Rapid characterization of feline leukemia virus infective stages by a novel nested recombinase polymerase amplification (RPA) and reverse transcriptase-RPA

Feline leukemia virus (FeLV) is a major viral disease in cats, causing leukemia and lymphoma. The molecular detection of FeLV RNA and the DNA provirus are important for staging of the disease. However, the rapid immunochromatographic assay commonly used for antigen detection can only detect viremia at the progressive stage. In this study, nested recombinase polymerase amplification (nRPA) was developed for exogenous FeLV DNA provirus detection, and reverse transcriptase polymerase amplification (RT-RPA) was developed for the detection of FeLV RNA. The approaches were validated using 108 cats with clinicopathologic abnormalities due to FeLV infection, and from 14 healthy cats in a vaccination plan. The nRPA and RT-RPA assays could rapidly amplify the FeLV template, and produced high sensitivity and specificity. The FeLV detection rate in regression cats by nRPA was increased up to 45.8% compared to the rapid immunochromatographic assay. Hence, FeLV diagnosis using nRPA and RT-RPA are rapid and easily established in low resource settings, benefiting FeLV prognosis, prevention, and control of both horizontal and vertical transmission.


Results
Development of the nRPA. The pGEM-T vector harboring the 145 bp of the U3LTR FeLV fragment was used as a template to optimize the nRPA assay. After completion of the first RPA reaction, the RPA product was diluted tenfold and subjected to a second RPA reaction under various conditions. The expected amplicon band of 101 bp was visible at all tested concentrations of the inner primer (0.24, 0.18, and 0.12 µM), temperature (37 °C, 39 °C, and 4 °C), incubation time (10,20, and 30 min) and volume of primary template (1, 2, and 3 µL) ( Fig. 1a-d). However, nonspecific bands appeared in the nontemplate control when the incubation time was longer than 10 min and when the template was present in high amounts. Consequently, the primary RPA product was serially diluted 100-fold from 10 −2 to 10 −10 , and the expected amplicon band was observed at a 10 −2 template dilution (Fig. 1e) without any nonspecific bands. Therefore, the optimal nRPA conditions was selected to be 37 °C for 10 min using 0.18 µM of inner primer concentration and 100-fold dilution of the primary RPA product.
Development of RT-RPA. FeLV RNA extracted from Nobivac ® Feline 2-FeLV vaccine was utilized for the RT-RPA optimization. The primers were the same as those used in the primary RPA and PCR reactions to amplify the 145 bp of exogenous U3LTR FeLV fragment. Although all primer concentrations tested (0.24, 0.12, and 0.06 µM), incubation times (10, 20, and 30 min), and temperatures (40 °C, 41 °C, and 42 °C) showed positive results (Fig. 2a-c), an incubation time of longer than 10 min caused the appearance of nonspecific bands in the no-template control. Hence, the appropriate conditions were selected as 40 °C, 10 min, using 0.12 µM primers, as this resulted in the most obvious positive amplicon band with no primer-dimer in the negative control.

Validation of nRPA in clinical samples.
From the 122 clinical samples tested, the FeLV provirus was detected by RPA and/or nRPA in all of the 90 samples found to be positive by PCR and/or nPCR. The positive RPA or PCR samples revealed a 145 bp amplicon, with the negative results being subsequently retested by nRPA or nPCR to yield the expected 101 bp amplicon in positive samples. There was an almost invisible positive band in seven negative nRPA samples with negative nPCR results and in four negative nPCR with negative nRPA results. These samples were retested using both techniques, and the negative results were confirmed. The results are summarized in Table 1. Detection of FeLV provirus by PCR revealed 73 positive and 49 negative results, and 17 of these 49 negative PCR were then found to be positive by nPCR, giving a total of 90/122 positive samples. Detection of FeLV provirus by RPA revealed 70 positive and 52 negative results, with 20 out of these 52 negative RPA samples being found to be positive by nRPA, giving a total of 90/122 positive samples. Thus, 90 samples were found to be positive by both the RPA/nRPA and the PCR/nPCR approaches, with nRPA and nPCR having a higher sensitivity than RPA and PCR.   www.nature.com/scientificreports/ Comparisons of the results obtained from RPA with those from PCR, and of the results obtained from RPA plus nRPA with those from PCR plus nPCR for FeLV provirus detection are summarized in Tables 2 and 3, respectively. The RPA assay resulted in three false negative results compared to the PCR assay, which was taken to be the gold standard. The diagnostic performance of RPA compared to PCR produced 95.89% (95% confidence intervals [CI] of 88.46-99.14) sensitivity, 100% (95% CI of 92.89-100) specificity, and a kappa value of 0.949 (CI of 0.893-1.000). The nRPA demonstrated one false negative and one false positive result compared to nPCR. The diagnostic performance of RPA plus nRPA compared to PCR plus nPCR was 98.89% (95% CI of 93.96-99.97) sensitivity, 96.88% (95% CI of 83.78-99.92) specificity, and a kappa value of 0.958 (CI of 0.899-1.000). The accuracies of the RPA vs. PCR and RPA plus nRPA vs. PCR plus nPCR were 97.54% and 98.36%, respectively. The results supported the contention that the novel nRPA assays increased the sensitivity of FeLV provirus detection.
Validation of RT-RPA in clinical samples. The detection of FeLV RNA by RT-RPA was compared with its detection by rapid immunochromatographic assay or RT-PCR, and the results are shown in Table 4. The RT-RPA of 53 clinical samples gave the expected 145 bp product, while seven samples gave two amplicon bands, comprised of the expected 145 bp and an unexpected 200 bp product (Fig. 3). Sequence analysis of the amplicons from all seven samples revealed that the larger amplicon was due to a triplet repeat of a 21 bp sequence insertion between nucleotide positions 84-146 of the FeLV U3LTR (GenBank accession number AY374189.1), accounting for the extra size of the larger amplicon. Of 64 (52.46%) samples found to be FeLV positive by RT-PCR, there was one false negative results compared with the rapid immunochromatographic assay, and four compared with RT-RPA. The diagnostic performance of RT-RPA compared to the rapid immunochromatographic methods ( Categorization of FeLV infective stages. The molecular diagnosis produced by all four methods was analyzed to determine the FeLV infective stage in symptomatic and healthy cats ( Table 5). The progressive stage,

