Enzyme-Linked Immunosorbent Assay

I. INTRODUCTION
The enzyme-linked immunosorbent assay (ELISA) (Engvall and Perlmann, 1971) is a highly versatile and sensitive technique that can be used for qualitative or quantitative determinations of practically any antigen or antibody (Berzofsky et al.,1999). Reagents are stable, nonradioactive, and, in most cases, available commercially. Its use ranges from testing of individual samples to fully automated systems for high throughput screening. In one of its simplest forms, the assay involves immobilization of one reagent (e.g., antigen) on a plastic surface, followed by the addition of test antibodies specific for the antigen and, after washing, enzyme-conjugated secondary antibodies against the test antibodies. The addition of substrate giving coloured, fluorescent, or luminescent reaction products makes it possible to determine the concentrations of the reactants at very low levels (Butler, 1994). Depending on the quality of the reagents used and the choice of substrate, sensitivities in the picogram or subpicogram per milliliter range can be obtained.

Of several enzymes suitable for ELISA, alkaline phosphatase (ALP) and horseradish peroxidase (HRP) are the most commonly used. Various methods of enhancing sensitivity may be employed, most of which, like the commonly used biotin-streptavidin system, are designed to amplify the signal by increasing the amount of enzyme bound (Ternynck and Avrameas, 1990).

This article provides three examples of ELISA protocols: an indirect ELISA to determine antibodies (the prototype for many serological assays) and two sandwich or catcher ELISAs designed for the detection of antigen and antibodies, respectively. For the many other variants and applications of ELISA, the literature should be consulted (e.g., see Butler, 1994; Maloy et al., 1991; Mark-Carter, 1994; Ravindranath et al., 1994; Zielen et al., 1996).

One important modification of the ELISA is the ELISpot, which instead of measuring an analyte in solution measures it at the site of a producing cell. This is made possible by using a precipitating rather than a soluble substrate, with the result being a visible imprint or spot, each representing an individual, producing cell. Enumeration of the spots gives the frequency of producing cells, which may be as low as 1 in 100,000 cells. Due to its very high sensitivity, the ELISpot is particularly well suited for measuring specific immune responses and it was originally developed for the detection of immunoglobulin production by specifically stimulated B cells (Czerkinsky et al., 1983; Sedgwick and Holt, 1983). However, today it is mainly used for the analysis of specific T-cell responses where the induced production of cytokines by antigentriggered T-cells is exploited (Lalvani et al., 1997; Larsson et al., 2002). Depending on the cytokine analysed the test may, apart from the number of responding cells, also give information about the type of responding cell (e.g., CTL, Th-1, or Th-2).

II. MATERIALS AND INSTRUMENTATION
A. Elisa
Flat-bottomed microtiter plates: Maxisorp from Nunc A/S or High Binding from Costar (Cat. No. 3590)
Round-bottomed microtiter plates for preparation of dilutions
Micropipette, multichannel pipette (Cat. 4540-500), and disposable pipette tips (Finnpipette)
Microplate washer (Skatron Instruments)
V_max kinetic microplate reader with computer program SOFTmax (Cat. No. 79-200 105, 79-200 100)

B. ELISpot
96-well PVDF (polyvinylidene) filter membrane plates (Millipore Corporation)
Sterile plastic vials for the handling of cells and preparation of dilutions
Cell culture medium RPMI 1640 with 10% fetal calf serum (FCS, Invitrogen)
Cell incubator with 5% CO₂ atmosphere
Dissection microscope (40×) or ELISpot reader (AID, Autoimmun Diagnosticka GmbH)

III. PROCEDURES
A. Biotinylation of Immunoglobulin
See Ternynck and Avrameas (1990).

Solutions

1. 0.1M NaHC0₃:84.01 g/1000 ml H₂O, pH 8.4
2. Biotin: Sulfo-NHS-LC-biotin [sulfosuccinimidyl-6- (biotinamido)hexanoate] (Cat. No. 2135) (Pierce)

Steps

1. Dialyze affinity-purified polyclonal antibodies or purified monoclonal antibodies against 0.1M NaHCO₃ at 4°C overnight and adjust to 2mg/ml.
2. Immediately prior to use, dissolve 1 mg of biotin in deionized water and add the biotin to the antibody solution at a molar ratio of 20:1. Incubate for 30-60 min.
3. Dialyze overnight at 4°C against phosphatebuffered saline (PBS) with 0.02% NaN₃.

