Development of molecular confirmation tools for swift and easy rabies diagnostics

Rabies is a lethal zoonotic disease caused by a group of 16 negative-strand RNA viruses of the genus Lyssavirus in the family Rhabdoviridae of the order Mononegavirales [1]. It is a societal tragedy that in the twenty-first century, rabies, a zoonosis that can easily be prevented in humans and controlled in domestic animal species is still neglected and continues to create a significant social and economic burden on a global scale [2–4]. In low-income countries where control efforts are lacking and awareness of the disease and access to appropriate preventive and post–exposure prophylaxis is limited or non–existent, rabies is estimated to still cause around 60,000 human deaths per year [5, 6].

The real impact of this neglected zoonotic disease is likely to be underestimated. One major problem here is surveillance, which unfortunately is inadequate or even non-existing in many Asian and African countries, where the burden of rabies is highest [6–9]. Adequate rabies surveillance in both humans and animals goes hand in hand with rapid and reliable rabies diagnosis. Currently, rabies diagnosis relies on laboratory tests for detection of viral antigens. Among those tests, the fluorescence antibody test (FAT) is the “gold standard test” in post mortem rabies diagnosis [6, 10]. Recently, alternative antigen detection methods such as ELISA, the direct rapid immunohistochemical test (DRIT) or the indirect rapid immunohistochemistry test (IRIT) have been developed which do not require expensive fluorescence microscopic equipment along with the expertise and financial input needed to maintain them [11–15]. The rabies tissue culture infection test (RTCIT) or the mouse inoculation test (MIT) are mainly used as confirmatory tests and based on the complex and time consuming propagation and isolation of the virus [16, 17]. All these test methods, however, are (i) dependent on the quality of the supplied sample, or (ii) require longer turnaround times [6, 18]. Easy to use and swift test systems for rapid, cost-efficient diagnosis, with no loss of sensitivity or specificity would therefore improve the diagnostic situation significantly [12, 18, 19]. In this respect, immunochromatographic strip tests, also called lateral-flow devices (LFDs) or antigen-capture point-of-care tests, have a great potential [20–26], however for rabies there is still a need for standardization and quality check before being an alternative for rapid and simple diagnostics in resource-limited settings [27].

In the course of the past three decades, molecular tools based on the detection of the genetic information of rabies virus have become more widely accepted for the diagnosis of rabies [18]. The development of reverse transcription polymerase chain reaction (RT-PCR) techniques provided an alternative method for post mortem rabies diagnosis [28, 29], and enabled rapid ante mortem diagnosis of human rabies [30–32]. Next to conventional or real-time RT-PCRs for rabies [18, 32–37], alternative rapid genome detection tests have been developed, i.e. nucleic acid sequence based amplification and reverse-transcriptase loop-mediated isothermal amplification (RT-LAMP) [34, 38–43]. Other nucleic acid based methods, i.e. HighSpeed RT-qPCR and isothermal recombinase polymerase amplification (RPA), successfully used for detection of other viral pathogens including Schmallenberg virus (SBV), bovine viral diarrhea virus (BVDV), or foot-and-mouth disease virus (FMDV) [44–46] have not been applied for RABV yet. Also, initial efforts to optimize and shorten the process of nucleic acid extraction [47] by remaining highly sensitive and specific have not been further explored.

Therefore, the aim of this study was to (i) design and test both a HighSpeed RT-qPCR and a RT-RPA assay for rapid detection of the RABV, (ii) test the suitability of novel RNA extraction methods that improve reproducibility, analytical sensitivity and operational performance for brain tissue and (iii) combine them with the new downstream assays using different amplification kits. Those tests could then be considered both in advanced as well as in standard laboratories for the detection and confirmation of rabies virus, and thus contribute to the development of better tests for rapid, economical diagnosis for rabies.

Although new techniques have been developed for the diagnosis of rabies in recent years, the number of laboratory confirmed human rabies cases from the most affected countries is limited and represents an underestimate of the real impact of the disease [19, 29]. Therefore, the WHO stresses the need for better tests for rapid and economical diagnosis, with no loss of sensitivity or specificity [6]. The quick and simple to perform diagnostic tests would be ideal for use particularly in countries where laboratory infrastructure is still unfavorable [18, 20]. In the range of diagnostic assays molecular tests are getting more and more attention [18, 29]. Because of the higher sensitivity and specificity over antigen directed assays the molecular nucleic acid amplification tests, i.e. PCRs, are increasingly used as comparative standard detection methods [18, 32–37], although they are not recommended for routine use by international organizations yet [6]. Interestingly, proficiency trials among European laboratories showed that PCRs had less false negative results compared to virus isolation using the standard RTCIT [17].

In this study, we developed and validated specific and sensitive rapid molecular detection methods for the detection of RABV, compared to a published assay and improved the performance of the assays by the combination with novel RNA extraction methods. The results show that simplification of this kind of assays with no relevant loss of sensitivity or specificity can be achieved; thus encouraging both their (i) acceptance as a rapid confirmatory test to first-line assays and (ii) extension of its application in many laboratories, thus improving the overall diagnostic capacities.

