Using fluorescence for detection offers improvements over traditional strip testing.
High sensitivity, ease of use, and on-site feasibility are key requirements for new tests in areas such as food safety or environmental analysis.
Immunoassays are tests that take advantage of the specific binding of an antibody to its antigen. The main characteristic of immunological techniques is the appropriate labeling of the antibody or the antigen. This label helps create a signal that correlates with the immunoreaction and allows the detection of the analyte of interest. In laboratory assays, various wash steps are required to remove free labeled or unlabeled reactants and allow the detection of the analyte by the bound and labeled reactants. Test results are then compared with a calibration curve to quantify them.
In contrast, well-known lateral-flow immunoassays (LFIAs), such as pregnancy tests, do not require washing steps. In a lateral-flow test system, the antibody-antigen reaction, as well as the removal of excess reactants, takes place by chromatographic separation. The detector reagent (i.e., the labeled antibody) and the sample wick through the pads or membranes of the test strip. At a capture line, the detector reagent interacts with the capture reagent, which has been immobilized on the membrane. The test result is evaluated visually, typically by two lines—a test line and a control line.
Compared with other immunoassay formats, the major advantages of LFIAs is their ease of use; aside from the dispensing of the sample, no additional handling steps are usually necessary. Thus, LFIAs can be easily performed by nontrained users and used on-site during sample collection. The simplicity of the tests, paired with their quick return of results (2–15 minutes), means that testing is cost-effective.
Despite these advantages, LFIAs are often limited to screening applications. This is because LFIAs, in their present form, are not easily quantifiable and are not sensitive enough for certain applications. This article will examine issues related to lateral-flow sensitivity and explore how special labeling methods can overcome the format's limitations.
The first immunoassays were performed in the late 1950s. They were classified as radioimmunoassays and used 131I, an isotope whose high sensitivity increases with test time, as a label.1 To perform these types of tests, scintilators and protective equipment were required. As a result, on-site testing was not feasible. Even so, in the late 1970s, a solid-phase immunoassay was developed using 131I as a marker.2 This test system, which was performed in a flow-communication device, could be considered a precursor of modern lateral-flow technology.
In the 1980s, enzyme labels emerged. These made highly sensitive tests possible without the need for extensive safety requirements.3 Today, we cannot imagine immunological standard analytics without such enzyme-linked immunosorbent assays (ELISAs). In general, peroxidase-labeled immunoreagents are used for these tests, and the signal is created after the immunoreaction by adding the appropriate substrate solution. The signal intensity can be measured by photometry and quantified by comparing it with a calibration curve. In membrane-based assays, on the other hand, alkaline phosphatase labels are often used because the color produced by substrate conversion can be detected visually.4 However, the results are usually difficult to quantify. In theory, it is possible to use enzyme labels for LFIAs, but the process necessary for developing a readable color is not practicable outside a laboratory.
Fluorescence and luminescence immunoassays were also developed, along with appropriate readers. Although these labeling techniques can be used for LFIAs, the complex and often expensive detection equipment required has limited the market for such tests. The same is true for using paramagnetic particles as labels.
For LFIAs, labels such as colloidal gold, carbon particles, or colored latex beads, which allow visual detection without additional readers, are used, but compared with the substrate conversion with enzyme labels, no enhancement effect is possible. Therefore, and because of the relatively high ratio of antibody to particle, only a comparatively low sensitivity is achieved with such labels.
As with other immunoassays, the sensitivity of an LFIA is influenced by the affinity of the antibodies used in the test. In optimal physiological conditions, the affinity constant of an antibody has a value that cannot be improved upon and is ultimately responsible for the sensitivity of the test. This is the case for biochemically unchanged antibodies. However, if antibodies are labeled, this can lead to steric restrictions, causing a change, or even a complete loss, of affinity.5
Biochemically directed labels, like enzymes or fluorophores, usually use reactive amino groups from the lysine residues of the antibody's primary structure. If this conjugation affects the affinity of the antibodies, the label can possibly be coupled by using carboxylic groups of the glutamate or aspartate residues.6 In the case of undirected labeling like gold, some antibodies will likely have a decreased affinity. Therefore, the best labeling conditions must be determined empirically. However, it is important to note that many antibodies, when using visual labeling methods, do not provide the necessary high-affinity constant (even unlabeled ones) needed to meet the required detection limits for various analytes.
