Vox Sang. 29: 1-14 (1975)
A New Solid Phase Radioimmunoassay (CLB-RIA) for the Detection of Hepatitis-B Antigen and Antibody W. J. DUIMEL and H. G. J. BRUMMELHUIS~
Abstract. A new competitive solid phase radioimmunoassay (CLB-RIA) has been developed for the detection of HBAg and HBAb in human serum and plasma. In the assay, sheep antibodies to HBAg, covalently linked to an insoluble carrier, highly purified lasI labelled HBAg and the serum or plasma sample are incubated for 20 h at room temperature. After incubation, the bound and the free fraction of the tracer are separated by centrifugation. Both the presence of HBAg and HBAb, result in a decrease of the amount of bound tracer, when compared with a negative control serum. Differentiation between HBAg and HBAb requires the use of another type of radioimmunoassay. For this purpose a sandwich solid phase radioimmunoassay, for the detection of HBAb only, has been developed (CLB-AURIA). In this, assay-purified HBAg is covalently linked to an insoluble carrier. Using a mixture of both immunosorbents (insolubilized HBAg and HBAb), it is possible to detect and to distinguish HBAg and HBAb in one single solid phase radioimmunoassay (CLB-MIRIA). The influence of three parameters on the CLB-RIA, the incubation time, the amount of tracer and the effect of Tween-20 has been studied. The sensitivity of the described solid phase CLB-RIA for the detection of HBAg is comparable to that of other radioimmunoassays reported in literature; its specificity is very high.
for her valuable The authors wish to thank Miss A. H. M. VAN GIJLSWIJK assistence.
Received: August 30, 1974; accepted: October 24, 1974.
2
Introduction Hepatitis-B antigen (HBAg) was first described by BLUMBERG [5] and later shown to be associated with viral hepatitis [6,19]. Since, several techniques have been developed for the detection of both antigen and corresponding antibody (HBAb). The techniques mainly used, are the double agar gel diffusion (AGD) technique, counterimmunoelectrophoresis and to some extent, the complement fixation test. However, the sensitivity of these techniques, in terms of the minimal detectable amount of antigen, is rather poor in comparison with other techniques, such as the haemagglutination inhibition and the radioimmunoassay [13, 16,22-241. Several types of radioimmunoassays have recently been described. In general two different techniques are used in which either soluble or insolubilized antibodies are used respectively. When using a soluble antibody, the separation between the antigen-antibody complexes and the free radioiodinated antigen can be obtained in different ways. WALSHet al. [28] separated the bound from the free fraction by means of electrophoresis. AACHet al. [l] and HOLLINGER et al. [14] separated the bound from the frez fraction, after addition of a second antibody, directed against the immunoglobulins of the species corresponding to the used anti-HBAg antibody. The complex was then precipitated by ultracentrifugation. COLLAR et al. [lo] added ammonium sulphate to the incubation mixture. AMOUCH et al. [4] diluted the incubation mixture after incubation to enhance the precipitation of the antigen-antibody complex. In general in a radioimmunoassay separation between the bound and the free fractions is greatly facilitated if the antibody or the antigen is made insoluble, e.g. by adsorption to a tube [9, 291 or by covalent linking to an insoluble carrier [29]. LINGand OVERBY [18] developed a sandwich radioimmunoassay in which the antibody was made insoluble by adsorption to a plastic tube and in which they used radioiodinated antibody as a tracer. AALBERSE et al. [2] described the possibility of the simultaneous detection of and distinction between IgA and anti-IgA and between IgE and anti-IgE antibodies, using a mixed suspension of insoluble antigen and insoluble antibody. Because of the facility of the separation and the speed of the determination, we decided to use a solid phase radioimmunoassay. In order to be able to detect both antigen and antibody, we chose a competitive assay. This paper describes the development and practical use of three different types of solid phase radioimmunoassays: (1) a competitive assay for the de-
New Solid Phase Radioimmunoassay (CLB-RIA)
3
tection of HBAg and HBAb; (2) a sandwich assay to detect HBAb, and (3) a combination of 1 and 2 in which HBAg and HBAb can be detected and distinguished.
Materials and Methods All chemicals used were reagent grade. Bufers. Buffer I =phosphate-buffered saline (PBS) (0.140 M NaCI, 0.0113 M Na,HP04, 0.0017 M NaH,PO,, pH 7.2); buffer II=PBS diluted with an equal volume of distilled water, pH 7.5 ;buffer 111=PBS containing 0.001 M EDTA (Nh) and 0.01% NaN,, adjusted to pH 7.8; buffer IV=buffer 111 containing 2.5 mg/ml bovine serum albumin; buffer V = buffer IV to which 0.1% Tween-20 was added; buffer VI = 0.05 M borax, adjusted with HCI to pH 8.6; buffer VII=O.O5 M acetic acid in saline (0.1548 M NaCI) adjusted to pH 4.5 with NaOH. Antisera. All antisera used were prepared at the Central Laboratory of the Netherlands Red Cross Blood Transfusion Service, Amsterdam. Purification Procedures Sheep HBAb. After dialysis of the sheep anti-HBAg antiserum against buffer 11, a partially purified IgG-fraction was obtained by column chromatography on Sephadex DEAE A-50 (Pharmacia, 5 x 40 cm) equilibrated with buffer 11, which buffer was also used for the elution. The fractions containing anti-HBAg antibody in the double AGD test, were pooled, concentrated by ultrafiltration and dialyzed against buffer VI. Human HBAb. A 16% immunoglobulin concentrate was prepared from pooled human plasma, according to a modified Cohn VI fractionation method [ l q . Before pooling, all plasmas were tested and found positive for HBAb in the double AGD technique. HBAg. Hepatitis-B antigen, purified by adsorption on Aerosil and gelfiltration on Sephadex G-200 as described previously [ l l ] was highly purified by isopycnic banding in a CsCl density gradient. Before centrifugation, the density (measured by refraction at 25 "C) of the HBAg-solution was brought to 1.20 g/ml by addition of solid CsCI. Centrifugation at 50,000 rpm was performed for 40 h at 4°C (rotor 65, Beckman ultracentrifuge). Labelling Procedure The highly purified HBAg was labelled with lZ6I according to HUNTER and GREENWOOD [15]. The reaction mixture contained 0.25 ml HBAg solution (E;km=LlO), 0.15 ml chloramine-T solution (4 mg/ml) and 0.01 ml KlP6Isolution (1 mCi la61). The reaction was stopped after 30 sec by the addition of 0.8 mg sodiummetabisulphite in 0.20 ml buffer I and of 0.15 ml KI solution (50 mg/ml). After dialysis against three charges of 500 ml buffer 111 to which 0.01 mg/ml KI had been added, the labelled HBAg was separated from the last reaction reagents and byproducts by gel filtration on a Sephadex G-200 column (Pharmacia, 2.4 x 25 cm), equilibrated and eluted at 4°C with buffer IV. Preparation of Insolubilized HBAg and HBAb Sheep HBAb. The partially purified antibodies against Hepatitis-B antigen after dialysis against buffer VI were linked covalently to an insoluble carrier, i.e. voluminous m-diazo-
DUIMEL/BRUMMELHUIS
4
benzyloxymethylene cellulose (0.83 mg N/g amino cellulose) according to a modification described by BRUMMELHUIS [8] of the original method of GURVICH et al. [12]. For the coupling of sheep HBAb, 15 m1 protein solution (E:Bcb"=23.7) was mixed with 15 ml (20 mg/ml) diazotized voluminous amino cellulose. The reaction was allowed to proceed for 24 h at 4°C under continuous stirring. To inactivate the diazo groups, which did not react, the suspension was then stirred for 4 h at room temperature. The insolubilized antibody preparation ((HBAb)-SP) was washed twice with saline, twice with buffer VII, once with buffer IV and twice with buffer V. Finally the (HBAb)-SP was suspended in buffer V (0.75 mg cellulose/ml) and stored at 4°C. It could be shown that about 2.5 mg protein were bound to 10 mg cellulose. Human HBAb. For the coupling of human HBAb to the cellulose carrier, after dialysis against buffer VI, 0.8 ml of the 16% immunoglobulin concentrate was incubated with 50 mg cellulose (total volume 3.6 ml). The incubation, the washing and the storage were performed as described above for sheep HBAb. Approximately 12mg protein were bound to 10 mg cellulose. HBAg. The coupling of the highly purified HBAg to voluminous m-diazobenzyloxymethylene cellulose was performed in the same way as the coupling of HBAb. However, because of the lower protein concentration, 12.6 ml HBAg solution (Elirn=0.26) were incubated with 30 mg diazotized voluminous amino cellulose (total reaction volume : 15 ml). About 0.1mg HBAg % = 3.726; VYASet al. [26]) was bound to 10mg cellulose. The stock suspension of these insoluble antigens ((HBAg)-SP) was 0.5 mg cellulose/ml in buffer V.
