Methods for reducing non-specific antibody binding in enzyme-linked immunosorbent assays

Enzyme-linked immunosorbent assays were used to study (i) binding of rabbit antibodies (raised against litter mate liver plasma membrane fraction) to the immunizing membrane fraction, and (ii) binding of human antibodies to liver membrane fractions and to liver-specific lipoprotein (a liver membrane-derived antigen complex). When assays were conducted using the non-ionic detergent Tween 20 as blocking agent, high non-specific binding was encountered. With the low titre rabbit antisera high binding of non-immune test antibody and of second antibody (anti-rabbit IgG) to the immunogen, and also directly to the solid phase, was found. This was abolished by replacement of Tween 20 in the antibody diluent buffers by a non-reactive protein, casein proving to be a more effective blocking agent than either bovine serum albumin or gelatin. With human sera, high binding of human IgG to the solid phase was noted. This too was blocked by casein, but only when the anti-microbial agent Thimerosal was included in the casein buffer, and when Tween 20 in the wash buffer was replaced by casein-Thimerosal so that the solid phase was exposed to casein before incubation with the test serum. The casein buffers described may prove of general value in solid-phase assays where high non-specific binding is encountered.     

Detection of vitronectin by ligand blotting with type 1 plasminogen activator inhibitor

Ligand blotting procedures were developed for the detection of type 1 plasminogen activator inhibitor (PAI-1) binding protein(s) (BPs) transferred to nitrocellulose sheets after sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Purified vitronectin (Vn), a well characterised PAI-1 BP, was employed to optimise this assay system. After blocking with casein, the sheets were washed and then incubated with either ¹²⁵I-labelled PAI-1 (direct assay), or with unlabelled PAI-1 followed by a polyclonal antiserum to PAI-1 (indirect assay). In the latter case, the bound antibody was detected by using an ¹²⁵I-labelled second antibody. Binding was dose-dependent with respect to both Vn and PAI-1, and only active PAI-1 bound to Vn (i.e. latent PAI-1 and PAI-1 in complex with tissue-type plasminogen activator (t-PA), did not bind to Vn in this system). Moreover, t-PA, arginine and acidic conditions dissociated PAI-1 from Vn adsorbed to nitrocellulose. Analysis of bovine plasma by these techniques revealed the presence of a single PAI-I BP, and this protein was recognised by antisera to Vn. These results indicate that Vn previously fractionated by SDS-PAGE and transferred to nitrocellulose, continues to bind to PAI-I in a manner that resembles its behaviour in plasma and on extracellular matrix. These ligand blotting procedures may thus represent useful new approaches for the detection of other SDS-stable PAI-1 binding proteins in biological samples.     

Factors influencing the quantitation of antibodies to influenza virus by indirect solid-phase radioimmunoassay

An indirect solid-phase radioimmunoassay involving purified influenza haemagglutinin bound to polyvinyl microtitre plates has been used to quantitate the humoral response to influenza infection in ferrets. Sialic acid residues, present on normal ferret immunoglobulin, caused non-specific binding of immunoglobulin to the haemagglutinin, resulting in a non-linear relationship between serum dilution and primary antibody bound to antigen. Direct proportionality was achieved by removal of sialic acid residues from serum immunoglobulin or by inclusion of free sialic acid in the incubation buffer. Absolute quantitation of IgG and IgM antibodies was achieved by determining the relationship between primary antibody bound to antigen and secondary anti-immunoglobulin antibodies using ¹³¹I-IgG or -IgM and the homologous ²⁵I-labelled monospecific antiserum. Examples of antibody quantitations in sequential post-infection sera are presented.     

The Non-Specific Binding of Immunoglobulins to Silicone Implant Materials: The Lack of a Detectable Silicone Specific Antibody

