Immunodiagnosis simplified: Memorandum from a WHO Meeting*

Memoranda are state- Les Memorandums ments concerning the exposent les conclu- /emorante a , conclusions or recom- sions et recomman-M e m oran da mendations of certain dations de certaines/ a t w / /WHO scientific meet- /reunions scientifiques ings; they are signed de /'OMS; ils sontMemoranaurns Iby the participants in signes par les partici- the meeting. pants a ces reunions. Bulletin ofthe World Health Organization, 62 (2): 217-227 (1984) © World Health Organization 1984 Immunodiagnosis simplified: Memorandum from a WHO Meeting* Technologies suitable for the development ofsimplified immunodiagnostic tests were reviewed by a Working Group of the WHO Advisory Committee on Medical Research in Geneva in June 1983. They included agglutination tests and use ofartificialparticles coated with immunoglobulins, direct visual detection of antigen-antibody reactions, enzyme- immunoassays, and immunofluorescence and fluoroimmunoassays. The use of mono- clonal antibodies, in immunodiagnosis and ofDNA/RNA probes to identify viruses was also discussed in detail. The need for applicability of these tests at three levels, i.e., field conditions (or primary health care level), local laboratories, and central laboratories, was discussed and their use at the field level was emphasized. Classical serological techniques have been used for a long time for diagnostic purposes, e.g., for con- firmation of clinical diagnoses, epidemiological studies, testing of blood donors, etc. Some of these techniques have been standardized to a high degree of reproducibility, e.g., the complement fixation test (CFT) and the indirect haemagglutination test, and have been used in spite of difficulties reported by many laboratories (e.g., the anticomplementary properties of sera for the CFT). Techniques using antibodies or antigens labelled with different markers have been introduced into immunodiagnosis during the past few decades and three of them have been quite widely used in routine testing, i.e., radio-immunoassay (RIA), immuno- fluorescence (IMF), and enzyme-immunoassays (EIA), e.g., ELISA. The high degree of immuno- chemical sensitivity of these tests makes them useful for the estimation of not only antibodies but also antigens or other substances that are present in minute quantities in the blood or in the urine, stool, etc. * This Memorandum is based on the report of the ACMR (WHO Advisory Committee on Medical Research) Working Group on Diag- nostic Tests for Use at the Primary Health Care Level, which met in Geneva on 14-16 June 1983. The participants at this meeting are listed on page 227. Requests for reprints should be addressed to Immunology, Division of Communicable Diseases, World Health Organization, 121 1 Geneva 27, Switzerland. A French translation of this Memorandum will appear in a later issue of the Bulletin. All these tests can be carried out in laboratories that are equipped with basic instruments as well as special- ized apparatus (e.g., gamma counters for RIA, ultra- violet microscopes for IMF, etc.), which are usually available only in the larger, central laboratories. The idea of developing simplified immunodiag- nostic tests for use at the primary health care level or for field studies was considered a long time ago. The introduction of artificial particles (latex, bentonite, beads) and stabilized red blood cells coated with anti- gens has initiated a wide range of simple tests (drop slide tests) for detection of rheumatoid factors and antinuclear antibodies, for the diagnosis of Chagas' disease, echinococcus, etc. Some of these tests are still in use as screening and/or qualitative tests. The visualization of antigen-antibody reactions has recently been achieved by other techniques, described in this report, which could lead to the development of new simple tests. The specificity of most immunodiagnostic tests depends on the purity of the reactants used for the estimation. Some tests have shown very high degrees of specificity (e.g., the estimation of hormones by RIA), while others have been less effective because of cross-reactivity (e.g., the estimation of antibodies to parasites using partially purified antigens). The intro- duction of monoclonal antibodies will undoubtedly improve the immunochemical specificity of some immunodiagnostic tests, as already demonstrated. 4395 -217 MEMORANDUM The ideal diagnostic test should be sensitive, specific, reproducible (precision of assay), inexpen- sive, rapid and simple to perform. In addition, in some situations, e.g., infectious diseases, it should be able to discriminate between active infections and past infections and indicate the effect of therapy and management of patients. The detection of antigens may be one of the ways to fulfil the latter require- ments. TECHNOLOGIES (AVAILABLE OR PROPOSED) Agglutination using microorganisms Simple agglutination reactions have been used for many years to identify various species of bacteria. Typhus fever (a rickettsiosis) is identified by the Felix- Weil tube-agglutination test and depends on agglutin- ation of a strain of Proteus bacillus containing an antigen that cross-reacts with the rickettsial antigens. Similarly, typhoid fever is detected by an agglutin- ation assay, the Gruber-Widal test, which depends on flagellar and somatic antigens on the Salmonella organisms. Brucellosis in infected cattle is diagnosed by a tube-agglutination ring test; milk from an infected animal is mixed with stained Brucella organisms and if antibodies are present in the milk, the organisms agglutinate and form a ring of clumps floating on the fat layer at the surface of the milk. All these tests are used in laboratories but not in the field despite their simplicity. Slide-agglutination tests are used for the identifi- cation of many microorganisms, such as Neisseria meningitis serogroups A, B, C and D, and are of great use in epidemiological studies. Typing of the Sal- monella serotypes is also routinely performed using slide agglutination with a battery of specific anti- sera. The bacterial agglutination tests are extremely simple, inexpensive to perform, and useful for the identification of many microorganisms. For these reasons they continue to be used today and serve as model systems for the development of other tests where simple assays are absolutely required. Indeed, if other diagnostic tests could be as simple as these ag- glutination reactions, a great need would be satisfied. A specific example of the need for simplifying diag- nostic tests is the surveillance for African trypano- somiasis. In the endemic areas, such tests have to be applied regularly on as large a proportion of the popu- lation at risk as