Immunoassays

RP Ekins, Division of Molecular Endocrinology, University College London Medical School, Mortimer Street, London, UK

Immunoassays fall within the broader class of techniques - frequently described as 'binding assays' -that rely on observation of the binding reaction between the analyte and a specific binding substance, the latter being generally of biological origin and typically comprising a specific binding protein. Antibodies constitute an especially important class of protein in this context, as they can be generated conveniently and in large amount (either by conventional in vivo methods or, more recently, by the in vitro techniques pioneered by Milstein and Köhler)

against an extensive range of substances of biological interest, including high molecular weight substances (e.g. proteins, polypeptides, viral particles, oligonucleotides, etc.) as well as substances of relatively small molecular size not themselves intrinsically immunogenic ('haptens'), such as drugs, steroid and thyroid hormones, vitamins, etc. Naturally-occurring specific binding proteins nevertheless continue to be used in 'protein binding' assays, e.g. in assays for vitamin D and its metabolites, vitamin B12, folic acid and (occasionally) thyroxine, albeit the availability of specific antibodies against most of these substances has resulted in the progressive abandonment of the binding proteins as analytical reagents. More recently, binding assays essentially identical in principle to immunoassay, but relying on DNA 'probes' as binding reagents, have been introduced, although these are restricted in application to the detection and/or measurement of complementary nucleotide sequences, and have not as yet achieved the widespread usage or commercial importance of antibody-based assay methodologies.

The binding reaction between high concentrations of analyte and antibody may be observed 'directly' in consequence, for example, of the formation of a visible precipitate (comprising the analyte-antibody complex), whose amount is indicative of the amount of analyte in the test sample (assuming a sufficiency of antibody to be present). This approach formed the basis of traditional methods used to quantify both antigens and antibodies, such as the immuno double-diffusion and immunoelectrophoresis techniques. However, for greater sensitivity, either of the reac-tants may be labeled using a marker more readily detectable than the reactants or reaction products themselves, enabling observation of binding reactions between lower reactant concentrations. One of the original labels used for this purpose comprised red cells, to whose surface antigen was linked. The presence of antibodies in a test sample was detected by cell agglutination (arising from cross-linkage by the antibody), resulting in a characteristic layer of agglutinated cells at the bottom of the incubation tube or well. This technique was also used to detect specific antigens in test samples by their inhibitory (competitive) effect on the agglutination reaction between antibody and (identical) antigen on the red cell surface. Such techniques were semiquantitative in that they yielded a somewhat imprecise measure of the maximal dilution of test samples causing visible agglutination.

The emergence of far more precise and sensitive immunoassays followed the adoption of radioisotopic labeling techniques for the observation of antibody-antigen reactions. The original methods of this genre relied on the use of labeled antigens (used essentially as tracers of unlabeled antigen present in test samples), and are now commonly known as radioimmunoassays (RIAs). The term is occasionally applied to methods relying on isotopically-labeled antibodies; however, the term immunoradiometric assay (IRMA) is conventionally applied to methods based on the latter approach.

Immunoassays characteristically display high structural specificity, reflecting the ability of appropriately selected antibodies to distinguish and bind to antigenic sites (epitopes) on particular analyte molecules. The use of high specific activity isotopes such as ,25I or !H (or, in the case of DNA probes, 12P) to label either the antibody or (tracer) analyte permits binding reactions between small numbers of molecules to be readily monitored, endowing these methods with exceedingly high sensitivity. These attributes, coupled with their relative simplicity, underlie the widespread use of isotopically-based immunoassays in medicine and biology and increasingly in many other related fields (e.g. the food industry, agriculture, forensic science, environmental monitoring, etc.). Since the early 1980s, however, increasing interest has focused on so-called alternative, nonisotopic techniques based on principles identical to those of RIA and IRMA, albeit differing in the reagent labels used. Various environmental, legal, economic and practical considerations underlie this development (e.g. the limited shelf-life of isotopically-labeled reagents, the problems of radioactive waste disposal, the complexity of radioisotope counting equipment, the demand for diagnostic kits for home use, etc.), most of which are primarily of social or logistic significance. Other factors are more fundamental, relating to the development of methodologies of greater sensitivity than the isotopic methods or which address other analytical objectives unattainable using isotopic labels. Such objectives include the development of transducer-based im-munosensors for directly measuring analyte concentrations in biological fluids, and multianalyte assay systems capable of measuring multiple analytes in the same sample.