Recent studies have suggested that anti-silicone antibodies develop in patients implanted with silicone materials. The majority of these studies have utilized enzyme-linked immunosorbent assay (ELISA) methodology with a silicone material substrate as a means to detect the presence of the anti-silicone antibody. The current studies were undertaken to determine whether the binding of IgG to a silicone substrate was consistent with an antigen-specific antibody interaction or the result of non-specific hydrophobic interactions. While significant differences were detected in serum from silicone antibody “positive” and “negative’’ patients when the ELISA was conducted using a phosphate buffered saline (PBS)-0.05% Tween 20 (Tween) blocking system, the difference in the responses was attenuated when protein blocking systems were used or when incubation times were decreased. Furthermore, ELISA studies, using purified mouse and human IgG, demonstrated a concentration-dependent binding of IgG to silicone elastomer substrate which was also attenuated when a protein blocking system was used in lieu of Tween. In controlled animals studies in which female B6C3F1 mice were implanted with silicone gel or silicone elastomer for 180 days, no difference was observed between the implanted animals and the PBS control animals with respect to binding of IgG to the silicone substrate. Similar studies in female Fischer 344 rats implanted with silicone gel for 84 days also failed to demonstrate the presence of anti-silicone antibody. Collectively, the results suggest that the binding of IgG to silicone implant materials is non-specific in nature, consistent with the well-recognized interactions between hydrophobic molecules (IgGs) and hydrophobic surfaces (silicones) in an aqueous-based system.     

The Non-Specific Binding of Immunoglobulins to Silicone Implant Materials: The Lack of a Detectable Silicone Specific Antibody

Recent studies have suggested that anti-silicone antibodies develop in patients implanted with silicone materials. The majority of these studies have utilized enzyme-linked immunosorbent assay (ELISA) methodology with a silicone material substrate as a means to detect the presence of the anti-silicone antibody. The current studies were undertaken to determine whether the binding of IgG to a silicone substrate was consistent with an antigen-specific antibody interaction or the result of non-specific hydrophobic interactions. While significant differences were detected in serum from silicone antibody “positive” and “negative' patients when the ELISA was conducted using a phosphate buffered saline (PBS)-0.05% Tween 20 (Tween) blocking system, the difference in the responses was attenuated when protein blocking systems were used or when incubation times were decreased. Furthermore, ELISA studies, using purified mouse and human IgG, demonstrated a concentration-dependent binding of IgG to silicone elastomer substrate which was also attenuated when a protein blocking system was used in lieu of Tween. In controlled animals studies in which female B6C3F1 mice were implanted with silicone gel or silicone elastomer for 180 days, no difference was observed between the implanted animals and the PBS control animals with respect to binding of IgG to the silicone substrate. Similar studies in female Fischer 344 rats implanted with silicone gel for 84 days also failed to demonstrate the presence of anti-silicone antibody. Collectively, the results suggest that the binding of IgG to silicone implant materials is non-specific in nature, consistent with the well-recognized interactions between hydrophobic molecules (IgGs) and hydrophobic surfaces (silicones) in an aqueous-based system.     

Protein-binding of secretin in human plasma

Protein-binding of endogenous plasma secretin and of 125I-labelled secretin incubated with charcoal-treated plasma examined by gel filtration on a Sephacryl S-200 Superfine column (16 X 980 mm) showed that secretin in plasma appears both to be bound to at least two different plasma proteins where albumin appears to be the major binding protein, and also to occur as a free molecular form. In addition, protein-binding studied by incubating 125I-labelled secretin with charcoal-treated plasma under various conditions followed by charcoal separation of bound from free label indicated the presence of more specific secretin-binding sites on the plasma proteins with an avidity comparable to that otherwise reported for albumin as a binding protein. The protein-binding of 125I-labelled secretin was optimal or reached equilibrium after 2 days incubation at 20 degrees C and first after 8 days incubation at 4 degrees C. Also, the protein-binding of 125I-labelled secretin was higher at an incubation temperature of 20 than of 4 degrees C; was optimal at pH 7.4; increased with increasing amounts of charcoal-treated plasma up to an amount of 800 microliters in our assay system before levelling off; and increased in a constant and predictable manner with increasing amounts of 125I-labelled secretin at least with the amounts of labelled secretin examined here.     

Enzyme-linked immunofiltration assay for rapid serodiagnosis of syphilis

A new rapid technique for detection of serum treponemal antibodies is described which is based on an enzyme-linked immunoassay using nitrocellulose as solid phase. With this technique antigen-antibody binding is accelerated by the filtration of the antibody solution through the antigen-coated nitrocellulose filter instead of its remaining over the solid phase for incubation. Test results are available in less than 15 min. Serum specimens from 255 syphilitics and 829 non-infected subjects were investigated. The sensitivity and specificity of the Treponema pallidum enzyme-linked immunofiltration assay were comparable to those of the Treponema pallidum haemagglutination assay and the fluorescent treponemal antibody absorption test.     

