In this article we will discuss about:- 1. Assay Development 2. Assay Evaluation 3. Assay Validation.
The main developmental goals of an assay system are its desired accuracy, precision, sensitivity and specificity. Therefore, the foremost consideration in an assay development and optimization should be the intended application of the assay which may further affect the developmental goals. Immunoassays requires a longer time period for assay development in comparison to other bio-analytical techniques.
During the immunoassay development, the generation of the reagent antibody (monoclonal or polyclonal antibody) takes the longest time. It can take even months and thereafter also there is no confirmation that a desired reagent antibody will be produced. To save time and efforts, we can opt for commercially available kits or reagent antibody/antiserum which can be used directly and therefore the longest step can be skipped off.
Academia can serve as the best source of reagent antibodies for novel biomarkers. There are a number of immunoassays which needs to be developed on some specific formats.
A broad range of different aspects that must be considered is as follows:
1. Selection of Assay Format:
The three most commonly used assay formats for immunodiagnostic products include:
i. Antibody Capture Assays:
These assays are used to quantitate the levels of antigens as well as antibody and also to compare antibody binding site. In antibody capture assays, an antigen is initially immobilized on a solid phase. A diluted serum sample containing antibody is then added. The antibody binds the immobilized antigen. The antibody can then be detected by labeling directly or by using a secondary reagent that specifically recognizes the antibody. For example, Direct ELISA, Indirect ELISA.
ii. Antigen Capture Assays:
Some small molecular weight analytes cannot be detected by two-site-immunometric sandwich assays. Therefore, antigen capture assays are primarily used for their detection. In antigen capture assays are limited-reagent assays which have no limitations on the nature of the analyte as long as a specific antibody is available. In antigen capture assays, the amount of antigen in a test sample is quantified by using a competition between labeled and unlabeled antigen. For example, Competition ELISA.
iii. Two Site Immunometric Sandwich Assay:
For the detection and quantification of large analytes (like proteins), the two site immunometric assays are most commonly used in the immunodiagnostic industry. Such large analytes have multiple epitopes that are spatially separated. Therefore, these assays require two antibodies that can bind to discrete non-overlapping epitopes on the antigen. Such matched pair of antibodies has a high affinity and specificity. For example, Sandwich ELISA.
2. Preparation of Solid Phase:
Solid phase serves as an anchor to which all the reactants bind subsequently. Therefore, preparation of solid phase is the most important step-while developing solid phase enzyme immunoassays. If the preparation of solid phase has not been optimized properly, then no matter how well the subsequent stages of the assay work, the assay will fail in any case. So, the following points should be considered during the preparation of solid phase.
3. Coating Process:
The most commonly used immobilizing reagent on the solid phase is a protein. It may an antigen or antibody.
(b) Method of Immobilization:
Reagent immobilization can be done by any one of the following approaches:
(i) Direct or Non-Covalent Binding:
It is the easiest method in which the passive adsorption of the capture reagent takes place on the surface of the solid supports such as plastics, glass, etc. The passive adsorption is a natural and spontaneous process. The basis of the passive adsorption is the hydrophobic interactions between the surface and the sub-surface hydrophobic amino-acids residues on the solid surface. In some cases, charged groups can be induced on polystyrene by gamma – irradiation which further promotes the binding of the some other proteins on the solid support.
(ii) Indirect Method:
In indirect immobilization methods, a second antibody (for example, Streptavidin, Protein G, Protein A, etc.) is initially immobilized non-covalently on the solid support and then the first antibody is allowed to adsorb specifically on the bound second antibody. These indirect methods have unique advantages of specificity, sensitivity and reproducibility.
(iii) Covalent Coupling Methods:
Covalent coupling methods were developed to overcome the limitations of the direct and indirect immobilization methods. It led to the development of polystyrene plates with chemically reactive surfaces. We have uniformly coated surfaces in covalent coupling methods which provide increased assay signal and improved assay precision. Non-specific binding is also lowered in these methods. Still there are some disadvantages like the use of hazardous materials, limited stability, etc.
The concentration of the protein should be optimized as such that it can saturate the solid surface matrix. Practically, 1µg/ml of protein is generally used to saturate the solid surface. This concentration can be increased to 10 µg/ml as per needs.
