ENZYME LINKED IMMUNOSORBENT ASSAYS REVISITED

Abstract

Enzyme linked immunosorbent assay (ELISA) has now gained wide acclaim in the immunodiagnosis of infectious diseases. ELISA is continuously evolving with newer and sensitive formulations being added to it. The basic tenets of ELISA are discussed. The role of ELISA in two important issues facing us today, namely tuberculosis and human immunodeficiency virus infection are reviewed.

Introduction

During the last two decades there has been a phenomenal increase in the number and variety of immunodiagnostic tests performed. Of these, the solid phase immunoassays employing non-isotopic labels (like enzymes) for antibodies and antigens have taken the trend away from radioisotopes in liquid phase assays. Automation or semiautomation of these assays, use of monoclonal antibodies, sophistication in methodologies of labels and solid phases, use of recombinant and synthetic peptides, new assay formulations and thorough validation criteria have made these assays extremely sensitive, specific, rapid and suitable for large volume testing. In the journey from the earlier laborious serological tests to the polymerase chain reaction (PCR) today, enzyme immunoassays (EIA) have gained a firm and pivotal place in diagnostic microbiology.

The use of enzymes as immunochemical labels in competitive binding assays was reported in 1971 by Avrameas [1]. The most widely used terms in enzyme based tests are enzyme immuno assay (EIA) [2] and enzyme linked immunosorbent assays (ELISA) [3].

General aspects

EIA is a generic term that encompasses all assays that use enzymes as labels. These enzymes are linked to either antibodies or antigens such that the complexes formed have both immunological and enzymatic activity. Degradation of a chromogenic substrate by the enzyme yields an amplification signal (colour), which enables accurate and sensitive detection of the presence of the enzyme and in turn the analyte in question. EIAs fulfill the following criteria. First, they employ immunological elements (enzyme linked tracers or conjugates) to detect the analyte of interest, and second, enzyme activity is used to quantitate the analyte. The primary function of the enzyme in the EIA is to act as an amplifier, thus increasing sensitivity.

EIAs fall under two major categories. The homogenous EIA is one in which a hapten (drug or hormone) is linked to an enzyme such that the enzyme activity is modulated when the hapten combines with the antibody. The assay system consists of an enzyme-labelled hapten + antibody (to the hapten) + substrate + test sample. If there is hapten in the sample it competes with the labelled hapten for the antibody. This will reduce the availability of the free antibody that may inhibit the enzyme activity of the labelled hapten, so leaving the labelled hapten free to degrade the substrate, with colour production that can be read in a spectrophotometer. Homogenous EIA is specially applied

for the rapid assay of low molecular compounds like hormones, chemotherapeutic agents and drugs of abuse [4]. On the contrary, in the heterogenous EIA (ELISA) the binding of antigen to enzyme-la-belled antibody does not modulate the activity of the enzyme label.

ELISA scores over immunofluorescence (IF) which is tedious, time consuming, not easily automated, unsuitable for large batches of tests and suffers from subjective interpretation bias. Radioimmunoassay (RIA) has its disadvantages in that the reagents used have a short shelf-life, sophisticated equipment is required and there is hazard of radioactivity. Above all. theoretical sensitivity of ELISA ranges from 10⁻⁴ to 10⁻¹⁶ moles per litre (nanogram quantities per mL) giving it the upper edge in clinical use [5]. ELISA has now been greatly prejudiced by availability of convenient, reliable, commercial kits [6].

ELISAs have been developed in several different configurations. In each of these, one of the reactants, either the antigen or the antibody, is immobilised on to a solid phase matrix. There are many commonly employed ELISA configurations [7] (Fig 1 A & B).

Fig. 1.

Format of direct & indirect EIAs. Antibody (Ab or Ab1) adsorbed to the well captures the antigen (Ag). Bound Ag is then detected directly by enxyme (E) labelled Ab as in A or indirectly with unlabelled Ab (Ab2) followed by enzyme labelled Ab (Ab3) as in B.

