Medical Device Link.

PROCESSING TECHNOLOGIES

Lateral-flow immunoassays (LFIAs) represent a well-established technology appropriate for use in a wide variety of point-of-care (POC) or field-use applications. However, this technology has not been applied much where very sensitive, highly reproducible, or quantitative results are required.

A range of new approaches to manufacturing lateral-flow assays have appeared in the past few years that do allow for the application of this technology in a wide variety of more-demanding applications. These innovative approaches compel a fresh look at the fundamentals of manufacturing LFIAs and an investigation of the resources that developers and manufacturers require in order to succeed as innovators in this area.

This article concludes a two-part series begun in the June issue. The first article focused on traditional manufacturing technologies and materials for LFIAs, discussed the limitations of each, and outlined some of the approaches that can be taken to overcome those limitations in order to create more-reproducible assays. This follow-up focuses on new approaches to materials, signal reagents, readers, and the whole LFIA concept that have the potential to result in more-quantitative, more-sensitive LFIA products. It also addresses some of the new manufacturing challenges that accompany these novel approaches.

Directions of Approach

A significant amount of development effort through the years has gone into finding ways to overcome many of the inherent weaknesses of the materials and designs used in traditional lateral-flow immunochromatographic assays. Paying careful attention to product design and manufacturing processes makes it possible to produce, by means of traditional methods, lateral-flow assays that have relatively low process-level signal coefficients of variation (CVs). These assays can have sensitivities approaching or bettering those of competing technologies. However, generating LFIAs that exhibit even higher reproducibility and sensitivity is likely to require some considerable changes in approach in the areas of materials science, labeling, and device design.

Materials Science. The materials primarily used in LFIAs were not designed specifically for use in those systems. Nitrocellulose, glass fiber, polyester, rayon, and other filtration media have all been adapted for use in these assays, but each exhibits inherent inhomogeneity that hinders the production of truly reproducible LFIAs in traditional configurations.

Novel materials are needed that can improve upon the functional performance of which existing materials are capable. Materials that can perform multiple functions in the assay system also are desirable. The application of new materials opens the door to entirely new concepts of LFIA device design.

Labels. The traditional use of colloidal gold or colored or fluorescent latexes dried onto conjugate pads as particulate labels results in high levels of variation due to instability and inconsistent release. The sensitivity of such labels tends to be somewhat limited owing to the absence of amplification and the loss of signal because of limited availability for detection. That is, in a visual assay in a three-dimensional matrix such as nitrocellulose, signal is measured from only approximately the top 10 µm of the membrane; the remaining signal in the system is lost.

Novel labels, coupled with alternative materials and reading systems, can yield significant improvements in performance, as can alternative approaches to the use of existing labels.

Device Design and Reader Integration. Innovative device designs that retain desirable features of the traditional LFIA but that attempt to avoid its major material- and process-related pitfalls are beginning to appear. Many of these designs rely on the integration of reader technologies. Consequently, they require a multidisciplinary development approach that adds complexity to the product design and development cycle.

Materials and Device Design

The use in lateral-flow assays of overlapping bibulous materials with multiple functionalities can introduce intellectual property issues, make the manufacturing process much more complicated, and increase variability in the assay. In order for the next generation of lateral-flow assays to reach the goals of high reproducibility and high sensitivity, it is critical that attention be paid to the development of new materials for use in those assays. This applies across the entire assay device, not just to the membrane system.

Although some attempts to innovate have been made over the past 10 years, truly novel approaches to LFIA materials, particularly reaction surfaces and conjugate application pads, have not been numerous. Those that have been tried have, for the most part, not achieved significant market penetration. Very recently, however, several innovations that have shown some promise have appeared on the market.

In considering alternative matrices for LFIAs, designers might look at how many of the following ideal characteristics they exhibit:

• A highly regular surface yielding very good line quality.

• A three-dimensional matrix with consistent pore size, thickness, and protein binding capacity.

• True capillary flow with a variety of wicking rates.

• The greatest thinness reasonably possible.

• Good fluid-flow characteristics over the entire shelf life, independent of treatment.

• Low CVs for capillary rise time over the entire shelf life, independent of treatment.

• Minimal contamination by metal.