B. Optimal Reagent Concentrations

1. Concentration of coating reagent: Most plates designed for ELISA have a protein-binding capacity of 300-600ng protein per well (96-well plates). In practice, if using purified antigens, affinity-purified polyclonal antibodies, or monoclonal antibodies, coating with 1 µg/ml (100 µl/well) is normally sufficient. Most proteins coat well in PBS but, if there are indications of poor coating efficiency, buffers with different pH should be tried. This may be particularly important when using smaller peptides for coating.
2. Concentration of test reagent: If possible, use one known positive, one known negative, and, if available, a standard. Find the concentration where the known positive sample gives a good positive reading and responds to dilution.
3. Concentration of developing reagent: Choose the lowest concentration assuring excess; the only limiting factor in the setup should be the amount of test reagent. Typically, the detecting reagents (e.g., biotinylated or enzyme-conjugated secondary antibody) are used at concentrations around 1 µg/ml but should be determined by titration.

C. ELISA and ELISpot Protocols
Solutions

1. Coating buffer: PBS, pH 7.4: 100ml PBS stock (10×), 200mg NaN₃, and H₂O to 1000ml or, alternatively, 0.1M NaHCO₃ pH 9.6: 1.59g Na₂CO₃, 2.93g NaHCO₃, 200mg NaN₃, and H₂O to 1000ml
2. Incubation (diluent) buffer: 100ml PBS stock (10x), 5g bovine serum albumin (BSA), 0.5ml Tween 20, 200 mg NaN₃, and H₂O to 1000 ml
3. Washing buffer: 45g NaCl, 2.5ml Tween 20 (0.05%), and H₂O to 5000 ml
4. Streptavidin-ALP: Sigma-Aldrich
5. Enzyme substrate buffer: 97ml diethanolamine, 800ml H₂O, 200mg NaN3, and 101mg MgCI₂·6H₂O (should be added last). Adjust finally to pH 9.8 with 1M HCl (~100ml).
6. Substrates for ALP: p-nitrophenyl phosphate (NPP), 5-mg tablets (Sigma-Aldrich), 1 tablet/5ml of enzyme substrate buffer (ELISA). BCIP/NBT-Plus (Moss, Inc.)

1. Indirect ELISA for Screening of Specific Antibodies in Serum or Hybridoma Supernatants
Reactants

1. Antigen
2. Antibody specific for test antigen
3. Anti-immunoglobulin enzyme conjugate
4. Substrate

Steps

1. Coat plate with 100µl/well of antigen, diluted in PBS, pH 7.2, overnight at 4°C.
2. Block with 100µl/well of incubation buffer for 1 h at room temperature (37°C). Wash four times with washing buffer.
3. Add 100µl/well of immune serum, e.g., diluted 1/1000, or hybridoma supernatant, e.g., diluted 1/5, in incubation buffer and incubate for 2h at room temperature. Wash four times with washing buffer.
4. Add 100µl/well of ALP-conjugated antiimmunoglobulin (e.g., goat antihuman γ chains, rabbit antimouse Ig) diluted in incubation buffer. Incubate for 1 h at room temperature (37°C)
5. . Wash four times with washing buffer.
6. Develop with fresh NPP, 100µl/well, and read absorbance at 405 nm.

