Rapid and simple detection of foot‐and‐mouth disease virus: Evaluation of a cartridge‐based molecular detection system for use in basic laboratories

Summary Highly contagious transboundary animal diseases such as foot‐and‐mouth disease (FMD) are major threats to the productivity of farm animals. To limit the impact of outbreaks and to take efficient steps towards a timely control and eradication of the disease, rapid and reliable diagnostic systems are of utmost importance. Confirmatory diagnostic assays are typically performed by experienced operators in specialized laboratories, and access to this capability is often limited in the developing countries with the highest disease burden. Advances in molecular technologies allow implementation of modern and reliable techniques for quick and simple pathogen detection either in basic laboratories or even at the pen‐side. Here, we report on a study to evaluate a fully automated cartridge‐based real‐time RT‐PCR diagnostic system (Enigma MiniLab®) for the detection of FMD virus (FMDV). The modular system integrates both nucleic acid extraction and downstream real‐time RT‐PCR (rRT‐PCR). The analytical sensitivity of this assay was determined using serially diluted culture grown FMDV, and the performance of the assay was evaluated using a selected range of FMDV positive and negative clinical samples of bovine, porcine and ovine origin. The robustness of the assay was evaluated in an international inter‐laboratory proficiency test and by deployment into an African laboratory. It was demonstrated that the system is easy to use and can detect FMDV with high sensitivity and specificity, roughly on par with standard laboratory methods. This cartridge‐based automated real‐time RT‐PCR system for the detection of FMDV represents a reliable and easy to use diagnostic tool for the early and rapid disease detection of acutely infected animals even in remote areas. This type of system could be easily deployed for routine surveillance within endemic regions such as Africa or could alternatively be used in the developed world.

Furthermore, new portable molecular platforms have also been developed for field-based real-time RT-PCR (rRT-qPCR) (King et al., 2008) which integrate nucleic acid extraction, thermal cycling and calling of results without user intervention and are appropriate for use by non-specialists (Madi et al., 2012). These rRT-qPCR platforms such as the Enigma Field Laboratory â (FL), which utilize lyophilized reagents, have been successfully used within endemic settings for the rapid, simple, on-farm detection of FMDV in a range of clinical samples (Howson et al., 2015). These proof-ofconcept studies provide evidence to support the use of highly sensitive molecular assays in field settings; however, there may be other scenarios where it would be beneficial to analyse the samples at higher throughput near farms such as local laboratories using similar fully automated platforms. In this study, we report the performance of a simple-to-use cartridge-based automated rRT-qPCR platform, the Enigma MiniLab â (ML), onto which the gold-standard OIE-recommended rRT-qPCR assay for detection of FMDV was transferred and evaluated for its analytical sensitivity and robustness in European and African laboratories. In contrast to the FL that can be used on a farm, the ML is designed for use in basic laboratories and is a modular system that could harbour up to six units for independent detection and reporting of FMDVpositive samples.

| The Enigma ML â system
The cartridge-based Enigma ML â system combines fully automated nucleic acid extraction with rRT-qPCR and autonomous result calling. The system consists of one control module and up to six processing modules. Each processing module can independently run one sample, with a maximum of six samples running simultaneously.
The cartridges can be stored at room temperature, are single-use and fully contained. They include tools and reagents (buffers and magnetic beads) for the robotic nucleic acid extraction, freeze-dried rRT-qPCR mixes and a polymer capillary for thermal cycling and fluorescence detection. The only user interaction required is loading 2 ml of sample and starting the run with the touch screen interface. The time from sample loading to reporting of results is less than 90 min.
As previously described for the Enigma FL â prototype for onfarm use (Howson et al., 2015), the FMDV assay on the Enigma ML â platform is a lyophilized form of the OIE-recommended "gold standard" rRT-qPCR that targets the conserved 3D region of the FMDV genome (Callahan et al., 2002). The cartridge also includes an internal control that is run together with the FMDV assay. To this means, MS2 phage RNA is added to the sample before the nucleic acid extraction and is subsequently detected along with the FMDV target in a duplex rRT-qPCR (assay modified from (Rolfe et al., 2007)).  were tested in parallel with the Enigma ML â and the standard rRT-qPCR at the FLI and TPI for the milk sample (Table 1). In addition, cattle epithelial suspensions prepared from field outbreak samples (n = 53), two positive control samples (tongue epithelial suspensions prepared from SAT1 and O naturally infected cattle) and two negative epithelial samples were tested with both systems at the SUA in East Africa.

| European inter-laboratory proficiency test
The performance of the Enigma ML â was further evaluated in a col-

| Ethics statement
All clinical samples used in this project were samples collected by local authorities in endemic countries and submitted to TPI or SUA, or archival samples from experimental studies at FLI that had been reviewed by an independent ethics commission and approved by the competent authority (State Office for Agriculture, Food Safety and Fisheries Mecklenburg-Vorpommern, Rostock, Germany).