A wide range of nonisotopic labels have been employed in this context, of which the most commonly used are enzymes, and chemiluminescent and fluorescent substances. Regrettably, the diversity of different labels now in use has given rise to a confusing range of acronyms by which the differing assay formats developed by individual researchers or immunoassay kit manufacturers are known. The conventional terminology distinguishing the original, isotopically based methods is often used; for example, methods relying on fluorescent labels are referred to either as fluoroimmunoassay (labeled antigen) or immunofluorometric assay (labeled antibody), on enzyme labels as enzymoimmunoassay and immunoenzymometric assay, etc. However, other terms - some of which depart from this convention - are also in common use, including, for example, enzyme-linked immunosorbant assay (ELISA; generally signifying the use of an enzyme labeled antibody and solid-phase antigen), time-resolved fluoroimmunoassay (TR-FIA; based on time-resolution methods of fluorescence measure ment), and competitive enzyme-linked immunoassay (CELIA; relying on the use of enzyme-labeled antibodies in a competitive assay mode).

The terms 'competitive' and 'noncompetitive' frequently feature in descriptions of immunoassay and other binding assay methodologies, as exemplified by CELIA and competitive protein-binding assay (CPBA). These terms (and similar descriptions, such as 'inhibition' assay) derive from the historical portrayal of certain binding assay methods (e.g. RIA) as dependent on competition between unlabeled and radiolabeled (tracer) analyte for a limited number of antibody (or protein) binding sites. For this reason, labeled antibody methods are frequently viewed as inherently noncompetitive, the terms being used virtually synonymously. However, certain labeled antibody methods are also described as competitive (e.g. CELIA), evidently on the grounds that, as with labeled analyte methods, maximal sensitivity is achieved using restricted amounts of antibody (see below).

The underlying reasons for the existence of these two broad types of assay may be clarified by brief consideration of the basic principles governing immunoassay. All such assays essentially rely on the measurement of (fractional) antibody binding site occupancy by analyte following reaction between the two. However, binding site occupancy may be determined either by direct measurement of occupied sites (noncompetitive assay), or by indirect measurement, i.e. by the measurement of unoccupied sites, from which binding site occupancy is inferred (competitive assay). The optimal amount or concentration of antibody yielding maximal assay sensitivity largely depends on which of these alternative strategies is adopted, tending to infinity in the first case, and to zero in the second. This difference in assay design - which is essentially unrelated to which component, if any, of the assay system is labeled -underlies the distinction between, and differing performance characteristics of, noncompetitive and competitive methods. For example, introduction of labeled analyte into the assay system either following, or during, the reaction between antibody and analyte provides a method of estimation of unoccupied antibody binding sites, implying that all labeled analyte methods may be classed as competitive. In contrast, if labeled antibodies are employed, the assay may be classed as competitive or noncompetitive, depending on the approach used to separate and measure binding sites occupied by analyte. Thus, if an immunosorbant is used to sequester unoccupied sites, and labeled antibody attached to the immunosorbant is determined, the assay is competitive. Conversely, if labeled antibody remaining in solution (i.e. antibody bound to analyte) is measured, the assay is noncompetitive. Assuming an efficient method of separation of occupied from unoccupied sites, the optimal antibody concentrations yielding maximal sensitivity will significantly differ, depending on which of these alternative approaches is adopted.

Irrespective of whether (tracer) analyte or antibody is labeled (but assuming the use of a label of infinite specific activity), the maximal sensitivity attainable using a competitive assay may be shown to be e/K, where e represents the experimental error (cv) in the measurement of the assay response variable (e.g. fraction of labeled analyte bound to antibody), and K = the antibody affinity constant vis-a-vis analyte. For example, if experimental errors in the measurement of the assay response approximate 1% (i.e. e = 0.01), and K = 1012 1 mol ', the maximal sensitivity attainable using a competitive approach is about 10 14 mol 1 ', i.e. about 6 X 10" molecules ml The use of radioisotopic labels (such as the iodine isotope l2,I) of finite, albeit high, specific activity, does not, in practice, significantly reduce assay sensitivity (assuming use of an antibody of affinity constant about 1012 1 mol 1 or less). This implies that no significant sensitivity improvement can result from the substitution of other labels in immunoassays of competitive design. However, in the case of noncompetitive immunoassays, similar theoretical analysis reveals that (assuming efficient separation of occupied from unoccupied labeled antibody binding sites), sensitivities of about 102-1()! molecules ml ' are, in principle, attainable, albeit only using labels of much higher specific activity than the commonly used radioisotopes. These conclusions underlie the current development of a variety of ultrasensitive, noncompetitive, labeled antibody methods relying on, for example, lanthanide chelate fluorophors, chemiluminescent substances such as the acridinium esters, and various enzyme labels.