5. Coating Buffer and pH:
The proteins bind to the solid support by mechanisms which involve charge and hydrophobicity. The charge on a protein molecule is a function of the pH of the buffer in which the protein is dissolved. The pH of the coating buffer is kept two units above the pI of the protein so that adsorption remains independent of the pH.
Most commonly used coating buffers are carbonate/bicarbonate buffers (pH = 9.6), phosphate and Tris buffers of pH 7.0 – 7.5. The concentration of the coating buffers ranges between 0.005 – 0.1 M. Buffers of low ionic strength are generally preferred. To avoid interference with the coating process, the buffers should also be free of any contaminating proteins.
6. Coating Conditions:
Uniform coating of the capture antibody on the solid support is the key to consistent and accurate performance of enzyme immunoassays. Solid phases can be stationary (microtitre plates, tubes, etc.) or mobile (agarose, latex beads, polystyrene balls, etc.). Stationary solid phase can be coated directly by adding an appropriate volume of the capture antibody solution whereas mobile phase can be coated by suspending and agitating in the capture antibody solution. Passive adsorption of the capture antibody is a diffusion controlled process.
Therefore, sufficient time should be given to saturate the surface of the solid support which will depend upon the concentration and temperature of the coating solution. Incubation for shorter periods may result in unequal coating or immobilization of the capture antibody.
Coating at 37°C for 1-2 hours is sufficient but it is generally recommended to coat for at least 16-24 hours at 4°C to obtain the best precision. The coefficient of variation (CV) should not be more than 5% and all the binding values should be within 10% of the mean.
After the immobilization of the capture antibody on the solid support, it is necessary to block the unoccupied sites on the solid support to prevent non-specific adsorption of the assay components.
An ideal blocking agent should not displace the immobilized component, should be able to bind to all the remaining protein binding sites, should not alter the immuno-reactivity of the immobilized component, should be inert, should not interfere to the selection signal. The blocking agents are broadly divided into the following categories – Proteins (BSA, Casein, Gelatin, Ovalbumin, etc.), Non-proteins (Polyvinyl alcohol, polyvinyl pyrrolidone) and Detergents (Triton X -100, Twenty 20).
8. Production of Reagent Antibody:
The key reagent in any immunoassay is the antiserum. The antiserum governs the sensitivity, specificity, accuracy and precision of the immunoassay. Production of monoclonal antibodies is best suited for large molecules to be used in two-site sandwich assays and polyclonal antibodies for small molecules. Production of monoclonal antibody involves a series of immunizations of mice with antigens for several weeks to enhance the activation and proliferation of the B-cells which produce antigen specific antibody titre. Spleen cells are then fused with myeloma cells and then cultured.
Once the desired antibody secreting cells are identified, the cells are expanded and the antibodies are harvested. This process takes nearly three months. But now-a-days, a novel repetitive, multiple site immunization strategy called RIMMS, can produce monoclonal antibodies in a month.
For the production of polyclonal antibody, animals are immunized with the immunogen and then booster doses are given several times. After attaining a suitable titre, the serum is harvested. Immunization and boosting is usually carried out intra-dermally, intra-muscularly or subcutaneously.
9. Selection and Production of Labels:
Selection and production of label is a labour intensive and technically difficult process. Therefore, commercial sources of labels are generally sought.
There are four main types of labels that are commonly used in pharmaceutical industries, which are:
(i) Radio Labels:
We can either radio label an antigen as in Radio immunoassay or the antibody as in immuno-radiometric assay (IRMA). Major radioisotopes used are 1251, 3H.
(ii) Enzyme Labels:
It is the most versatile and popular class of labels. Enzymes are covalently coupled to a protein which then leads to amplification of the signal. Now-a-days, a number of antibody-enzyme conjugates are available commercially that can be used as detection reagents. The enzyme labels used must have high immuno-reactivity, high specificity and low non-specific binding.
(iii) Fluorescent Labels:
It is a viable alternative to radio labels and enzyme labels for immunoassays with the potential for lower background values and greater sensitivity. For DELFIA (Dissociation Enhanced lanthanide Fluorescence Immuno-Assay) lanthanide metal ions (Europium) are used to label the molecule of interest.
(iv) Chemiluminescent Labels:
These labels are popular in immuno-diagnostic industry and supports clinical chemistry applications. Presently these labels are not widely used in conventional bio-analysis departments other than in Enzyme immunoassays to quantitate enzyme labels. For example, streptavidin-biotin system.
