Medical Device Link.

POINT-OF-CARE TECHNOLOGIES

Figure 1. The FluoSens sensor (top). The 90-g optoelectronic unit is implemented in a housing as a commercial product (bottom).

This article describes an approach to developing point-of-care (POC) nucleic acid–based testing platforms that provide rapid results on a subtyping specificity level and are sensitive and mobile, and shows how the systems can be applied to MRSA testing. The underlying assay works from blood directly, is able to be scaled up, and makes most IVD sample-preparation steps unnecessary.

In the previous installment of this discussion of new POC techniques, handheld devices for lateral-flow immunodiagnostic applications (including detection of MDMA, or Ecstasy) were considered.² The case was made that combining devices with assays that work from clinical samples directly without extensive sample preparation is a favorable approach to rapid and cost-efficient POC and other field-based testing. It is certainly simpler and more efficient than trying to miniaturize complex procedures that require multiple steps.

Furthermore, mobile and handheld devices designed for on-site testing render obsolete the need to send the sample to a testing laboratory. Bringing the test to the sample source saves time and cost. It also reduces the risks of sample contamination and sample degradation due to incorrect or prolonged storage, and can therefore decrease the incidence of false-positive and false-negative results.

Handheld and portable testing platforms for nucleic acid–based tests using recombinase polymerase amplification (RPA) have been developed on the same modular system as was described in Part 1 of this article. These devices provide documentation of results and quantitative data. They are economically priced and well suited to serve both the POC and home-care test markets due to their specificity, speed, and simplicity.

Limits of Current Techniques

Nucleic acid amplification can provide accurate and rapid detection of viral, bacterial, and fungal species on a subtyping level of specificity. For example, the dangerous H5N1 subtype of avian flu can be distinguished from other, less dangerous, influenza virus subtypes. Bird flu is just one prominent case. There are currently more than 1400 infectious diseases known to modern medicine. Nucleic acid– based testing is now routinely applied to food pathogens, genetically modified organisms, and antibiotic-resistant pathogen strains such as MRSA.

Although the polymerase chain reaction (PCR) nucleic acid amplification method is extremely sensitive and specific, it is not suitable for use in a handheld or mobile format, particularly when rapid turnaround is desired. The extensive necessary heating and cooling steps require sophisticated and heavy instrumentation such as a thermal cycler, consume a lot of power, and, most important, involve time-consuming and troublesome sample-preparation procedures.

In addition, temperature homogeneity is very critical for achieving accurate results with PCR, and a high degree of sample purity is required.

RPA, an isothermal nucleic acid amplification technique, avoids these limitations.³ The use of this innovative technology in nucleic acid–based testing platforms is discussed later in this article.

Benefits of Larger Sample Volumes

The nature of the inhibitors in samples, as well as approaches to reducing their effects, may have to be rethought once novel nucleic amplification methods have been assessed. For example, rather than purify the nucleic acid to remove potential inhibitors, the inhibitors may be rendered inconsequential through dilution. This approach, however, requires a method that can be scaled up in terms of volume.

It is difficult to perform PCR in high volume, as the heating and cooling steps are not compatible with the time-to-result requirement of volume testing. Additionally, working from clinical samples at high temperatures may lead to coagulation of proteins and outgassing of the samples. With such methods, extensive sample preparation is a must. But every step is less than 100% efficient, which results in a loss of targets. This is of particular concern when dealing with low copy numbers.

However, if a method can be scaled up, then inhibitors can be diluted, no concentration procedure has to be applied, and the entire sample volume can be used. Thus, not a single target molecule will be lost. Enrichment steps, such as growth of cell cultures, may be significantly shortened. Fewer steps also means a smaller risk of sample contamination.

Higher sample volumes are desirable also for addressing the target delivery problem, which is to deliver a statistically relevant and representative number of target molecules to the test. Blood tests, for example, often require that a sample volume of 0.5 ml be drawn in order to avoid missing targets in extremely low concentrations and to conclude with meaningful results. Regulatory requirements for food tests require even as much as 25 g of sample in order to eliminate the target delivery problem.

