Clinical Chemistry. Prague, 1976. 10 Kruse, V. Dissociation rate constants and fractional binding of tracer estimated for three antibody populations in unstripped and stripped antiserum. Scand. J. din. Lab. Invest. 39, 215, 1979. 1 I Kruse, V. Selective removal of fast-dissociating antibodies from a high-affinity antiserum. Molec. Immun. (In press). 12 Kruse, V. & Lind, 0. A rapid and precise sequential saturation radioimmunoassay for thyroxine. Scand. J. d i n . Lab. Invest. 31, 149, 1977. 13 Nieschlag, E., Usadel, K.H., Schwedes, U., Kley, H.K., Schoffling, K. & Kruskemper, H.L. Alterations in testicular morphology and function in rabbits following active immunization with testosterone. Endocrinology 92, 1142, 1973. 14 Niswender, G.D. p. 128 in Cameron, E.H.D., Hillier, S.G. & Griffiths, K. (eds) Steroid Immunoassay. Alpha Omega Publishing Ltd., Cardiff, U.K., 1975. 1 5 Oliver, L.K. & Cano, C. Removal of an endogenous antigen from an antibody to increase its effective affinity constant, as illustrated by triiodothyronine assay. Clin. Chem, 23,2039, 1977. 16 O’Neill, S.P. & Robb, T.E. Hapten concentrations in the circulation of animals actively immunized with hapten-protein conjugates. J . Immun. 116, 363, 1976. 17 Pratt, J.J. & Woldring, M.G. Radioimmunoassay specificity and the ‘first-come, first-served effect’. Clin. Chim. Acta 68,87, 1976. 18 Rodbard, D. Statistical quality control and routine data processing for radioimmunoassays and imrnunoradiometric assays. Clin. Chem. 20, 1255, 1974. 19 Seth, J., Rutherford, F.J.’ & McKenzie, I. Solidphase radioimmunossay of thyroxine in untreated serum. Cfin. Chem. 21, 1406, 1975. 20 Skrabanek, P., Kirrane, J. & Powell, D. Apparent improvement of antisera for radioimmunoassay by treatment with sodium iodide. J. Immun. Merh. 16, 331, 1977.
Received 26 May 1978 Accepted 6 March 1979
ISSN: 0036-5513 (Print) 1502-7686 (Online) Journal homepage: http://www.tandfonline.com/loi/iclb20
Removal of endogenous ligands from a highaffinity antiserum for radioimmunoassay V. Kruse To cite this article: V. Kruse (1979) Removal of endogenous ligands from a high-affinity antiserum for radioimmunoassay, Scandinavian Journal of Clinical and Laboratory Investigation, 39:6, 533-541, DOI: 10.1080/00365517909108831 To link to this article: http://dx.doi.org/10.1080/00365517909108831
Published online: 14 Feb 2011.
Submit your article to this journal
Article views: 3
View related articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=iclb20 Download by: [137.189.171.235]
Date: 16 June 2016, At: 11:38
Scand. J. din. Lab. Invest. 39, 533-541, 1979.