@!d
Assay Procedures Double AGD was performed according to BRUMMELHLIIS [A.Counter immunoelectrophoresis was performed according to ALTERet al. [3]. The coated tube radioimmunoassay (AUSRIA-125, Abbott Laboratories), was performed according to the instructions of the manufacturer. Procedure of the CLB-Radioimmunoassays CLB-RIA. A competitive solid phase radioimrnunoassay. In a polystyrene tube (10 x 40 mm) 0.1 ml of the sample, 0.2 ml of the tracer and 0.2 ml of the immunosorbent were incubated for about 20 h at room temperature. The radioactivity of the tracer ( T HBAg) was about 50,000 cpm/ml, counted in a well-type automated gamma-counter (Wallac GTL 300). The immunosorbent suspension was a mixture of (HBAb)-SP (0.03mg cellulose/ml) and of normal voluminous aminocellulose (0.3mg cellulose/ml) in buffer V. The amount of immunosorbent was chosen in such a way that after incubation for 96 h (=at equilibrium) 30% of the tracer had been bound. After an incubation time of 24 h about 15% of the tracer was bound; this implicates that the counting level of negative samples is about 1,500cpm. During incubation the tubes were rotated end over end. After incubation, to separate the bound (laSI-HBAgbound to (HBAb)-SP) from the free fraction of lasI-HBAg, the tubes were centrifuged for 2 min at 3,OOOg. The supernatant was discarded, the remaining immunosorbent was washed six times with 1.2 ml PBS for each washing. Thereafter, radioactivity of the bound fraction was counted. CLB-A URIA. A sandwich solid phase radioimmunoassay and CLB-MIRIA, a combination of CLB-RIA and CLB-AURIA. With the exception of the immunosorbents the performances of the CLB-AURIA and the CLB-MIRIA were identic to the CLB-RIA
New Solid Phase Radioimmunoassay (CLB-RIA)
5
performance. In the CLB-AURIA the immunosorbent consisted of 0.2 ml of a mixture of (HBAb)-SP and (HBAg)-SP (0.03, respectively, 0.5 mg cellulose/ml). Samples were regarded as positive in the CLB-RIA and the CLB-MIRIA if the amount of bound tracer differed more than 15% of that of the mean of eight negative control sera. (15% was about 5 standard deviations of the mean of the control sera.) Samples which differed 10-15% were called dubiously positive. In the CLB-AURIA, samples were regarded as positive if the amount of bound tracer was at least twice the amount of the control sera mean; between 1.5 times and twice this mean the sample was called dubiously positive.
Results PuriJication and labelling of HBAg After isopycnic banding the HBAg-positive fractions were further examined. No human serum proteins could be detected in the AGD against horse anti-total human serum, nor against several specific antiserums (anti-IgG, anti-IgM, anti-a,M, anti$ lipoprotein and anti-transferrin). After labelling and dialysis, gelfiltration of this purified HBAg preparation on a Sephadex G-200 column resulted in three peaks containing radioactivity (fig. 1). Of the radioactivity in the first peak fractions (eluted at V,) 97% could be precipitated with trichloroacetic acid (TCA) (final TCA concentration 10%). The TCA-precipitable radioactivity in the fractions of the second and the third peak was 80 and 30%, respectively. The total amount of radioactivity in the fractions of the first peak was about 120,uCi (net yield of labelling procedure was 12%). A sample out of the pooled fractions 14-16 was examined for isopycnic homogenicity. More than 90% of the radioactivity were recovered in one single band with a density of 1.23 g/cmS(fig. 2). Radioimmuno Procedures CLB-RIA. In order to undo the influence of the protein concentration of the sample to be tested, it was necessary to add Tween-20 to the incubation mixture. This addition of Tween-20 prevented adsorption of the tracer to the plastic tube. It was found that the adsorption to the tube depended strongly on the protein concentration in the mixture. Therefore the amount of ‘bound’ tracer counted was about 80% more in saline than in serum. The difference in ‘bound’ tracer was almost entirely due to adsorption to the tube. It was found that this adsorption could be prevented by the addition of Tween-20 (0.02-0.04%). This had no influence on the amount of the bound radioactivity in serum (protein concentration in reaction mixture about 14 mglml). It could be shown that after the addition of Tween-20 the sensi-
6
a45ml
Fig. I . Elution pattern of 149-labelledHBAg. Column: Sephadex G-200. For details
see text.
Fig. 2. Ultracentrifugation in CsCl density gradient of purified 0=Radioactivity; A =density. For details see text.
HBAg.
tivity of the test was independent of the protein concentration, which means that comparable dose-response curves were obtained with serum dilutions and with buffer (PBS) dilutions of HBAg-positive sera (table I). The influence of the incubation time on the sensitivity of the assay was studied. Using two different amounts of tracer, the B/Bo was calculated as a function of the incubation time (B/Bo is the ratio of the amounts of bound tracer between a HBAg- positive and a HBAg-negative sample). The sensi-
New Solid Phase Radioimmunoassay (CLB-RIA)
7
Table I. Titers obtained by testing twofold serial dilutions of 12 HBAg-positive serum samples in the CLB-RIA No. sample
Titer when diluted with PBS
Titer when diluted with HBAgand HBAb- negative serum
1 2 3 4 5 6 7 8 9 10 11 12
128 1,024 128 1,024 512 512 16 32 8 8 1 2
256 1,024 128 2,048 256 1,024 16 128 32 16 2 4
Table ZZ. CLB-RIA. Effect of the incubation time and the amount of tracer on the ratios B Bo-B -and Bo S Time, h
B Bo
2 8 20
T = 16,000 cpm/ml
T = 5,500 cpm/ml
0.27 0.16 0.13
Bo-B S
12 21 29
B Bo
0.28 0.16 0.13
Bo-B S
20 36 44
tivity increased with the increase of the incubation time (table 11). At this level there was no influence of the amount of tracer on B/Bo. The influence of the amount of tracer became manifest when the parameter (Bo-B)/s was used instead of B/Bo (table 11). (Bo-B)/s is the ratio between the difference of the amounts of bound tracer of a HBAg-negative and a HBAg-positive sample and the statistical counting error (s) at the Bo level. The ratio (Bo-B)/s gives an impression of
DUINIELlBRUMMELHUIS
8
Fig. 3. CLB-RIA dose-response curve of HBAg (correlation coefficient=0.989) (RIA). For details see text.