Aside from the affinity of the antibody, there are a number of other conditions that may limit immunoassay sensitivity.7 In addition to the sample matrix and antibody label, the membranes or pads used can have a significant effect on test results. In nitrocellulose membranes, for instance, wicking rate plays a critical role in the test process. Ordinarily, the higher the wicking rate, the lower the sensitivity, especially when using antibodies with a low affinity constant.
Sensitivity can also be influenced by test design issues such as membrane thickness, the pretreatment of the membrane, and the concentration of the immobilized and the labeled reactants. This means, for instance, that a loss of sensitivity due to membrane issues can be compensated for by increasing the concentrations of the immobilized antigen on the membrane and the labeled antibodies. However, this usually requires increased reagent volume and therefore, increased cost. This might also result in a loss of specificity and a higher assay background.
Although many different LFIAs have been developed, pregnancy tests and drugs-of-abuse tests remain the most common and widely known. For a good reason, these popular tests often use urine as the sample matrix. Due to its salt content and neutral pH, urine offers almost optimal stability and affinity conditions for antibodies. Complex matrices like blood or saliva also offer physiological salt and pH conditions, but can contain a high amount of proteins, lipids, or polysaccharides. These substances can be separated with laboratory sample preparation methods, but for on-site tests, the need for sample prep should be kept to a minimum.
One alternative is the use of special sample pads that can remove interfering substances. For the testing of food or environmental samples, in which greater sample purity is needed, special buffer conditions must create the necessary physiological conditions for the test. The easiest way to achieve these is by diluting the sample in buffer. Doing so, however, also often dilutes the analyte of interest, leading to a loss in sensitivity.
In overcoming the sensitivity limitations of LFIAs, it is also necessary to match different test components. Other than this, since the affinity of the antibodies is limited, the sensitivity of a test can only be increased by using different labels. The development of new and highly sensitive readers for fluorescence, luminescence, or paramagnetic labels has created new opportunities. But these tests tend to be more expensive and, due to the size and complexity of the readers, are limited for on-site testing.
A different approach to achieving a stronger signal is to increase the labeling density of the antibodies. The following description of the fluorescence-labeled optical-read immunodipstick assay (FLORIDA) technology, a novel fluorescence labeling method, shows how a sensitivity increase of up to three orders of magnitude compared with traditional lateral-flow labeling methods can be achieved while still being able to detect results with a simple, portable handheld lamp.
Figure 1. (click to enlarge) Test results on FLORIDA test strips. The strips represent indirect immunoassays with control signals (upper lines). The test signal (lower line) on the first strip (a) indicates that no analyte is present in the sample. The appearance of only a control line on the second strip (b) indicates that the analyte is present.
The FLORIDA format is based on a special method of labeling biomarkers using the protocols of Cibitest GmbH (Neu-Ulm, Germany). The test strips are manufactured in the same way as traditional LFIAs, using standard equipment such as the BioDot XYZ3050 platform by BioDot Inc. (Irvine, CA) and standard test-strip materials such as the Immunopore nitrocellulose membranes by Whatman International Ltd. (Maidstone, Kent, UK). However, in contrast to traditional LFIAs, in which gold particles or latex beads are used as labels, FLORIDAs use fluorescent dyes. With this labeling technology, a detection limit of only a few parts per trillion can be achieved both with competitive immunoassays as well as with direct formats like sandwich assays, while still obtaining visual signals.
As shown in the case of indirect immunoassay (e.g., LFIAs using colloidal gold), no significant changes in the test results evaluation are necessary (see Figure 1). When two lines—the test line and the control line—appear, no analyte is present in the sample. If only the control line appears, corresponding to the cutoff level, the analyte is present.
Figure 2. (click to enlarge) Equipment for FLORIDA test evaluation. After running the test, strips are inserted into a handheld lamp (a), and results are evaluated visually (b).
