Development of a recombinase polymerase amplification lateral flow assay for the detection of active Trypanosoma evansi infections

Background Animal trypanosomosis caused by Trypanosoma evansi is known as “surra” and is a widespread neglected tropical disease affecting wild and domestic animals mainly in South America, the Middle East, North Africa and Asia. An essential necessity for T. evansi infection control is the availability of reliable and sensitive diagnostic tools. While DNA-based PCR detection techniques meet these criteria, most of them require well-trained and experienced users as well as a laboratory environment allowing correct protocol execution. As an alternative, we developed a recombinase polymerase amplification (RPA) test for Type A T. evansi. The technology uses an isothermal nucleic acid amplification approach that is simple, fast, cost-effective and is suitable for use in minimally equipped laboratories and even field settings. Methodology/Principle findings An RPA assay targeting the T. evansi RoTat1.2 VSG gene was designed for the DNA-based detection of T. evansi. Comparing post-amplification visualization by agarose gel electrophoresis and a lateral flow (LF) format reveals that the latter displays a higher sensitivity. The RPA-LF assay is specific for RoTat1.2-expressing strains of T. evansi as it does not detect the genomic DNA of other trypanosomatids. Finally, experimental mouse infection trials demonstrate that the T. evansi specific RPA-LF can be employed as a test-of-cure tool. Conclusions/Significance Compared to other DNA-based parasite detection methods (such as PCR and LAMP), the T. evansi RPA-LF (TevRPA-LF) described in this paper is an interesting alternative because of its simple read-out (user-friendly), short execution time (15 minutes), experimental sensitivity of 100 fg purified genomic T. evansi DNA, and ability to be carried out at a moderate, constant temperature (39°C). Therefore, the TevRPA-LF is an interesting tool for the detection of active T. evansi infections.