10. Preparation of Sample and Standards:
The sample and standard preparation steps are the largest source of errors. The sample extraction or dilution buffer should be as such that it provides efficient extraction of the analyte from the sample into a liquid phase and minimizes the background noise. The most commonly used diluents are assay buffers containing protein, zero standard, hormone free serum, stripped serum, etc. Standards are often prepared in analyte free human serum.
11. Assay Condition:
To obtain desirable results within given assay specifications, several factors have to be considered to optimize assay conditions which are as follows:
i. Incubation Temperature:
Lower temperature is generally recommended while performing assays for temperature-sensitive analytes. If the assay temperature is increased, it will increase the rate of the reaction between antigen and the antibody. It may also lead to proteolysis or other adverse-affects. The sensitivity, reproducibility and precision of the assay are improved by extended incubation at lower temperatures.
ii. Incubation Time:
Short incubation periods can be used for the assays of the analytes that occur at higher concentrations and longer incubation periods for low analyte concentrations to reach equilibrium. For maximum signal sizes, it is usually recommended to incubate for a sufficient time to allow all reactions to achieve equilibrium.
12. Buffers and Incubation pH:
If the antigen-antibody binding is pH dependent then the composition of the buffer and its pH will influence the antigen-antibody reaction. The efficiency of the binding reaction depends on the pH, ionic strength, and the presence of additives (detergents, carrier proteins, inhibitors, etc.).
Buffers should include factors that stabilize the immuno-reactants and promote their interactions and also increase the signal-to-noise ratio. The most commonly used buffer systems are phosphate buffered saline (pH = 7.2, 0.15M), Carbonate buffer (pH = 9.6, 0.05M), Tris buffered saline (pH = 7.5, 0.1M), etc.
13. Reagent Mixing:
Uniform speed of mixing is important particularly for non-equilibrium immunoassays to obtain good assay precision. Mixing of the reactants also reduces the incubation time because the rate of the reaction is proportional to the diffusion distance raised to the inverse of third power.
14. Separation Steps:
The binding reaction is greatly influenced by the separation method used. There are three factors for the solid phase immunoassays – Composition of wash solution, gentle and vigorous shaking; and aspiration versus hand decanting. The wash solution should always contain a physiological or enzyme friendly buffer. Vigorous washing can inactivate the bound enzyme label or strip off the bound complex. Hand decanting often gives better results.
15. Color Development:
This step involves selection of suitable substrates, timing of reaction and selection of developmental conditions. For example, 3, 3’, 5, 5’-tetramethylbenzidine (TMB) is suited for kinetic analyses because of its rapid reaction rates with peroxidase. It also shortens the incubation time and results in greater assay sensitivity.
After the development of a new immunoassay, technical evaluation is the first step in the assessment of its performance. The evaluation methods vary with the varying assay format. The assay should always be tested with respect to sensitivity, precision, accuracy and specificity. It should be compared with a reference method and should also be tested for batch to batch variation. The main aim of clinical evaluation is to assess the ability of a system to provide accurate test results in a timely fashion.
Some of the important parameters for clinical evaluations are described below:
The most important aspect of an immunoassay performance is its precision. It is sometimes referred to as reproducibility which is a statistical measure of the variation between repeated determinations on the same sample. If a method cannot give reproducible results in a given laboratory then the method will not be acceptable for routine use. The estimate of the true precision will improve with the number of replications.
Therefore, a precision profile of the assay should be prepared with 15-20 replications at every standard concentration. This precision profile will contain useful information on the working range of the assay which defines the range of analyte concentrations over which measurements have sufficient precision for a given application. Especially for clinical decisions, high precision is important. There are some sources of imprecision in immunoassays such as pipetting errors, binding reaction, reagents, color reaction and detection, data processing, etc.
Accuracy measures the true value of the analyte. In analytical measurements, accuracy is defined as how close the average measured value is to the true value of the analyte. Accuracy is affected by every component of the immunoassay such as cross reactivity, lack of specificity, etc. Practically, accuracy is evaluated in terms of inaccuracy or bias which is a measure of the difference between the measured and the true value. Assay bias may be positive or negative.