Removal of endogenous ligands from a highaffinity antiserum for radioimmunoassay V. KRUSE
National Institute of Animal Science, DK-1958 Copenhagen, Denmark
Downloaded by [] at 11:38 16 June 2016
Kruse, V. Removal of endogenous ligands from a high-affinity antiserum for radioimmunoassay. Scand. J . clin. Lab. Invest. 39, 533--541, 1979. A method for removal of endogenous ligands from high-affinity antisera (stripping) is described. A thyroxine antiserum of very high affinity was used to develop the method. The antiserum was incubated at 50°C in a glutamate buffer at pH 4.4 together with some ethanol and methyl cellulose and a large amount of activated charcoal. After incubation for up to 2 days, the stripped antibodies were separated from the ligand adsorbed to the charcoal by centrifugation. The estimated titre increased several fold by the stripping, although the stripping method caused a loss of about 13% of the total number of binding sites over a period of 2 days. When stripped antiserum was used instead of unstripped antiserum in an assay system, the sensitivity was up to 3 times better, and the concentration, which could be measured with the best relative precision, was 3 times lower. The stripped antiserum showed a poorer specificity than the unstripped antiserum when a short incubation period was used. However, when a long incubation period was used, the specificity was nearly the same. Keywords: antiserum; charcoal ; endogenous ligand ; radioimmunoassay ; removal; stripping of antibodies; stripping of antiserum; thyroxine
V. Kruse, Ph.D., Novo Research Institute, Novo Alle, DK-2880 Bagsvaerd, Denmark
It has been known for some years that antibodies produced against a ligand will be partly saturated if the ligand is a normal component of the blood of the immunized animal [2, 7, 8, 13, 151. Ligand concentrations in immunized animals can be up to 1000 times higher than in non-immunized animals [2, 151. It has recently been reported that ligands derived from the ligand-protein conjugates used for immunization may be a significant source of saturating ligand [16]. 0036-55 13/79/ 1000-0533$02.00 0 1979 Medisinsk Fysiologisk Forenings Forlag
Several attempts have been made at removing endogenous Iigands from antiserum [3, 6, 14, 15, 19, 201. However, only a few investigators have reported some degree of success with the treatment [3, 6, 151. Our own attempts at improving a triiodothyronine (T3) antiserum by the method of Fyhrquist & Wallenius [6] were unsuccessful (Avivi & Kruse, unpublished). Other investigators have also found the method to be unsatisfactory [20]. It is necessary for an antiserum to contain at least some antibodies of high affinity when it is to be used for radioimmunoassay purposes. If
533
534
V. Krirse
not, the standard curves will be of no use. The high affinity of the antibodies makes it difficult, however, to remove endogenous ligands from radioimmunoassay antisera. This paper reports a simple and efficient method for the removal of endogenous ligands from highattinity antibodies [9].
MATERIAL AND METHODS
Downloaded by [] at 11:38 16 June 2016
Chemicals. Activated charcoal, Med. Norit, Ph. Nordica, was obtained from Norit, Amersfoort, Holland; methyl cellulose, type low substitution from BDH Chemicals, Poole, England; and L-glutamic acid from Riedel-de H a h , SeelzeHannover, W. Germany. Reagents. Glutamate buffer 0.04 molil, pH 3.4; glutamate-EDTA buffer: glutamate buffer with 3 mmol/l EDTA and 0.12 mmol/l merthiolate; acetate buffer 0.14 mol/l with 0.15 mol/l NaCl and 20 mmol/l EDTA, pH 5.0; phosphate buffer 0.2 mol/l with 10 mmol/l EDTA, pH 7.0; tris buffer 0.12 mol/l with 20 mmol/l NaN,, pH 8; and trichloroacetic acid (TCA) 0.6 mol/l. Antiserum. Thyroxine (T4) antiserum from three bleedings of rabbit number 8 (Ref. 8) was used. The bleeding from 22 July 1975 contained 350 nmol/l endogenous T4. It was taken after 10 months of T3 suppression [8]. The bleedings from 12 and 23 March 1976 were obtained after a period without Ts suppression and contained 1600 and 1500 nmol/l endogenous T4, respectively. The last bleeding contained T4 antibodies with dissociation rate constants of 0.550, 0.045 and 0.006 h-' at 23°C [lo]. These values correspond to T4-antibody complex half-lives of 1 h, 15 h and 115 h (4.8 days). The raw antisera were stored at about -20°C until use. Removal of endogenous ligands from arttisera (stripping) Principle. The stripping method is based on three principles: (I) a high dissociation rate is brought about by physico-chemical means, (2) the dissociated ligand is adsorbed to particles, and (3) the stripped antiserum is separated from the particles by centrifugation.