• Low background fluorescence.

• Noninterference.

• Stability during storage.

• Nonflammability.

• Low cost.

• An ability to be activated for covalent linkage.

• Multifunctionality, i.e., the ability of a single matrix to act as conjugate application area, sample application area, reaction surface, separation medium, and wick.

Figure 1. Illustration of the rapid development of an hCG assay on Fusion 5 from Whatman (Florham Park, NJ): (a) 0 seconds, (b) 15 seconds, (c) 1 minute.

Several new materials and related design approaches that take these basic principles into consideration are under development, although none has yet reached the market that fulfills all of the listed specifications. Descriptions of some of these follow.

An Innovative Matrix. One material that has been developed to meet at least some of the specifications above is the Fusion 5 matrix from Whatman (Florham Park, NJ). This large-pore, single-layer matrix is hydrophilic and non–protein binding in nature. The material is intended to be used to fulfill single-handedly the functional requirements of all of the materials used in a traditional lateral-flow device. That is, it performs as sample pad, conjugate pad, membrane, and wick.

Because it is non–protein binding, Fusion 5 cannot accommodate a method traditionally used to lay test and control lines onto a substrate. A strategy of placing boulders in the stream, so to speak, is used instead. Large-diameter beads (they measure approximately 2 µm across) are conjugated to the test- and control-line proteins and then dispensed onto the material at the appropriate locations. The large beads become immobilized in the matrix and form the test- and control-line areas. When the sample and conjugate flow past the so-called boulders, binding and signal formation occur at those locations.

The open-pore nature of this system means that assays can be extremely fast, a capability that has both positive and negative aspects. Speed, or reaction time, can be important in LFIAs, but in many instances it bears an inverse relationship to sensitivity. The bead approach used with Fusion 5 largely overcomes this owing to its lower inherent background and the increased surface area for ligand binding it provides. Figure 1 illustrates an extremely rapid assay for hCG based on Fusion 5.

Incorporation of Fusion 5 into an LFIA product has the capacity to greatly simplify a manufacturing process. Multiple dipping, drying, and lamination steps that are basic to producing standard LFIAs are eliminated. The material does impose a requirement—the deposition of tight lines of relatively large particulates at the test and control areas—that can be technically challenging. However, this deposition process has been demonstrated to be technically feasible, using both contact and noncontact methods.

Figure 2. The micropillar-based 4CastChip from Amic AB (Uppsala, Sweden).

A Micropillar Substrate. A second novel assay and material design approach is represented by the 4CastChip from Amic AB (Uppsala, Sweden). That company has developed an assay substrate that consists of a highly ordered array of micropillars on a plastic slide (see Figure 2).

Figure 3. Regular pillar structure of the Amic device flow path.

The micropillars of the 4CastChip are hydrophilized by dextran, and act to drive capillary flow of sample and reagents in the flow path. Also, the pillars provide a biocompatible surface for the attachment of capture ligands at test and control lines. The material is highly regular in comparison with a standard nitrocellulose material (see Figure 3). As with the Fusion 5 approach, the Amic device offers the capacity for multiple functionality; the pillar-defined flow path can act as sample application area, reaction surface, and wick.

Figure 4. (click to enlarge) Troponin I (TnI) data generated on the 4CastChip from Amic, quantified on a GenePix scanner from Molecular Devices Corp. (Sunnyvale, CA).

Linkage of proteins to the surface of the substrate is by means of covalent attachment of amines to the aldehyde groups on the chip surface. Protein is dispensed onto the surface and then reacted in a humid environment for a short time to enable the linkage to occur. This substrate can be employed to generate sensitive assays that use fluorescent labels and whose sensitivity is evident when the assay is linked to a reader system (see Figure 4).

Figure 5. An AD 3200 dispenser from BioDot Inc. (Irvine, CA) laying down test and control lines onto Amic chips in a production-scale unit.

In terms of manufacturability, this device, like the Fusion 5, has the potential to remove several steps from the production process. At the same time, it introduces the need for discontinuous dispensing of proteins onto discrete substrates, something that can be achieved only through a noncontact technique. The requirement for chips to be handled individually could be considered a drawback from a processing standpoint, by contrast with the ability of materials such as nitrocellulose to be processed in-line. Dispensing methods must be highly regular, reproducible, and carefully controlled so as to ensure that line widths are consistent (see Figure 5).

