Finally, after years of promise, an aptamer-based test has been developed for commercial practice.
|Figure 1. Loading results for analysis.|
Aptamers were first discovered in 1989 by Craig Tuerk and Larry Gold. The classic definition of an aptamer is a single-stranded oligonucleotide (RNA or DNA) that has the capacity to mimic an antibody in terms of binding to specific analytes. Many have predicted that this discovery represented a technology with the potential to rival the antibody revolution in molecular diagnostic testing. Years have passed, and while many aptamers have been discovered and reported in academic journals, the aptamer-based diagnostic revolution has not materialized.
One issue that has delayed the revolution is the difficulty associated with the identification of aptamers that bind to a target analyte. For antibodies, the identification process has been well developed and is relatively routine. For aptamers, the reliance on polymerase chain reaction (PCR) based amplification of very small concentrations of short target DNA molecules can easily lead to artifacts. In addition, without effective counter measures, the immobilization of the analyte onto resin for selection purposes can easily lead to the identification of aptamers for components of the resin, rather than the analyte. It’s been my experience that many published aptamers exhibit significant binding to various resins, and very little to the analyte of interest.
In general the comparison between aptamers and antibodies shown in Table I is a fair assessment of the current state of the art.
The problems associated with identification of aptamers and the alternative strategy that we have developed are beyond the scope of this article. Here, we will focus on the issues involved in developing an aptamer that has been identified, and that binds to the target analyte with high affinity and specificity into a commercial diagnostic system.
For any technology to be commercially viable, at least four factors must be satisfied:
1. The production cost must be less than the price that the market will bear for the product.
2. The system’s sensitivity must be consistent across production batches.
3. The system must provide sufficient sensitivity to comply with regulatory standards.
4. The system must provide a clear advantage over existing products.
Consideration of these points resulted in my focus on DNA rather than RNA as the basis for my development efforts. RNA is more difficult to synthesize chemically, and is much less robust than DNA in terms of shelf life. Each of these factors has proven to be important in the commercial development process.
Antibodies have been around for several decades and the cost of producing them on a large scale is not the problem that it once was. The amount of an antibody or aptamer required per diagnostic kit is a function of the binding affinity for the target molecule; that means higher affinities translate directly into higher profit margins. In general, antibodies have much stronger binding affinities than aptamers; unfortunately, this means that it is necessary to use more aptamers to do the same job. Aptamers generally have higher specificity for their targets than corresponding antibodies, and this has significant advantages when it comes to detection. With more specific recognition, results are more reliable. In practice, this means that the testing system can be scaled down and the speed of testing increased without a loss in sensitivity.
Other than the direct cost comparison between DNA and antibodies, there are many other cost factors involved in a diagnostic system. A key advantage of the use of DNA over antibodies is that DNA enables a much broader selection of resins. DNA can be conjugated to a large number of surfaces, including glass. The key constraint is the tendency for the DNA to form aspecific bonds with the surface on which it is immobilized. In the case of microarray analysis, this isn’t a significant issue. The binding affinity between two single-stranded oligonucleotides is extremely high and will easily overcome these aspecific bonds. However, this is a problem with aptamers with disassociation coefficients for their targets (Kd) in the low nanomolar range. Another key issue is the reliability of the conjugation. For discovery purposes, the variation in efficiency of chemical conjugations of an oligonucleotide to a resin is not a significant problem. For commercial purposes, the conjugation’s efficiency must be high to reduce the cost of the DNA being used, and the rate of conjugation must be reliable.
For commercial applications, DNA aptamers are synthesized in large-scale batches. Note that the number of suppliers for large-scale DNA synthesis has declined in recent years. Moreover, the existing quoting structure for third-party synthesis is based on the starting amount targeted rather than on the amount actually synthesized. As there can be significant differences in product yield from one production run to another, this quoting system is unacceptable in terms of managing profit margins with a commercial product. Terms with guarantees for minimum yield need to be negotiated.
Fortunately for aptamer-based systems, this isn’t a key concern. As long as the DNA purification specifications are met, and the internal quality-control standards for the amount of DNA conjugated to resin are achieved, product performance is consistent. This translates into consistent batch-to-batch recovery rates of the target molecule.
A diagnostic kit can only be commercially viable if it can detect the analyte at the appropriate regulatory approved levels. Hence, we must identify aptamers that bind to the target analyte with a sufficiently low coefficient of disassociation (Kd). However, this is only a start toward creating a commercial test. The system must also deliver an adequate sensitivity level in relevant backgrounds. In agri-food testing, this background (or matrix) can vary from flour to juices to fermented products like wine or beer. Even after purification in an affinity column, matrix effects persist and differ depending on the matrix.