Introduction Trypanosoma evansi is a haemoflagellate parasite which is closely related to T. brucei, the causative agent of human sleeping sickness and nagana in animals [1]. T. evansi is the causative agent of "surra" or "mal de caderas", which is the most common and widespread trypanosomal disease of domestic and wild animals and is characterized by high morbidity and mortality. The parasite is mechanically transmitted by biting flies and is found in many regions around the globe [2][3][4][5][6]. Outbreaks of surra have been reported in all types of ungulates (camels, cattle, buffaloes, horses, pigs, and deer) in Africa [7], Asia [8][9][10], Latin America [11][12][13] and recently Europe [14][15][16]. While T. evansi is commonly known as non-infective to humans, human infections were recently reported and confirmed in India and Vietnam, indicating that T. evansi may be emerging as a potential human pathogen [17][18][19][20]. Control of T. evansi trypanosomosis is mainly accomplished by drug treatment, but resistance of T. evansi to trypanocidal compounds has been reported in Africa [21,22] and in the far east of Asia [23].
T. evansi parasites are classified into two groups based on their kDNA minicircle type [24], which are characterised by the presence (Type A) or absence (Type B) of the gene encoding the RoTat1.2 variant surface glycoprotein (VSG) [25,26]. T. evansi Type B are less commonly found and have only been reported to occur in certain regions in Africa [27][28][29][30][31][32]. In contrast, T. evansi Type A are widespread. Many diagnostic methods are available to detect T. evansi infections and include parasitological, serological, and molecular assays [33]. While some methods detect both T. evansi Types A and B, others are specific to one of both types. Conventional blood smear examination technique is widely used in the field and detects both T. evansi Type A and B. However, it can only diagnose clinical stages of infection and not latent or chronic infection [34]. In addition, it is time consuming and requires both the presence of microscopy equipment and specifically trained personnel at the screening site. To overcome these shortcomings, the T. evansi card agglutination test (CATT/T. evansi) was developed. It is a standard test for epidemiological field studies of T. evansi Type A since it is based on the use of the T. evansi RoTat 1.2 VSG antigen as an agglutination agent for host antibodies [35]. The advantage of this technique is that it is fast, easy to execute and suitable for field diagnosis. The main disadvantage of the technique is the lack of discrimination between previous exposure and current infections. Indeed, the host antibodies that drive the reaction can be a result of an active infection, a past infection, repeated exposure without necessarily initiation of successful infection, or even polyclonal B cell activation by other infectious agents such as helminths [36].
The diagnosis of trypanosomosis has been improved by the development and application of DNA-based techniques such as PCR, which is a very sensitive and effective method for the detection of chronic infections or prepatent period of disease [37,38]. The DNA of killed trypanosomes does not remain in the blood for more than 24 to 48 hours, thus PCR-based assays are highly suitable for the detection of active infections [39]. Several genes have been investigated as targets for the PCR-based diagnosis of T. evansi; these include the RoTat1.2 VSG gene (Type A specific) [40][41][42], ribosomal DNA [43], a region from r-RNA internal transcribed spacer 1 (ITS-1) [44], the gene encoding the invariant surface glycoprotein ISG-75 [45], and the VSG JN 2118Hu gene (Type B specific) [26,28,46,47]. The drawback of PCR-based methods is that they require well-trained and experienced personnel and a laboratory environment suitable for correct protocol execution. Hence, they are difficult to deploy and maintain under most field conditions. An interesting alternative to PCR is the so-called Recombinase Polymerase Amplification (RPA) [48]. The reaction mechanism of RPA has been reviewed elsewhere [49,50] and is summarized in Fig 1 (the figure legend contains a detailed explanation of the RPA reaction). This isothermal nucleic acid amplification technology is simple, fast, cost-effective and is suitable for minimally equipped laboratories as well as for use in the field [51]. Hence, RPA is especially useful in infectious disease diagnostics and epidemiological studies [52][53][54][55]. The RPA reaction can be completed in 10 to 20 minutes at temperatures between 24˚C to 45˚C [56]. The amplification product can be visualized by gel electrophoresis or in real-time by the inclusion of a nucleic acid dye. The specificity and sensitivity of RPA are typically enhanced by probe-based methods, which (depending on the type of probe) allow amplicon detection based on fluorescence or a lateral flow (LF) assay [48]. To date, RPA has been successfully applied for the detection of bacteria [57,58], foodborne pathogens [59,60], parasites [61,62], and viruses [63,64].
In this present study, we describe the development of the first recombinase polymerase amplification lateral flow assay for the detection of active Type A T. evansi infections (TevR-PA-LF). The T. evansi RoTat1.2 VSG gene was chosen as the target for the TevRPA-LF for the following reasons: i) to ensure high specificity of the TevRPA-LF for T. evansi as this parasite is closely related to T. brucei, ii) T. evansi Type A are most commonly encountered and widespread, and iii) to allow comparison with the previously described PCR targeting the T. evansi RoTat1.2 VSG gene [33]. We demonstrate that the TevRPA-LF assay is highly specific for T. evansi since no cross-reactions with the closely related parasite T. brucei could be observed. In addition, we have tested the TevRPA-LF in an experimental mouse model and demonstrate that it can be used as a test-of-cure tool. The TevRPA-LF described here has a processing time of 15 minutes and can be performed at a constant temperature of 39˚C. Combined with the simplicity, robustness and reliability of the RPA-FL principle, the findings presented in this paper show that the TevRPA-LF can be a promising tool for the detection of active T. evansi infections.

Ethics statement
All experiments, maintenance and care of the mice complied with the European Convention for the Protection of Vertebrate Animals (ECPVA) used for Experimental and Other Scientific Step 4: the generated amplicons are again invaded by primer-recombinase complexes in a self-perpetuating cycle fueled in ATP by creatine kinase.
Step 5: an oligonucleotide (FAM-probe) carrying a 5' FAM tag, a spacer sequence and a 3' blocking group forms a complex with the recombinase and invades the biotinylated amplicon generated in the previous steps.
Step 6: only when the FAM-probe has successfully invaded the biotinylated amplicon and bound its complementary sequence, can the Nfo endonuclease bind and cleave the spacer region and 3' blocking group.
Step 7: after removal of the 3' region of the FAM probe, the Nfo endonuclease dissociates. This allows the DNA polymerase to employ the cleaved FAM-probe as a forward primer. Together with the biotinylated reverse primer (TevRPA-Rv-biotin) this leads to the formation of an amplicon bearing both the FAM and biotin tags. B: Read-out of the RPA via LF. The FAM-and biotin-tagged RPA product is mixed with the LF buffer, loaded onto the sample pad and is transported to the adsorbent pad through capillary flow. The RPA product is first bound by goldlabeled rabbit anti-FAM antibodies and later captured by a streptavidin-coated test line (TL). The control line (CL) is coated with anti-rabbit antibodies. While a valid negative test only contains a reddish band at the CL, a valid positive test will display bands at both the TL and CL. Purposes guidelines (CETS n˚123) and were approved by the Ethical Committee for Animal Experiments (ECAE) at the Vrije Universiteit Brussel (Permit Number: 14-220-31).