The 4CastChip represents a shift in thinking with regard to LFIA substrates, being as it is effectively a two-dimensional substrate without discrete pores. However, the device does exhibit true capillary flow, has an extremely regular hydrophilic surface, and generates visually acceptable lines at sensitivities (in systems tested to date) that are comparable to those of existing clinical and POC systems. The material itself is highly stable and keeps proteins in a stable condition for extended periods. As such, it meets many of the criteria for an ideal device listed near the beginning of this section.

Labels and Reader Systems

Several types of particulate conjugate are used in LFIA systems. The choice of label in many ways defines the achievable sensitivity, the ease of conjugate manufacturability, and the necessity for a reader system to be used in conjunction with the test. If a reader is necessary, the label used will determine the type of reader, as well. Standard systems utilize colloidal gold or latex particles containing either visual or fluorescent dyes.

Assay systems configured that way suffer from several drawbacks. For example, when a visual label is assessed, either by eye or using a reader, the signal that is visualized is present in only about the top 10 µm of the membrane. A significant portion of the captured signal is therefore lost to the reader’s perspective. Also, lateral-flow assays typically do not allow signal amplification, as multistep assays are not usually performed in these systems owing to considerations of design and user-friendliness.

However, some novel approaches to particulate labeling have been introduced in recent years. These include paramagnetic particles (PMPs), as used in the magnetic immunochromatographic test (MICT) technology of Magna BioSciences (San Diego), and innovative fluorescent labels, as used in the fluorescent-labeled optically read immunodipstick assay (FLORIDA) technology developed by Cibitest GmbH (Neu-Ulm, Germany).

PMP Technology. Magna Biosciences’ MICT technology has been described previously in IVD Technology.¹ To recapitulate, the advantages of using paramagnetic particles in LFIA systems include the following:

• PMPs are a direct replacement for colored or fluorescent conjugated colloidal particles.

• The conjugation chemistry is well known and effectively standardized.

• The reagent is stable.

• The signal is stable after development.

• Biological matrices have low magnetic backgrounds.

• Magnetic fields are unimpeded by visual barriers; thus, tests of samples that stain or color a membrane can be run, and test and control lines do not have to be visualized.

• The reader can detect all of the signal that is in the volume underneath it.

Figure 6. (click to enlarge) Data from a cardiac troponin I (cTnI) study performed using the Magna BioSciences system. (a) A cTnI standard curve showing linearity from 0 to 10 ng/ml; (b) a breakout of this curve from 0 to 1 ng/ml.

The last item indicates that the problem of detector particles captured beneath the level of the substrate that can be seen by eye or a visual reader is overcome. The higher available signal strength can translate into significant benefits in terms of assay sensitivity.

The feasibility of the MAR system has been demonstrated in a variety of systems (see Figure 6). The system has demonstrated high sensitivity and the capacity for application in a variety of analyte systems and matrices.

Figure 7. Image of the signal generated on test and control lines using FLORIDA fluorophore technology from Cibitest GmbH (Neu-Ulm, Germany).

Fluorophore Technology. Another sensitive labeling and reading method for lateral-flow assays, Cibitest’s FLORIDA technology, employs as the signal-generating reagents proprietary fluorophores conjugated in high density to carrier molecules (see Figure 7). The fluorophore and carrier are conjugated in turn to specific antibodies for use in the assay. This means that the fluorophore-to-antibody ratio is very high. Consequently, signal intensity is greater than with a standard conjugation of antibody to fluorophore.

An additional advantage with this technology is that the labels fluoresce in the visible wavelength, which makes signal generation and detection significantly simpler than is the case with standard fluorophores whose detection requires a complex reader. The Cibitest device calls for only an excitation lamp, as reading can be performed by eye. Highly sensitive assays therefore can be generated, although quantification of the signal remains an issue.

Fluorescent Reader Technologies. For any LFIA developer whose test requires a reader system, one of the major issues that must be dealt with is access to an appropriate system. Access to fluorescent readers has been problematic for these assay developers because of the large up-front investment in the licensing and development of readers that has been necessary before strip development could be completed. The market has suffered somewhat from the absence of a low-cost, easily customized reader for lateral-flow assays that could quickly be converted into a customized product.