A number of proposed methods have been tested, including electro-chemical detection, evanescent lightwave
|Table 1. Comparison of strengths and weaknesses of antibodies versus aptamers in diagnostic applications.|
detection, gold nanoparticles, and lateral flow separation. None of these approaches satisfied the criteria for sufficient (ppb) sensitivity in different matrices. Lateral flow separation did, but only after the product had been purified in an affinity column. Instead, an innovative approach (described later) was developed. All aptamers that we have identified have optimum pH ranges outside of which their performance drops off sharply. Fortunately, within this optimum pH range, performance doesn’t vary significantly. However, this means that the matrix’s pH must be adjusted by adding a universal buffer to ensure that all samples are well within the aptamer’s operating range. This posed a particular problem for detecting analytes in wine, because of the strength of its acidity. This has been solved by creating a proprietary buffer solution made of several components.
More Than a Me-Too Product
It’s important to either develop a product that performs as well as existing antibody-based products but costs less, or performs better for the same price. I achieved success by developing a system with similar speed, complexity, and cost to ELISA-based assays, but with the sensitivity of a combination of an antibody-based affinity column and fluorescent HPLC
measurements. The market niche for this product is for customers who want a quantitative assay, 96-well format for high-throughput analysis, with HPLC quality data. The detection system’s strength represents a key milestone in terms of testing systems. Because of the specificity of the aptamer and the sensitivity of the testing system, the volumes required could be reduced for analysis, greatly increasing the speed of analysis without sacrificing the reliability of the results. This breakthrough also paves the way for the development of integrated automated systems for mycotoxin detection.
Ochratoxin A (OTA) is produced by fungi that grow on cereals, dried fruits, coffee, grapes, and other food products. It is one of the most abundant food-contaminating mycotoxins, and is considered a carcinogen to humans. The European Union has set maximum allowed levels of 5 ppb for cereals, 2 ppb for wine, and 0.5 ppb for baby food. I have developed an aptamer-based diagnostic kit for detecting OTA. The OTA-Sense System consists of two components: an aptamer-based cleanup column and a detection solution. The first step is to extract the toxin, where grain is ground and extracted with four volumes of 60% acetonitrile in water to one part grain. A portion of this grain extract is diluted with a binding buffer to ensure that the pH is consistent across tests.
|Figure 2. Fluorescence and retention time for a) zearalenone, b) aflatoxin M1, and c) ochratoxin A.|
The diluted sample is filtered through glass wool, and then loaded onto the OTA-Sense affinity column. These columns have our ochratoxin aptamer immobilized onto a resin. After a wash step, the bound ochratoxin is eluted with a buffer that doesn’t contain cations. The eluted sample is then combined with a detection solution that contains free aptamer and terbium. Terbium acts as a cation bridge between the aptamer and the ochratoxin. This facilitates binding and, more importantly, enhances the terbium’s fluorescence by providing a microenvironment where water is excluded.
The amount of ochratoxin present is measured by exciting the ochratoxin molecules at a 380-nm wavelength, and measuring fluorescence emitted by terbium at 540 nm. The fluorescence measurement is time resolved in that measurements don’t start until 70 µs after excitation. A third party has successfully validated the performance of this aptamer in an affinity column compared to an antibody column followed by fluorescent HPLC analysis. Including the free aptamer/terbium detection step in the OTA-Sense System provides the targeted market advantage described earlier. The OTA-Sense System is the only system that can support a 96-well-plate, high-throughput format with a high level of sensitivity and resolution for ochratoxin A analysis. The kit format provides advantages over ELISA systems by eliminating the cost of running the necessary wells for calibrants and eliminating process time sensitivity. Applications are being developed in a similar format for the other major mycotoxins (zearalenone, fumonisin and deoxynivalenol) based on identified aptamers. Application of the OTA-Sense System has been extended to include all alcoholic beverages, including beer and white and red wine.
Aflatoxins are another group of mycotoxins produced by different species of Aspergillus. Aflatoxins are potent toxic and carcinogenic agents produced as secondary metabolites by the above fungi. Among the different types of aflatoxins identified, major members are aflatoxin B1, B2, G1, and G2. Crops affected by the toxin range from corn to peanuts to spices and nut trees. Aflatoxin M1 and M2 are metabolites of B1 and B2 respectively, found in milk of animals that have consumed contaminated feed.
An aptamer-based detection system for the aflatoxins has been developed. The system employed for toxin purification is the same as for ochratoxin A. An aptamer identified for aflatoxin B1 was immobilized on streptavidin agarose using a biotin moiety on the aptamer’s 5' end. The immobilized aptamer is used in an affinity column under identical conditions to ochratoxin A for purification. An iodine derivatization is used to quantify the amount of aflatoxin present following purification. This improves the fluorescence of aflatoxin B1 to a level equal to that of B2. This level of fluorescence intensity allows for the detection of total aflatoxins in grain samples (corn and peanuts) down to at least 1 ppb. Observed recovery rates reach 90% in corn and close to 100% in peanut matrices. This system doesn’t require time-resolved fluorometer for analysis. The system has also been optimized for the detection of aflatoxin M1 in milk.