Due to the fact that the FLORIDA technology uses fluorescent dyes, evaluating the test strip requires the additional excitation of the fluorophores. However, because the special labeling method results in a very strong fluorescence signal, a simplified reader is sufficient for this task. After inserting the test strip into a handheld lamp, the test signals generated at the test and the control lines can be detected visually (see Figure 2).
Due to its bright signal, FLORIDA produces a rapid immunoanalysis of strongly colored specimens. And due to the evaluation of the results with a handheld lamp, the results can be seen under bad lighting, or even in complete darkness.
The extreme sensitivity of FLORIDAs can be achieved by conjugating certain fluorophores in high densities to carrier molecules and then conjugating the antibodies to this fluorescent complex. The enhancing effect results from the very high ratio of fluorophores conjugated to a single antibody. Compared with the direct conjugation of fluorescent dyes to antibodies (e.g., via NHS-esters of the dyes), the molar ratio of fluorophore to antibody is raised by about 100–1000 times. Thus, an increase in sensitivity of three orders of magnitude can be achieved.
Figure 3. (click to enlarge) Comparison of the detection limit of an antiampicillin antibody used in an indirect, competitive ELISA and in a FLORIDA immunoassay.
To demonstrate, this sensitivity was determined with an affinity-purified antiserum raised against ß-lactam antibiotics. As shown in Figure 3, this antiserum was used in a competitive ELISA as well as in the FLORIDA format for the detection of ampicillin in buffered solution. The detection limit of the ELISA is approximately 0.5 ng/ml, and the midpoint of the test is 10 ng/ml (10 ppb), which correspond well to the expectation for this kind of assay. For the semiquantitative FLORIDA format, a cutoff value of 10 pg/ml (10 ppt) was determined. This finding differs approximately three orders of magnitude from the average sensitivity of traditional gold-labeled lateral-flow assays.8 Comparable sensitivities can routinely be achieved only by radioimmunoassays in a laboratory environment. But the FLORIDA system is an on-site test, and the latest data show that this unexpected increase in sensitivity is connected to the special method for conjugating fluorophores to antibodies.
Stability of Fluorescent Signals
The photostability of the fluorophore used in the FLORIDA system is not a limiting factor for the test performance. The test cups with the lyophilized fluorophores are stored in a dark environment before use. The test can be performed in daylight without a problem. Even a long chromatography of up to 20 minutes does not cause photobleaching. After evaluation using the handheld lamp, the strips can be dried and easily archived. So long as they are stored under dark conditions, the signal can be visually read, even after more than 18 months.
Figure 4. (click to enlarge) Top-down scan along the detection zone of three test strips. One test (null) was performed without any fluorescent label. The other two tests were conducted with fluorescent-labeled antibody, one with analyte in the sample (positive) and the other without (negative).
It is also possible to evaluate the signal optoelectronically (e.g., using a charge-coupled device camera together with the appropriate software). As shown in Figure 4, compared with the zero signal, an increased fluorescent background can be detected. This background signal usually does not interfere with the visual evaluation of the test. When using an optoelectronic detection system, the choice of appropriate cutoff will separate the background signal from the test signal completely.
By exciting the fluorophores with blue light instead of ultraviolet light, possible interferences can be reduced. However, there might still be some specimens whose autofluorescence will cause a stronger background signal. For instance, one fluorophore currently used with the FLORIDA technology is not suitable for use with milk. This is because a test strip used with a milk sample and excited with blue light will exhibit a green autofluorescence that is not easily distinguishable.
Another example is the testing of acridin. This substance, which is used in selective enrichment supplements for certain microorganisms (e.g., listeria), also has an autofluorescence that interferes with the test signal. To be able to use these matrices with the FLORIDA format, other fluorophores with a greater Stokes shift are being developed. Such fluorophores emit in a red spectral range and could be distinguished easily from green autofluorescence.
FLORIDA Compared with Existing Labeling Techniques
Developments in labels continue to improve detection systems (for instance, by creating readers that allow objective evaluation of the results). However, one issue that remains difficult, or in some cases impossible, to solve is the increase of label sensitivity beyond a certain limit.