The use of Tween 20 alone as a blocking agent for immunoblotting can cause artefactual results

The use of Tween 20 as a suitable blocking agent in immunoblotting studies was evaluated by screening a panel of monoclonal antibodies (MoAbs) against a selection of blotted proteins which were unrelated to the antigens used to raise the MoAbs. Using Tween 20 alone to block the nitrocellulose membranes clear reactions were observed between the panel of MoAbs and several components of the blotted protein mixture. In contrast, when haemoglobin was used to block the membranes such reactions were not observed. In the absence of added protein the use of Tween 20 alone as a blocking agent for immunoblotting appears to lead to false positive reactions by non-specific antigen-antibody complexes.     

Line immunoassay and enzyme-linked line immunofiltration assay for simultaneous detection of antibody to two treponemal antigens

Two enzyme immunoassays, the line immunoassay (LIA) and the enzyme-linked line immunofiltration assay (ELLIFA), were studied for suitability in the serodiagnosis of syphilis. In both assays, antibody to treponemes was detected using the recombinant DNA derived treponemal protein TmpA and the purified axial filament derived from the Reiter treponeme. The antigens were applied in parallel lines onto nitrocellulose membranes. The sensitivity and specificity of both assays were compared with that of theTreponema pallidum hemagglutination assay (TPHA), the fluorescent treponemal antibody absorption test, and the axial filament and TmpA enzyme-linked immunosorbent assays. The sensitivity and specificity of the LIA and the ELLIFA were found to be comparable to that of the TPHA using serum samples from 65 untreated syphilitic patients, 95 patients treated for syphilis and 60 blood donors, except in the case of the LIA using axial filament. This latter test was slightly less sensitive in primary and early latent syphilis than the TPHA. In the LIA procedure, serum antibodies to two antigens could be detected simultaneously within two hours. This assay may be useful for fieldwork. In the ELLIFA procedure, antibodies to the two antigens could be detected simultaneously within 15 minutes. The ELLIFA procedure may provide a multiple antigen test with a very short assay operation time.     

Comparison of methods for the analysis of outer membrane antigens of Neisseria meningitidis by western blotting.

The method of extraction of outer membrane proteins (OMPs), the conditions of electrophoretic transfer, and the conditions of antibody binding, were compared in Western blotting studies of Neisseria meningitidis outer membrane antigens. The OMP profiles obtained by SDS-PAGE of outer membrane vesicles extracted with lithium chloride/sodium acetate were compared with profiles obtained by Sarkosyl extraction; these profiles were further compared with the patterns obtained by 125I-labelling of surface-exposed proteins. Sarkosyl extracts gave profiles most closely resembling those of 125I-labelled whole-cells and gave the best resolution of the major proteins. After transfer in Tris-glycine-methanol buffer some proteins, including the major proteins, were not completely transferred and remained in the gel, with the class 2/3 and 5 proteins not effectively detected on nitrocellulose by amido black staining. There was weak antibody recognition of the class 1 and 4 proteins but good recognition of lipooligosaccharide (LOS) and H8 antigen. Empigen BB had no effect on renaturation of the class 1 protein. When 0.1% SDS was incorporated in the same buffer all of the proteins were removed from the gel, and although the major proteins bound to nitrocellulose other proteins did not. There was weak antibody recognition of the class 1 and 4 proteins, stronger reaction to the class 5 protein, but no recognition of the class 2 protein, LOS or H8 antigen, Empigen BB slightly enhanced antibody recognition of the class 1 protein. After transfer in Tris-glycine buffer, all the major proteins were transferred and bound to nitrocellulose and, other than the class 2 protein, were recognised by antibody, both in the presence or absence of Empigen BB, as were LOS and the H8 antigen. Differences existed in the patterns of antibody recognition between the lithium and the Sarkosyl extracts; additional proteins were recognised in the lithium extracts. The surface-labelling studies indicated, however, that some of these proteins were not surface-exposed. Some minor proteins appeared to be more highly immunogenic than the major proteins.     

