How to Choose and Optimize ELISA Reagents and Procedures

ELISAs have been used for detection of biological molecules since the 1970s. Over time, scientists have developed and utilized many variations of this assay format. The various types of ELISAs (Antibody detection, Antigen capture, Sandwich, Competitive, etc.) are really variations on a theme. The overarching themes of ELISAs are to specifically and sensitively detect a biological molecule (antibody, cell receptor, cytokine, etc.) in a biological matrix. Virtually any type of molecule (protein, lipid, carbohydrate, nucleic acid, etc.) can be detected by an appropriately developed ELISA.

Briefly, the ELISA works by attaching a capture moiety to a microtiter plate, allowing a sample to bind, and detecting the sample via another binding moiety coupled to a signal generating molecule. The capture and detector components are usually added at saturating concentrations and allowed to come to binding equilibrium. Excess unbound material is washed out of the wells in between each assay step. The sample concentration will, of course, be variable. This article will describe general concepts to keep in mind as you choose and optimize ELISA reagents and procedures.

In this day and age it is usually possible to find many commercially available ELISA reagents. These include a wide variety of antibodies to many different biologically important molecules, multiple buffers, and a wide assortment of detection and signal generation molecules. However, sometimes appropriate reagents are not available and will need to be developed. The rest of this article will assume the development of an antigen capture ELISA for detection of a serum cytokine. In general try to maximize desired binding events while minimizing the unintended binding events. Optimization of reagents is accomplished by maximizing “signal to noise” ratios between positive and negative samples. Using the example assay, key areas that need attention in the development of every ELISA assay will be addressed.

The Target

First and foremost one must understand the target of the ELISA. By this I mean more than just what the target is, although that is definitely an important first step. One must understand the biological and chemical characteristics of the target molecule. Both aspects of a target cytokine can affect the ELISA. Understanding the chemical nature of the target molecule is essential to choosing appropriate sample buffers to obtain, store, and test the target molecule. Obtaining some purified target material will assist with preliminary stability testing in likely buffers and matrices. On the biological front, some easy questions, which might not have easy answers, need to be asked. First, how much target cytokine is likely to be present in the desired sample? If there is very little target present, such as picograms or less, then one will need to spend time and effort to find or develop sensitive capture and detector reagents. If the target is present at milligram quantities or greater, then it will probably be easier to find or develop the capture and detection reagents.

Secondly, how does the target cytokine exist in its normal matrix? Does the target have any interactions with other biological molecules and to what extent? Does it exist as whole molecules or as various metabolic fragments? Cytokines frequently exist in serum bound to other molecules. These other binding partner molecules may confound the binding of capture and/or detection reagents to the cytokine. Capture and detector reagents that work in the presence of these target binding partners need to be carefully chosen or developed to accurately detect the target cytokine. If the target can exist in the sample matrix as a whole molecule along with some number of metabolic fragments, then internal competition of fragments with whole molecules may occur within the ELISA. One fragment may bind nicely to the capture reagents, but not present a suitable epitope for the detector reagents. Many variations of this problem will impact the accuracy of the ELISA. In essence, one needs to develop and optimize the capture and detector reagents as a pair, not as stand-alone molecules.

The Capture

Assuming one has the target identified and well characterized; it is time for a first go at the cytokine detection assay. It is necessary to line up the anticipated capture antibody, samples, detector antibody, signal generator, and all the appropriate buffers to go along with them. Choose an assay format that can fit with the available laboratory tools and instrumentation (i.e. pipetters, plate washers, plate readers, etc.). The 96 well polystyrene microtiter plate is generally a great place to start. There are many variations of microtiter plate with different binding characteristics. For initial attempts a generic ELISA plate is recommended. Once the assay is established, one can decide if the more expensive high binding plates will improve the assay or not.

The cytokine antigen assay will utilize an anti-cytokine polyclonal antibody (PAB) to capture the target cytokine. The choice of capture reagent will change depending on the needs of each assay. It is important to optimally attach the capture PAB to the plate. This can usually be accomplished via passive adsorption. A fairly minimalist buffer such as phosphate buffered saline (PBS pH 7.4), or carbonate/bicarbonate buffer (pH 9.5), is typically used to dissolve the capture PAB. Adjusting the pH is an easy way to alter the charge on the capture PAB or other protein that one may be adsorbing to a plate. Try a few different concentrations of capture PAB because the buffer/antibody/plate combination may vary in binding capacity. A good PAB concentration to bracket around is 1 ug/ml. This may vary for other proteins. It is recommended that you test the binding of the PAB to the plate by completing the assay. Since the rest of the assay has yet to be optimized you need to use reasonable reagent concentrations and chemistries for choosing the remaining reagents. These will subsequently be optimized. Initial capture PAB binding to the microtiter plate should incubate for at least 1 hour. Incubation times may need further optimization to ensure complete binding has occurred to the microtiter plate.