Currently, time-consuming concentration or enrichment steps are applied, since there are not many rapid, sensitive, and scalable methods that can detect pathogens at a subtyping level of specificity. Lateral-flow tests and a cell-based test called the Canary Technology hold promise in this regard.⁴ Besides these, scalable isothermal nucleic acid amplification tools have the greatest potential to overcome the bottlenecks in developing simple, rapid, sensitive, and specific diagnostic tests. Such tests are desired particularly in the field, in POC and home-care settings, where bulky and sophisticated instrumentation is not available or cannot be operated, and in every healthcare setting where cost is an issue.

Now, however, most isothermal methods lack either the necessary scalability, sensitivity, rapid time to result, or specificity, or some combination of these parameters. The remainder of this article discusses devices and methods toward solving these challenges.

Device Design and Function

Both tubes and lateral-flow strips have been used to read out the results of nucleic acid–based tests in colorimetric and fluorescent modes.^3,5,6 In addition to the lateral-flow readers described in Part 1 of this article, ESE Embedded System Engineering GmbH (Stockach, Germany) has also built handheld and mobile devices for the readout of nucleic acid tests directly in tubes, both for real-time monitoring and for end-point reading. One such device, the FluoSens Reader, is a fluorescence sensor that illuminates a reaction tube from the side, the bottom, or the top, and has been used in conjunction with RPA to monitor the nucleic acid amplification reaction in real time. The sensor includes two excitation sources and two detectors for different wavelengths. The mini­aturized sensor permits mobility (see Figure 1).

Figure 2. (click to enlarge) Principles of confocal (left) and off-axis (right) sensor design compared. The confocal geometry allows for higher tolerances in positioning accuracy. If Position A moves vertically, the off-axis system goes out of focus and will not yield a signal.

In order to maximize their positioning tolerances and enable them to be adjusted to different focal lengths, ESE designed the sensors according to confocal principles rather than an off-axis geometry (see Figure 2). Nonstationary sensors, such as those used for surface measurements in field-based tests, need to accommodate higher positioning tolerances. Confocal technology makes this possible. The confocal devices are capable of measuring femtoamperes of current (see Figure 3) and picomolar solutions without integration, and will give equally accurate results whether exposed to ambient light or not.

Figure 3. (click to enlarge) The voltage output of the electronics element of the FluoSens sensor relative to ambient temperature, measured at a 10-GΩ amplification factor. Voltage of 1.5 V was introduced through a 10-GΩ series resistance, being 150 pA of current. The measured output is in the range of 1.65 V, which, at 10-GΩ resistance, translates to about 165 pA. The optical system using these electronics provides typical outputs down to around 0.1 mV, or 16 fA of current, which the sensors measure in the standard setup. Temperature dependence is surprisingly flat.

In liquids, the sensors have been shown to detect 0.5-pM solutions of fluorescein in a standard cuvette, thus demonstrating very high sensitivity (see Figure 4). In combination with RPA, this system exhibited sensitivity down to a few molecules, with results obtained, as discussed below, in about 20 minutes.

Figure 4. (click to enlarge) Sensitivity (detection limit) of the confocal sensors developed by ESE has been shown to be in the 0.5-pM range for fluorescein, pH 12 in a standard glass cuvette.

Despite the sensitivity of the system in terms of analyte concentration, its electronic components must handle currents in the femtoampere range. The standard optoelectronic unit can measure down to the order of 10 fA, as demonstrated in the experiment represented by Figure 3. The devices are therefore very well suited to serve for other analytical techniques also, such as electrochemical measurement, offering them the highest sensitivity and accuracy. The dependence of the sensors’ data output on their ambient temperature is low, as can be seen from Figure 3.

Light-emitting diodes (LEDs) are used as light sources in these devices, and photodiodes as detectors. Closed-loop feedback is designed into the sensors to provide a constant illumination output independent of the LED’s temperature. The lengths of the electronic circuits are kept extremely small. The sample is detected once with the LED turned on and once with the lamp turned off, in order to allow subtraction of ambient light (see Figure 5). This measurement principle results in fast measurements and low noise both with and without ambient light and on surfaces as well as in liquids.