The OTA-Sense and AFLA-Sense affinity columns can also be purchased as one component for use as a cleanup step for HPLC testing. The advantages over antibody-based columns are price and speed. Only 1 to 3 ml of grain extract is needed for cleanup, reducing assay time to one to three minutes. Antibody-based columns take 30 minutes, as 10 ml of solution need to pass through the purification column at a 1 drop/s rate. It’s possible to use the mobile phase for HPLC analysis for elution of mycotoxins from aptamer-based cleanup columns. This results in cleaner HPLC chromatographs and eliminates the need to dry down samples to change solvent systems. (See Figure 2.)
The most promising approach of which I am aware has been the efforts of Barthelmebs et al., with the development of enzyme-linked aptamer assays (ELAA). This group has performed a re-selection of aptamers for ochratoxin A, and used the selected aptamers, as well as ours, in a competitive ELISA format. It’s interesting to note the similarity in sequence among the aptamers that they identified and the sequences that we identified. They immobilized ochratoxin A on a surface and determined the amount of aptamer retained in solution in the presence of free ochratoxin A. They went on to demonstrate the efficacy of this approach in spiked red wine samples, but this approach didn’t work as well as existing antibody-based ELISA analyses, as red wine is not a trivial matrix. We obtained recovery rates through our aptamer-based affinity columns of 80% ochratoxin across a range of red wines.
Detecting ochratoxin eluted from these columns isn’t compromised by the remaining presence of contaminants from the wine. Protocols have been developed and these aptamer-based affinity columns can be used with all wine and beer samples with expected recovery rates greater than 85%. Using aptamer-based affinity columns decreases the time needed for cleanup of these samples prior to HPLC analysis. The volume of solution that needs to be passed through an antibody-based affinity column, including loading, washing, and elution, is over 30 mL. For aptamer-based columns, one needs to pass a total of only 3 ml. A larger sample volume from a homogenous solution containing the mycotoxin, such as a grain extract or mycotoxins present in wine or milk, doesn’t provide a better statistical sampling of the material. The reliability of the determinations of mycotoxin concentration isn’t compromised by smaller sample sizes, as long as the system is sufficiently sensitive for detection.
The current mycotoxin testing system is based on quality control rather than quality assurance (QA). Testing of grain that’s been blended at the point of delivery is akin to looking for horses after they’ve left the barn. A QA program requires the capacity to prevent the horses from leaving the barn in the first place. This means that to support a QA program, it’s necessary to detect the presence of mycotoxins at the point of entry into the food production system. The entry point globally is the delivery of grain from the fields to grain-receiving terminals or the direct delivery to processing plants. Analysis at this stage globally relies entirely on visual inspection of grain for fungal damage. It’s understood that there is a poor correlation between visual inspection and toxin presence. Moreover, this correlation is worse for the mycotoxins that are regulated at the ppb level (ochratoxin A, aflatoxins) rather than ppm level (fumonisins, deoxynivalenol, zearalenone) because of the distribution heterogeneity of the fungus in the grain, and potentially even greater distribution heterogeneity of the mycotoxin.
At the point of grain delivery, there are two primary constraints to test delivery, time, and environment. On average, it takes a grain delivery truck 10 to 15 minutes to unload. After this, a decision must be made for which bin the grain has been assigned. It’s now possible to determine moisture and protein content by NIR analysis, but the time required to grind and homogenize flour from grain samples imposes another constraint even before the time required for existing di agnostic tests is reached. Moreover, the environment is limiting in terms of the level of dust present and the lack of analytical equipment. Providing a technology to meet this opportunity is similar to the need for NASA scientists to solve the problems associated with the Apollo 13 flight, using only the equipment that exists onboard. We need to work with the constraints of delivering a technology at a grain elevator, not in a laboratory environment with fume hoods and toxic waste disposal.
A starting point is using much lower volumes of extraction solvent in aptamer-based affinity columns than with similar antibody-based affinity columns. This presents an opportunity to rethink the approach to mycotoxin sampling at the point of grain delivery. We have a vision for an integrated system where flour homogenization is replaced by the exposure of a small amount of extraction solvent to a large amount of grain, allowing all mycotoxins to rapidly homogenize in the liquid phase. We are now developing an integrated approach to on-site mycotoxin detection, based on aptamers for all five major mycotoxins. The goal is to create an automated process within a machine that requires minimal operator expertise.
Gregory Penner, PhD, is president and CEO of NeoVentures Biotechnology Inc. He can be reached at firstname.lastname@example.org.