The Wash

After the capture PAB incubation one needs to utilize a wash buffer. The purpose of the wash buffer is to remove assay components that are not bound to the plate in an intended fashion. Wash buffers are usually formulated at physiologic pH and salt concentrations (e.g. PBS). The goal is to be gentle and prevent denaturation of any assay components while at the same time removing excess material not specifically and tightly bound to the well. To accomplish this it may be necessary to add some mild detergents to the wash buffer. A good place to start is with Tween-20, or something similar, at 0.5% or thereabouts.

The type of detergent and its concentration must be compatible with all assay reagents. Too much detergent, or the wrong one, may wash capture PAB off the plate or inhibit binding of assay components. Sometimes additional proteins may be added to help stabilize assay components. It will be necessary to do some empirical testing to make sure of complete assay compatibility. For ease of use a single wash buffer is generally used for a given ELISA. Rarely certain assay reagents may require different wash buffers. Understanding the chemistry and biology is a must. The proper wash will greatly stabilize the assay leading to sensitive and accurate results.

The number of washes is arrived at empirically. In essence, washing is just diluting the soluble components of the well. Think about how much of a dilution is needed at each step in the ELISA. One needs to balance washing out the undesired components with reducing the amount of the desired components in the assay. Stability of the wash fluid is often overlooked, especially in laboratories with low throughput. Wash fluids can begin to grow unwanted “things” if left sitting. Adding a preservative or ensuring the wash is made fresh for each assay will likely prevent any stability problems. Don’t forget plate washer maintenance to prevent clogging of tubes. Nothing creates false positives faster than a faulty plate washer.

The Block

After the first wash to remove excess capture molecules, it is necessary to block the well. The purpose of the blocking step is to prevent any non-specific (i.e. unintended) binding of other assay components to the well. The blocking agent usually consists of some assay neutral protein that is unlikely to interfere with any of the other binding events in the assay. Try readily available solutions first, such as 1-5% albumin or casein, before investigating more exotic blocking mixes. There are many commercial blocking mixes that work very well and can sometimes serve as a starting point for further modification to achieve optimal blocking. A good blocking buffer may also contain detergents. The detergents may be similar to the wash buffer or may vary. Understanding the interactions of a detergent with the chosen microtiter plate and other assay reagents is necessary to help in the selection. Be careful with detergents as they sometimes have a tendency to remove desired components from the well, thereby decreasing sensitivity. All components of a blocking solution should be tested to ensure they are not inhibiting any of the subsequent binding events in the assay. The last step of the block is to wash out the excess blocking solution in preparation for sample addition.

The Sample

The sample is, of course, the key component to any ELISA. Take care in acquiring, transporting, and storing samples. These are critical components of any assay. It is essential that everything under your control happen as you planned. Things out of your control need to be documented and the impact on the sample well understood. For instance, one should be able to tell that the samples sat under the equatorial sun for three days and therefore may not be suitable for the assay. But, let’s get back to the things happening in the lab that are under your control.

When the sample arrives in the lab it should be stored in a manner that will stabilize the target molecules and minimize any other degradation that could interfere with the assay. Many samples are not assay ready as collected. They will need to be further processed to alter the concentration and sample matrix/buffer. This buffer should be well thought out to prevent any degradation to the target molecule in the sample. In addition, the sample buffer must be compatible with all other components in the assay. A good first choice of sample buffer is, in fact, the blocking buffer. Depending on the sample, the buffer may be adjusted (i.e. pH, salts, detergent, stabilizers) based on the biologic and chemical constraints of the target molecule. The sample incubation must achieve binding equilibrium with the capture molecules. The timing of this will vary depending on sample concentration and affinity for the capture reagents. As always the microtiter plates are washed prior to the next step.

The Detector

There are many different detection systems employed by ELISAs. The goal of each is sensitive and specific detection of the target molecule. The cytokine detection assay will employ a standard monoclonal anti-target Monoclonal antibody (MAB). There are many fine commercially available antibodies to many biologic molecules. Of course, you will have to test a given MAB in the assay to ensure that it meets your needs. The detector MAB is usually unlabeled. This ensures that the MAB has not been altered in any way and will provide the desired sensitivity and specificity. The detection antibody is usually diluted in the blocking buffer, or perhaps in a manufacturers recommended buffer. Either buffer is a great place to start, but always keep in mind that modifications may be necessary based on the chemical needs of the target and other assay components. The proper dilution is usually arrived at empirically by titering the antibody and comparing relative sample to background signal strength between known positive and negative samples. The goal of the assay to measure very low concentration or perhaps high concentration samples may drive the final choice for detector MAB concentration and buffer. After suitable sample incubation which allows for binding equilibrium, excess non-bound detector is washed from the wells.