possible. This implies that most of the tests will be used by mobile teams or by medical per- sonnel of the small rural health centres. In the latter, there is usually no laboratory know-how nor strong motivation for laboratory work. The current practice is mainly to use mobile teams for screening the popu- lation on the basis of lymph node enlargement and presence of trypanosomes in the blood and lymph node juice. These methods are time-consuming and require an important investment in terms of equip- ment, skilled personnel, and running costs. In add- ition, the proportion of false-negative results from this type of screening can be considerable - accord- ing to a recent experiment in West Africa, as high as 80%. With the introduction of the fluorescent anti- body test, it was demonstrated that serodiagnosis would increase the efficacy of screening remarkably (threefold increases of confirmed positive cases were reported in Zaire and Congo). However, these tests are usually carried out in a laboratory far away from the endemic areas and so there is a delay between the time of sampling and obtaining the results. It is thus difficult to retrace later the patients with positive diagnoses and refer them to hospitals for treatment. It also adds a further burden on the limited personnel and transport resources. What is needed for the diagnosis of trypanosomi- asis, therefore, is a simple technique that will give results rapidly and will permit the recognition of trypanosomiasis suspects on the spot. The direct agglutination test for Trypanosoma brucei gambiense (or Card Agglutination Test for Trypanosomiasis (CATT)) is a remarkable improve- ment and has been adopted by WHO as the new screening system. The antigen used is a carefully selected trypanosome clone with a variant antigen type which is seen frequently in many strains of T.b.gambiense in West and Central Africa. It may also cross-react With other trypanosomes. The organ- isms are highly concentrated, stained blue, and lyo- philized. The procedure is performed simply by mix- ing a drop of whole blood (fingerprick) with the re- suspended trypanosomes, shaking it for 3 minutes (preferably in a standard fashion by using a mech- anical device), and then reading it. For users who only rarely apply the test (dispensaries, for instance), the use of plasma or serum is recommended since it makes reading easier. From preliminary results of a series of several hundred tests in various foci of T. b. gam- biense, it seems that the number of false-negative and false-positive reactions is below 20o. The cost of the test will be approximately US$ 0.11 per test when large-scale application permits regular production. The use of artificial particles Artificial particles coated with immunoglobulins have been used for many years in agglutination reac- tions to detect and measure antiglobulins. In one sense, the sheep red blood cell, coated with rabbit antibody, can be considered a readily available "arti- ficial" particle since its underlying biological prop- 218 IMMUNODIAGNOSIS SIMPLIFIED erties are not central to the intended use in this system, which is the detection of rheumatoid factors. The sub- stitution of human Cohn-fraction II for the rabbit immunoglobulin, and polystyrene latex particles of microscopic dimensions for the red cell somewhat improved "standardization" of such tests but hardly changed the fundamental nature of these immuno- chemically sensitive tests for rheumatoid factor. Many obvious extensions of such agglutination tests have been described wherein artificial particles have been coated with microbial antigens of interest in order to assay specific antibodies, or with specific antibodies to assay for microbial antigens. Although such assays certainly can be described as simple to use since little other than the reagents are required to per- form rapid tests, at the molecular level quite compli- cated reactions may be represented, with assay results importantly influenced by the presence of rheumatoid factors, immune complexes, C I q, staphylococcal protein A, the valency of antibodies, and physico- chemical properties of the buffers and specimens. Also, an accurate determination of end-point titres may be difficult in some instances, leading to reduced clinical sensitivity and specificity and consequently a poor predictive value of positive test results. The diagnostic consequences of false-positive and false- negative assay results must always be considered when using such "simple" tests. Apart from their use in agglutination assays, arti- ficial particles have been extensively used to provide one form of a solid phase reactant to facilitate the separation of immunochemical bound antibody (or antigen) from "free" antigen (or antibody) in ligand binding assays of soluble analytes (either antigens or antibodies). The surface of the artificial particle is analogous to the surface of the plastic microtitre plate in ELISA assays. As with microtitre plates, the binding of a reagent to the "clean" surface of an artificial particle may be passively obtained through unknown "ab- sorptive forces", or more actively pursued through knowledge of some specific chemistry; in either case, a desired or undesired selective macromolecular orientation may result. Obviously, different indi- cators (radioisotopes, fluorochromes, enzyme-sub- strates) can be used to follow the bound virus-free reactants in artificial particle systems. Various kinds of artificial particles have been used for ligand assays of soluble antigens, e.g., bentonite surfaces, the surface of microscopic fibres of re- generated cellulose, the relatively large internal pore surface of microscopic (approximately 30 Am di- ameter) agar-gel beads (or Defined Antigen Substrate Spheres (DASS)), the outer surface of somewhat smaller (approximately 10 Am diameter), small pore polyacrylamide-gel beads, and the outer surface of microscopic (approximately lym diameter) and macroscopic (approximately 8 mm diameter) poly- styrene beads that are internally impervious to reac- tants. The DASS system is only mentioned in this dis- cussion of "simple" assays for historical reasons because it necessitates the non-automated use of a prohibitively expensive (approximately $40 000) microfluorometer. This is used to make time-con- suming measurements on many individually viewed microscopic beads of variable brightness that are slow to equilibrate and wash because the immunochem- istry occurs within the pores rather than at the surface of relatively large beads. With the solid (hard) macroscopic bead (approx- imately 8 mm), a single bead per test specimen is usually used; it can be separated easily from the incub- ation mixture by decantation and washed by