Protocol. Over a number of experiments, the following procedure has proved the most efficient one: mix 300 mg methyl cellulose with 3 ml ethanol, add 26 ml glutamate buffer, add 1 ml of the antiserum to be stripped, and adjust to about pH 4.4. Heat to 50"C, and add 500 mg of activated charcoal. Keep the teniperature at 50°C and mix gently. Stop the stripping after about 6 h by removing the charcoal by high-speed centrifugation ( 12,000 g ) . Treated antiserum was stored at 4-10°C for a few days without further dilution. In case of longer storage, the antiserum was diluted with 6 volumes of tris buffer. Evalmtion of stripping eflicienry Stripping curves. In order to ensure complete mixture of labelled T4 with the whole pool of endogenous Tq,the antiserum was incubated with lo6 cpm/ml of [1251]T4for 1 day at 50°C plus 4 days at room temperature before stripping. Sodium azide 15 mmol/l was used as a preservative (Caution: poisonous gases are liberated at low pH!). The equilibration procedure did not cause any detectable loss of binding sites. The stripping process was followed by analysis of samples of the stripping suspension taken after I min to 2 days of stripping. The charcoal was removed from 2 ml of stripping suspension by high-speed centrifugation and precipitable radioactivity in the supernatant was determined. The precipitation was carried out on 150 ,uI supernatant plus 400 pl normal bovine serum with 4 ml TCA. The precipitate contained predominantly labelled T4-antibody complexes and the supernatant labelled iodide. A control tube with normal rabbit serum, instead of antiserum, was stripped according to the protocol. Less than 1:4 precipitable radioactivity was found after 10 min of stripping. Assay for thyroxine in thyroxine antisera. The antisera were diluted 25 to 100 times, and T,-free serum was added to make the unknown samples similar to the standards in respect of serum proteins. Samples and standards were boiled for 15 min [4] in glutamate-EDTA buffer. The T4 concentration was measured by a sequential saturation assay [ 121. Non-specific binding was measured for all dilutions of
Stripping of antisera
boiled antiserum. No antibody binding activity was detected after boiling.
Downloaded by [] at 11:38 16 June 2016
Eflcct of c tripping on atitiseriim characteristics Titre. Curves showing the titre as a function of stripping time were constructed on the basis of the samples taken of the stripping suspension. The titre was taken as the factor with which the antiserum must be diluted to yield a 50% binding of the tracer under the given titration conditions. The incubates had a volume of 1 ml and were made in barbita;l-buffer with 40 pl T,-free serum, 1.25 pmol ammonium 8-anilino1 -naphthalene-sulphonate (displaces T, from serum proteins), 110 fmol labelled T4 and diluted antiserum. Incubation was carried out at 23°C for 4 h, and separation of bound and free T4 was performed with charcoal at room temperature [8]. Sensitivity and precision. Standard curves were made on the basis of equilibrium-type assays.
0
I
2
535
The incubates had a volume of 1.25 ml and were made in barbital buffer, with reagents added in the following order: 50 rcl serum standard, 1.6 pmol ammonium 8-anilino-lnaphthalene-sulphonate, 100 fmol labelled T, and diluted antiserum. The incubation time was 6 days at 23°C. Bound and free T, were separated as for the titration. The assay sensitivity was estimated as 2 x SD at zero concentration. The sensitivity was compared for unstripped and stripped antiserum on the basis of curves showing the same initial binding (Bo). All assays were run simultaneously, with the same batch of tracer. A precision profile was constructed for each of the standard curves. A precision profile shows the coefficient of variation as a function of the measured T, concentration. The concentrations measurable with the lowest coefficient of variation (the best relative precision) were read off from these curves. An example of a precision profile has been published by Rodbard (Fig. 6G in Ref. 18). The calculations
3
4
5
6
Time ( h )
FIG.1 . Stripping curves. The effect of 50 (. . . . .), 150 (- - -),
500 (----) and 1500 (.--.-.) nig charcoal per ml antiserum on the removal of labelled thyroxine from a thyroxine antiserum is shown as a function of stripping time. Acetate buffer (pH 5.0) without methyl cellulose was used. Other reagents and conditions were as described in the protocol. Antiserum from rabbit number 8, bled 12 March 1976.