However, for detection and amplification of lyssaviral target genomes high quality extraction of RNA is crucial. Manual RNA extractions are prone to errors and relatively time-consuming, but also conventional automated extraction techniques often have lengthy procedures and need trained staff. Recently, automated rapid magnetic bead-based RNA extraction methods have been described but only optimized for blood and serum samples [47]. Therefore, three commercial rapid extraction methods were tested regarding their suitability for brain tissue as the material of choice for post-mortem rabies diagnosis and compared to standard extraction method. In terms of reproducibility, the TRIzol & RNeasy method showed the highest intra-run variations (6.6%), and the three rapid extraction methods delivered nearly comparable results. While in the latter case, the EZ1 method as a fully-automated platform was assumed to show lowest coefficients of variation, interestingly best results were obtained with the manual SpeedXtract (SXT) method (2%) (Table 1). Although variations due to variable sample input or samples taken from a different brain region cannot be completely ruled out, sensitivity and efficiency of RNA recovery was acceptable for all rapid RNA extraction methods, with highest rates for SXT as well as KF Duo (Fig. 1). All three tested rapid RNA extraction methods used magnetic particles for nucleic acid extraction but request different instrument equipment. The advantage of the reverse extraction procedure of SXT is that it only requires a heat block and a magnetic stand, whereas the other two rapid methods employ technically sophisticated instrumentation. Both instruments differ enormously in size and weight: the KF Duo is much smaller (40 cm × 46 cm × 34 cm; 17 kg) than the EZ1 (51 cm × 57 cm × 57 cm; 48 kg) making it more attractive for use in smaller laboratories or integration into a mobile laboratory. Regarding the processing time, again SXT and KF Duo performed best (Table 2). The longer total hands-on time for the EZ1 is due to the extensive pre-filling step of cartridges, hence, commercially available cartridges are not suitable for application with shortened extraction protocol. It must be emphasized, however, that the EZ1 method developed by Aebischer et al. includes a self-assembly step, where empty EZ1 cartridges are filled with optimized extraction buffers [47]. Consequently, this EZ1-based method currently is not commercially available and its use depends on the supply of eligible kits (Table 4).

Reduced turnaround times and improved applicability of molecular detection methods can be achieved in different ways. Usually, HighSpeed RT-PCRs take advantage of special but expensive PCR machines with fast heating and cooling ramps, but even with a standard thermocycler (CFX96) an enormous reduction of reaction time for an established detection protocol (R14 RT-qPCR to HighSpeed RT-qPCR) can be achieved.

While the use of thermocyclers depend on a permanent power supply, isothermal amplification methods like RPA can be run, for example, on the ESEQuant TS2; a portable stand-alone, battery powered instrument with multi-channel real-time fluorescence detection capabilities. Highly specific loop-mediated isothermal amplification (LAMP), rolling circle amplification (RCA), strand-displacement amplification (SDA) or helicase dependent amplification (HDA) are either relatively cumbersome in terms of primers design, not rapid enough or costly despite relatively easy primer design and fast amplification [44]. RPA is based on the formation of a recombinase filament, strand displacement and abasic nucleotide analogons and has a broad temperature spectrum [50]. Even successful amplification of target RNA by RPA using body heat has been shown [51]. A further advantage of the RPA method is that in contrast to RT-qPCRs the dried reagents of RT-RPA do not require a cold chain. Unfortunately, the lower analytical sensitivity of RT-RPA (1000 target RNA copies per μl reaction) with a smaller dynamic range seems to run the advantages futile (Fig. 3). This analytic sensitivity seems to be comparable to other described isothermal amplification methods for rabies [39, 42]. However, these methods are proposed as suitable for rabies diagnostics. Although the tested RNAs represent different lyssavirus species and major RABV lineages from different regions of the world, the genetic diversity is still not fully covered and the methods requires further validation. However, at least for the HighSpeed RT-qPCR, which is basically a high speed version of the standard R14 RT-qPCR [34] a combination as multiplex or parallel assays with the R13 RT-qPCR, a slightly modified version of the Wakeley protocol [52], could help to overcome the diversity among RABVs and limit the chance of false negatives tremendously [34].

Despite lower analytical sensitivity, RPA in combination with SpeedXtract would be a promising candidate for simplified molecular diagnosis of rabies. Furthermore, molecular detection methods seem to be more reliable than the diagnostic standard FAT in decomposed animals [53]. As the viral load in brain tissue of rabid animals is relatively high, the methods can aid and enhance the passive laboratory confirmed surveillance under resource-limited settings [27]. Furthermore, the simple test approaches with no requirement for cold chain, could even be integrated into mobile test systems, as shown for food-and-mouth-disease virus or Ebola virus [45, 54] (Fig. 5; Tables 4 and 5).

Commercial magnetic bead-based rapid RNA extraction methods are suitable for brain tissue and show high sensitivity as well as a high level of reproducibility compared to standard RNA extraction methods and thus, can help standardizing RNA extraction and molecular diagnostics under routine conditions.

Furthermore, the here presented proof-of-principle simplified rapid rabies virus molecular detection methods (HighSpeed RT-qPCR and RPA) showed almost no loss of diagnostic sensitivity or specificity compared to validated standard molecular assays. Thus, the simple, quick and sensitive virus RNA extraction from brain samples combined with fit-for–purpose detection methods may improving the reliability and acceptance for the rapid molecular diagnosis of rabies. Particularly the high analytical sensitivity makes the HighSpeed RT-qPCR a potential candidate as a method of choice as rapid (i) ante mortem diagnostics for human rabies, (ii) differential diagnostics for organ transplantations and (iii) confirmatory diagnostics to first-line assays including integrated bite management and subsequent post-exposure prophylaxis. Especially under resource-limited settings, the SpeedXtract based RNA extraction combined with the RPA detection of the rabies genome could be a useful and robust molecular diagnostic approach.