Preparation of purified genomic DNA
Total genomic DNA of the different parasites used in this study (Table 1) was extracted and purified from infected mouse whole blood using a DNeasy Blood & Tissue Kit (Qiagen, Germany) according to the manufacturer's instructions. The DNA was eluted in 50 μl nucleasefree water and stored at -20˚C until further use. The concentration and quality of the purified DNA were determined by gel electrophoresis (1% agarose gel run in TBE buffer at 110 V for 30 min) and spectrophotometric analysis (measurement of the absorbance at 260 nm, A260; examination of the ratio of the absorbances at 260 nm and 280 nm, A 260 /A 280 ; performed on a NanoDrop-2000/2000c).

Preparation of crude genomic DNA
Genomic DNA was robustly extracted by boiling. Briefly, 50 μl of blood was mixed with 10 μl nuclease-free water (Thermofisher). The sample was heated at 100˚C for 5 minutes followed by centrifugation at 20000 g for 5 minutes, and the supernatant was applied as a crude DNA template. The DNA template was kept at -20˚C until use.

RPA primers and probes design
The primers and probes were manually designed based on the gene sequence of the Rode Trypanozoon antigenic type 1.2 VSG (RoTat 1.2 VSG) of T. evansi (GenBank accession code: AF317914.1). The NCBI's nucleotide BLAST tools combined with Primer 5 were used to search for primers specific to T. evansi without significant overlap with other genomes. The TwistAmp LF Probe oligonucleotide backbone includes a 5'-antigenic label FAM group, an internal abasic nucleotide analogue 'dSpacer' and a 3'-polymerase extension blocking group C3-spacer. The details of the primers and probes used are given in Table 2.

Development and optimization of the TevRPA assay
The RPA reactions were conducted with the TwistAmp Basic kit (TwistDx, Cambridge, UK). A 47.5 μl reaction mixture containing the following components was prepared in a 1.5 ml tube: 2.4 μl of both forward and reverse primers (final concentration: 480 nM), 29.5 μl Reactions were halted by placing the tubes on ice. The amplified products were first purified using the GenElute PCR Clean-Up kit (Sigma-Aldrich) and visualized on a 2% agarose gel.

Development and optimization of the TevRPA-LF
LF-RPA assays were performed following the indications provided in the TwistAmp nfo kit (TwistDx, Cambridge, UK). Briefly, the RPA reaction was assembled as described above (Materials and Methods subsection 'Development and optimization of the TevRPA assay') with the exception of the addition of 2.1 μl of both forward and reverse primers (final concentration: 420 nM) and 0.6 μl probe (final concentration: 120 nM) to the reaction mixture. The amplified DNA was detected using LF strips (Milenia Hybridtech 1, TwistDx, Cambridge, UK) following the instructions indicated in the kit. Briefly, 1 μl of the amplified product was diluted with 99 μl LF buffer. Ten μl of this diluted sample was then loaded on the sample application area according to the manufacturer's instructions. The final result was visually read out after incubation for 2 minutes at room temperature. A testing sample was considered positive when both the detection line (biotin-ligand line) and the control line (anti-rabbit antibody line) were visible. A testing was considered negative when only the control line was visible (Fig  1). The amplicons could be analyzed on a 2% agarose gel after purification with the GenElute PCR Clean-Up kit (Sigma-Aldrich) to further confirm the testing result.

Evaluation of sensitivity and specificity of the TevRPA-LF
The specificity of the TevRPA-LF was assessed by employing 20 ng of purified genomic DNA isolated from various parasites (Table 1). Samples containing only nuclease-free water were used as negative controls.
The sensitivity of the TevRPA-LF was tested by employing the following concentrations of T. evansi purified genomic DNA as templates for the RPA reaction: 10 ng μl −1 , 1 ng μl −1 , 100 pg μl −1 , 10 pg μl −1 , 1 pg μl −1 , 100 fg μl −1 , 10 fg μl −1 and 1 fg μl −1 . The results were analyzed by lateral flow and agarose gel electrophoresis. The remaining mouse in each group was used as a negative control and was not infected. The mice were bled at different times post-infection. The mice in Group 1 were bled at days 1, 3, 5 and 6 post-infection. The animals in Group 2 were bled at days 0, 2, 4, 6, 8, 10 and 12 post-infection. All individuals from Group 2 were treated with Berenil (40 mg per kg), administered intraperitoneally at day 5 post-infection. For both groups, at each time point, 102.5 μl of whole blood was collected from the tail of each individual using nuclease-free tubes with 30 ml heparinized saline (10 units/ml; Sigma-Aldrich) to prevent coagulation. 2.5 μl of the collected blood was used to follow-up mice parasitemia by diluting the sample 200-fold (during high parasitemia periods) and 100-fold (during low parasitemia periods) in PSG buffer and counting the parasites under the light microscope. The rest of the collected blood (100 μl) was split into two parts to evaluate the samples using the TevPCR and TevRPA-LF. Fifty μl of collected blood was employed to prepare purified genomic DNA for the TevPCR, whereas the remaining 50 μl of collected blood was used to obtain crude genomic DNA for the TevRPA-LF. The TevPCR was performed as described in [40] with the following modifications: the amount of purified genomic DNA as starting material (250 ng vs. 3000 ng) and the addition of 10% DMSO to the reaction mixture.