536
V. Kruse Effect of buffer and p H . Preliminary studies showed that it was essential to perform the stripping under acid conditions. Stripping in acetate buffer at pH 5 was very effective (Fig. 3). However, about 15% of the T 4 antibodies were lost by precipitation when this buffer was used at 50°C. Precipitation could be avoided to a large extent if use was made of glutamate buffer. Figure 3 shows that a high stripping rate combined with a small loss of titre units was achieved when glutamate buffer was used at pH 4.4.
of sensitivity and precision were made according to the principles of Ekins [5].
Cross-reaction studies. Calculations were made according to Abraham [ I ] on a molar basis.
Downloaded by [] at 11:38 16 June 2016
RESULTS E j e c t of charcoal concentration on stripping rate and titre. Figure 1 shows that an increase in charcoal concentration causes an increase in the disappearance rate of labelled Ta-antibody complexes. However, the titre curves in Fig. 2 clearly show that there was a great loss of antibodies when 1500 mg charcoal was used per ml antiserum. When only 50 or 150 mgiml was used, the titre increased by a factor of about 5 . The loss of binding sites found at high charcoal concentrations was most likely caused by adsorption of antibodies to the charcoal. When no charcoal was used (Fig. 2) there was a 14%, loss of binding sites per 2 days of stripping.
Efect of ethanol. It was observed that the
stripping rate could be enhanced considerably by including some ethanol in the stripping medium (Fig. 4). Use of 10% ethanol caused a loss of about 13% of the original titre when the antiserum was incubated for 2 days with all the reagents except charcoal (Fig. 5 ) . The loss was unacceptable when higher concentrations of ethanol were used (Fig. 5 ) . The destruction of binding sites appeared to take place
2oooc
............ .............?,- - -.-.-
.-.-
- ................... - - - -- - - ...... .................
~ r r ;
I
1500C
cE
I0000
5000 I
--.-_
-.--
\.
-
-_
L .\
L 0
I
3
2
4
5
6
Time f h )
FIG. 2. Titre of antiserum stripped with 50 (. ' .. .), IS0 (--), 500 (-) and 1500 (.--.--.) mg charcoal per ml antiserum as shown in Fig. 1 . Antiserum incubated under stripping conditions without charcoal is included for comparison (. --- ). +
Downloaded by [] at 11:38 16 June 2016
Stripping of antisera
I 0
1
I
I
l
2
3
I 4
I
I
5
6
537
Time ( h )
FIG.3. Stripping curves. Effect of buffer and pH on the removal of labelled thyroxine from a thyroxine antiserum. Phosphate buffer, pH 7.2 (.. ".), glutamate buffer adjusted to p H 5.0 (- - -), pH 3.8 (-) or pH 4.4 (-) and acetate buffer, pH 5.0 (.--.--.) were used. The titre of samples taken after 3.5 h of stripping was not determined (. . . .), 18000 (-- -), 15000 (--), 18000 (--) and 13000 (.-.-.). No methyl cellulose was used. The antiserum was from rabbit number 8, bled 12 March 1976.
especially during the first few hours of incubation (Fig. 5 ) . Effect of methyl cellulose. The problem of adsorption of antibodies to charcoal at high charcoal concentrations (Figs 1 and 2) could be overcome by adding 300 mg of methyl cellulose to the stripping suspension. A suspension was made as described in the stripping protocol and compared with one made without methyl cellulose. The methyl cellulose had practically no effect on stripping and denaturation curves. The titre, however, was 17% higher when methyl cellulose was used. Eflect of serum proteins. Normal rabbit serum was added to stripping suspensions to see if this would increase the titre by preventing the adsorption of antibodies to charcoal. The suspensions were made according to the protocol, except that 0, 1 or 4 ml serum was substituted for glutamate buffer. No effect was seen on stripping and titre curves.