Figure 5. (click to enlarge) Oscilloscope plot of one measurement period for the optoelectronic unit of the FluoSens, showing truly parallel fluorescence detection in two channels suitable for the measurement of two different dyes and wavelengths. Only one value is taken after a specified delay, so there is no integration of signal. Measurements without ambient light (a) and with ambient light (b) are equally accurate, the signal offset in channel 2 caused by ambient light being compensated for through subtraction of the value recorded once with the light source switched on and once with it switched off.

Another ESE device, the handheld TubeScan reader (in prototype), contains the same microoptoelec-tronic unit as the FluoSens. It can measure several tubes at a time (see Figure 6). This scanner moves a temperature-controlled tube rack over the sensor and illuminates the tubes from below. The fluorescence intensity of each tube is plotted versus time and displayed in one graph. The device can read two colors.

Figure 6. The TubeScan reader provides readout of real-time, end-point, and other fluorescence or colorimetric reactions and kinetics in tubes, such as SYBR-Green assays, PCR end-point determination, and DNA and protein quantification based on fluorescence.

The TubeScan runs off batteries for more than 2 hours at a charge, and has an internal memory. The tube rack heats 200-µl tubes to 37°C, the optimum temperature for RPA, from a starting point of 20°C within 2 minutes.

Reading out real-time, end-point, and other fluorescence or colorimetric reactions and kinetics in tubes, the device can be used to monitor isothermal DNA/RNA amplification at the point of care or in the field. Due to its modular design, it can be configured for either fluorescence or colorimetric readout, and can be adjusted to various formats, such as tubes, cuvettes, microfluidic chips, capillaries, gels, microscope slides, and others, because the carriage is interchangeable. The modular approach allows for rapid, cost-efficient customization and application versatility.

Application Example: MRSA

Molecular diagnostic techniques have gained a substantial share of the market for routine IVD tests. The appearance of antibiotic-resistant strains of pathogens at alarmingly increased levels is one reason for the greater demand for such tests. In reference laboratories, DNA- and RNA-based tests constitute one of the fastest-growing categories of products and services due to their high levels of sensitivity and specificity, which can allow for subtyping of species. Such high specificity can be matched only by time-consuming microbiological procedures that require days or weeks to complete. Current nucleic acid–based diagnostic tests, by contrast, give rapid access to results, taking about an hour.

Figure 7. (click to enlarge) Direct detection of DNA in whole blood using recombinase polymerase amplification. One hundred copies of MRSA DNA were specifically amplified, as is evident in the gel electrophoresis image. Total sample volume was 100 µl (70 µl of blood and 30 µl of reagents. M = size marker. Experiment and images are courtesy of ASM Scientific Ltd. (Cambridge, UK).

This section describes nucleic acid–based tests using RPA that produce results in 20 minutes or less while maintaining high sensitivity and specificity and adding mobility as a feature. Experimentation has shown that RPA yields amplification in whole blood and that the procedure accommodates such clinical samples without the need for preparatory steps (see Figure 7).

MRSA DNA was spiked into a blood sample in various concentrations (1000, 100, 10, and 0 copies) and then subjected to RPA at 37°C. Total sample volume in the experiment was 100 µl, and the blood was diluted by a factor of 30% via reagents added directly. This feature of RPA suits the procedure well for field-based tests: field tests using RPA can provide a rapid turnaround because they avoid the need to store, transport, or extensively prepare the sample. Efforts are underway at ESE to adapt the handheld devices described here to fluorescence measurement from blood directly.

RPA and the TubeScan and FluoSens devices together offer a handheld mobile diagnostic nucleic acid–testing platform for real-time and end-point detection.

Figure 8. (click to enlarge) Real-time detection of MRSA III to determine antibiotic resistance using RPA and the mobile FluoSens device. The sensor took a little more than 11 minutes to detect 2000 copies of MRSA.

The devices described in this article add another dimension to the groundbreaking RPA technology by making real-time measurements mobile while further speeding up the process. Because of the low noise and high sensitivity of the handheld fluorescence readers, it proved possible to demonstrate experimentally the detection of 2000 isolated DNA copies of MRSA in just above 11 minutes (see Figure 8). An end-point assay showed that the system was capable of detection down to 17 copies (see Figure 9). Interestingly, at these copy numbers—that is, fewer than 104 copies—the end-point assay provides a quantifiable result.