60-
5
10
50
100
200
500
-+ dllutlons
Fig. 4. Exact dilutions in negative serum of a HBAg- ( 0 )and a HBAb- ( A ) positive serum tested in the CLB-RIA (IS = HBAb-SP). For details see text.
the significance of the difference between Bo and B, with the higher amount of tracer this significance was greater. Besides the minimal amount of HBAg that gave a positive reaction was determined. In figure 3 the dose-response curve is shown. The smallest amount of antigen that gave a significant decrease in B, was about 2-4 ng (HBAg concentration in the sample: 20-40 nglml). Not only HBAg, also HBAb-positive samples gave a decreased amount
New Solid Phase Radioimmunoassay (CLB-RIA)
5
9
10
Fig. 5. Exact dilutions in negative serum of a HBAb-positive serum tested in the CLBAURIA (IS = HBAg-SP). For details see text.
510-
1B x100% '0
210 -
---;-1
-
-&tion 5
10
50
100
200
500
Fig. 6. Exact dilutions in negative serum of a HBAg- ( 0) and a HBAb-positive ( A )
serum tested in the CLB-MIRIA.
HBAg SP HBAb SP
. For details see text.
of bound tracer. It is therefore impossible to distinguish between HBAgand HBAb-positive samples in this radioimmunoassay performance (fig. 4). CLB-A URIA. Using insolubilized HBAg as an immunosorbent instead of insolubilized HBAb, it was possible specifically to test it for the presence of HBAb (fig. 5). CLB-MIRIA. When a mixed suspension of insolubilized HBAg and insolubilized HBAb was used as an immunosorbent, it was possible not only
DUIMEL/BRUMMELHUIS
10
to detect HBAg, but also to distinguish between HBAg- and HBAb-positive samples. As in the CLB-RIA, a HBAg-positive sample gave a decrease and as in the CLB-AURIA a HBAb-positive sample gave an increase of the amounts of bound tracer (fig. 6). Specijicity The specificity of samples which were found positive for HBAg in the CLB-RIA was confirmed in an assay system in which human antibody to HBAg was used instead of sheep antibody. The assay itself was performed in the same way as the CLB-RIA using sheep antibody. The results of tests for specificity are summarized in table 111. The 16,330 samples were tested in pools, composed of 10 donations each. Of these 1,633 tested pools, 43 were found positive, or dubiously positive, for HBAg. To identify the samples which were responsible for the positive reactions of the pools, all 430 samples of these 43 pools were tested separately. This resulted in 34 positive samples and in 9 dubiously positive for HBAg. Of the 34 samples found HBAgpositive in the CLB-RIA, 4 samples were negative and 30 were positive when tested for specificity. When tested in CEP, only 11 of these 34 samples were CEP-positive, 23 were CEP-negative. The 4 CLB-RIA nonspecific positive samples were all negative in CEP. Among the 9 samples found dubiously positive for HBAg in the CLB-RIA, only 2 were positive and 7 were negative when tested for specificity in the assay in which human HBAb is used. In the CEP all 9 samples were negative.
Table III. Results of testing the specificity of the 43 samples which were found HBAgpositive or dubiously positive in screening 16,330 blood donations with the CLB-RIA No samples
CLB-RIA Sheep HBAb
-
7
+
2
9 4 19 11
-
34
CLB-RIA human HBAb
+ + +
+ 4-
CEP
New Solid Phase Radioimmunoassay (CLB-RIA)
11
Discussion
For this radioimmunoassay of HBAg it is essential to use highly purified antigen, since the presence of impurities may decrease the sensitivityand may cause false-positive results. Therefore, it was necessary to extend the previously described purification procedure [161. After labelling and dialysis of the highly purified HBAg, 75% of the radioactivity were eluted in the first peak fractions from a Sephadex G-200 column. The maximum binding of the tracer to insolubilized HBAb was about 80%, measured directly after gel filtration. By ageing, this percentage decreased to about 60%. The binding could be inhibited to less than 5% by the addition of HBAg-positive sera. The binding of the tracer to insolubilized negative serum was about 4%. Approximately 5% of the tracer were bound by the insolubilized IgG-fraction of a horse anti-human total antiserum. From these data we concluded that the binding of labelled HBAg is specific, whereas aspecific binding of the tracer to the immunosorbent hardly occurs. The addition of Tween-20 to the incubation mixture permitted the assay of HBAg and HBAb in samples independent of the protein concentration, at least if the protein concentration in the incubation mixture was less than 3%. This is in contrast to other sensitive tests such as those based on haemagglutination reactions [21]. Since the sensitivity of the assay was found to be time dependent, it was necessary to incubate for 18-20 h. Although longer incubation times did increase the amount of bound tracer, the sensitivity had not increased significantly. Therefore, a test could be performed within 24 h at a tracer concentration of about 20.000 cpm/ml. Although the (Bo-B)/s was not optimal at this tracer concentration (its optimum was at 80,000 cpmlml) we chose this concentration because the ratio B/Bo was really optimal. The reproducibility of the test was very good. The standard deviation of the amounts of bound tracer of the controls was rarely more than 3-4%, at a counting level of about 1,500 cpm. False-positive results may occur, due to interactions which disturb the reaction between HBAg and the insolubilized HBAg. These interactions may be caused by the presence of proteins which react, or cross-react, with the insolubilized sheep antibodies. PRINCE et al. [20] and VYASet al. [27] considered samples truly positive only if their reactivity could be specifically inhibited. This inhibition was carried out with human HBAb. In terms of competitive radioimmunoassays this means, that samples are truly positive only if they react positively with both human and
12
DUIMEL/BRUMMELHUIS
heterologous HBAb. The results of the test for specificity of the samples found HBAg-positive in the CLB-RIA were very good. The percentage of 85 real positives is high in comparison with the data of PRINCE et al. [20] and VYASet al. [27], who reported 21 and 8%, respectively, using a sandwich solid phase radioimmunoassay. The disagreements are probably caused by the fact that these investigators used a sandwich technique, whereas we used a competitive technique. False-positive reactions in the sandwich solid phase radioimmunoassay may be caused, according to PRINCC et al. [20] by binding (sandwich formation) of (cross-reacting) anti-guinea pig globulin antibodies with the insolubilized and with the labelled guinea pig anti-HBAg antibodies. All insolubilized guinea pig antibodies may cause false-positive reactions. In the CLB-RIA this phenomenon can not occur since in this radioimmunoassay the tracer consists of labelled antigen instead of labelled antibody. Falsepositive reactions in the CLB-RIA can be caused by binding of (cross-reacting) anti-sheep globulin antibodies, causing inhibition of tracer binding by steric hindrance. Because only a small part (less than 