Direct ELISA : used for antigen detection by employing antibody-adsorbed wells, which capture the antigen (analyte). This bound antigen is then detected directly by a secondary antibody labelled with enzyme (Ab with E in Fig 1A).

Indirect ELISA : used mainly for antibody detection by employing antigen-coated wells which capture the antibody in the sample (Ab2). This bound analyte is detected indirectly by enzyme labelled anti-human globulin (AHG) (Ab3 with E in Fig 1B).

In both the above, the final end product is development of colour following addition of the substrate (Fig 1). Two more modifications of ELISA need to be mentioned. In the antibody class capture ELISA [8], one can also ascertain the class of antibody besides detecting its presence. For IgM class capture and detection, the wells need to be coated with anti-μ chain (M antibody capture ELISA or MACELISA) and for IgG class capture, anti-gamma chain is coated on to the wells (GACELISA). After adding the test sample to look for a particular class of antibody (IgG or IgM), the other steps are as for an indirect ELISA. In the competition ELISA [7] for antibody detection, the test sample (containing free antibody) and enzyme-labelled antibody conjugate are simultaneously added to the antigen coated wells. They compete with each other for the antigen. If free antibody is present in the test sample it binds to the antigen and subsequent addition of substrate, after washing, does not elicit colour development. Hence, colour development in this assay is inversely proportional to the quantity of specific antibody in the test sample. Sera with preservatives like sodium azide are unsuitable for competition ELISA format, since the preservative may inactivate the enzyme.

Technical aspects [9]

Solid phase: Amongst the wide range of solid phases, the commonly used ones are beads or wells (popularly a 96-well microtitre plate) made of polypropylene or polysterene.

Enzymes: Alkaline phosphatase (AP) and horseradish peroxidase (HRPO) are commonly used.

Substrates: Appropriate to the above named enzymes are para-nitrophenyl phosphate (PNPP) for AP and tetra methyl benzidine (TMB) or ortho phenylenediamine hydrochloride (OPD) combined with H₂O₂ for HRPO. These chromogens are light sensitive; it is necessary to incubate the substrate-chromogen reaction in the dark.

Washing solutions: After addition of sample and reagents, and incubation, the wells are washed so as to remove the unbound entity. This prevents any non-specific binding and erroneous resuls. The washing solution used is phosphate buffer solution (PBS) containing 0.05 per cent Tween 20 which is a non-ionic detergent and prevents non-specific binding of test materials by reducing surface tension.

Conditions of the test: It is necessary to ascertain the optimum conditions for each stage of the test. These are optimum concentrations of antigen or antibody coating the solid phase, appropriate dilution/concentration of all the reagents, optimum incubation times and temperature, and thorough washing after each step. The enzyme action on the substrate is halted after a specified period (depending on the kit used) by using 1N sulphuric acid (stopping solution).

Results : The colour produced is measured at a specific wavelength using monochromatic filters (complementary to the colour produced by the substrate) using a spectrophotometer. The absorbance value of the colour can be expressed by means of optical density (OD) which is proportionate to the intensity of the colour, which in turn is proportionate to the concentration of the analyte in the standard ELISAs.

Interpretation of results: Reference standards, positive and negative controls, are always put up along with the test samples in the same test run. A cut off value (COV) is then determined for each test run using ODs of reference standards (controls). A common method to calculate COV is to take average absorbance values (ODs) of 3 known negative controls from within the test run plus a constant factor, which is usually 3 or 4 standard deviations of the mean absorbance values from a large number of known negative samples (usually > 100). Alternatively, the COV is calculated from a large number (statistically significant) of known negative samples by determining the highest absorbance obtained within the distribution curve. It is necessary to re-calculate the constant factor whenever a new batch of reagents are used. Ideally, the COV is about twice the OD given by the negative reference standard. Values of the test samples above the COV are read as positives and those below are considered negative. Each test run is put through validation criteria (pertaining to the expected ODs of the reference controls). Any test run failing to meet the validation criteria is null and void and must be repeated.