Endogenous thyroxine in thyroxine antiserum. The estimated concentration of Tgin unstripped antiserum is given with the description of the antisera (see Material and methods). The concentration of T4 in the stripped antiserum was too low to be measured by the described method. Sensitivity and precision. Figure 6 shows standard curves for unstripped and stripped antiserum. The assay sensitivity varied from 0.09 to 0.57 nmol/l, and was up to 3 times better for stripped antiserum than for unstripped antiserum. The largest difference in sensitivity was seen when curves with high initial binding (Bo) were compared. The lowest coefficients of variation were achieved when concentrations of about 1.5 and 5 nmol/l were measured with stripped and unstripped antiserum, respectively. Cross-reactivity. In general, stripped antiserum gave higher cross-reactions than unstripped
538
V. Kruse
Downloaded by [] at 11:38 16 June 2016
I00
0
I
I
I
I
I
I
2
3
4
5h'
&
I 2d
Time
FIG. 4. Stripping curves. Effect of 0 (.....), 5 (- - -), 10 (-), 15 (.--.--.) and 20 (-) % ethanol on the removal of labelled thyroxine from a thyroxine antiserum. The antiserum was from rabbit number 8, bled 22 July 1975. 100000
EOOCIO
---___ \
GOO00 '.
-. -. -. - -
'.
-,
- . -.'. '.
?
\
40000
.-
L
k
!
\
I
L
I
I
I
I
0
I
2
3
4
h:
'%d
Time
(...**), 5 (- - -), without charcoal.
FIG.5. Titre of antiserum incubated under stripping conditions with 0 10 (-),
I5 (--.--.) or 20 (--
--)
% ethanol as in Fig. 4-but
Stripping of antisera
Downloaded by [] at 11:38 16 June 2016
80
0
539
-
L/
I
I
I
I
0.5
I
2
4
I 8
I
16
Thyroxine (nrnol /I)
FIG.6. Standard curves made with unstripped o r stripped thyroxine antiserum in different dilutions. The unstripped antiserum was used in final dilutions between I :2,500 a n d 1 :12,600 (- - -). Antiserum from the same bleeding was stripped for I day and used in final dilutions between 1 :13,000 and I :40,000 (--). Antiserum from rabbit numher 8, bled 23 March 1976.
antiserumwith the T4 analogues tetraiodothyropropionic acid, tetraiodothyroacetic acid, 3,3’, 5’-triiodo-L-thyronine (reverse T3), the X-ray contrast medium a-ethyl-3-hydroxy-2,4,6triiodohydrocinnamic acid (Teridax) and 3,5,3’triiodo-L-thyronine (T3). However, the crossreactions estimated for reverse T3, Teridax and T, were highly dependent on incubation time, especially when stripped antiserum was used, The cross-reactions were in the range of 0 to 19% after a 2 h incubation period. When incubation was prolonged to 24 h, the crossreactions were only 0 to 4% and almost the same for unstripped and stripped antiserum. The effect of incubation time was not seen for the tetraiodo analogues. These analogues showed cross-reactions in the range of 22-81 Further details will be published elsewhere [I I].
z.
DISCUSSION In principle, the only criterion of successful stripping is that the T4 has been removed from
the antiserum (without harm to the antibodies). The effectiveness of stripping can be measured direct by determining the concentration of iodine, T4 or labelled T4 in the antiserum before and after stripping. This has only been done in rare cases [9, 191. If the concentration of labelled T4 is to be taken as a measure of the efficiency of stripping, it is of course essential that the labelled T4 is in equilibrium with the whole pool of endogenous T4. There has been a tendency to evaluate the effectiveness of a stripping method on the basis of estimates of titre [3, 6, 15, 201, apparent equilibrium constant [6, 151, assay sensitivity [6, 15, 201 or assay specificity [6]. However, such indirect estimates of stripping efficiency should be regarded as unreliable indicators of stripping, as already pointed out by Skrabanek et al. [20]. The estimates depend to a large extent on the conditions under which they are measured, as shown for specificity in Results and as discussed for titre and sensitivity below. Some of the T4 antibodies in the antiserum used for development of the stripping method
Downloaded by [] at 11:38 16 June 2016
540
V. Kruse