1%) of the insolubilized sheep immunoglobulin is directed against HBAg, only a small part of the (cross) reactions will cause steric hindrance with regard to the HBAb binding sites and thus cause false-positive results. For this reasons, in our opinion the CLB competitive solid phase radioimmunoassay may be more specific than a sandwich solid phase radioimmunoassay. Although the mechanisms are quite different, it was impossible to distinguish between an antigen and an antibody positive sample in the CLBRIA. When antigen is present there is a competition between the tracer and the antigen with respect to the insolubilized antibody. When the sample is antibody positive, there is a competition between the insolubilized and free antibody with respect to the tracer. Part of the tracer will react with the free antibody and form soluble complexes which are washed out after the incubation. Both competitions result in a decrease of tracer binding to the insolubilized antibodies. Addition of insolubilized antigen together with insolubilized antibody (CLB-MIRIA) permitted to detect and to distinguish between HBAg and HBAb. The sensitivity for detecting HBAg was the same in the CLB-RIA and CLB-MIRIA, in contrast to the sensitivity for detecting HBAb which was slightly less in the CLB-MIRIA. In the CLB-AURIA, the sensitivity for detecting antibody was the same as in the CLB-MIRIA. This means that with our methods the sensitivity for detecting HBAb, with labelled HBAg as a tracer, is better in a competitive assay with insolubilized antibody than in a sandwich assay with insolubilized
New Solid Phase Radioimmunoassay (CLB-RIA)
13
antigen. By screening samples with the CLB-RIA test and retesting the positive samples in the CLB-MIRIA test, even samples weakly positive for HBAg and HBAb can be detected. Samples will be found positive for HBAg in the CLB-RIA and the CLB-MIRIA at HBAg concentrations in the actual samples as low as 20-40 ng/ml. References 1 AACH,R. D.; GRISHAM, J. W., and PARKER,C. W.: Detection of Australia antigen by radioimmunoassay. Proc. nat. Acad. Sci., Wash. 68: 1056 (1971). R. C.; LOCHEMVAN, ERNA;MUNSTER VAN, P. J. J., and NADORP, J. H. J. : 2 AALBERSE, Human antibodies to IgE: a simple method to avoid ambiguous results of the radioimmunosorbent test (RIST). J. Lab. clin. Med. 83: 831 (1974). R. H.: Counterelectrophoresis for 3 ALTER,H. J.; HOLLAND,P. V., and PURCELL, detection of hepatitis-associated antigen: methodology and comparison with gel diffusion and complement fixation. J. Lab. clin. Med. 77: 1000 (1971). P. J. et DROUHET, J. : Dosage radioimmunologique de l'antigen Australie. 4 AMOUCH, Nouv. Rev. franc. HCmat. 11: 656 (1971). 5 BLUMBERG, B. S.: Polymorphisms of the serum proteins and the development of isoprecipiuns in transfused patients. Bull. N.Y. Acad. Med. 40: 377 (1964). 6 BLUMBERG, B. S.; SUTNICK,A. I.; LONDON, W. T., and MILLMAN, I.: Current concepts. Australia antigen and hepatitis. New Engl. J. Med. 283: 349 (1970). H. G. J. : Technical aspects and results of the determination of hepatitis 7 BRUMMELHUIS, associated antigen (HAA) and antibody. Vox Sang. 19: 228 (1970). 8 BRUMMELHUIS, H. G. J. :Het isoleren van specifieke antistoffen met behulp van immuun sorbentia; thesis Amsterdam. Hemstetstad Amsterdam. 9 CATT, K. J. and TREGEAR, G. W.: Solid phase radioimmunoassay in antibody-coated tubes. Science 158: 1570 (1967). 10 COLLAR, J. A. ; MILLMAN, I. ; HALBHERR, T. C., and BLUMBERG, B. S. : Radioimmunoprecipitation assay for Australia antigen antibody and antigen-antibody complexes. Proc. SOC.exp. Biol. Med. 138: 249 (1971). 11 DUIMEL,W. J.; BRUMMELHUIS, H. G. J., and KRIJNEN,H. W.: The purification of hepatitis-associated antigen (HAA/SH) from human serum and the preparation of a sheep anti-HAA serum. Vox Sang. 23: 249 (1972). 0. B., and TUMANOVA, A. E.: Production of protein12 GURVICH,A. E.; KUZOVLEVA, cellulose complexes (immunosorbents) in the form of suspensions able to bind great amounts of antibodies. Translated from Biokhimiya 26: 934 (1961). 13 GUST,I. D.: Evaluation of the techniques used for the detection of Australia antigen and antibody. Pathology 3: 151 (1971). F. B.; VORNDAM, V., and DREESMAN, G. R.: Assay of Australia antigen 14 HOLLINGER, and antibody employing double-antibody and solid-phase radioimmunoassay techniques and the comparison with the passive haemagglutination methods. J.Immunol.107: 1099 (1971). 15 HUNTER,W. M. and GREENWOOD, F. C.: Preparation of iodine-131 labeled human growth hormone of high specific activity. Nature, Lond. 194: 495 (1962).
14
DUIMEL/BRUMMELHUIS
16 KAPP, J. F.: Sensitive Australia antigen complement fixation test (CFT) and its comparison with Ouchterlony and countercurrent electrophoresis techniques. Med. Microbiol. Immunol. 157: 58 (1971). H. G. J., and BEENTJES,S. P.: Production of specific 17 KRIJNEN,H. W.; BRUMMELHUIS, antibody concentrates; in MERLERImmunoglobulins, p. 332 (National Academy of Science, Washington D.C. 1970). 18 LING,C. M. and OVERBY,L. R.: Prevalence of hepatitis B virus antigen as revealed by direct radioimmunoassay with (lza1) antibody. J. Immunol. 109: 834 (1972). 19 PRINCE,A. M.: An antigen detected in the blood during the incubation period of serum hepatitis. Proc. nat. Acad. Sci., Wash. 60: 814 (1968). A. M.; BROTMAN, B.; JASS, D., and IKRAM, H.: Specificity of the direct solid 20 PRINCE, phase radioimmunoassay for the detection of hepatitis B antigen. Lancet i: 1346 (1973). H. G. J.: Evaluation of a new 21 REESINK,H. W.; DUIMEL,W. J., and BRUMMELHUIS, haemagglutination technique for the demonstration of hepatitis B antigen. Lancet ii: 1351 (1973). J. P. ; CHICOT,D., et SALMON,C. : Comparative study of four 22 ROPARS,C. ; CARTRON, techniques for the detection of Australia antigen and a dextran sulfate test. Nouv. Rev. franc. Hemat. 12: 217 (1972). 23 SCHAFFER, W. E.; VYAS,G. N.; SHAKED, A.; CHEN,E., and PERKINS,H. A.: Comparison of counterelectrophoresis and haemagglutination inhibition test for hepatitis associated antigen. Vox Sang. 22: 366 (1972). E. H. : Relative sensetivity of gel diffusion 24 SCHMIDT,N. J. ;GEE,G. S., and LENNETTE, complement fixation and immunoelectroosmophoresistests for detection of hepatitis associated antigen and antibody. Appl. Microbiol. 22: 165 (1971). 25 VYAS,G. N. and SCHULMAN, N. R.: Haemagglutination assay for antigen and antibody associated with viral hepatitis. Science 170: 332 (1970). E. W. ;KLAUS,G. G. B., and BOND, H. E. :Hepatitis associated 26 VYAS,G. N. ;WILLIAMS, Australia antigen-protein, peptides and amino acid composition of purified antigen with its use in determining sensitivity of the haemagglutination test. J. Immunol. 180: 1114 (1972). 27 VYAS,G. N. : Personal communication. 28 WALSH,J. H.; YALOW, R. S., and BERSON, S. A.: Detection of Australia antigen and antibody by means of radioimmunoassay techniques. J. infect. Dis. 121: 550 (1970). Radio29 WIDE, L.: Solid phase antigen-antibody systems; in KIRKHAMand HUNTER immunoassay methods, p. 405 (Churchill/Livingstone, London, Edinburgh 1971).