Sensitivity and specificity issues: Sensitivity is the ability of a test system to detect the minimal possible quantity of the analyte in a test sample. Specificity connotes the ability of a test to detect none other than the analyte of interest. Present ELISA formats have sensitivity in the range of 98-100 per cent, and specificity of 95-98 per cent.

Applications of ELISAs

ELISA has not yet become popular in a bacteriological laboratory. It is likely that it will continue to occupy pride of place in viral serology, and it is also proving useful in parasitic diseases [10]. The common infectious diseases where ELISAs are applied are outlined in Table 1.

TABLE 1.

Common applications of ELISA in infectious diseases

	Organisms
Bacteriology	M.tuberculosis, Yersinia
Virology	Hepatitis viruses Human immunodeficiency viruses Measles, Mumps, Respiratory syncytial viruses Herpes viruses Japanese encephalitis. Dengue, West Nile viruses
Mycology	Aspergillus, candida
Protozoology	Toxoplasm gondii, Leishmania, Trypanosoma

ELISA in tuberculosis

ELISA in tuberculosis

No fully reliable serological test exists for tuberculosis. The obvious reasons are the ubiquitous nature of the organism and prior exposure to BCG. The role and fallacies of ELISA in diagnosis of tuberculosis are discussed in brief.

Specimens: Serum is the usual specimen. For local antibody production, CSF, pleural fluid, and bronchial washings have been used, but only CSF antibody detection holds some diagnostic promise. Development of PCR has undermined the role of ELISA in local antibody detection. The major problem with using serum is the necessity to predilute it to 1:100-1:500, so as to counter non specific binding of immunoglobulins. Dilutions of the samples do decrease the sensitivity of the test.

Antigens for ELISA: Even purified antigens (Ag) contain multiple cross-reactive epitopes interfering with the specificity. An ideal antigen would be one with species-specificity and strong immunogenecity. Predominantly protein antigens like Ag5 and Ag6, and A60 are used [11, 12, 13]. Ag5 being the most extensively tried since it contains a highly immunogenic 38 KD protein [11]. A60 contains, in addition to proteins, both free and bound lipids and polysaccharides representing the main thermostable components of PPD [13]. Several distinct protein antigens which are restricted to M. tuberculosis, M. bovis, M. africanum and BCG strains are minimally shared by atypical mycobacteria. These have been identified by using monoclonal antibodies. Of these, the 38 KD has 7 monoclonal antibody defined epitopes [14]. It has been the most useful single antigen in the diagnosis of multibacillary pulmonary tuberculosis but it has no potential to diagnose extrapulmonary tuberculosis [15].

Sensitivity and specificity issues: Using Ag5 with 38 KD protein, a sensitivity of 85 per cent and specificity of 97 per cent can be achieved [15]. Sera negative for antibodies to 38 KD protein are rarely positive for other mycobacterial epitope specific antibodies [16]. Using A60 antigen sensitivity, specificity and positive predictive value have been seen to be 48 per cent, 71 per cent and 50 per cent respectively with only IgG estimation, whereas using IgM estimation alone the parameters were 76 per cent, 98 per cent and 95 per cent respectively. Combining the results of IgG and IgM, the above figures were 68 per cent, 100 per cent and 100 per cent respectively [12].

Smear positive tuberculosis: The indication for serology in a smear positive patient is restricted to ruling out atypical mycobacteria. Anti-38 KD antibodies are the first to appear after commencing treatment. High titres of these antibodies are associated with extensive disease and poor prognosis [17].

Smear negative pulmonary tuberculosis: There is weak humoral response in this group; hence ELISA results are less satisfactory, even 38 KD Ag based ELISA is insensitive in this group [15].