Development and optimization of the TevRPA
The first requirement of the TevRPA-LF is a high specificity for the detection of T. evansi. This parasite is closely related to T. brucei and thus the selection of an appropriate nucleotide sequence that is unique to T. evansi is crucial. This is the case for a specific region (bp 1 to bp 1300) of the T. evansi RoTat1.2 VSG gene [40][41][42], which forms the target of the TevRPA-LF for T. evansi detection (Fig 1). This limits the use of the TevRPA-LF described here to the detection of Type A T. evansi, and not Type B. Based on this particular region, a primer pair was designed for the TevRPA such that the resulting amplicon does not exceed 500 bp (as suggested by the RPA manufacturer instructions). As can be seen from Fig 2A, an RPA with this primer pair (initially incubated at 37˚C for 30 minutes) on T. evansi purified genomic DNA extracted from infected mice blood yields an amplicon of around 289 bp. The reaction was also performed on genomic DNA purified from a naive mouse to exclude the possible lack of specificity due to cross-reactivity. No amplification could be observed in this negative control sample (Fig 2A). Next, the assay conditions were optimized by allowing the RPA reaction to proceed at various incubation temperatures and amplification times. First, a range of incubation temperatures between 25˚C and 50˚C were tested at a constant amplification time of 30 minutes. As can be seen from Fig 2B, 39˚C represents the most optimal incubation temperature as it produces the highest amount of amplicon. In a second phase, the RPA was performed at a constant incubation temperature of 39˚C while varying the amplification times from 5 to 40 minutes in 5 minute increments (Fig 2C). Although the TevRPA can be performed within 10 minutes, longer incubation times clearly yield a higher signal. The amplification time of 15 minutes was selected in an effort to maintain a balance between providing maximum sensitivity and obtaining a minimal reaction time. In conclusion, these experiments demonstrate that the TevRPA may be reliably performed with an amplification time of 15 minutes and an incubation temperature of 39˚C. These conditions were maintained for all subsequent experiments.

The TevRPA can be translated into a specific and sensitive TevRPA-LF
The visualization of the RPA amplicon via agarose gel electrophoresis requires an additional purification step to avoid smeared bands on the gel due to the presence of enzymes and crowding agents [50]. This additional handling step is not necessary if the assay's read-out is performed via a lateral flow (LF) device [48,49]. However, the translation of an RPA to an RPA-LF necessitates the addition of a labeled probe to the RPA reaction mixture and the biotinylation of the RPA reverse primer (Fig 1). Two candidate probes were screened for their potential to generate an RPA-LF for T. evansi detection (from here on referred to as TevR-PA-LF). Although both probes gave rise to positive signals when tested on T. evansi purified genomic DNA in both agarose gel electrophoresis and lateral flow detection formats, probe 1 clearly generates false positives while probe 2 does not (Fig 3A, right and left panels, respectively). Therefore, probe 2 was selected to be incorporated in the RPA assay to allow postamplification detection of the amplicon via the TevRPA-LF.
Next, the specificity of the TevRPA-LF was evaluated by employing purified genomic DNA of various Trypanosoma and one Leishmania species as starting material for the amplification reaction. Only T. evansi genomic DNA resulted in visible bands at the test line, while the genomic material of other trypanosomatids did not result in any detection (Fig 3B).
Finally, the detection limit of the TevRPA-LF was compared to the sensitivity of amplicon visualization via agarose gel electrophoresis by performing the TevRPA on a 10-fold dilution Development of a RPA-LFA for T. evansi series ranging from 10 ng to 1 fg T. evansi purified genomic DNA per reaction (Fig 3C). When visualized using agarose gel electrophoresis, the lowest amount of genomic DNA that produces an amplicon that can be detected is 100 pg. In contrast, the TevRPA-LF allows amplicon detection at an amount of 100 fg genomic DNA, which is 1000-fold more sensitive compared to agarose gel electrophoresis. The loss of sensitivity during post-amplification visualization via agarose gel electrophoresis is most probably related to the additional required purification step [65]. Hence, for the TevRPA, the extra purification step comes at the cost of sensitivity, which advocates the use of the TevRPA-LF over the TevRPA followed by agarose gel electrophoresis.