pipetting the buffer over its surface. However, a very much larger total surface is available per unit test volume and the average diffusion distance for soluble antigen molecules to a bead surface is greatly diminished if many microscopic beads (instead of one macroscopic bead) are used per test. Typically, immunofluoro- metric assays may employ 1.5 x 108 polystyrene beads (1.1 tm diameter) per 0.1 ml test specimen. Separ- ation of bound from free antigen is by low-speed centrifugation, in a centrifuge that can be used for many routine purposes. Magnetic separation systems also are available that use magnetic artificial particles to separate bound from free reagents. Direct visual detection of antigen-antibody reactions In the methods mentioned below involving identifi- cation and quantitation of antigens or antibodies, the test materials used are serum, plasma, whole blood, and other biological fluids sampled by, for example, venepuncture or capillary tubes with anticoagulant or finger-tip blood collected and dried on filter-paper discs. Several methodological variants have been devel- oped. They have, however, the following principal steps in common, (a) primary coating of solid flat surface, (b) application of sample, (c) visualization of antigen-antibody reaction. (a) Various materials giving a solid flat surface have been tried for their suitability. At an early stage, untreated glass was found unsatisfactory unless pre- coated with a thin layer of metal, e.g., indium, gold- indium, or tantalum. Since then, the general trend has been to turn to a surface of polymer material, e.g., polystyrene and polyvinyl or to employ metal- based silicone oxide or dioxide. These have not so far been applied extensively in clinical situations. The primary coating of the surface by the relevant serological reagent (antigen or antibody) is achieved by either spontaneous adsorption (hydrophobicity) or 219 MEMORANDUM by physicochemical binding (covalent). A method- ological requirement is the establishment of a thin ''mono' molecular layer of the reagent adequately attached to the surface, with low serological reactivity and with a sufficient density of reacting sites. The material used for coating is either a solution of anti- gen (preferably purified) or a solution of antibody, preferably an affinity-purified Ig-preparation from hyperimmune serum. For special purposes specific ligands may be used for the primary coating. (b) There are in principle two ways of placing a test sample on the primary coated surface. One is to do it dropwise; a single drop for a qualitative assay or a series of drops representing a serial dilution for quantitation (end-point titration). The other way of placing the sample is to use an agar layer established on top of the coated surface. The sample is placed into punched holes in the agar and dilution of the reagent occurs by the concentration-gradient principle. (c) Visualization of an area where binding of sero- logically corresponding reagents (antigen and anti- body) has taken place can be achieved by various techniques: - In the water vapour condensation technique, the cool coated surface is exposed to an increased humid- ity, by vapour from heated water or, as a simpler alternative, the investigator's breath. In an area of antigen-antibody binding an increased wettability is visualized as drops of greater size than those of the background layer area with unaltered primary coat- ing. Thus, this type of differentiation is based on a simple physical phenomenon which is easy to create and to record. - Another and somewhat more sensitive mode of visualization can be obtained by applying to the surface a suspension of small polymer beads, either uncoated (nonspecific adsorption) or coated with a suitable serological reagent or combination of re- agents, e.g., according to the mixed haemadsorption principle. - For the identification of bound antibodies and their Ig isotype differentiation, a reinforcement visualization technique can be applied. Prior to the visualization an additional step is introduced where the surface is exposed to anti-Ig or class-specific anti- Ig reagents. - Still another mode of visualization of areas where an antigen-antibody reaction has taken place on the solid surface is the use of a sero-enzymatic in- dicator system, e.g., horse-radish peroxidase linked anti-Ig or isotype-specific anti-Ig. An area of antigen- antibody binding induced by a test sample is regis- tered as an easily recognized change of colour. - When metal or metal-based silicone surfaces are employed in thin layer analyses, the identification of areas with bound antigen-antibody is also based on colour observation, e.g., for indium plates, grades of darkness of the colour brown, which result from effects obtained with thin films. In the metal-based silicone dioxide system, detection is dependent on a series of colour changes. However, these are some- what difficult to differentiate. There are quite a few reports published about the use of these methods for assaying specific antibodies in serum or blood samples in relation to various infec- tious and non-infectious diseases. Less experience has been gained, so far, about the suitability of these methods for detection of antigens. As the assaying techniques mentioned are quite simple, versatile, and inexpensive, and as no particular instrumentation is needed, they should be suitable for serodiagnostic use in, for example, field investigations in developing countries at the level of local community health centres. Further trials of their application in this way should be promoted if a constant supply of the suit- able slides (e.g., indium) could be arranged. Enzyme-immunoassays Enzyme-immunoassays (EIA), e.g., enzyme-linked immunosorbent assays (ELISA), are techniques rou- tinely used for the detection of minute amounts of constituents that are present in biological fluids or cells. They are based on antigen-antibody reactions in which one of the constituents has been linked to an enzyme. The enzymes are not directly detectable but can be detected indirectly after reaction with appro- priate substrates. This step offers the advantage of making it possible to visualize the enzyme activity through a coloured product. Enzyme-immunoassays have been devised either for quantitative tests, using procedures similar to those described for radio-immunoassays, and for the detection and localization of material in tissues or cells, as with fluorescent reagents. Two types of assays have been devised. (1) The homogeneous assays, which make use of antigen coupled to enzyme, are based on antibody-mediated changes of enzyme activity and do not require the separation of the antigen-antibody