Extrapulmonary tuberculosis: Sensitivity of ELISA is poor in this group also, even with crude antigens like lipoarabinomannan (LAM) or with A60 Ag (50% and 52% respectively) [18]. Diagnosis remains a major challenge here. Recently, Munshi et al [19] studied 100 clinically suspected cases of extrapulmonary tuberculosis using A60 antigen based ELISA and found specificity of 92 per cent and sensitivity of 100 per cent.

Primary tuberculosis and tuberculosis in children: Serodiagnoses using crude and semipurified antigens have been less fruitful. A 14 KD antigen has been tried without much success.

Monitoring of disease progression: During treatment, there is initially a rise in antibody titres due to release of antigens from killed bacilli [20]. Both specific and non-specific IgG responses may persist for many years [21]. Hence, ELISA should not be applied as a test of cure.

Antigen detection assays heve been used for antigen detection in sputum, bronchial lavage, serum and other body fluids. In general, there has been low sensitivity reflecting the difficulty in detecting the antigen in such samples. More recently, antigen capture ELISAs have been developed but are yet to be tried extensively. Desai et al [22] have tried to detect antigen in CSF using antibody enzyme conjugates against M. tuberculosis H37RV and M. bovis BCG strains. Both gave specificity of 100 per cent but positivity was 79.7 per cent and 67.5 per cent respectively. Diagnostic utility of the estimation of mycobacterial Ag A60 specific immunoglobulins IgM, IgA and IgG in the sera was tried by Gupta et al [23]. They concluded that a very high sensitivity (91.6%) and specificity (90.0%) can be obtained if combined IgA and IgG antibody titres are considered, to detect cases of adult tuberculosis. The role of IgM estimation can be restricted to the detection of cases of reactivation of tuberculosis.

It is unlikely that any test will displace the conventional smear examination for the bacilli. As yet, ELISA is not a reliable tool in diagnosis of tuberculosis. The available studies, however, point to its potential as an adjunct to conventional means of diagnosis. Further advances in ELISA formats may prove beneficial in areas where specimen collection is difficult, e.g. childhood tuberculosis, or where bacterial identification difficult such as in inaccessible walled-off lesions and in tuberculosis of CNS, bone, gastrointestinal tract and genitourinary tract.

ELISA in HIV infection

The first screening assay for HIV-1 became available for screening blood and blood products on Mar 1985 [24]. Today, the indications for HIV testing by ELISA has widened to include individuals with risk to contract HIV infection [25]. At present serological screening by ELISA forms the cornerstone for routine clinical and blood bank purposes.

Serologic response to HIV-1 infection : This can be understood by appreciation of the major protein products of the HIV genome [26]. The envelope (env) region has 2 glycoproteins, gp120 and gp41. Beneath the envelope, the 2 capsid proteins (gag) protecting the nucleic acid are p17 and p24. The polymerase and reverse transcriptase (pol) enzyme complex is intimately associated with the nucleic acid and contains proteins, p66, p51 and p31. It takes 4-6 weeks for anti-p24 and anti-gp41 antibodies to become detectable after infection. This is called the ‘serological-window’ period. In this period p24 antigen is detectable by ELISA. HIV-2 has gp36 envelope protein instead of gp41 of HIV-1. The advent of HIV-2 infection outside Africa has resulted in a strategic change in HIV testing policy which is to test for both HIV-1 and HIV-2 by combination ELISAs since 01 June 1992 as mandated by FDA [27].