| RESULTS AND DISCUSSION
Rapid and reliable diagnostic systems can play an integral role in the detection, monitoring, control and subsequent eradication of animal diseases and pathologies such as FMD. Especially in non-specialized, front-line laboratories with only basic equipment, for example in rural Africa, a demand exists towards test systems that are simple to perform, robust, inexpensive, and that yield unambiguous results for easy reporting (Knight-Jones et al., 2016). In the recent past, several capacity building projects have sought to provide basic laboratories all over the world with reliable diagnostic systems for notifiable animal diseases. During these projects, real-time PCR machines and other technology have been transferred with initial training. However, unfortunately, the availability and expense of consumables within the donor countries have often been inhibitive to the sustainability of these initiatives. For this reason, reliable and easy to handle test systems without delicate technology are still needed. In this context, different candidate systems were explored within the EUfunded RAPIDIA-Field project (grant agreement no. 289364), among them a new automated molecular diagnostic platform, the Enigma ML â instrument, which combines RNA extraction, FMDV rRT-qPCR and a sample reporting pipeline suitable for deployment into simple field laboratories. Its evaluation in different laboratories and within an inter-laboratory comparison test is reported here.
Initial tests were carried out to determine the limit of detection of the Enigma ML â compared to a standard RNA extraction and rRT-qPCR. To this means, a 10-fold dilution series of FMDV O/UAE 2/2003 was used (Figure 1). When testing RNA extracted with the MagNA Pure LC robot in the rRT-qPCR on the Mx3005P, the last positive dilution, that is the limit of detection, was a 10 À6 dilution of the original material. RNA extracted by the Enigma ML â -tested positive to 10 À5 dilution. This one-log reduction in analytical sensitivity observed when the RNA was extracted on the Enigma ML â and the rRT-qPCR was performed on the Mx3000P, and was also observed when both steps were performed on the Enigma ML â . Previous studies have shown that the analytical sensitivity of rRT-qPCR assays is maintained when lyophilized reagents are used (Howson et al., 2015). Therefore, it is likely that the extraction protocol on the Enigma ML â is less efficient than a standard laboratory extraction robot, but the reason for the reduced efficiency and possible remedies is unclear. It has to be noted that the limit of detection was not fully quantified in terms of genome copies, and that differences between the methods mostly occurred in dilution steps that were only weakly positive in the standard method (C T -values > 30).
Nevertheless, samples of clinically diseased animals contain higher genome loads (Murphy, Bashiruddin, Quan, Zhang, & Alexandersen, 2010;Pacheco, Stenfeldt, Rodriguez, & Arzt, 2016), and thus, the apparent lower sensitivity of the Engima ML â platform is unlikely to affect the ability of this platform to reliably confirm FMDV infection in clinical animals.  F I G U R E 2 Diagnostic performance of the Enigma ML â compared to standard rRT-qPCR using field samples of cattle epithelium in an African laboratory. Open circles were reported as negative on the Enigma ML â printout; closed circles were reported as positive on the Enigma ML â printout (n = 51 FMDV positive, n = 2 FMDV negative) along with two positive (tongue epithelial suspensions from FMDV SAT1 and O naturally infected cattle) and two negative controls (tongue epithelial suspensions healthy cattle) were tested on the Enigma ML â and compared to the standard laboratory RNA extraction and rRT-qPCR pipeline used at SUA (Figure 2). There was 100% concordance between the Enigma ML â and the standard rRT-qPCR for the four control samples.
The Enigma ML â also correctly classified 48 of the 53 epithelial samples (90.6%), but five samples which were considered positive by standard rRT-qPCR were reported as negative by the Enigma ML â printout. These samples when assayed using the gold standard qRT-PCR had an average C T value of 24, and therefore, training the algorithms of the Enigma ML â to recognize these samples as positive is required for future improvement of the system to ensure reliability of automated text only result calling.
When the proficiency test panel was analysed in four European laboratories, the relative sensitivity was 100% with all positive samples being accurately reported by the Enigma ML â in all laboratories (Table 2). One participant reported a positive result in one duplicate for the negative sample 12, resulting in a relative specificity of 95% (Table 2). Based on all aliquots tested, the overall accuracy was 99%, and 83 of 84 aliquots were correctly identified across four laboratories, underscoring the robustness of the Enigma ML â platform.
Overall, the Enigma ML â displayed good concordance when compared to the standard laboratory RNA extraction and RT-qPCR pipeline for clinical and laboratory-prepared samples. Provided that future validation and optimization of the RNA extraction step can bring its performance to match that of standard laboratory extraction robots, the comparative data for the lyophilized reagents indicate that it is possible to generate a stabilized assay with equivalent (or better) performance than the wet-assay format.
The data presented here show that it is possible for a fully automated molecular diagnostic platform, which is easy to use and requires minimum input from the operator, to rapidly detect FMDV We are thankful to Sandra Blome and Michael Eschbaumer for critically revising the manuscript. Sadly since completion of this article B.
Haas has passed away. We would like to dedicate this short communication to him.

CONFLI CT OF INTEREST
James Wood and Paul Martin were employees of Enigma Diagnostics and the reported work was partially funded by the RAPIDIA- O R C I D