Dr. Jr. W. J. DUIMEL,Centraal Laboratorium van de Bloedtransfusiedienst, Plesmanlaan 125, Amsterdam- W (The Netherlands)
A New Solid Phase Radioimmunoassay (CLB-RIA) for the Detection of Hepatitis-B Antigen and Antibody W. J. DUIMEL and H. G. J. BRUMMELHUIS~
Abstract. A new competitive solid phase radioimmunoassay (CLB-RIA) has been developed for the detection of HBAg and HBAb in human serum and plasma. In the assay, sheep antibodies to HBAg, covalently linked to an insoluble carrier, highly purified lasI labelled HBAg and the serum or plasma sample are incubated for 20 h at room temperature. After incubation, the bound and the free fraction of the tracer are separated by centrifugation. Both the presence of HBAg and HBAb, result in a decrease of the amount of bound tracer, when compared with a negative control serum. Differentiation between HBAg and HBAb requires the use of another type of radioimmunoassay. For this purpose a sandwich solid phase radioimmunoassay, for the detection of HBAb only, has been developed (CLB-AURIA). In this, assay-purified HBAg is covalently linked to an insoluble carrier. Using a mixture of both immunosorbents (insolubilized HBAg and HBAb), it is possible to detect and to distinguish HBAg and HBAb in one single solid phase radioimmunoassay (CLB-MIRIA). The influence of three parameters on the CLB-RIA, the incubation time, the amount of tracer and the effect of Tween-20 has been studied. The sensitivity of the described solid phase CLB-RIA for the detection of HBAg is comparable to that of other radioimmunoassays reported in literature; its specificity is very high.
for her valuable The authors wish to thank Miss A. H. M. VAN GIJLSWIJK assistence.
Received: August 30, 1974; accepted: October 24, 1974.
2
Introduction Hepatitis-B antigen (HBAg) was first described by BLUMBERG [5] and later shown to be associated with viral hepatitis [6,19]. Since, several techniques have been developed for the detection of both antigen and corresponding antibody (HBAb). The techniques mainly used, are the double agar gel diffusion (AGD) technique, counterimmunoelectrophoresis and to some extent, the complement fixation test. However, the sensitivity of these techniques, in terms of the minimal detectable amount of antigen, is rather poor in comparison with other techniques, such as the haemagglutination inhibition and the radioimmunoassay [13, 16,22-241. Several types of radioimmunoassays have recently been described. In general two different techniques are used in which either soluble or insolubilized antibodies are used respectively. When using a soluble antibody, the separation between the antigen-antibody complexes and the free radioiodinated antigen can be obtained in different ways. WALSHet al. [28] separated the bound from the free fraction by means of electrophoresis. AACHet al. [l] and HOLLINGER et al. [14] separated the bound from the frez fraction, after addition of a second antibody, directed against the immunoglobulins of the species corresponding to the used anti-HBAg antibody. The complex was then precipitated by ultracentrifugation. COLLAR et al. [lo] added ammonium sulphate to the incubation mixture. AMOUCH et al. [4] diluted the incubation mixture after incubation to enhance the precipitation of the antigen-antibody complex. In general in a radioimmunoassay separation between the bound and the free fractions is greatly facilitated if the antibody or the antigen is made insoluble, e.g. by adsorption to a tube [9, 291 or by covalent linking to an insoluble carrier [29]. LINGand OVERBY [18] developed a sandwich radioimmunoassay in which the antibody was made insoluble by adsorption to a plastic tube and in which they used radioiodinated antibody as a tracer. AALBERSE et al. [2] described the possibility of the simultaneous detection of and distinction between IgA and anti-IgA and between IgE and anti-IgE antibodies, using a mixed suspension of insoluble antigen and insoluble antibody. Because of the facility of the separation and the speed of the determination, we decided to use a solid phase radioimmunoassay. In order to be able to detect both antigen and antibody, we chose a competitive assay. This paper describes the development and practical use of three different types of solid phase radioimmunoassays: (1) a competitive assay for the de-
New Solid Phase Radioimmunoassay (CLB-RIA)
3
tection of HBAg and HBAb; (2) a sandwich assay to detect HBAb, and (3) a combination of 1 and 2 in which HBAg and HBAb can be detected and distinguished.
Materials and Methods All chemicals used were reagent grade. Bufers. Buffer I =phosphate-buffered saline (PBS) (0.140 M NaCI, 0.0113 M Na,HP04, 0.0017 M NaH,PO,, pH 7.2); buffer II=PBS diluted with an equal volume of distilled water, pH 7.5 ;buffer 111=PBS containing 0.001 M EDTA (Nh) and 0.01% NaN,, adjusted to pH 7.8; buffer IV=buffer 111 containing 2.5 mg/ml bovine serum albumin; buffer V = buffer IV to which 0.1% Tween-20 was added; buffer VI = 0.05 M borax, adjusted with HCI to pH 8.6; buffer VII=O.O5 M acetic acid in saline (0.1548 M NaCI) adjusted to pH 4.5 with NaOH. Antisera. All antisera used were prepared at the Central Laboratory of the Netherlands Red Cross Blood Transfusion Service, Amsterdam. Purification Procedures Sheep HBAb. After dialysis of the sheep anti-HBAg antiserum against buffer 11, a partially purified IgG-fraction was obtained by column chromatography on Sephadex DEAE A-50 (Pharmacia, 5 x 40 cm) equilibrated with buffer 11, which buffer was also used for the elution. The fractions containing anti-HBAg antibody in the double AGD test, were pooled, concentrated by ultrafiltration and dialyzed against buffer VI. Human HBAb. A 16% immunoglobulin concentrate was prepared from pooled human plasma, according to a modified Cohn VI fractionation method [ l q . Before pooling, all plasmas were tested and found positive for HBAb in the double AGD technique. HBAg. Hepatitis-B antigen, purified by adsorption on Aerosil and gelfiltration on Sephadex G-200 as described previously [ l l ] was highly purified by isopycnic banding in a CsCl density gradient. Before centrifugation, the density (measured by refraction at 25 "C) of the HBAg-solution was brought to 1.20 g/ml by addition of solid CsCI. Centrifugation at 50,000 rpm was performed for 40 h at 4°C (rotor 65, Beckman ultracentrifuge). Labelling Procedure The highly purified HBAg was labelled with lZ6I according to HUNTER and GREENWOOD [15]. The reaction mixture contained 0.25 ml HBAg solution (E;km=LlO), 0.15 ml chloramine-T solution (4 mg/ml) and 0.01 ml KlP6Isolution (1 mCi la61). The reaction was stopped after 30 sec by the addition of 0.8 mg sodiummetabisulphite in 0.20 ml buffer I and of 0.15 ml KI solution (50 mg/ml). After dialysis against three charges of 500 ml buffer 111 to which 0.01 mg/ml KI had been added, the labelled HBAg was separated from the last reaction reagents and byproducts by gel filtration on a Sephadex G-200 column (Pharmacia, 2.4 x 25 cm), equilibrated and eluted at 4°C with buffer IV. Preparation of Insolubilized HBAg and HBAb Sheep HBAb. The partially purified antibodies against Hepatitis-B antigen after dialysis against buffer VI were linked covalently to an insoluble carrier, i.e. voluminous m-diazo-
DUIMEL/BRUMMELHUIS
4
benzyloxymethylene cellulose (0.83 mg N/g amino cellulose) according to a modification described by BRUMMELHUIS [8] of the original method of GURVICH et al. [12]. For the coupling of sheep HBAb, 15 m1 protein solution (E:Bcb"=23.7) was mixed with 15 ml (20 mg/ml) diazotized voluminous amino cellulose. The reaction was allowed to proceed for 24 h at 4°C under continuous stirring. To inactivate the diazo groups, which did not react, the suspension was then stirred for 4 h at room temperature. The insolubilized antibody preparation ((HBAb)-SP) was washed twice with saline, twice with buffer VII, once with buffer IV and twice with buffer V. Finally the (HBAb)-SP was suspended in buffer V (0.75 mg cellulose/ml) and stored at 4°C. It could be shown that about 2.5 mg protein were bound to 10 mg cellulose. Human HBAb. For the coupling of human HBAb to the cellulose carrier, after dialysis against buffer VI, 0.8 ml of the 16% immunoglobulin concentrate was incubated with 50 mg cellulose (total volume 3.6 ml). The incubation, the washing and the storage were performed as described above for sheep HBAb. Approximately 12mg protein were bound to 10 mg cellulose. HBAg. The coupling of the highly purified HBAg to voluminous m-diazobenzyloxymethylene cellulose was performed in the same way as the coupling of HBAb. However, because of the lower protein concentration, 12.6 ml HBAg solution (Elirn=0.26) were incubated with 30 mg diazotized voluminous amino cellulose (total reaction volume : 15 ml). About 0.1mg HBAg % = 3.726; VYASet al. [26]) was bound to 10mg cellulose. The stock suspension of these insoluble antigens ((HBAg)-SP) was 0.5 mg cellulose/ml in buffer V.