Radioallergosorbent test (RAST) for measurement of IgG antibodies to Aspergillus fumigatus in sera of patients with different lung diseases

A radioallergosorbent test (RAST) for measuring human anti-Aspergillus fumigatus (Af) antibodies of the IgG class is described. The use of 125I-labelled animal antibodies against human IgG is compared with the use of 125I-labelled protein A. Under optimal conditions the radioactivity binding ratio between pooled patients' serum and pooled healthy persons' serum is 8-11.5. The immunoblotting technique was used to investigate so-called non-specific binding. The results obtained show that most if not all human sera contain anti-Af antibodies of the IgG type. The difference between pathological and normal immunological response to Af antigens seems to be in the antibody titres rather than in the presence or absence of antibodies to these antigens.     

Analysis of serum antibody repertoires by isoelectric focusing and capillary blotting onto nitrocellulose paper

A rapid method for the analysis of antigen-specific human serum antibodies is described. After isoelectric focusing on thin agarose gels, serum proteins are transferred to nitrocellulose paper by capillary blotting. The kinetics of transfer indicated that ca. 90% of 125I-labelled isolated human immunoglobulins (IgG, IgA and IgM) are transferred to the nitrocellulose paper within 10 min. Probing of the nitrocellulose blot with 125I-labelled tetanus toxoid antigen followed by autoradiography reveals the total serum anti-tetanus toxoid antibody repertoire, including high molecular weight IgM. This procedure is rapid (results can be obtained in less than 12 h), sensitive and should prove very useful for detailed studies and analysis of antibody repertoires in body fluids and extracts     

Tween 20 removes antibodies and other proteins from nitrocellulose

It has generally been assumed that the binding of most proteins to nitrocellulose is stable in the non-ionic detergent Tween 20. However, the following paper demonstrates that in immunoassays where antibodies or other proteins are bound directly to the nitrocellulose, 0.05% Tween 20 may dissociate these proteins from the membrane. The degree of dissociation appears to be dependent on the individual protein studied. Some antibodies and other proteins bind tightly to nitrocellulose and dissociation of these proteins by Tween 20 is barely detectable. In contrast, other proteins are nearly completely stripped from the nitrocellulose by the same detergent. Therefore, unless it is known from control experiments what proteins will or will not be dissociated from nitrocellulose by Tween 20, the routine use of Tween 20 in the development of Western blots, native blots and dot-blots should be discontinued.     

Macrophage phagocytic recognition sites. Demonstration of selectivity by hetero- and alloantisera.

Separate and independent phagocytic recognition sites have been identified on mouse peritoneal macrophages through the use of xenogeneic antimacrophage serum (AMS) and allogeneic anti H-2 antisera. Anti H-2d and anti H-2b antisera inhibit the binding and ingestion of opsonized erythrocytes (EA) by macrophages bearing H-2 haplotypes d and b, respectively. AMS and its F(ab')2 and Fab fragments inhibit the binding and ingestion of EA, and the ingestion but not the binding of 125I-labelled Shigella by macrophages. Neither antiserum inhibited the binding or ingestion of latex particles by macrophages. The results suggest that particulate binding to macrophages can be inhibited by two different mechanisms: a non-specific one where antibody bound to certain cell-surface antigens can mediate either directly or indirectly, and a specific interaction with Fc receptors. The possible mechanisms of non-specific antibody mediated phagocytic inhibition are discussed.     

The use of tween 20 as a blocking agent in the immunological detection of proteins transferred to nitrocellulose membranes

The determination of the immunoreactivity of protein antigens in complex mixtures has been greatly facilitated by combining their separation via sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) with electrophoretic transfer to nitrocellulose membrane (NCM), and probing of bound proteins with specific antisera. Methods using various buffers and blocking agents have been published, but no studies have been published which compare these methods with each other or with others of potential merit. We have performed such a comparative study using protein antigens from Chlamydia trachomatis and Neisseria gonorrhoeae. In addition, we describe a method that blocks unoccupied protein binding sites on NCM with the nonionic detergent Tween 20, rather than proteins. This system proved to be equivalent or superior to other methods evaluated in the detection of immunoreactive proteins, and permitted staining of the NCM for protein after immunological probing. Such staining allowed precise identification of immunoreactive proteins. In addition, individual stained proteins could be excised and assessed for bound antibody in an indirect radioimmunoassay.     

A Modified Crossed Radioimmunoelectrophoretic Method for Determination of Allergen-Specific IgG Antibodies

A crossed radioimmunoelectrophoretic (CRIE) method for detection of specific IgG antibodies in patients' sera against horse hair and dander was developed. The unacceptably high non-specific binding encountered when substituting 125I-labelled antihuman IgG for 125I-labelled antihuman IgE in an ordinary CRIE was eliminated by the combined use of 125I-labelled Protein A as detector, and F(ab')2-fragments of the allergen-specific rabbit antibodies. The low background binding thus obtained makes the method useful for detection of specific IgG in sera where the ratio between specific and non-specific IgG is low. Therefore the method should also be applicable to other antigen/allergen systems.     