The Signal

At this point the signal generation components are added. For the cytokine assay an anti-species antibody directed against our detector MAB is utilized. The anti-species antibody is conjugated to an enzyme. Some popular enzymes are alkaline phosphatase and horseradish peroxidase. There are many commercially available antibody-enzyme signal generation combinations for a multitude of species. However, they are not all created equal, and it is usually prudent to test several. In the end price and convenience may also drive the choice of signal system. Test the signal system by bracketing test dilutions around the manufacturer’s recommendation. Frequently you can choose lower concentrations and save some money, but sometimes you may need to have a higher concentration to get the desired assay signal strength. If the manufacturer has a recommended buffer compatible with your assay then start with that. If you need to use your own buffer, a good place to start is with the blocking buffer. As with all buffer decisions, ensure compatibility with all other assay components.

After signal reagent incubation and wash, an appropriate substrate is added. There are many commercially available stable substrates. It is recommended to go with the commercially available substrates because most of the stability issues have already been solved by the manufacturer. The substrate incubation time will need to be arrived at empirically depending on the amount of target molecule in the samples and the absolute sensitivity of the assay. The substrates will undergo a distinct colorimetric change in the presence of the enzyme. The substrate should show little to no conversion in blank or negative wells. Some substrate systems require a stop solution to be added to the wells after the incubation period. This has the advantage of stopping the reaction across multiple plates at approximately the same time, thereby preventing the last plates read from incubating longer than the first plates read. The plate is now read using any one of many suitable microtiter plate spectrophotometers to measure the optical density of the wells. The optical density is proportional to the amount of target captured in the wells.

There are many different molecules besides enzymes that can generate signals in ELISAs. Some of these include; fluorescent molecules, chemiluminescent molecules, and even colloidal gold. Discussion of the relative merits of the different signal generation molecules is best left to another article.

The Incubations

The term incubation has been mentioned several times above. Clearly each step of the ELISA has an incubation component. For an ELISA to work, the various components of the capture, sample, detector, and signal generation must be allowed to come together. Incubation time and temperature are two additional components that greatly affect the outcome of the ELISA. Sufficient time for binding equilibrium to occur must be achieved at each step. The higher the affinity between components the faster this equilibrium can be achieved. Typical ELISA steps require a minimum 30 minutes of incubation and may easily require more. Sufficient time to achieve equilibrium between surface bound and soluble components at each step of the ELSIA must be achieved.

It is much better to use some sort of controlled temperature incubator rather than to trust the building maintenance man to keep the room air at a constant temperature. Even 2-3⁰ C differences will affect ELISA results, especially if one is trying to quantitate or measure samples near the lower cutoff of the assay. Time and temperature optimization and consistency will greatly add to assay stability over time or between labs.

In addition, gentle mixing of the microtiter plate is sometimes recommended. If mixing is needed it can best be addressed with an optimized assay. Remember that the conditions the assay has been optimized under need to be replicated whenever and wherever the assay is performed. Will the likely users of the assay have all the necessary equipment and reagents to successfully execute the ELISA? Keep this in mind if your assay users will be in clinically regulated laboratories or out in the bush.

The Analysis

During the reagent optimization the assay signal from known positive samples is continuously compared with the signal from known negative samples. Assay reagent concentrations, buffers, and incubations are adjusted to maximize this “signal to noise” ratio. Now one has an optimized assay format along with all key reagents and buffers. The assay is ready to be put to the test. Can the ELISA distinguish between true positive and negative samples? What are the upper and lower limits of the assay? How reproducible is the assay? If necessary, work with your organization’s internal review board to ensure that appropriate consent has been obtained if you are planning to use human samples.

It is essential to obtain or create a panel of samples consisting of positives, negatives, and blanks. The panels should be 10 or more samples from populations likely to serve as actual sample sources. If the ELISA will be used on many different populations then these populations should be represented in the panels. The positives need to cover the expected range of target molecules one is likely to acquire under real use conditions. In the case of the cytokine we are seeking, it may not be possible to obtain strong enough positives unless we acquire samples from ill individuals or “create” them by spiking known amounts of purified cytokine into negative samples. On the other hand, it may not be possible to obtain negative samples if the cytokine is normally at detectable levels in all people. In that case some thought may be needed on how to create a negative that is as similar to the real sample matrix as possible. Depleting samples of the target may be a viable option to create negative samples. This needs to be done in a manner that does is not harmful to the sample matrix. Substituting with other species samples can be done but should be approached with caution due to specificity concerns.

However one obtains or creates the appropriate positive and negative samples, it is now time to test the system. Consulting with a statistician trained in the art of assay development would be wise at this juncture. The statistician can help with plans to show the limits of the assay and how to set a cutoff. Repeat assay runs can easily document the true reproducibility of the assay. That being said, a good place to start with a cutoff is 2-3 standard deviations above the average response of your negative samples. Multiple runs will be necessary to accurately determine the reproducibility of the ELISA. The goal is to set a cutoff that minimizes false negatives and positives.

With all that has been discussed, one should be well equipped to develop and optimize a reasonable ELISA. Remember that every target/Ab combination is unique. If one is ready for surprises they can usually be addressed.

Written by James E. Drummond Ph.D