@!d
Assay Procedures Double AGD was performed according to BRUMMELHLIIS [A.Counter immunoelectrophoresis was performed according to ALTERet al. [3]. The coated tube radioimmunoassay (AUSRIA-125, Abbott Laboratories), was performed according to the instructions of the manufacturer. Procedure of the CLB-Radioimmunoassays CLB-RIA. A competitive solid phase radioimrnunoassay. In a polystyrene tube (10 x 40 mm) 0.1 ml of the sample, 0.2 ml of the tracer and 0.2 ml of the immunosorbent were incubated for about 20 h at room temperature. The radioactivity of the tracer ( T HBAg) was about 50,000 cpm/ml, counted in a well-type automated gamma-counter (Wallac GTL 300). The immunosorbent suspension was a mixture of (HBAb)-SP (0.03mg cellulose/ml) and of normal voluminous aminocellulose (0.3mg cellulose/ml) in buffer V. The amount of immunosorbent was chosen in such a way that after incubation for 96 h (=at equilibrium) 30% of the tracer had been bound. After an incubation time of 24 h about 15% of the tracer was bound; this implicates that the counting level of negative samples is about 1,500cpm. During incubation the tubes were rotated end over end. After incubation, to separate the bound (laSI-HBAgbound to (HBAb)-SP) from the free fraction of lasI-HBAg, the tubes were centrifuged for 2 min at 3,OOOg. The supernatant was discarded, the remaining immunosorbent was washed six times with 1.2 ml PBS for each washing. Thereafter, radioactivity of the bound fraction was counted. CLB-A URIA. A sandwich solid phase radioimmunoassay and CLB-MIRIA, a combination of CLB-RIA and CLB-AURIA. With the exception of the immunosorbents the performances of the CLB-AURIA and the CLB-MIRIA were identic to the CLB-RIA
New Solid Phase Radioimmunoassay (CLB-RIA)
5
performance. In the CLB-AURIA the immunosorbent consisted of 0.2 ml of a mixture of (HBAb)-SP and (HBAg)-SP (0.03, respectively, 0.5 mg cellulose/ml). Samples were regarded as positive in the CLB-RIA and the CLB-MIRIA if the amount of bound tracer differed more than 15% of that of the mean of eight negative control sera. (15% was about 5 standard deviations of the mean of the control sera.) Samples which differed 10-15% were called dubiously positive. In the CLB-AURIA, samples were regarded as positive if the amount of bound tracer was at least twice the amount of the control sera mean; between 1.5 times and twice this mean the sample was called dubiously positive.
Results PuriJication and labelling of HBAg After isopycnic banding the HBAg-positive fractions were further examined. No human serum proteins could be detected in the AGD against horse anti-total human serum, nor against several specific antiserums (anti-IgG, anti-IgM, anti-a,M, anti$ lipoprotein and anti-transferrin). After labelling and dialysis, gelfiltration of this purified HBAg preparation on a Sephadex G-200 column resulted in three peaks containing radioactivity (fig. 1). Of the radioactivity in the first peak fractions (eluted at V,) 97% could be precipitated with trichloroacetic acid (TCA) (final TCA concentration 10%). The TCA-precipitable radioactivity in the fractions of the second and the third peak was 80 and 30%, respectively. The total amount of radioactivity in the fractions of the first peak was about 120,uCi (net yield of labelling procedure was 12%). A sample out of the pooled fractions 14-16 was examined for isopycnic homogenicity. More than 90% of the radioactivity were recovered in one single band with a density of 1.23 g/cmS(fig. 2). Radioimmuno Procedures CLB-RIA. In order to undo the influence of the protein concentration of the sample to be tested, it was necessary to add Tween-20 to the incubation mixture. This addition of Tween-20 prevented adsorption of the tracer to the plastic tube. It was found that the adsorption to the tube depended strongly on the protein concentration in the mixture. Therefore the amount of ‘bound’ tracer counted was about 80% more in saline than in serum. The difference in ‘bound’ tracer was almost entirely due to adsorption to the tube. It was found that this adsorption could be prevented by the addition of Tween-20 (0.02-0.04%). This had no influence on the amount of the bound radioactivity in serum (protein concentration in reaction mixture about 14 mglml). It could be shown that after the addition of Tween-20 the sensi-
6
a45ml
Fig. I . Elution pattern of 149-labelledHBAg. Column: Sephadex G-200. For details
see text.
Fig. 2. Ultracentrifugation in CsCl density gradient of purified 0=Radioactivity; A =density. For details see text.
HBAg.
tivity of the test was independent of the protein concentration, which means that comparable dose-response curves were obtained with serum dilutions and with buffer (PBS) dilutions of HBAg-positive sera (table I). The influence of the incubation time on the sensitivity of the assay was studied. Using two different amounts of tracer, the B/Bo was calculated as a function of the incubation time (B/Bo is the ratio of the amounts of bound tracer between a HBAg- positive and a HBAg-negative sample). The sensi-
New Solid Phase Radioimmunoassay (CLB-RIA)
7
Table I. Titers obtained by testing twofold serial dilutions of 12 HBAg-positive serum samples in the CLB-RIA No. sample
Titer when diluted with PBS
Titer when diluted with HBAgand HBAb- negative serum
1 2 3 4 5 6 7 8 9 10 11 12
128 1,024 128 1,024 512 512 16 32 8 8 1 2
256 1,024 128 2,048 256 1,024 16 128 32 16 2 4
Table ZZ. CLB-RIA. Effect of the incubation time and the amount of tracer on the ratios B Bo-B -and Bo S Time, h
B Bo
2 8 20
T = 16,000 cpm/ml
T = 5,500 cpm/ml
0.27 0.16 0.13
Bo-B S
12 21 29
B Bo
0.28 0.16 0.13
Bo-B S
20 36 44
tivity increased with the increase of the incubation time (table 11). At this level there was no influence of the amount of tracer on B/Bo. The influence of the amount of tracer became manifest when the parameter (Bo-B)/s was used instead of B/Bo (table 11). (Bo-B)/s is the ratio between the difference of the amounts of bound tracer of a HBAg-negative and a HBAg-positive sample and the statistical counting error (s) at the Bo level. The ratio (Bo-B)/s gives an impression of
DUINIELlBRUMMELHUIS
8
Fig. 3. CLB-RIA dose-response curve of HBAg (correlation coefficient=0.989) (RIA). For details see text.