complexes. How- ever, the present state of development of these assays limits their use to substances of low relative molecular mass, like drugs and hormones. (2) In the heter- ogeneous assays, the antigen-antibody complexes have to be separated from free antigen and/or anti- body, and thus solid-phase bound antigen or anti- body is necessary. These assays allow the quantitation of both antigen and antibody. For the quantitation of antigens, competitive or non-competitive procedures are available. Satisfac- tory results are obtained by the non-competitive "sandwich" procedure, using an immobilized anti- body to trap the antigen from the biological fluid and 220 IMMUNODIAGNOSIS SIMPLIFIED the same or another enzyme-labelled antibody for measurement. The quantitation of antibodies is performed using immobilized antigen. The amount of antibody bound to the antigen is then evaluated with an enzyme- labelled anti-immunoglobulin. The values obtained are referred to a standard curve for quantifying. Thus the sensitivity and reproducibility of the vari- ous enzyme-immunoassays depend mainly on four "more or less controllable" parameters: enzymes, substrates, conjugates, and the solid-phase used for immobilization. Enzymes and conjugation procedures have been well investigated. Mainly, four enzymes of high turn- over activity are currently used: horse-radish per- oxidase, alkaline phosphatase from Escherichia coli or calf intestine, and beta-galactosidase from E. coli. Glucose-oxidase from Aspergillus niger is less commonly used. Conjugates prepared with antibodies directed against the immunoglobulins of many species (and their isotypes) are commercially available. The quality of antibody conjugates depends mainly on the avidity or affinity of the antibody. They have to be purified by affinity chromatography or immuno- adsorbents in order to obtain conjugates which give the best results in terms of positive signal over background. However, immunoglobulins prepared by salt fractionation can in some cases be useful, especially when monoclonal antibodies obtained from ascites are used. Antigen-enzyme conjugates have been less commercialized, but some already exist in kits. Conjugates prepared under the same conditions with all these different enzymes give the same range of detection, but the time for the enzyme activity reac- tion varies. There exist some alternatives to the use of enzyme- antibody conjugates which in some cases increase the sensitivity of the test. These include: - use of peroxidase-antiperoxidase (PAP) com- plexes; - use of synthetic hybrid antibodies with double specificity directed against immunoglobulins and either enzyme or another marker such as red blood cells; - use of enzyme protein A conjugate; - use of biotinylated antibody or antigen and enzyme-labelled avidin. Various substrates for each enzyme are available: ,chromogenic substrates, the product of which can be either soluble (useful for spectrophotometric measurement) or insoluble (giving a coloured pre- cipitate allowing the precise localization of the site of the enzyme reaction). For a more sensitive test, fluor- ogenic substrates exist which require the use of special ultraviolet light-emitting equipment. The substrates used for the detection of peroxidase activity are the most rapid, but they can be easily oxidized by any traces of oxidizing agents (other than the enzyme and hydrogen peroxidase) in the vessels and they are light- sensitive. Beta-galactosidase and alkaline phos- phatase substrates are relatively stable. The solid-phase is the major problem. The im- mobilization of the antigen or antibody can be made either by covalent binding or by adsorption through non-covalent interactions. Supports can be agarose, cellulose polyacrylamide, or polyacrylamide-agarose particles. They are separated from the liquid phase by centrifugation. More convenient are particles rendered magnetic, which can be easily and quickly isolated with a magnet. The most widely used sup- ports are polystyrene plates with 96 wells for which automatic readers are commercially available. Poly- styrene balls have also been used. All these supports have their advantages and disadvantages: particles give good reproducibility but necessitate multiple handlings, magnetic polyacrylamide beads are useful and can be re-used after regeneration, but the lack of homogeneity in the bead size may produce irregular results. In conclusion, evaluation of antigen or antibody in biological fluids can be performed by various immunoassays. The choice of the test must be dic- tated by individual considerations. In the current state of development of these tech- niques, the following simplifications are already prac- tised in some laboratories: (i) Simplification in terms of "reading the test". In many instances there is no need for a sophisticated instrument for detecting the reactions. A naked-eye appreciation can be roughly estimated by reference to coloured standards. These standards have, however, to be prepared for the test since the reproducibility of the enzyme assays depends on the time and tempera- ture at each step. (ii) Simplification in terms of rapidity. The differ- ent times of incubation (immune-complex formation and enzyme activity) can be shortened without dram- atically decreasing sensitivity by slightly increasing the concentration of the enzyme-conjugate. (iii) Simplification concerning the solid-phase. The most convenient are the 96-well polystyrene micro- plates. They offer the following advantages: - they can be previously coated with almost all antigens or antibodies at low concentrations; - they do not require large amounts of material (50-200 Id); - they are easily washable (by immersion for example); 221 MEMORANDUM - they are easily used with commercially-available multiple pipettes adapted to this kind of plate. But there are some disadvantages: - reproducibility varies between different lots; - they are expensive; - small amounts of antigen must be coated to avoid release during further incubations (this could be a limitation when the antigen is only part of the mix- ture); - they require a large volume of washing medium (PBS-Tween), which may be considered quite compli- cated for field use. The simplifications already described should be adapted because they have so far not been well studied in routine test procedures. (iv) Use ofpaper sheets. Nitrocellulose sheets have been found suitable for spotting antigens (various proteins, liposaccharides, or DNA), which are after- wards detected by antibody-enzyme conjugates. These could be used either as antigen carriers for anti- body diagnosis or as material carrier (for the collec- tion of blood or gland extract, etc). The advantages are: - they