2. Sandwich ELISA for Detecting Antigens
Reactants

1. Capture antibody specific for test antigen
2. Antigen
3. Biotinylated antibody specific for test antigen; See Section IV
4. Streptavidin-ALP
5. Substrate

Steps

1. Coat plate overnight at 4°C with 100µl/well of antigen-specific monoclonal or polyclonal antibody diluted in PBS.
2. Block with 100µl/well of incubation buffer for 1 h at room temperature. Wash four times with washing buffer.
3. Add 100µl/well of test reagent (e.g., cell culture supernatant or diluted serum sample) and, if available, a standard antigen in serial dilutions. Incubate for 1 h at 37°C. Wash four times with washing buffer.
4. Add 100µl/well of biotinylated monoclonal antibody specific for a different determinant of the antigen or biotinylated polyclonal antibody of the same antigen specificity as used for coating (see Section IV) diluted in incubation buffer. Incubate for 1 h at 37°C. Wash four times with washing buffer.
5. Add 100µl/well of ALP-conjugated streptavidin in incubation buffer for 1 h at 37°C. Wash four times with washing buffer.
6. Develop with fresh NPP, 100µl/well, and read absorbance at 405 nm.

3. Sandwich ELISA for Detection of Specific Antibodies
Several recently developed serological assays are similarly based on the sandwich ELISA principle but use antigen rather than antibodies for coating and detection. Exploiting the fact that antibodies, also after binding to an antigen immobilized on an ELISA plate, usually retain a free antigen-binding site (IgG) or sites (IgM), labeled antigen may be used as a detecting agent. Being less prone to background problems often encountered in conventional serological assays it is more reliable and, as it allows the testing of less diluted samples, also more sensitive.

Reactants

1. Capture antigen to which antibody reactivity is to be tested
2. Biotinylated form of the same antigen
3. Streptavidin-ALP
4. Substrate

Steps

1. Coat plate overnight at 4°C with 100µl/well of relevant antigen in PBS.
2. Block with 100µl/well of incubation buffer for 1 h at room temperature. Wash four times with washing buffer.
3. Add 100µl/well of serum to be tested. Incubate for 1 h at 37°C. Wash four times with washing buffer.
4. Add 100µl/well of the same antigen as used for coating in biotinylated form (see Section IV) diluted in incubation buffer. Incubate for 1 h at 37°C. Wash four times with washing buffer.
5. Add 100µl/well of ALP-conjugated streptavidin in incubation buffer for 1 h at 37°C. Wash four times with washing buffer.
6. Develop with fresh NPP, 100µl/well, and read absorbance at 405 nm.

4. Cytokine ELISpot
Different types of cells can be analysed in the ELISpot. Typically, for the analysis of specific T-cell responses in humans, peripheral blood mononuclear cells (PBMC) containing both T-cells and antigen presenting cells are used.

Reactants

1. Antibody specific for the test cytokine
2. Biotinylated anticytokine antibody: Of noncompeting specificity
3. Streptavidin-ALP
4. Substrate

Steps

1. Before coating with antibody, activate the PVDF membrane by adding 100µl/well with 70% EtOH and incubate for 1-2min. Wash plates with 4 × 200 µl sterile-filtered PBS and, without letting the membrane dry, add 100µl of antibodies against test cytokine (15µg/ml in sterile PBS) and incubate overnight at 4°C.
2. Remove excess antibody by washing (4 × 200 µl/ well) and block the membrane by adding 150µl/well with serum containing (e.g., 10% FCS) tissue culture medium. Incubate for 1 h at room temperature.
3. Add PBMC together with the antigen to be tested in triplicate. Use cells without antigen as control for spontaneous cytokine production and a polyclonal activator (e.g., 1-10µg/ml of phytohemagglutinin) as a positive control for activation. The suitable number of cells analysed should be determined for each situation but do not add more than 300,000 cells/well, as this will result in multiple cell layers leading to diffuse spots. Depending on the kinetics of the cytokine analysed, incubate for 12-48 h at 37°C in a cell incubator with a humid atmosphere and containing 5% CO₂. Do not move the plates during incubation.
4. Wash the plates with filtered PBS (4× 200 µl/well).
5. Add 100µl of biotinylated anticytokine antibody in filtered PBS. Incubate for 2h at room temperature.
6. Wash as described earlier and add 100µl streptavidin-ALP, diluted 1/1000 in filtered PBS. Incubate for 1 h at room temperature.
7. Wash as described previously, add substrate (BCIP/NBT-plus), and incubate at room temperature until dark spots emerge (10-30min). Stop color development by washing with 4 × 200µl/well of tap water.
8. Leave plates to dry and count spots in a dissection microscope or an ELISpot reader.
9. Store plates in the dark at room temperature (Fig. 1).