60-
5
10
50
100
200
500
-+ dllutlons
Fig. 4. Exact dilutions in negative serum of a HBAg- ( 0 )and a HBAb- ( A ) positive serum tested in the CLB-RIA (IS = HBAb-SP). For details see text.
the significance of the difference between Bo and B, with the higher amount of tracer this significance was greater. Besides the minimal amount of HBAg that gave a positive reaction was determined. In figure 3 the dose-response curve is shown. The smallest amount of antigen that gave a significant decrease in B, was about 2-4 ng (HBAg concentration in the sample: 20-40 nglml). Not only HBAg, also HBAb-positive samples gave a decreased amount
New Solid Phase Radioimmunoassay (CLB-RIA)
5
9
10
Fig. 5. Exact dilutions in negative serum of a HBAb-positive serum tested in the CLBAURIA (IS = HBAg-SP). For details see text.
510-
1B x100% '0
210 -
---;-1
-
-&tion 5
10
50
100
200
500
Fig. 6. Exact dilutions in negative serum of a HBAg- ( 0) and a HBAb-positive ( A )
serum tested in the CLB-MIRIA.
HBAg SP HBAb SP
. For details see text.
of bound tracer. It is therefore impossible to distinguish between HBAgand HBAb-positive samples in this radioimmunoassay performance (fig. 4). CLB-A URIA. Using insolubilized HBAg as an immunosorbent instead of insolubilized HBAb, it was possible specifically to test it for the presence of HBAb (fig. 5). CLB-MIRIA. When a mixed suspension of insolubilized HBAg and insolubilized HBAb was used as an immunosorbent, it was possible not only
DUIMEL/BRUMMELHUIS
10
to detect HBAg, but also to distinguish between HBAg- and HBAb-positive samples. As in the CLB-RIA, a HBAg-positive sample gave a decrease and as in the CLB-AURIA a HBAb-positive sample gave an increase of the amounts of bound tracer (fig. 6). Specijicity The specificity of samples which were found positive for HBAg in the CLB-RIA was confirmed in an assay system in which human antibody to HBAg was used instead of sheep antibody. The assay itself was performed in the same way as the CLB-RIA using sheep antibody. The results of tests for specificity are summarized in table 111. The 16,330 samples were tested in pools, composed of 10 donations each. Of these 1,633 tested pools, 43 were found positive, or dubiously positive, for HBAg. To identify the samples which were responsible for the positive reactions of the pools, all 430 samples of these 43 pools were tested separately. This resulted in 34 positive samples and in 9 dubiously positive for HBAg. Of the 34 samples found HBAgpositive in the CLB-RIA, 4 samples were negative and 30 were positive when tested for specificity. When tested in CEP, only 11 of these 34 samples were CEP-positive, 23 were CEP-negative. The 4 CLB-RIA nonspecific positive samples were all negative in CEP. Among the 9 samples found dubiously positive for HBAg in the CLB-RIA, only 2 were positive and 7 were negative when tested for specificity in the assay in which human HBAb is used. In the CEP all 9 samples were negative.
Table III. Results of testing the specificity of the 43 samples which were found HBAgpositive or dubiously positive in screening 16,330 blood donations with the CLB-RIA No samples
CLB-RIA Sheep HBAb
-
7
+
2
9 4 19 11
-
34
CLB-RIA human HBAb
+ + +
+ 4-
CEP
New Solid Phase Radioimmunoassay (CLB-RIA)
11
Discussion
For this radioimmunoassay of HBAg it is essential to use highly purified antigen, since the presence of impurities may decrease the sensitivityand may cause false-positive results. Therefore, it was necessary to extend the previously described purification procedure [161. After labelling and dialysis of the highly purified HBAg, 75% of the radioactivity were eluted in the first peak fractions from a Sephadex G-200 column. The maximum binding of the tracer to insolubilized HBAb was about 80%, measured directly after gel filtration. By ageing, this percentage decreased to about 60%. The binding could be inhibited to less than 5% by the addition of HBAg-positive sera. The binding of the tracer to insolubilized negative serum was about 4%. Approximately 5% of the tracer were bound by the insolubilized IgG-fraction of a horse anti-human total antiserum. From these data we concluded that the binding of labelled HBAg is specific, whereas aspecific binding of the tracer to the immunosorbent hardly occurs. The addition of Tween-20 to the incubation mixture permitted the assay of HBAg and HBAb in samples independent of the protein concentration, at least if the protein concentration in the incubation mixture was less than 3%. This is in contrast to other sensitive tests such as those based on haemagglutination reactions [21]. Since the sensitivity of the assay was found to be time dependent, it was necessary to incubate for 18-20 h. Although longer incubation times did increase the amount of bound tracer, the sensitivity had not increased significantly. Therefore, a test could be performed within 24 h at a tracer concentration of about 20.000 cpm/ml. Although the (Bo-B)/s was not optimal at this tracer concentration (its optimum was at 80,000 cpmlml) we chose this concentration because the ratio B/Bo was really optimal. The reproducibility of the test was very good. The standard deviation of the amounts of bound tracer of the controls was rarely more than 3-4%, at a counting level of about 1,500 cpm. False-positive results may occur, due to interactions which disturb the reaction between HBAg and the insolubilized HBAg. These interactions may be caused by the presence of proteins which react, or cross-react, with the insolubilized sheep antibodies. PRINCE et al. [20] and VYASet al. [27] considered samples truly positive only if their reactivity could be specifically inhibited. This inhibition was carried out with human HBAb. In terms of competitive radioimmunoassays this means, that samples are truly positive only if they react positively with both human and
12
DUIMEL/BRUMMELHUIS
heterologous HBAb. The results of the test for specificity of the samples found HBAg-positive in the CLB-RIA were very good. The percentage of 85 real positives is high in comparison with the data of PRINCE et al. [20] and VYASet al. [27], who reported 21 and 8%, respectively, using a sandwich solid phase radioimmunoassay. The disagreements are probably caused by the fact that these investigators used a sandwich technique, whereas we used a competitive technique. False-positive reactions in the sandwich solid phase radioimmunoassay may be caused, according to PRINCC et al. [20] by binding (sandwich formation) of (cross-reacting) anti-guinea pig globulin antibodies with the insolubilized and with the labelled guinea pig anti-HBAg antibodies. All insolubilized guinea pig antibodies may cause false-positive reactions. In the CLB-RIA this phenomenon can not occur since in this radioimmunoassay the tracer consists of labelled antigen instead of labelled antibody. Falsepositive reactions in the CLB-RIA can be caused by binding of (cross-reacting) anti-sheep globulin antibodies, causing inhibition of tracer binding by steric hindrance. Because only a small part (less than 1%) of the insolubilized sheep immunoglobulin is directed