can be stored dry or wet in small, capped tubes; - they can be washed easily with a small volume of washing medium; - they can be the carrier of many different anti- gens and can be incubated in one step with the patient's serum; - they can be used with precipitating substrates and stored. The disadvantages are: - they are very expensive at present; - they are difficult to handle as a sheet and should be adapted to something solid (a stick, for in- stance); - no systematic work has been done concerning the stability of the spotted material. New developments of tests based on competition to eliminate one step of incubation and to shorten the assay should be encouraged. In addition, the develop- ment of heat-stable antibodies and enzymes would be an advantage. Immunofluorescence andfluoroimmunoassays None of the tests are useful in the field and most likely will not be in the near future. Immunofluor- escence can be used in the local diagnostic labora- tories. Fluoroimmunoassays require the facilities of a central laboratory, including maintenance. However, the tests, even if carried out in central laboratories, may have a great impact on primary health care. Immunofluorescence. Indirect immunofluores- cence is one of the best methods available today for the rapid diagnosis of many organisms. The reagents are often made commercially and are quality con- trolled. Good fluorescence microscopes are absol- utely essential for reading the tests and their main- tenance is a necessary condition for their use. The training of experts is also needed. A training course of 2-3 weeks may not be sufficient for beginners, but further training may be organized by making avail- able positive and negative control specimens (fixed on slides) for staining and reading in local laboratories. High quality reagents are also absolutely mandatory for reliable immunofluorescence tests. It is thought that, in the future, many immuno- fluorescence assays may be replaced by enzyme- immunoassays. Fluoroimmunoassays. It is possible that in a few years fluoroimmunoassays will replace radio- immunoassays and enzyme-immunoassays in well- equipped laboratories, though this will mean new investment in equipment. The following types of fluoroimmunoassays (FIA) are available: (1) Solid-phase FIA. This is similar to the two-site, "sandwich", immunoradiometric assay (IRMA) in that the antibody carries the label, in this case a fluorochrome. For viral antigen detection, "capture" antibodies are bound on solid phase, usually a plastic material, which is efficient for binding the antibody. High backgrounds, including that from the plastic may result in unacceptably low sensitivities for anti- gen detection. However, the sensitivity may be high enough for the assay of viral antibodies. (2) Fluorescence-polarization method. In this assay, the labelled antigen is excited with polarized light and the degree of polarization of the fluorescent emission is measured. If antibodies bind to the anti- gen, brownian rotation diminishes and the degree of polarization increases. The assav is simple and rapid, requiring no separation steps, but unfortunately it has some major disadvantages. Sensitivity is limited by background problems, the polarization response is non-linear, and the relative molecular mass of the detectable antigen cannot be larger than about 20 000 (e.g., whole viruses cannot be detected by this method). (3) Fluorescence-quenching methods. In direct fluorescence quenching, a fixed amount of antigen that is labelled with a fluorescent probe (e.g., fluor- escein isothiocyanate) is mixed with an antibody that is labelled with a fluorescent quencher (e.g., rhod- 222 IMMUNODIAGNOSIS SIMPLIFIED amine) which adsorbs the fluorescent emission from a proximal probe. If unlabelled antigen (from the specimen) is present, it binds the quencher-labelled antibody to prevent quenching. A modification is that half of the antibody molecules are labelled with a fluorescent probe and the other half with a quencher. When unlabelled antigen is present, the two types of antibodies are brought together causing quenching of fluorescence. Another type of quenching makes use of fixed amounts of: (i) antigen labelled with a fluor- escent probe, (ii) antibody to the fluorescent probe which quenches its fluorescence, and (iii) antibody to epitopes on the antigen that are close enough to the fluorescent probe to sterically block the action of the antibody to the fluorescent probe. Excess antigen added to the system removes the blocking antibodies resulting in quenching of fluorescence. (4) Fluorescence-enhancement methods. In excep- tional cases, an increase in the fluorescence of labelled antigens has been found after binding to antibody but there are no reports indicating that this method is applicable to microbial antigens. (5) Time-resolved fluoroimmunoassays. A new type of FIA, called time-resolved fluoroimmunoassay (TR-FIA), has recently been developed and applied to viral antigen and antibody assays. The TR-FIA is based on a fluorescence probe with a long life-time. It is excited by a short light pulse and the specific fluorescence is measured after a selected time delay. During the delay the background fluorescence, which has a short decay time, is eliminated and the specific fluorescence is measured with similar or higher sensi- tivity than obtained by radio-immunoassays or enzyme-immunoassays. Rare earth metals, lanthanides, e.g., europium (Eu), usually have a long decay time, 100 to 1000 As. An additional advantage of lanthanide chelates is that there is a large difference between the excitation and emission wavelengths (Stokes' shift), which causes further reduction in background and improves the possibility of detecting specific fluorescence. Binding of Eu to antibody or antigen can be done via EDTA which binds efficiently to lanthanides, and an amino- phenyl derivative of EDTA has been synthesized which can be linked to proteins by the standard coupling reactions used in protein chemistry. Solid-phase assays have been used in TR-FIA for the measurement of viral antigens and antibodies. After the immunoreactions have been completed, the solid phase is washed, a counting solution (2-naph- toyltrifluoroacetate in aqueous detergent solution) is added, and the fluorescence is measured in a single- photon counting fluorometer equipped with a xenon flash lamp. The excitation wavelength is 340 nm and the length of the excitation pulse is lts. After a time delay of 400 As, the single-photon emission is counted for 500 As at 613 nm. After another delay of 100 As, the cycle is repeated for about 1000 times during the total counting time of 1 second. The results are ex- pressed as counts per second (cps) values. In