A disadvantage of the use of high antibody concentrations in noncompetitive assay designs is a concomitant loss in assay specificity, arising from the increased relative potency of substances other than the analyte as the antibody concentration is raised. (Note that all substances capable of reaction with an antibody become equipotent in an immunoassay when the amount of antibody is infinite.) This problem is minimized in so-called 'sandwich', or 'two-site' assays, relying on dual antibodies. The first, linked to a solid support - often termed the 'capture' antibody - is generally directed against one epitope characterizing the analyte molecule and is used to isolate molecules bearing the epitope from the medium. The second (labeled) antibody, directed against a second epitope, enables determination of the number of captured analyte molecules adhering to the solid support which carry both epitopes. Sandwich assays thus combine the sensitivity of the noncompetitive approach with the specificity improvement inherent in dual-epitope recognition, and -subsequent to the development of in vitro techniques of monoclonal antibody production - have emerged as the dominant methodology for the assay of ana-lytes of large molecular size. (Note, however, that isoforms of many analytes (e.g. erythropoietin) differ in both structure and biological activity, notwithstanding their possible possession of many identical epitopes. Thus, sandwich immunoassays cannot be assumed to be totally specific for analyte molecules of a single, unique, structure, and may therefore be analytically invalid. This phenomenon contributes to the problems of immunoassay standardization.)

For the two-site dual antibody approach to be successful, the analyte molecule must be of sufficient size to permit its simultaneous binding to two different antibodies without steric impediment. Attempts to develop sandwich assays for analytes of small molecular size (such as steroid hormones) by generating antibodies reacting exclusively with occupied capture-antibody binding sites have so far failed, largely because of the propensity of antibodies generated against occupied sites also to bind to unoccupied sites. However, a recently described technique relies on blocking of unoccupied sites by an anti-idiotypic antibody prior to exposure of capture antibody-ana-lyte complexes to labeled antibody. The latter - prohibited from binding to unoccupied sites - thus binds only to capture antibody-analyte complexes. The technique thus constitutes a noncompetitive assay for analytes of small molecular size, and is thus potentially capable of yielding greater sensitivity. However, it remains questionable whether the approach provides the specificity advantages of a genuine sandwich assay, and overcomes the loss in specificity that the use of large amounts of labeled antibody to maximize sensitivity implies.

A further reason for the adoption of nonisotopic labels is the necessity, in the case of the isotopically-based methods, for physical separation of the immunological reaction products, whichever is labeled. Certain nonisotopic labels permit the reaction products to be distinguished without such separation, methods based on their use being generally described as homogeneous. In practice, such methods have in the past been largely confined in their application to analytes of small molecular size, relying on a labeled form of the analyte, the signal from which is affected by its binding to antibody. Certain commercial assay kits rely on the use of enzyme labels, whose activity is either restricted or enhanced when the labeled analyte is antibody bound. Another tech nique relies on fluorescent labels, and exploits alterations in the plane of polarization of the fluorescent emissions when the labeled antigen molecule is in free or antibody-bound states. Such techniques are of relatively low sensitivity, and their use has been confined to the assay of analytes present in biological fluids at relatively high concentration, such as therapeutic drugs. Recently, however, homogeneous sandwich methods of high sensitivity for the assay of analytes of large molecular size have been developed by various manufacturers and are beginning to appear on the immunodiagnostic market. One such method relies on a time-resolved fluorescence detection system involving the transfer of laser excitation energy from one antibody (labeled with a fluorescent europium cryptate) to the second, the latter (labeled with the fluorophor allophycocyanine) forming the complementary part of the sandwich. Such energy transfer essentially only occurs when both antibodies are simultaneously bound to individual analyte molecules and are therefore in close proximity. Thus only analyte molecules simultaneously bound to the two antibodies generate a detectable signal.

Neither analyte nor antibody is labeled in a conventional sense in the case of the newer immuno-sensor technologies now under development. Such devices comprise a solid probe on whose surface antibody is coated; antibody occupancy following exposure to an analyte-containing medium is monitored by an internal transduction system, yielding an observable electronic or optical signal. A number of such transduction systems, such as field-effect transistors (FFTs), have been proposed, but most have proved either too low in sensitivity or vulnerable to nonspecific effects to be of practical use. A promising contender in this field relies on surface plasmon resonance (SPR), an electronic effect generated in an ultrathin conducting metal (e.g. gold or silver) film on glass (or quartz) when light impinging at a particular angle on the glass-air interface is internally reflected. The resonance angle depends on the mass of material present at the interface; it thus alters in response to changes in the occupancy of antibody (or antigen) molecules coated to the metallicized/glass surface. Instruments relying on these principles are now commercially available, albeit they are primarily intended for research studies (such as epitope mapping) rather than for routine immunoassay use. Although likely to prove of great value in certain situations, the transducer-based sensor approach to antibody-occupancy measurement possesses a number of potential drawbacks which makes it uncertain that it will substantially replace conventional label-based methodologies in the near future. For example, the specificity improvements offered by two-site sandwich assays (particularly in the context of high-sensitivity, noncompetitive assay designs) makes large-scale reversion to single-site immunoassays unlikely. Although sensor techniques such as SPR are capable of detecting second (i.e. sandwich-forming) antibodies bound to analyte molecules adhering (via capture antibody) to a glass surface (particularly if the second antibodies are labeled with, for example, colloid metal particles to increase their mass), the inclusion of second antibodies in an assay system narrows the distinction between sensor-based and conventional label-based methods of antibody-occupancy measurement. Thus the future impact of SPR and other transducer-based techniques is likely to depend largely on the reduction in operational costs, if any, that they offer as compared with current methodologies. Meanwhile, the slow rates of association and dissociation of antigen-antibody complexes imply that the ultimate goal of immunosensor development - the ability to monitor changing analyte concentrations in a medium - remains far from fulfillment.