against HBAg, only a small part of the (cross) reactions will cause steric hindrance with regard to the HBAb binding sites and thus cause false-positive results. For this reasons, in our opinion the CLB competitive solid phase radioimmunoassay may be more specific than a sandwich solid phase radioimmunoassay. Although the mechanisms are quite different, it was impossible to distinguish between an antigen and an antibody positive sample in the CLBRIA. When antigen is present there is a competition between the tracer and the antigen with respect to the insolubilized antibody. When the sample is antibody positive, there is a competition between the insolubilized and free antibody with respect to the tracer. Part of the tracer will react with the free antibody and form soluble complexes which are washed out after the incubation. Both competitions result in a decrease of tracer binding to the insolubilized antibodies. Addition of insolubilized antigen together with insolubilized antibody (CLB-MIRIA) permitted to detect and to distinguish between HBAg and HBAb. The sensitivity for detecting HBAg was the same in the CLB-RIA and CLB-MIRIA, in contrast to the sensitivity for detecting HBAb which was slightly less in the CLB-MIRIA. In the CLB-AURIA, the sensitivity for detecting antibody was the same as in the CLB-MIRIA. This means that with our methods the sensitivity for detecting HBAb, with labelled HBAg as a tracer, is better in a competitive assay with insolubilized antibody than in a sandwich assay with insolubilized
New Solid Phase Radioimmunoassay (CLB-RIA)
13
antigen. By screening samples with the CLB-RIA test and retesting the positive samples in the CLB-MIRIA test, even samples weakly positive for HBAg and HBAb can be detected. Samples will be found positive for HBAg in the CLB-RIA and the CLB-MIRIA at HBAg concentrations in the actual samples as low as 20-40 ng/ml. References 1 AACH,R. D.; GRISHAM, J. W., and PARKER,C. W.: Detection of Australia antigen by radioimmunoassay. Proc. nat. Acad. Sci., Wash. 68: 1056 (1971). R. C.; LOCHEMVAN, ERNA;MUNSTER VAN, P. J. J., and NADORP, J. H. J. : 2 AALBERSE, Human antibodies to IgE: a simple method to avoid ambiguous results of the radioimmunosorbent test (RIST). J. Lab. clin. Med. 83: 831 (1974). R. H.: Counterelectrophoresis for 3 ALTER,H. J.; HOLLAND,P. V., and PURCELL, detection of hepatitis-associated antigen: methodology and comparison with gel diffusion and complement fixation. J. Lab. clin. Med. 77: 1000 (1971). P. J. et DROUHET, J. : Dosage radioimmunologique de l'antigen Australie. 4 AMOUCH, Nouv. Rev. franc. HCmat. 11: 656 (1971). 5 BLUMBERG, B. S.: Polymorphisms of the serum proteins and the development of isoprecipiuns in transfused patients. Bull. N.Y. Acad. Med. 40: 377 (1964). 6 BLUMBERG, B. S.; SUTNICK,A. I.; LONDON, W. T., and MILLMAN, I.: Current concepts. Australia antigen and hepatitis. New Engl. J. Med. 283: 349 (1970). H. G. J. : Technical aspects and results of the determination of hepatitis 7 BRUMMELHUIS, associated antigen (HAA) and antibody. Vox Sang. 19: 228 (1970). 8 BRUMMELHUIS, H. G. J. :Het isoleren van specifieke antistoffen met behulp van immuun sorbentia; thesis Amsterdam. Hemstetstad Amsterdam. 9 CATT, K. J. and TREGEAR, G. W.: Solid phase radioimmunoassay in antibody-coated tubes. Science 158: 1570 (1967). 10 COLLAR, J. A. ; MILLMAN, I. ; HALBHERR, T. C., and BLUMBERG, B. S. : Radioimmunoprecipitation assay for Australia antigen antibody and antigen-antibody complexes. Proc. SOC.exp. Biol. Med. 138: 249 (1971). 11 DUIMEL,W. J.; BRUMMELHUIS, H. G. J., and KRIJNEN,H. W.: The purification of hepatitis-associated antigen (HAA/SH) from human serum and the preparation of a sheep anti-HAA serum. Vox Sang. 23: 249 (1972). 0. B., and TUMANOVA, A. E.: Production of protein12 GURVICH,A. E.; KUZOVLEVA, cellulose complexes (immunosorbents) in the form of suspensions able to bind great amounts of antibodies. Translated from Biokhimiya 26: 934 (1961). 13 GUST,I. D.: Evaluation of the techniques used for the detection of Australia antigen and antibody. Pathology 3: 151 (1971). F. B.; VORNDAM, V., and DREESMAN, G. R.: Assay of Australia antigen 14 HOLLINGER, and antibody employing double-antibody and solid-phase radioimmunoassay techniques and the comparison with the passive haemagglutination methods. J.Immunol.107: 1099 (1971). 15 HUNTER,W. M. and GREENWOOD, F. C.: Preparation of iodine-131 labeled human growth hormone of high specific activity. Nature, Lond. 194: 495 (1962).
14
DUIMEL/BRUMMELHUIS
16 KAPP, J. F.: Sensitive Australia antigen complement fixation test (CFT) and its comparison with Ouchterlony and countercurrent electrophoresis techniques. Med. Microbiol. Immunol. 157: 58 (1971). H. G. J., and BEENTJES,S. P.: Production of specific 17 KRIJNEN,H. W.; BRUMMELHUIS, antibody concentrates; in MERLERImmunoglobulins, p. 332 (National Academy of Science, Washington D.C. 1970). 18 LING,C. M. and OVERBY,L. R.: Prevalence of hepatitis B virus antigen as revealed by direct radioimmunoassay with (lza1) antibody. J. Immunol. 109: 834 (1972). 19 PRINCE,A. M.: An antigen detected in the blood during the incubation period of serum hepatitis. Proc. nat. Acad. Sci., Wash. 60: 814 (1968). A. M.; BROTMAN, B.; JASS, D., and IKRAM, H.: Specificity of the direct solid 20 PRINCE, phase radioimmunoassay for the detection of hepatitis B antigen. Lancet i: 1346 (1973). H. G. J.: Evaluation of a new 21 REESINK,H. W.; DUIMEL,W. J., and BRUMMELHUIS, haemagglutination technique for the demonstration of hepatitis B antigen. Lancet ii: 1351 (1973). J. P. ; CHICOT,D., et SALMON,C. : Comparative study of four 22 ROPARS,C. ; CARTRON, techniques for the detection of Australia antigen and a dextran sulfate test. Nouv. Rev. franc. Hemat. 12: 217 (1972). 23 SCHAFFER, W. E.; VYAS,G. N.; SHAKED, A.; CHEN,E., and PERKINS,H. A.: Comparison of counterelectrophoresis and haemagglutination inhibition test for hepatitis associated antigen. Vox Sang. 22: 366 (1972). E. H. : Relative sensetivity of gel diffusion 24 SCHMIDT,N. J. ;GEE,G. S., and LENNETTE, complement fixation and immunoelectroosmophoresistests for detection of hepatitis associated antigen and antibody. Appl. Microbiol. 22: 165 (1971). 25 VYAS,G. N. and SCHULMAN, N. R.: Haemagglutination assay for antigen and antibody associated with viral hepatitis. Science 170: 332 (1970). E. W. ;KLAUS,G. G. B., and BOND, H. E. :Hepatitis associated 26 VYAS,G. N. ;WILLIAMS, Australia antigen-protein, peptides and amino acid composition of purified antigen with its use in determining sensitivity of the haemagglutination test. J. Immunol. 180: 1114 (1972). 27 VYAS,G. N. : Personal communication. 28 WALSH,J. H.; YALOW, R. S., and BERSON, S. A.: Detection of Australia antigen and antibody by means of radioimmunoassay techniques. J. infect. Dis. 121: 550 (1970). Radio29 WIDE, L.: Solid phase antigen-antibody systems; in KIRKHAMand HUNTER immunoassay methods, p. 405 (Churchill/Livingstone, London, Edinburgh 1971).
Dr. Jr. W. J. DUIMEL,Centraal Laboratorium van de Bloedtransfusiedienst, Plesmanlaan 125, Amsterdam- W (The Netherlands)