each assay, the background fluorescence of the counting solution is measured in triplicate and the mean value may be deducted from the cps values in the assay. In a further developed, direct, one-incubation TR-FIA, the specimen is added simultaneously with the labelled viral antibody to polystyrene tubes or micro- titre strips coated with the same antibody as in the label. After a one-hour incubation and washing, the counting solution is added and the fluorescence is measured. The sensitivity of the direct TR-FIA with one incubation is the same as the sensitivity of the in- direct TR-FIA or the indirect radio-immunoassay with three incubations and washings. The TR-FIA is a good candidate for replacing radio- and enzyme-immunoassays. The disadvantage of the technique, especially in developing countries, is that it requires a single-photon counting photometer and even simple non-automated equipment without printer may be more expensive than spectrophoto- meters used for EIA reading. In all the technologies discussed so far, it is clear that the sensitivities and specificities of immuno- assays, whether applied for antigen or antibody assays, are not dictated by the probes used (isotope, enzyme, fluorochrome) but by the quality of the immunoreagents used. The main difference in these assays is that radio-immunoassays and fluoro- immunoassays, where the labelled antigen or anti- body is measured directly after the immunological reactions have completed, can be standardized more accurately. DNA/RNA probes The techniques on how to use the analysis of DNA and/or RNA probes of viruses or microorganisms for diagnostic purposes have not yet been worked out to such an extent that they can be applied in field labora- tories. However, there exist relatively simple and rapid methods to identify or characterize virus strains by analysing their genomes. In the following, various techniques will be mentioned which are applicable to different virus systems or to microorganisms and which are summarized in Table 1. (1) RNA-viruses (a) Migration rate analysis by polyacrylamide gel electrophoresis (PAGE) Viruses with a segmented genome exhibit a specific pattern of their RNA segments after PAGE. The RNA can be labelled either in vivo or in vitro prior to 223 MEMORANDUM Table 1. Molecular techniques for characterization of virus genomes and their applications Virus Main fields Special Technical genomes Method of application requirements complexity Usesa RNA (i) migration rate influenza virus, radioisotopes relatively D, E, R analysis in poly- rotavirus not simple acrylamide gels essential (ii) T,-oligonucleotide influenza virus, isotopes moderately D, E, R mapping poliovirus, complex flavivirus (iii) molecular influenza virus isotopes moderately D, E, R hybridization complex DNA (i) endonuclease herpes virus, radioisotopes, relatively D, E, R restriction cytomegalovirus, specialized simple mapping, varicella virus, enzymes polyacrylamide papilloma virus, gel migration poxvirus, analysis plasmids (ii) molecular herpes virus, radioisotopes, complex E, R hybridization hepatitis B, specialized onc-genes of enzymes RNA tumour viruses a D = diagnostic use; E = use in epidemiological studies; R = use in research work. PAGE (influenza viruses), or can be detected directly by staining the gel with ethidium bromide (rota- viruses). (b) Oligonucleotide fingerprinting of viral RNA Any viral RNA can be characterized by this tech- nique. The RNA is digested by Tl-RNase, which cleaves the RNA after G, and the oligonucleotides are separated by two-dimensional PAGE. The larger oligonucleotides are unique for each RNA and are used for characterization and determination of gen- etic relatedness. (c) Molecular hybridization If virus RNA (vRNA) is hybridized to its hom- ologous complementary RNA (cRNA), forming a double-stranded RNA, it is resistant against treat- ment with RNase A. In heterologous hybrids (vRNA of one strain, cRNA of a genetically-related strain), the regions of mismatching (indicating genetic di- versity) will be digested. Thus, the RNase-protection is a relative measure of genetic relatedness. Various combinations of this technique are available: (i) direct RNA-RNA hybridization; (ii) compet- itive hybridization (iii) RNA-DNA hybridization; (iv) PAGE analysis of hybrids after treatment with SI nuclease. (2) DNA viruses (a) Restriction endonuclease mapping of viral DNAs The isolated DNA of these viruses are treated by specific bacterial restriction endonucleases, and the resulting fragments are separated by PAGE and visualized either by staining with ethidium bromide or by autoradiography. DNA cleavage patterns have allowed the identification of different papilloma viruses, the typing of herpes simplex viruses (HSV) and the identification of genetic variants of HSV types 1 and 2 and cytomegalovirus, and the character- ization of plasmids. (b) DNA hybridization procedure DNA is fragmented by specific restriction endo- nucleases, and the fragments are separated by PAGE and transferred by "blotting" onto nitrocellulose membranes, to which they are firmly bound. The immobilized DNA is identified by hybridization with radio-labelled virus-specific DNA or RNA probes and visualized by autoradiography. For "dot hybrid- ization" the DNA is directly immobilized on nitro- cellulose filters and is detected using radio-labelled DNA or RNA probes, while in the sandwich hybridiz- ation method, one of the reagents is on the filter. These procedures have been used, for example, to detect virus-specific sequences integrated into cellular genomes. DNA can be hybridized also in situ, which allows the detection and localization of virus-specific sequences in tissue sections or cells. Future trends, applying DNA and/or RNA probes in diagnosis, are to prepare probes for hybridization, which are not labelled by radioisotopes but by other 224 IMMUNODIAGNOSIS SIMPLIFIED 225 markers, e.g., enzymes. Furthermore, the availability of individual genes, specific for the material under investigation, might help in the application of methods that have been worked out for viruses to more complex organisms such as parasites. MONOCLONAL ANTIBODIES: THEIR USE IN IMMUNODIAGNOSIS Monoclonal antibodies have already proved their value in biology as highly specific probes for antigens in complex mixtures. Their application to immuno- diagnosis is thus twofold: (i) as probes for identifi- cation, purification, and characterization of relevant antigens, and (ii) as reagents in diagnostic assays themselves. Monoclonal antibodies have an excep- tional degree of specificity and promising sensitivity, and they could make a great impact on immuno- logical diagnosis; ways to adapt them to the simplest possible