FIGURE 1 PBMC from a vaccinated donor was exposed in vitro for 16h to the antigens purified protein derivative (PPD) or tetanus toxoid (TT) and investigated for the presence of IFN-γ-producing cells.

IV. COMMENTS
A. Blocking
1. ELISA
Aftercoating is required to block vacant proteinbinding sites on the plastic surface. BSA, casein, milk powder, or gelatin is commonly used. To avoid crossreactive antibody binding to the blocking protein, the same protein as used for dilution of the reagents should be included in the incubation (diluent) buffer.

2. ELISpot
Although the filter membranes used in ELISpot have a much higher binding capacity than ELISA plates (up to 100 times higher binding), aftercoating is not always required due to the blocking effect of serum components in the cell culture medium. However, to avoid possible triggering of cell surface receptors through interactions with the highly adsorptive membrane, blocking for 1 h in culture medium before the addition of cells is usually recommended.

B. Controls
1. ELISA
Wells containing substrate but no reagents serve as a general background to be subtracted from all measured values. All samples are set up in duplicates. Negative controls with incubation buffer replacing the reagents to be tested should always be included and should give readings well below OD 0.100. Particularly in the two-site sandwich applications, it is essential that the capture antibody does not bind to the second antibody and vice versa. For screening of unknown samples, and comparison between different runs, include a known positive sample as a reference. For estimation of background, include expected negatives, e.g., nonimmune sera, cell culture medium, or irrelevant antigen.

2. ELISpot
Controls usually comprise wells with unstimulated cells, giving the frequency of spontaneously producing cells. In the case of cytokine-producing cells, the levels of spontaneous production may differ significantly for different cytokines and are usually higher in individuals with acute infections. Positive controls in the form of polyclonally activating reagents (e.g., PHA or anti-CD3) or specifically activating peptides from common infectious agents can be used (Currier et al., 2002).

C. Incubations
Coating the plates overnight is often practical and the coated plates can be stored for several weeks at 4°C, wrapped in plastic film. Coating for 3-4 h at room temperature or 1 h at 37°C may often be enough. The same holds true for the specific binding of the reagents. The signal may be increased significantly by longer incubations but shorter incubations at 37°C may suffice as well. Development of colour with the substrate varies in time but requires usually between 10 and 60min depending on the enzyme system used.

D. Amplification
Amplification of the signal in ELISA may be obtained by various means. Using an extra layer such as enzyme-conjugated anti-immunoglobulin as developing reagent or biotinylated antibody followed by the streptavidin-enzyme conjugate can lead to increased sensitivity. Similarly, using fluorescent (Rodriguez et al., 1998) or chemiluminscent (Tatsumi et al., 1996; Fukuda et al., 1998) substrates instead of chromogenic substrates, sensitivity can be improved 10- to 100-fold. Due to the increased sensitivity, amplified assays are often more prone to variations and background problems and usually require more careful handling and extensive washings.