The TevRPA-LF can detect active T. evansi infections in an experimental mouse model
Next, the TevRPA-LF was evaluated for its potential to differentiate between ongoing and past infections in an experimental mouse model. In this experiment, C57BL/6 mice infected with T. evansi RoTat1.2 were divided into two groups and the presence of parasites was analyzed by microscopy, the previously described TevPCR [40] and the TevRPA-LF at various time points. Group 1 was left untreated, while Group 2 was treated with Berenil at 5 days post-infection.
As shown in Figs 4 and 5, all three techniques yielded identical results for most of the collected samples. A discrepancy between the detection methods was only observed at 3 days post-infection in Group 1; while parasites could only be detected in 3 out of 5 mice by microscopy, all samples were found to be positive when tested by the TevPCR and TevRPA-LF (Figs 4A and 5A). It is noteworthy to mention that in Group 1 only 4 samples from infected mice were available for testing at day 6 post-infection due to the premature death of one mouse. As expected, all infected mice in Group 1 succumbed to the infection at 7 days post-infection. In contrast, the mice in Group 2 survived day 7 post-infection indicating successful parasite clearance after Berenil treatment at day 5 post-infection. One mouse in Group 2 did not display  Development of a RPA-LFA for T. evansi any signs of infection (4 days post-infection) and was scored as negative by all three methods. Importantly, no amplicons could be detected post-treatment by either the previously validated TevPCR [40][41][42] or the TevRPA-LF described in this work (Figs 4B and 5B). This demonstrates that the TevRPA-LF is a suitable 'test-of-cure' assay. While both the TevPCR and TevR-PA-LF display identical positive and negative score rates under these experimental conditions, the advantage of the TevRPA-LF is that it is effective when performed with crude genomic  Fig 4A). B: TevPCR and TevRPA-LF results for the mouse infection trial of Group 2 mice (corresponds to the data set shown in Fig 4B). In all panels Lane M indicates the molecular mass marker, Lanes 1-6 indicate the individual mice (mouse 6 was used as a negative control within each data set and was not infected), Lane N is a negative control sample (no template DNA) and Lane P is the positive control (T. evansi purified genomic DNA). CL and TL refer to the control and test lines, respectively. https://doi.org/10.1371/journal.pntd.0008044.g005 Development of a RPA-LFA for T. evansi DNA, whereas execution of the TevPCR requires additional purification of the isolated genomic DNA.

Conclusion
T. evansi is the one of the most widespread causative agents of animal trypanosomosis in the world [6]. An essential part of parasite control is the availability of reliable, quick, and userfriendly diagnostic methods. In this paper, we have described the development of a TevR-PA-LF, a test that specifically detects active Type A T. evansi infections by amplifying a region in the T. evansi RoTat1.2 VSG gene. While the T. evansi RoTat1.2 VSG is also targeted by the T. evansi CATT [35] and TevPCR [40][41][42] at the protein and DNA levels, respectively, the TevRPA-LF presents some interesting advantages: i) compared to antibody-based tests (RoTat 1.2 CATT, Surra Sero K-Set, and T. evansi trypanolysis) the TevRPA-LF can be employed to detect active parasitaemia and also serves as a test-of-cure tool since it is not hampered by the presence of infection-induced antibodies that could be the result of past infections or repeated parasite exposure without active infection and ii) the TevRPA-LF combines the RPA format with a dipstick read-out, which outperforms a regular PCR in terms of user-friendliness and field applicability. While it can be argued that LAMP [66] offers the same advantage, the proposed LF format offers an advantage in terms of user friendliness as it visually resembles an antibody-test format that is already in place, while offering the advantage of detecting active infections. Based on the above-mentioned findings, the newly developed TevRPA-LF presented in this paper provides a proof-of-concept with the potential of becoming a valid alternative for currently used screening tools. Its further development will require an additional evaluation of its performance in both experimental and clinical animal infection models.