diagnostic tests must be considered as a priority. One of the most positive features of monoclonal antibodies is also responsible for great difficulties in their use: that is, their individuality. Each mono- clonal antibody is unique, not only in antigen-binding specificity but also in its biochemistry. It is therefore important that the selection of monoclonal antibodies should be performed at several levels if ultimately they are to be used as a diagnostic reagent: first, selec- tion has to be for the desired specificity; second, for their ability to function in a particular assay; and third, for their stability, ease of purification, and ability to be manipulated in biochemical reactions such as those encountered during coupling to enzymes. Selection of monoclonal reagents The majority of monoclonal antibodies utilized so far have been selected primarily on the basis of specificity, often after primary screening only on the immunizing antigen. The search for diagnostically useful monoclonal antibodies should be approached from two directions: (i) by selecting from the battery of reagents produced by fusions, and (ii) by specifi- cally selecting reagents using the exact assay in which they will ultimately be employed, including testing on samples in the situation of intended use. In addition, in either approach, the use of selected detection reagents can aid in obtaining antibodies that are likely to have the functional properties desired for the ultimate assay and that are easy to purify and handle biochemically. Another level of selection which is important for the development of assays is concerned with the stability of the reagents, which are known to vary markedly in their tolerance to heat, desiccation, and stability in solution, unlike antisera which can often withstand harsh treatment. Selection of test systems Diagnostic assays are necessarily of three kinds: those used in the central laboratory, those used in local laboratories, and those used in the field. In this order, these are expected to have a decreasing com- plexity and decreasing dependence on expensive equipment. Because a good diagnostic assay should have high specificity and high sensitivity, these should be a minimum prerequisite; the degree of simplicity and the cost are next in importance. Fortunately, many immunological tests using monoclonal anti- bodies can be made relatively simple and inexpensive and thus will have all the attributes of a desirable assay. Many monoclonal antibodies have already been shown to have immunodiagnostic potential. Some have been extensively tested using actual clinical samples (although not in the field), and some use in- novative assays which can probably be modified for use in simple field tests for large numbers of samples. (a) Direct binding assays.a Suitable monoclonal reagents can be used to detect antigens that are bound to a solid phase (latex, microplate wells, nitrocellulose paper). In most assays, the monoclonal antibody must be labelled with 1251 or with appropriate enzymes for ELISA procedures. Selection of the monoclonal reagents will thus depend on high affinity for the antigen, and for their ability to cause direct ag- glutination, for example, in assays using direct ag- glutination of latex particles to which the antigen is adsorbed. Direct binding of specific monoclonal antibodies will also be useful in "sandwich" assays in which the monoclonal reagents are first used to bind the antigens from a mixture and then to detect the bound antigen, as previously mentioned. In some situations, antibodies specific for two different epitopes will be required. This type of assay is likely to prove of great value when the antigen to be measured is present in small quantities in crude mixtures. (b) Inhibition assays. Both antigen and antibody can be detected in assays based on the inhibition by test samples of binding of labelled monoclonal reagents to antigen bound to solid phase. Here it is important to select monoclonal antibodies of the "correct" affinity for antigen because it is possible a Direct binding assays in this report refer only to those assays in which antigen is detected by an unlabelled antibody (in the case of thin-layer assays), or labelled antibody with no need for a second antibody directed to the first, as is often the case with "indirect" assays. MEMORANDUM that antibodies of high affinity will compete too successfully with the inhibitor antibody in the test sample. If antigen is being detected, this may not be a problem. In inhibition assays, consideration must be given to the selection of monoclonal reagents which will allow the assay to be performed in as few steps as possible, e.g., by including the inhibitor and detecting antibody and antigen in one reaction mixture. A special type of assay based on anti-idiotype monoclonal antibodies can be developed which de- pends on the inhibition by antigen of the binding of two monoclonal antibodies, one of which is specific for an epitope in or near the antigen binding site of the other. The advantages of this type of assay are that the reagents (both monoclonals) can be obtained in extremely large quantities, there is no need for coupling of antigen to a solid phase, and the assay is suitable for detecting antigens present in small amounts in complex mixtures. Improvement of assays using monoclonal antibodies Recent theoretical considerations and some practical demonstrations suggest that the immuno- chemical sensitivity and specificity, the stability, the precision, and the rapidity of routine immunoassays can simultaneously be improved, possibly by several orders of magnitude, if monoclonal antibodies that have non-radioisotopic labels displaying higher equivalent specific activities are used rather than the currently used radioisotopes. It is claimed that con- ventional competitive binding radio-immunoassays (which employ a fixed amount of radioisotope- labelled antigen and saturate binding sites on a limited number of antibody molecules with excess antigen) have approached a theoretical limit of immuno- chemical sensitivity of approximately 107 antigen molecules per test. In practice, this limiting sensitivity has not been improved by increasing the specific ac- tivity of the labelled antigen, nor by replacing it with a signal-amplifying non-radioactive labelled antigen. However, the theoretical limiting sensitivity for non- competitive, two-site "sandwich" immunometric assays is one antigen molecule per test! For this type of assay, ideally, an unlabelled monoclonal antibody of high avidity for one antigen site (epitope) and present in large excess over antigen concentration is bound to a solid-phase surface to "immuno-extract" essentially all antigen