E. Sandwich ELISA
Antibody sandwich ELISAs are sensitive and very useful for the detection of antigen, e.g., cytokines in cell culture supernatants. To work, the analysed substance needs to have at least two separate binding sites for antibodies and is therefore not amenable to the analysis of small molecules. Both polyclonal and monoclonal antibodies may be used. As an example, for determining human interleukin (IL)-4 we use monoclonal antibodies of two different specificities against IL-4. For immunoglobulin isotype determinations in human serum or lymphocyte culture supernatants we use goat or rabbit antibodies made highly Fc specific by affinity purification. The same polyvalent antibody preparation can be used both as capture antibody and as enzyme-conjugated or biotinylated second antibody. For IgG subclass determinations, we use subclass-specific monoclonal antibodies as capture antibodies and Fc-specific, affinity-purified (depleted of antimouse Ig reactivity) goat antihuman IgG as the secondary antibody. For quantitation of isotypes or IgG subclasses, standard immunoglobulin preparations or myeloma protein solutions of known concentrations are available commercially.

F. Quantitation
For quantitation of specific antibodies, a standard curve with serial dilutions (e.g., 300, 100, 30, 10, 3, and l ng/ml) of a relevant standard immunoglobulin is prepared in wells coated with affinity-purified antiimmunoglobulin instead of the antigen. The linear range from a log/log curve is used for interpolation of the experimental values. Similarly, the amount of antigen can be determined in a sandwich ELISA with the help of a standard curve with known amounts of the antigen run in parallel. One standard curve can be used for several plates if care is taken that all plates are developed at the same time.


V. PITFALLS
A. Background
In both ELISA and ELISpot, a high purity and specificity of the reagents are basic requirements for reliable determinations. As much of the sensitivity of the assays depends on having a low background, low readings of the negative controls are absolutely essential. With appropriate controls it is usually possible to identify reagents giving rise to unwanted binding of the enzyme conjugate. However, this is not always the case as, when analyzing serum/plasma samples in sandwich ELISA, the presence of naturally occurring anti-immunoglobulin in some individuals may lead to cross-linking of the coating and detecting antibodies and thereby cause false-positive signals not easily discernible from a true positive signal. This effect can be prevented or minimized by the inclusion of irrelevant blocking antibodies in the incubation buffers.

Another phenomenon that may affect both ELISA and ELISpot is the so-called "edge effect," which refers to the often lower reproducibility obtained in the outer wells of a 96-well plate. Being at least partly due to a greater evaporation in these wells, it is essential that all incubations are performed in a way that evaporation is minimized.

A particular background problem that may occur in the ELISpot assay is the presence of artifactual spots or spot-like formations. Normally caused by debris from the cells or the buffers, the problem may be largely alleviated by careful washing of the plates and the filtration of the buffers and substrate used. As the artifactual "spots" are usually smaller and lack the diffuse rim that is characteristic for true spots, they can usually be discriminated from real spots when evaluated in the microscope. Similarly, in most of the existing ELISpot readers a similar function to discriminate artifacts and true spots has been incorporated. In the ELISpot, it is also important that the plate is not moved or shaken during incubation in the cell incubator, as this may result in "blurry" spots and/or an uneven distribution of spots in the wells.

B. Adsorption-Induced Protein Denaturation
Lost functional activity of antibodies is not an uncommon problem and can seriously affect the sensitivity of an assay. As much as 90% of the activity of polyclonal antibodies and all for some monoclonal antibodies can be lost due to the absorption to a solid support (Butler, 2000). Similarly, a changed conformation of antigens at coating may lead to masking of native epitopes, as well as to exposure of "new" epitopes. Varying the binding conditions by using coating buffers of different pH, the addition of stabilizing agents, or the use of alternative binding strategies (e.g., binding biotinylated antibodies to immobilized streptavidin) may help solve this type of problem.

References
Berzofsky, J. A., Berkower, I. J., and Epstein, S. L. (1999). Antigen-antibody interactions and monoclonal antibodies. In "Fundamental Immunology" (W. E. Paul, ed.), 4th Ed., pp. 88-91, Lippincott-Raven, Philadelphia.

Butler, J. E. (1994). Enzyme linked immuno-sorbent assay. In "Immunochemistry" (Oss C. J. Van, Regenmortel Van, eds.), pp. 759-803. Dekker, New York.