Figure 9. (click to enlarge) Results of an end-point assay in which 0, 17, 170, and 1700 copies of isolated MRSA III DNA were amplified for 60 minutes using RPA and detected using the mobile FluoSens system.

The handheld TubeScan prototype provides a kinetic readout of tubes while measuring the increase in fluorescence generated by the reaction over time as induced by a fluorescence resonance energy transfer (FRET) mechanism.³ The tubes are moved over a light source and illuminated from below. Due to the confocal geometry of the device, the recorded fluorescence values are measured through the same optics as were involved in the excitation. Scans can be done every 5 seconds for several hours.

Figure 10. (click to enlarge) Fluorescence real-time nucleic acid amplification in the handheld TubeScan device as a kinetic plot of raw data. The horizontal axis of the graphs represents the scanner or tube position, and the vertical axis represents fluorescence intensity. Experiment and images are courtesy of ASM Scientific Ltd. (Cambridge, UK).

In the handheld TubeScan, readings of the peak values are saved in the internal memory. Quantitave data appear in the internal display as numbers, or positive/negative, or high/medium/low results. A graphical user interface can be added, if desired. Although this is a battery-operated stand-alone device, it can be connected to a computer via a USB port for assay development purposes and to illustrate the data graphically—for example, as multiple plots proceeding through time (see Figure 10).

In an experiment, 2000, 200, and 0 copies of an undisclosed bacterial DNA strain were subjected to RPA at 37°C. (The zero-copy sample was the no-template control.) Fluorescence real-time data were recorded using the battery-operated handheld TubeScan developed in prototype by ESE. After 19 and 26 minutes of reaction time (see Figure 10), the 2000- and 200- copy samples show clear peaks, well above noise level, whereas the no-template control remains low.

Figure 11. (click to enlarge) Representative scan of reaction tubes after amplifying MRSA III DNA for 22 minutes, produced by the handheld TubeScan device.

Examination of the plot of a representative scan for an RPA reaction of MRSA III DNA reveals clearly the quantitative sensitivity of the system (see Figure 11). During the reaction, tubes were kept at a constant temperature of 37°C. Each peak in the scan reproduced in the figure represents the fluorescence intensity in a tube at the time of the scan. There were three tubes per scan: tube 1 contained 2000 DNA copies, tube 2 contained 200 copies, and tube 3, containing 0 copies, was the no-template control. No peak is visible in tube 3, as was expected. Each scan took about 1 second. In reading the reaction, the operator scans the tubes multiple times, as often and as frequently as he or she may specify, in order to obtain the kinetics of the reaction in each tube.

Conclusion

Battery-operated handheld and mobile diagnostic testing platforms for nucleic acid–based testing as well as for lateral-flow immunodiagnostics have been built. They have been shown to provide sensitive, accurate, and specific results, as well as rapid turnaround. The stand-alone devices demonstrate operational and physical robustness, and they can be manufactured to be affordable. The underlying assays work directly from such clinical samples as urine or blood.

Due to their mobility, the platforms require no sample storage or transportation. Due to their ability to work from clinical samples directly, they avoid most of the trouble of sample preparation. Sample prep, storage, and transportation are the most time-consuming aspects of every diagnostic test procedure. Therefore, with the test platforms described here, turnaround times can be short; they have been demon­strated to be in the range of a few minutes (rather than hours or days) for both immunodiagnostic methods and nucleic acid–based tests.

The platforms are also very well suited for volume scale-up, thus addressing the target delivery problem, and provide very high sensitivity combined with subtyping specificity. Sensitivity down to a few molecules on a subtyping specificity level has been proven experimentally.

Konrad Faulstich, PhD, MBA, is head of business development at ESE Embedded System Engineering GmbH (Stockach, Germany). He can be reached at konrad.faulstich@ese-gmbh.de.	Michael Eberhard is chief electronics engineer at ESE Embedded System Engineering GmbH. He can be reached at michael.eberhard@ese-gmbh.de.	Roman Gruler is head of R&D at ESE Embedded System Engineering GmbH. He can be reached at roman.gruler@ese-gmbh.de.	Klaus Haberstroh is CEO and cofounder of ESE Embedded System Engineering GmbH. He can be reached at klaus.haberstroh@ese-gmbh.de.