Typically, the detection limits are found in the lower parts-per-billion range, with an extraordinary antibody perhaps somewhat lower. In contrast, it can be shown using a morphine lateral-flow assay that by changing only the label from colloidal gold to the FLORIDA method, the sensitivity of the test increases by a magnitude of 100, without any further optimization. The only difference is the requirement of a simple reader.
Due to the strong signal created by the FLORIDA labeling method, there is no need for highly sophisticated and expensive readers, as is the case with other labeling methods being developed. Existing fluorescence labeling techniques, paramagnetic particles, and other more- exotic labels offer promising sensitivities and quantification possibilities, certainly superior to gold or latex. However, the costs of the test systems, as well as of the detection equipment, make these technologies available only for high-cost analytics. The FLORIDA method, on the other hand, enables a visual readout with a basic reader, useful in low-cost screening applications with a need for high sensitivities.
Table I. (click to enlarge) Possible applications of the FLORIDA method.
Due to the described advantages, the FLORIDA technology is applicable and useful in all fields with a need for highly sensitive, on-site analysis. As shown in Table I, FLORIDAs create new opportunities in the examination of food and environmental samples, as well as in medical diagnostics.
Important applications in the sector of food and feed analysis include the residue analysis of veterinary drugs (e.g., in freshly slaughtered animals), as well as the on-site control of mycotoxins and pesticides in fruits and vegetables. Pesticides often have very low maximum residue limits (e.g., in drinking water and wastewater) and, therefore, these analytes have not been easily accessible to rapid immunoanalysis. FLORIDA systems offer the potential to overcome this problem. As a result, the analysis of pesticides on-site, and the detection of warfare agents in soil, will be possible.
In addition, new analytes in medical diagnostics, such as hormones or cancer biomarkers, become attainable when using FLORIDAs. In this field, the analysis of problematic specimens such as whole blood or saliva, which often need to be diluted, is required. FLORIDA systems make on-site lateral-flow testing possible and still meet the requirements for a highly sensitive analysis.
That said, this novel labeling method is not limited to use in LFIAs. Although investigations are ongoing, it can be expected that the FLORIDA method also offers advantages in microplate assays or even microarrays.
Performing meaningful on-site tests in any diagnostics area is becoming increasingly important. LFIAs currently serve as the easy-to-use platform for this type of testing. However, the sensitivity needed to detect many analytes has made testing outside the laboratory not always feasible. If the best available antibody and the best combination of materials are already being used, an increase in sensitivity can typically only be achieved by changing the detection label. More than a few promising technologies are being developed, although these tend to require sophisticated and costly equipment. The FLORIDA technology, on the other hand, creates the opportunity to develop a highly sensitive, but at the same time low-cost, LFIA system.
1. SA Berson and RS Yalow, “Quantitative Aspects of the Reaction Between Insulin and Insulin-Binding Antibody,” Journal of Clinical Investigation 38 (1959): 1996–2016.
2. H Park, “New Technique for Solid-Phase Immunoassay: Application to Hepatitis B Surface Antigen,” Clinical Chemistry 25, no. 1 (1979): 178–182.
3. D Catty and C Raykundalia, “ELISA and Related Enzyme Immunoassays,” in Antibodies Vol. II: A Practical Approach, ed. D Catty (Oxford, UK: IRL Press at Oxford Unversity Press, 1989), 97–154.
4. JP Gosling, “Enzyme Immunoassay,” in Immunoassay, ed. EP Diamandis and TK Christopoulos (San Diego: Academic Press, 1996), 287–308.
5. C Davies, “Immunoassay Design,” in The Immunoassay Handbook, ed. D Wild (New York: Stockton Press, 1994), 15–48.
6. GT Hermanson, Bioconjugate Techniques (San Diego: Academic Press, 1996).
7. Rapid Lateral Flow Test Strips, Considerations for Product Development (Bedford, MA: Millipore Corp., 2002).
8. R Verheijen, “Immunological Strip Tests” in Analytical Biotechnology, ed. TGM Schalkhammer (Basel, Switzerland: Birkhäuser, 2002), 132–166.