molecules from the assay specimen, whereupon they are detected by a second monoclonal antibody that is labelled and is of high avidity for a second antigen site (epitope), sterically remote from the first site; this labelled monoclonal antibody also is present in large excess over antigen. If these ideal conditions are met, the assay specimen can be added to an appropriate mixture containing both the solid phase and the labelled monoclonal anti- bodies, which can interact only in the desired manner through the two sterically remote epitopes on the anti- gen molecule. Following this one-step incubation period, which is much shorter than with sensitive competitive binding assays (because here the reaction is driven by antibody excess rather than deficiency), the immune complex bound to the solid-phase is read with an appropriate sensor. By making use of two different epitopes, immunochemical specificity can be markedly enhanced, as has been demonstrated by an immunoradiometric assay (IRMA) employing two monoclonal antibodies, specific for the alpha and beta chains of human chorionic gonadotrophin (HCG). Biologically functional HCG is accurately assayed in the presence or absence of other hormones having the same alpha chain, or free alpha and beta chains. Although many IRMAs employing polyclonal antisera have been reported since the late 1960s when IRMA was introduced as an alternative to RIA, generally they have had the same order of immuno- chemical sensitivity to RIA. In general, the same anti- sera are used for the solid phase and the labelled anti- body in two separate incubations with intervening wash. Experimental evidence shows that if both anti- bodies have the same epitope specificity, labelled anti- body can compete with solid-phase antibody during the second incubation to release, and thereby lose, antigen from the system. Additional problems with IRMAs that employ the same polyclonal antibody for both incubations are that wasteful amounts of very pure antigen (when available) are needed to make polyclonal antisera monospecific (by affinity chrom- atography), and IRMA uses much more antibody and its specificity requirements are much greater than needed for RIA. In principle, monoclonal antibodies would solve all these problems. However, the partici- pants at the meeting were unaware of any practical use as yet of these interesting theoretical consider- ations that would simultaneously and markedly im- prove the sensitivity, specificity, stability, precision, and rapidity of routine immunoassays of antigens. Until more experimental evidence is forthcoming, a judicious balance between caution and enthusiasm is needed. Although these non-isotope immunometric assays cannot be considered as "simple", they could greatly simplify immunoassays. Monoclonal antibodies, when properly selected, thus have a tremendous potential for use in diagnostic assays and may allow development of extremely simple tests suitable for field use. They are of course required (rather than antisera) in those many situ- ations where antigens cannot be purified and used to make antisera. Monoclonal antibodies have a high degree of specificity and the ability to be standardized and made in large quantities, so that effort should be spent in their selection and adaptation towards simpli- fication of the tests. 226 IMMUNODIAGNOSIS SIMPLIFIED CONCLUSIONS The Working Group reviewed the technologies available for performing several simple immunodiag- nostic tests and discussed possible simplifications and/or improvements of other technologies. Direct agglutination of microorganisms (bacteria, parasites) and indirect agglutination of particles coated with antigens are the only simple tests widely used for routine diagnosis. Several recently developed simple tests, based on direct visualization of the antigen-antibody reactions either on solid modified surfaces or in gels, are still at the research stage of development and have not been validated in con- trolled trials with the exception of the indium slide test, which has shown promising results in the diag- nosis of schistosomiasis. The technologies using different markers (isotopes, fluorochromes, enzymes) represent the main diag- nostic methods at present; however, they can only be performed in well-equipped laboratories. Possible simplification and/or improvements have been dis- cussed. The enzyme tests (e.g., ELISA) are the only candidates for simple tests to be developed for the field level (naked-eye reading of the results in com- parison with coloured standards). Many modifi- cations discussed could improve these tests in terms of rapidity, sensitivity, specificity and economy, but they will need complicated equipment. The RNA/DNA probes have been successfully applied for diagnostic purposes in several infectious diseases and could be developed into field-level tests in the future, especially if the isotopes could be replaced by other markers, e.g., enzymes. Monoclonal antibodies have already proved their value in diagnostic procedures by increasing the specificity of the tests. In addition, they can be prepared in large quantities at any time and place worldwide by distibuting the clones which is an important factor for standardization. When properly selected, monoclonal antibodies may allow develop- ment of extremely simple test kits for detection of antibodies as well as antigens suitable for field use. * * P. Halonen, Department of Virology, University of Turku, Turku, Finland 0. Ouchterlony (Chairman), Department of Bac- teriology, Goteborg University, Goteborg, Sweden T. W. Pearson (Rapporteur), Department of Bio- chemistry and Microbiology, University of Vic- toria, Victoria, BC, Canada C. Reimer, Biological Reagents Section, Center for Disease Control, Atlanta, GA, USA C. Scholtissek, Institute for Virology, Giessen, Federal Republic of Germany T. Ternynck, Immunocytochemistry unit, Depart- ment of Molecular Biology, Pasteur Institute, Paris, France WHO Secretariat V. Houba (Secretary), Immunology, Division of Communicable Diseases F. A. Assaad, Division of Communicable Diseases B. Mansourian, Office of Research Promotion and Development V. R. Oviatt, Special Programme on Safety Measures in Microbiology T. A. Bektimirov, Virus Diseases, Division of Com- municable Diseases G. Causse, Bacterial and Venereal Infections, Division of Communicable Diseases R. H. Morrow, Parasitic Diseases Programme P. de Raadt, Trypanosomiasis and Leishmaniasis, Parasitic Diseases Programme L. Martinez, Malaria Action Programme B. 0. L. Duke, Filarial Infections, Parasitic Diseases Programme S. K. Noordeen, Leprosy, Division of Communi- cable Diseases A. Moncayo, Trypanosomiasis and Leishmaniasis, Parasitic Disease Programme 227