have a dissociation rate constant of 0.006 h - ’ at 23°C (see Material and methods). This value corresponds to a half-life of 4.8 days for the T,-antibody complexes. It is evident that with such stable complexes in the antiserum, it was necessary to take quite drastic measures in order to obtain a fast and complete stripping. However, these measures all tend to cause a loss of antibody activity. As the aim is a rapid and complete stripping, and as any loss of binding sites should be avoided, the stripping protocol must necessarily represent a compromise solution to the problem of removing ligands from high-affinity antisera. The method described does give a rather fast and practically complete stripping, albeit at the expense of about 13 ”/, of the total binding sites. A thorough study of the loss has been done for one antiserum only (Fig. 5). The question naturally arises as to what types of antisera can be stripped by the method. It was found possible to strip antisera against glucagon, ACTH, angiotensin 1, angiotensin IT, thyrotropin releasing hormone, Ts, progesterone, testosterone, estradiol, cortisol and corticosterone. The initial part of the stripping curves was steeper for all these antisera than for the T4 antiserum. This indicates that the antibodies dissociated relatively fast. About 5 % of the labelled glucagon could not be removed from the antiserum. The removal of all the other tracers was practically complete. It seems likely that the more stable the complexes of endogenous ligand and antibody, the greater the advantages gained by stripping. If it takes 4 5 days for half of the ligand-antibody complexes to dissociate, and if such stable complexes constitute a major part of the antibodies, it is clear that a titration involving incubation for 2, or even 24 h, does not give a measure of all the binding sites. The endogenous ligand would simply block a large fraction of the binding sites for days, and the tracer would not be distributed over all binding sites during the incubation period. The T4 antiserum was found to have more available binding sites after stripping in spite of the 13% loss of total binding sites caused by the stripping. Some bleedings from the rabbit showed as much as an 8-fold increase in titre after stripping. Only a few of the other hormone antisera, which were stripped in the same way as the T4
antiserum, showed a gain in titre (unpublished). The titre of antisera stripped by other methods is reported to have either increased considerably [3], remained the same [19], or have decreased [6, 15, 201. It has been shown [lo] that the good antibodies in an antiserum, i.e. the slow-dissociating antibodies, are utilized much better as binders in the stripped form than in the unstripped form under radioimmunoassay-like conditions. This improvement in utilization of good binders is considerable for high-affinity antisera, but slight for antisera of lower affinity [lo]. The assay sensitivity was up to three times better with the stripped T, antiserum. Other investigators have also found an increase in sensitivity after stripping [6, 151. The reason for the improved sensitivity is probably that endogenous ligand does not dilute the labelled ligand in the assay tube when stripped antiserum is used, in contrast to the use of unstripped antiserum. This may be illustrated by the following example. The standard curves for unstripped antiserum shown in Fig. 6 were set up with 100 fmol labelled T, and antiserum in final dilutions between 1 :2,500 and 1 :12,600. The amount of endogenous T4 added to each assay tube together with the unstripped antiserum was between 150 and 750 fmol. The specific radioactivity of the labelled T, was thus reduced to between 40 and 12% of that in the tubes with stripped antiserum. A similar reduction to 23 % of the original specific radioactivity of the tracer has been reported in the description of a T 3 assay [2]. The specificity of the T, antiserum did not appear to be improved by the stripping. Theoretically, the specificity is highest when the incubation i s carried out to equilibrium [17]. Unfortunately, the specificity was not estimated at equilibrium in the present case. However, the finding that the cross-reactions with reverse TB,Teridax and T 3 lessened the longer the incubation period, and that the decrease in cross-reactivity was most pronounced for the stripped antiserum, indicates that the specificity at equilibrium is at least as good for the stripped as for the unstripped antiserum. Incubation to equilibrium would have little effect on the cross-reactions with the tetraiodo analogues because no effect of incubation time on cross-reactivity was detected for these substances.
Stripping of antisera No problems have been encountered with the stability of stripped antiserum when neutralized with 1 volume of tris buffer and stored at 410°C. Freezing in 10 ml tubes up to six times did not cause any detectable loss of titre.
ACKNOWLEDGMENTS
Downloaded by [] at 11:38 16 June 2016
I wish to thank Karen Saxtorph Jensen for excellent technical assistance. Antisera were generously provided by G.D. Niswender, L.G. Heding, A.M. Kappelgaard, A. Veje, H. Worsaae and L. Hummer. This work was supported by The Scandinavian Contact Agency for Agricultural Research as Project 34.