Evolution of HIV ELISAs : The ‘first generation ELISA’ for HIV infection used whole viral lysates from cultures of HIV-1 from lymphocytes [28]. They suffered from the main limitation of a high rate of false positivity in low risk populations, due to reactions of serum antibodies to cellular contaminants such as those from major histocompatibility complex [29]. This problem led to the development of ‘second generation ELISAs’ that measure anti-HIV-1 antibodies using recombinant HIV-1 proteins of env and gag products and later to the ‘third generation assays’ that utilize chemically synthesized peptides corresponding to the highly conserved regions of gp41 (env) and p24 (gag) [30, 31, 32]. These newer assays have increased specificity, earlier detection of seroconversion or shortening of seroconversion window period and less number of indeterminate tests due to increased sensitivity of peptide based assays related to detection of anti-HIV IgG and IgM [28]. The various assay configurations are illustrated in Fig 2. Some theoretical limitations of the recombinant and synthetic EIAs include: lack of glycosylation of recombinant proteins which may limit interactions with serum antibodies, and lack of cross-reaction with HIV-1 variants. Their advantages are significant including safety, reproducibility of antigen content, and lack of contaminating cellular proteins that cause false positivity [33].

Fig. 2.

Showing three generation formats for HIV.

There are atleast 9 licensed standard EIAs available for detection of antibody to HIV [27]. While some use first generation viral lysates, 2 use recombinant antigens and the one made by United Biomedical Inc. (used at our centre) relies upon synthetic peptide antigens. Most EIA assays are indirect, while we have used competition assays (Veronostika) in the past.

Reporting of HIV results : No positive result from a single ELISA should be reported [34]. It is mandatory that a second sample has to be collected by the assayist himself. If the second ELISA, carried out by a test having different EIA principle, is positive the test result can be taken as confirmatory [34], in a developing country like ours where western blot (WB) facility may not be consistently available. When the repeat second test is by the same test format but on a second sample, it is necessary to confirm the test result at a referral centre. Following parameters should be sent to the referral centre for confirmation along with the second sample that tested positive : particulars of the patient that includes the risk group of the individual, ODs of the first ELISA and second ELISA tests, ratio of the sample OD to the cut-off OD and, if any preservative was added in the sample sent. Positive sera in an indirect/direct ELISA give a ratio of the sample to the cut-off OD of 1 or more, whereas this ratio is less than 1 in competition formats.

Sensitivity and specificity issues : In one study [28] on 575 sera the concordance between first, second and third generation assays reached 97.2 per cent with specificities 98.1 per cent, 99.9 per cent and 98.4 per cent respectively. The sensitivity and specificity for most manufacturers’ assays are 97.2-100 per cent and 99.6-100 per cent respectively, when using sera from high risk groups [35].

Specificity and positive predictive value of an ELISA test for HIV are dependent on the prevalence of HIV infection in the population tested. Number of false positives remain the almost the same per 1000 tests [27] for any population tested. To further comment on false positivity, American Red Cross data shows that among 2.58 million blood donors, only 3.5 per cent initially positive EIAs and only 13 per cent repeatedly reactive EIAs were reactive by WB.

The false positive and false negative results obtained in HIV-EIAs are due to many clinical and technical variables, specially seen in viral lysate EIAs [27]. False positive reactions can be caused by autoimmune disorders, IgM anti-HAV (hepatitis A virus), IgM anti-HBc (hepatitis B core), anti-HLA class II antibodies (observed more in multiparous and multiply transfused), alcoholic liver disease, primary biliary cirrhosis and sclerosing cholangitis, rapid plasma reagin test positivity of the serum, haemotological malignancies & lymphomas, renal transplantation and chronic renal failure and passively acquired HIV antibodies.

False negative reactions can occur due to incubation period or acute disease before seroconversion (window period), malignancy, long-term immunosuppresive therapy, replacement transfusion, bone marrow transplantation, kits that primarily detect anti-p24 antibodies, B-cell dysfunction and rheumatoid factor (in competition ELISA).

Future of ELISAs in HIV infection : The recombinant and synthetic protein EIAs for antibodies are endowed with enormous technical reliability. It has been aptly suggested that interposition of the recombinant or synthetic peptide assays could reduce the need for confirmatory WB assays by up to 47 per cent [28,36]. It may not be an overstatement to say that they may replace the expensive WB in days to come.