Butler, J. E. (2000). Solid supports in enzyme-linked immunsorbent assay and other solid-phase immuno-assays. Methods 22, 4-23.

Currier, J. R., Kuta, E. G., Turk, E., Earhart, L. B., Loomis-Price, L., Janetzki, S., Ferrari, G., Birx, D. L., and Cox, J. H. (2002). A panel of MHC class I restricted viral peptides for use as a quality control for vaccine trial ELISPOT assays. J. Immunol. Methods 260, 157-172.

Czerkinsky, C. C., Nilsson, L. A., Nygren, H., Ouchterlony, O., and Tarkowski, A. (1983). A solid-phase enzyme linked immunospot (ELISPOT) assay for enumeration of specific antibody secreting cells. J. Immunol. Methods 65, 109-121.

Engvall, K., and Perlmann, P. (1971). Enzyme-linked immunosorbent assay (ELISA): Quantitative assay of immunoglobulin G. Immunochemistry 18, 871-874.

Fukuda, S., Tatsumi, H., and Maeda, M. (1998). Bioluminescent enzyme immunoassay with biotinylated firefly luciferase. J. Clin. Ligand Assay 21,358-362.

Lalvani, A., Brookes, R., Hambleton, S., Britton, W. J., Hill, A. V., and McMichael, A. J. (1997). Rapid effector function in CD8+ memory T-cells. J. Exp. Med. 186, 859-865.

Larsson, M., Wilkens, D. T., Fonteneau, J. E, Beadle, T. J., Merritt, M. J., Kost, R. G., Haslett, P. A., Cu-Uvin, S., Bhardwaj, N., Nixon, D. E, and Shacklett, B. L. (2002). Amplification of low-frequency antiviral CD8 T cell responses using autologous dendritic cells. AIDS 25, 171-180.

Maloy, W. L., Coligan, J. E., and Paterson, Y. (1991). Indirect ELISA to determine antipeptide antibody titer. In "Current Protocols in Immunology" (J. E. Coligan, A. M. Kruisbeek, D. H. Margulies, E. M. Shevach, W. Strober, eds.), pp. 9.4.8.-9.4.11. Greene and Wiley-Interscience, New York.

Mark-Carter, J. (1994). Epitope mapping of a protein using the Geysen (Pepscan) procedure. In "Methods in Molecular Biology" (B. M. Dunn, M. W. Pennington, eds.), Vol. 36, pp. 207-223.

Ravindranath, M. H., Ravindranath, R. M. H., Morton, D. L., and Graves, M. C. (1994). Factors affecting the fine specificity and sensitivity of serum antiganglioside antibodies in ELISA. J. Immunol. Methods 169, 257-272.

Rodriguez, C. D., Fei, D. T., Keyt, B., and Baly, D. L. (1998). A sensitive fluorometric enzyme-linked immunosorbent assay that measures vasculare endothelial growth factor16s in human plasma. J. Immunol. Methods 219, 45-55.

Sedgwick, J. D., and Holt, P. G. (1983). A solid phase immuno enzymatic technique for the enumeration of specific antibodysecreting cells. J. Immunol. Methods 51, 301.

Tatsumi, H., Fukuda, S., Kikuchi, M., and Koyama, Y. (1996). Construction of biotinylated firefly luciferases using biotin acceptor peptides. Anal. Biochem. 243, 176-180.

Ternynck, T., and Avrameas, S. (1990). Avidin-biotin system in enzyme immunoassays. In "Methods in Enzymology" (M. Wilchek E. A. Bayer, eds.), Vol. 184, pp. 469-581. Academic Press, San Diego.

Zielen, S., Broker, M., Strnad, N., Schwenen, L., Schon, P., Gottvald, G., and Hofmann, D. (1996). Simple determination of polysaccharide specific antibodies by means of chemically modified ELISA plates. J. Immunol. Methods 193, 1-7.