Another development - that of multianalyte immunoassay systems - reflects an increasing need to measure many different analytes in the same sample, both in medicine and in many biologically-related fields, e.g. the food industry. The possibility of developing such technologies also emerges as a consequence of the availability of very high specific activity, nonisotopic, reagent labels, e.g. fluorescent labels. This permits the development of high sensitivity 'microspot' immunoassays in which a capture antibody is located on the surface of a probe within an area no more than 50-100 pm2. This in turn permits construction of multimicrospot arrays, each microspot directed against a different analyte. Provided the total number of capture antibodies located within each microspot is 'small' (i.e. about 0.05 "IK or less, where v = the volume of the analyte-contain-ing medium to which the microspot is exposed), fractional occupancy of capture antibody binding sites is solely dependent on the analyte concentration in the medium. Moreover, by labeling both capture antibody and a second 'recognition' antibody (directed against either occupied or unoccupied sites) with different labels, e.g. two fluorescent labels, the analyte concentration can be derived by observation of the ratio of signals emitted from the microspot when exposed, for example, to laser illumination. Thus, by utilization of laser scanning techniques comparable to those employed in other fields, arrays of many different antibody microspots may be rapidly examined, thus permitting the development of simple multianalyte immunoassay methodologies capable of measurement of many different analytes in the same small biological sample.

Microarray techniques of this kind (bur relying on oligonucleotide arrays rather than antibody arrays) are also applicable to DNA analysis; such development constitutes the objective of the US Genosensor Project, established in 1992.

Antibody microspot arrays may be constructed on a commercial scale using, for example, ink-jet spotting techniques. Researchers at Boehringer Mannheim GmbH have described the use of this approach to construct arrays comprising 196 microspots located on the flat circular base (3 mm in diameter) of a plastic sample container, these being produced at a rate of approximately 5000 chips an hour. A somewhat different approach has been adopted by the Californian company, Affymetrix Inc., for the construction of oligonucleotide arrays for use in DNA analysis. This relies on the use of so-called combinatorial chemical techniques to construct (in situ) the different oligonucleotide sequences within the individual spots that make up the oligonucleotide array. The technique involves sequential masking and exposure of selected areas on a silicon base using photolithographic techniques similar to those used in the electronics industry for the construction of semiconductor chips. Such technology -although applicable to binding agents such as oligonucleotides and polypeptides which can be built up sequentially - are obviously not applicable to antibodies, which cannot be synthesized in this fashion.

The current intense interest in microarray technologies being displayed both in Europe and in the USA, and in the development of both manufacturing techniques and appropriate instrumentation, suggests that antibody microspot arrays will form the basis of the next major revolution in the field of immunoassay. But whether or not this prediction is fulfilled, binding assay methodology of one form or another is clearly likely to retain a dominant position in the microanalytical armamentarium for the foreseeable future. In vitro techniques of monoclonal antibody production have removed much of the element of chance associated with in vivo antibody production methods, and increased control over the nature, properties, and long-term availability of binding reagents. The recent introduction of single domain antibodies (dAbs) by Ward and colleagues, which rely on genetic engineering techniques for the in vitro production of the heavy chain variable (V,,) domain of selected antibody molecules, represents a further step along a path leading ultimately to the synthesis of entirely artificial analyte-recognition binding molecules.

See also: Affinity; Agglutination; Antibodies, detection of; Antibody-antigen intermolecular forces; Antiserum; Enzyme labeling of antibodies and antigens; Enzyme-linked immunosorbent assay (ELISA); Fluor-ochrome labeling; Immunocytochemistry and enzyme markers; Monoclonal antibodies (mAbs); Precipitation reaction; Radiolabeling; Surface plas-mon resonance; Valency of antigens and antibodies; Western blotting.

How To Bolster Your Immune System

How To Bolster Your Immune System

All Natural Immune Boosters Proven To Fight Infection, Disease And More. Discover A Natural, Safe Effective Way To Boost Your Immune System Using Ingredients From Your Kitchen Cupboard. The only common sense, no holds barred guide to hit the market today no gimmicks, no pills, just old fashioned common sense remedies to cure colds, influenza, viral infections and more.

Get My Free Audio Book


Post a comment