Table I. (click to enlarge) Results of an HIV mixed-titer panel tested with the HIV 1/2 fluorescent LFIA from the One Step One Solution program (OSOS) and the Embedded Systems Engineering handheld fluorescent sensor. Cutoff was determined by averaging five confirmed HIV 1/2 negative samples and three standard deviations.

A fluorescent assay reader manufactured by Embedded Systems Engineering GmbH (ESE; Stockach, Germany) attempts to address this issue. ESE has created a miniature confocal optical sensor (see Figure 8) that can be adapted for integration into a variety of housings, including handheld and laboratory-based units (see Figure 9). This sensor can be modified quickly for use with any of a wide variety of fluorescent labels. The system exhibits high sensitivity and a large dynamic range in a variety of test systems (see Tables I and II). The availability of this type of device and service represents a key building block in the toolkit of LFIA developers.

Multidisciplinary Platform Development

Figure 8. A miniaturized modular fluorescence sensor from Embedded Systems Engineering GmbH (Stockach, Germany).

One of the major challenges facing developers of highly sensitive or reproducible quantitative LFIAs is that assay systems of this type require a more multidisciplinary approach than do standard lateral-flow assays using traditional technology. Input is needed from a range of developmental disciplines, including materials science, chemistry, biology, optics, and software and hardware engineering, as well as process design, equipment design, and project management. The expertise required to integrate all of these elements and ensure that a viable product is delivered can be difficult to obtain, particularly for smaller companies.

Therefore, developing more-quantitative LFIAs calls for a more collaborative approach. Initiatives being undertaken to enable the cross-pollination of ideas, technologies, and expertise that drives this process are exemplified by the One Step One Solution (OSOS) program operated by Diagnostic Consulting Network (Carlsbad, CA) in collaboration with BioDot Inc. (Irvine, CA). The OSOS program is designed to provide developers with a service that allows them to assess certain of the newer ICT-related assay technologies that are in development; to network with the providers of reader technologies, new substrates, and new label technologies; and to assess the utility of those state-of-the-art technologies in various systems.

Table II. (click to enlarge) Results of an HIV seroconversion panel tested with the OSOS HIV 1/2 fluorescent LFIA and the ESE handheld fluorescent sensor. Cutoff was determined by averaging five confirmed HIV 1/2 negative samples and three standard deviations

It also includes an educational component through which companies can learn firsthand the basics of LFIAs and stay abreast of cutting-edge developments. The microarray and microfluidics segments of the diagnostics industry rely heavily on collaborative, multidisciplinary approaches to development and manufacturing. This has involved considerable investment over the past decade. Adopting the concept of outsourcing components of multidisciplinary assay development programs will likely be critical to their success, particularly for smaller companies.

Figure 9. A handheld fluorescent LFIA reader.

Conclusion

The acid test for these newer technologies will be market acceptance, which in turn will be driven by factors such as utility, relevance to the application and the end-user, cost, manufacturability of the product, and the availability of clear license for use. A major factor in determining that this technology will be continually improved rather than replaced by a completely novel approach is its demonstrated utility and breadth of application.

The 2005 U.S. clinical diagnostic POC market has been estimated at $8.3 billion, approximately $3 billion of which involved POC tests not related to glucose monitoring.² The potential for significant reward accruing to a broadly applicable POC test that offers the positive attributes of an LFIA but has better performance characteristics is large.

And the clinical POC market is only the tip of the iceberg for this technology. Alternative market segments such as veterinary care, food microbiology, agriculture, aquaculture, environmental, and industrial health and safety all show signs of marked growth and great potential. To meet the needs of many new applications in these areas probably will require industrializing the types of approaches discussed in this article.

Brendan O’Farrell, PhD, is a principal at Diagnostic Consulting Network (Irvine, CA) and vice president of technology development at BioDot Inc. Jeff Bauer is also a principal at Diagnostic Consulting Network and is principal research scientist at BioDot Inc. The authors can be reached at bofarrell@dcndx.com and jbauer@dcndx.com, respectively.