Examples of bias in immunoassay measurements are:
Curve A – Perfect correlation between reference and test method.
Curve B – Constant error in measurement.
Curve C – Proportional error in measurement.
Sensitivity of an assay means the smallest amount of an analyte that can be detected under the specific assay conditions. The Lower Detectable Dose (LDD) is commonly used to define sensitivity. It is measured by assaying 10-20 replicates of the zero standards and then calculating the mean and standard deviation. The mean is used for the standard curve and the response, mean ± 2 standard deviations, read in dose from the standard curve is LDD. It is also referred to as analytical sensitivity.
4. Specificity (Cross Reactivity):
The accuracy of an assay system also depends on the specificity of the assay. Assay specificity is defined as the ability to detect a specific or particular analyte in a heterologous mixture which will depend on the antibody properties. Every antibody has a unique specificity for a particular epitope.
The structural similarity of epitopes on different antigens can create true site-specific competition between the antigens and the labeled analyte used in the assay. This competition is known as cross-reaction. Therefore, cross-reactivity is the measurement of antibody response to substance other than the analyte of primary interest.
Percent cross-reactivity = Concentration of standard at 50% B/Bo (S)/Concentration of cross-reactant giving 50% B/Bo (C) x 100
Once the desired assay system has been optimized, the validation can commence. Validation criteria revolve around the key concepts of precision and accuracy. Biological or clinical validation is concerned with demonstrating that the procedure returns results that are biologically, physiologically and diagnostically relevant.
The fundamental parameters for the bio-analytical method include accuracy, selectivity, precision, stability, reproducibility and sensitivity. The acceptability of analytical data corresponds directly to the criteria used to validate the method.
Different types and levels of validation are characterized as follows:
1. Full Validation:
It is important while developing and implementing a bio-analytical method for the first time, for a new drug entity or if metabolites are added to an existing assay for quantification.
2. Partial Validation:
It includes modifications of already validated bio-analytical methods.
It is a comparison of the validation parameters when two or more bio-analytical methods are used to generate data within the same study or between different studies.
Some of the important concepts involved in the process of clinical validation of immuno-diagnostic kits are as follows:
It is also known as reference value or normal range. It must be established for every new developed assay system.
The reference range should address the following three specific issues:
i. How to define the normal population?
ii. What would be the sample size?
iii. What statistics to apply?
Predictive Value of a Diagnostic Test:
Clinical sensitivity, specificity and efficiency are used to give the predictive value of a diagnostic test. Clinical sensitivity refers to the ability of the diagnostic test to detect the patients suffering from a disease. When the diagnostic test is positive, the result is said to be true positive (TP).
Efficiency = [(TP + TN)(TP + FP + TN + FN)] x 100
Clinical sensitivity or the positivity in the disease can be calculated as:
Clinical sensitivity = TP/(TP + FN) × 100
Clinical specificity refers to the diagnostic test that correctly identifies the patients who do not have the disease. The false positive (FP) are the number of healthy patients who are mis-classified by the test and true negatives (TN) are the patients correctly classified by the test.
Clinical sensitivity = TN/(TN + FP) × 100
i. Therefore, the predictive value of a positive test gives the percentage of patients suffering from the disease with positive test results and is given by –
TP (TP + FP) × 100
ii. The predictive value of a negative test is defined as the percentage of healthy patients with negative test results and is given by –
TN (TN + FN) × 100
iii. The overall efficiency of a diagnostic test defined as the percentage of patients correctly classified as diseased or healthy is given by –
Efficiency = [(TP + TN) (TP + FP +TN + FN)] × 100
Receiver Operating Characteristic (ROC) Curve:
Clinical sensitivity and specificity occur in pairs. A given test may have one set of sensitivity- specificity pair in one clinical situation, but it may have a different set of pairs when applied to different clinical situations. These two parameters can be evaluated in terms of the receiver operating characteristic curve. The ROC curves are constructed by plotting the relationship between the false-positive rate (specificity) and the sensitivity at various cut-off values.
These curves display assay performance over the entire decision levels. The ROC curve is extremely valuable for the comparison of two competing tests and in assessing the clinical utility of screening tests where the specificity is low. Therefore, by superimposing the ROC curves of more than one test method, we can select the best method for a given clinical need. A better test is one that displays a higher TP rate and a lower FP rate.