REFERENCES 1 Abraham, G.E. Solid-phase radioimmunoassay of
estradiol-17p. J. din. Endocr. 29, 866, 1969. 2 Alexander, N.M. & Jennings, J.F. Radioimmunoassay of serum triiodothyronine on small, reusable Sephadex columns. Clin. Chem. 20, 1353, 1974. 3 Auletta, F.J., Caldwell, B.V. & Hamilton, G.L. Androgens: Testosterone and dihydrotestosterone. p. 359 in Jaffe, B.M. & Behrman, H.R. (eds) Methods of’ Hormone Radioimmunoassay. Academic Press, New York and London, 1974. 4 Dunn, R.T. & Foster, L.B. Radioimmunoassay of thyroxine in unextracted serum, by a single-antibody technique. Clin. Chem. 19, 1063, 1973. 5 Ekins, R.P. Basic principles and theory. Br. med. Bull. 30, 3, 1974. 6 Fyhrquist, F. & Wallenius, M. Improvement of antisera against polypeptide hormones. Nature (Lond.)254,82, 1975. 7 Hillier, S.G., Cole, E.N., Groom, G.V., Boyns, A.R. & Cameron, E.H.D. Effect of active immunisation against testosterone-3-BSA on circulating levels of testosterone, LH, prolactin and testosterone antibody titre in the male rat. Steroids 22, 227, 1973.
541
8 Kruse, V. Production and evaluation of highquality thyroxine antisera for use in radioimmunoassay. Scand. J. din. Lab. Invest. 36, 95, 1976. 9 Kruse, V. A method for removal of endogenous ligands from high-affinity antibodies to be used in radioimmunoassay. p. 57 in Abstracts 2nd European Congress on Clinical Chemistry. Prague, 1976. 10 Kruse, V. Dissociation rate constants and fractional binding of tracer estimated for three antibody populations in unstripped and stripped antiserum. Scand. J. din. Lab. Invest. 39, 215, 1979. 1 I Kruse, V. Selective removal of fast-dissociating antibodies from a high-affinity antiserum. Molec. Immun. (In press). 12 Kruse, V. & Lind, 0. A rapid and precise sequential saturation radioimmunoassay for thyroxine. Scand. J. d i n . Lab. Invest. 31, 149, 1977. 13 Nieschlag, E., Usadel, K.H., Schwedes, U., Kley, H.K., Schoffling, K. & Kruskemper, H.L. Alterations in testicular morphology and function in rabbits following active immunization with testosterone. Endocrinology 92, 1142, 1973. 14 Niswender, G.D. p. 128 in Cameron, E.H.D., Hillier, S.G. & Griffiths, K. (eds) Steroid Immunoassay. Alpha Omega Publishing Ltd., Cardiff, U.K., 1975. 1 5 Oliver, L.K. & Cano, C. Removal of an endogenous antigen from an antibody to increase its effective affinity constant, as illustrated by triiodothyronine assay. Clin. Chem, 23,2039, 1977. 16 O’Neill, S.P. & Robb, T.E. Hapten concentrations in the circulation of animals actively immunized with hapten-protein conjugates. J . Immun. 116, 363, 1976. 17 Pratt, J.J. & Woldring, M.G. Radioimmunoassay specificity and the ‘first-come, first-served effect’. Clin. Chim. Acta 68,87, 1976. 18 Rodbard, D. Statistical quality control and routine data processing for radioimmunoassays and imrnunoradiometric assays. Clin. Chem. 20, 1255, 1974. 19 Seth, J., Rutherford, F.J.’ & McKenzie, I. Solidphase radioimmunossay of thyroxine in untreated serum. Cfin. Chem. 21, 1406, 1975. 20 Skrabanek, P., Kirrane, J. & Powell, D. Apparent improvement of antisera for radioimmunoassay by treatment with sodium iodide. J. Immun. Merh. 16, 331, 1977.
Received 26 May 1978 Accepted 6 March 1979