Charge heterogeneity of human secretory component: immunoglobulin and lectin binding studies.

Free secretory component purified from human milk showed considerable charge heterogeneity in the pH range 4.7-6.5 when analysed by isoelectric focusing in thin layer agarose gels. Unlike rabbit secretory component, allotypic variation could not be identified by comparing samples from different individuals. Treatment with neuraminidase resulted in a basic shift of secretory component charge isomers. The charge heterogeneity appeared to be unrelated to the immunoglobulin binding property of secretory component since all charge isomers bound 125I-labelled IgM. 125I-labelled secretory component bound more strongly to purified IgM compared with polymeric IgA, and predominantly to the IgM-containing region of focused normal human serum. Lectin-secretory component interaction was demonstrated by concanavalin A and wheat germ agglutinin binding in a dot-blot nitrocellulose assay and precipitation with concanavalin A by Ouchterlony immunodiffusion. Despite the relatively high carbohydrate composition of both secretory component and the polymeric immunoglobulins, no evidence for lectin-like binding was obtained by sugar inhibition and sugar desorption studies. Similarly, desialation of secretory component did not prevent secretory component-IgM binding. These observations suggest that the charge heterogeneity and sugar composition of secretory component are unrelated to immunoglobulin binding in vitro.     

A quantitative dot immunobinding assay for human HLA class II antigens using nitrocellulose membrane filters

A quantitative method for the evaluation of human HLA-DR antigen expression has been developed. Cell membrane proteins were solubilized in Nonidet P-40 or deoxycholic acid detergent and diluted in a Triton X-100 containing sample buffer. The samples were subsequently spotted on a nitrocellulose membrane filter and fixed by immersion in isopropyl alcohol-acetic acid solution. The membrane was saturated in a 5% BSA blocking buffer and sequentially incubated with specific monoclonal anti-HLA-DR antibody, and 125I-labelled protein A. Each spot was then assayed for radioactivity in a gamma scintillation counter. Immunoadsorbant purified HLA-DR antigen was used to standardize the method and a reference dosage curve was established with serial dilutions of the purified HLA-DR antigen. The method permitted the detection of HLA-DR antigens with reproducibility in the ng range, in cellular extracts, physiological and pathological fluids, and in fractions eluted from affinity columns.     

A sensitive radioimmunoassay for the detection of monoclonal anti-idiotope antibodies

A radioimmunoassay was developed in order to detect anti-idiotypic antibodies in the supernatants of hybrid cells. This assay is both sensitive and specific for anti-idiotypic (but not anti-allotypic) antibodies. Monoclonal antibodies present in test supernatants are bound by an anti-immunoglobulin coated solid phase. Subsequent incubation with a source of mouse immunoglobulin 'blocks' unreacted anti-immunoglobulin antibodies on the solid phase. Anti-idiotypic antibodies are then detected by their ability to bind 125I-labelled idiotype-bearing antibody. This paper describes the use of this assay to detect monoclonal anti-idiotypic antibodies in 2 systems: the cross-reactive idiotype of A/J anti-ABA antibodies, and the idiotype expressed by the myeloma protein HOPC 8. Similarly, 125I-labelled anti-idiotype antibodies may be used in this assay to detect monoclonal idiotype-bearing antibodies. Further modifications are described which would allow the detection of monoclonal anti-allotype antibodies.     

The use of 125I-labelled protein A for the detection of humoral immunity to gross murine leukaemia virus

A solid-phase radioimmunoassay utilising bind of 125I-labelled protein A to antibodies bound to virus adsorbed onto microtitre plates was shown to be suitable for detection of humoral immunity to Gross murine leukaemia virus (MuLV). The specificity of the reaction was shown by the fact that only homologous or closely related viruses effectively inhibited binding of antibodies to adsorbed virus. With this method a low level of spontaneous humoral immunity was demonstrated in sera from AKR/Crc mice, a strain with high concentrations of endogenous virus, whereas little or no anti-viral activity was found inCBA/H-T6Crc, a subline that does not appear to express MuLV.