Discussion
In order for veterinarians to effectively manage the treatment of FeLV, and advice cat owners how best to take care of their pets, characterization of the different FeLV infective stages is useful. In this study, we developed novel nRPA and RT-RPA approaches to detect FeLV DNA provirus and RNA in circulating blood. The use of nRPA for FeLV provirus amplification increased the detection rate by about 45.8% (27/59) over that of p27 antigen detection. This increase would improve clinical diagnosis, because antigen detection has limited sensitivity for the detection of viremia in the progressive stage, resulting in a loss of latent detection. The detection of FeLV provirus using nucleic acid amplification based methods has been recommended as the gold standard due to its higher sensitivity and specificity 12 . The presence of endogenous FeLV was related to disease progression, viral recurrence, and vertical transmission in cats' progeny.  www.nature.com/scientificreports/ FeLV-A is the most common subgroup in cats 24 . The nRPA and RT-RPA primers were designed to target U3LTR, which is the conserved fragment in all subgroups of FeLV 25 . nRPA targeting the U3LTR showed very high sensitivity and specificity, with almost perfect agreement with the results of nPCR and rapid immunochromatography detection methods. The concordance between nPCR targeting the pol gene and ELISA has previously been found to be substantial 26 , and nPCR has a 100-fold greater sensitivity than conventional PCR 27 . The sensitivity and specificity of RPA amplification could be enhanced, as in nPCR based methods. However, carry-over contamination must be prevented by physically separating the area of the primary PCR reaction from that of the secondary reaction. The primary RPA products were diluted to optimum concentration to produce high sensitivity and specificity. The dilution of the initial amplification products also reduced the carry-over problem. One negative control/16 nRPA reactions/nested nRPA run was checked for carry-over contamination. The positive control was used at the minimum concentration, and added as the last tube to each nested nRPA run to prevent contamination 28 .
The RT-RPA technique amplified FeLV RNA within 10 min. Longer incubation times yielded nonspecific bands only in the negative control. Bands with a smeared appearance were suggested to be primer dimers, which did not interfere with the interpretation of FeLV detection. Although unexpected 200 bp products were identified in seven samples, their sequences were similar to those of the expected amplicon, except for the presence of three 21 bp repeat sequences similar to those in FeLV subgroup A and the FeLV v-myc gene. The repeat elements naturally occur between the enhancer and promoter in the U3LTR during FeLV replication through error-prone reverse transcription, and by endogenous provirus recombination, including c-myc transduction. The LTR variants was associated to particular disease outcomes, especially thymic lymphomas, multicentric lymphomas, and other non-T cell diseases 29,30 .
The perfect kappa value indicated a significant correlation between the RT-RPA results with either the rapid immunochromatography or the RT-PCR assays. Similarly to insulated isothermal (ii)PCR and RT-iiPCR for point-of-need FeLV diagnosis, few discrepancies were revealed, especially in FeLV RNA detection 31 . Because RNA is fragile, some false negative results may result from poor quality or degenerated RNA, or a low viral load in the sample. The sensitivity and specificity of the three point of care FeLV p27 antigen test kits have previously been evaluated using qrtPCR as the gold standard, producing 63% sensitivity and 94% specificity for the SNAP Combo, 57% sensitivity and 98% specificity for Witness, and 57% sensitivity and 98% specificity for AntigenRapid 15 . The serological methods were unable to detect regressive disease in cats, and some kits produced false positive results due to interference by the antimouse IgG antibodies in cat sera 15 .
One of the advantages of the detection of FeLV provirus and RNA using nRPA and RT-RPA is the ability to achieve rapid amplification using general heating tools, with no need for thermal cyclers. However, the major drawback, the need for post amplification detection using gel electrophoresis, could be improved in the future by the use of SYBR green I or lateral flow strip (LFA) detection. A simple post RPA detection step in FeLV diagnosis is needed for FeLV diagnosis at a molecular level in veterinary clinics. Multiplexed RPA-LFA has been shown to be able to detect ESBL genes in E. coli isolated from pork meat and pig feces 32 . Recently, multidrug-resistant tuberculosis was rapidly detected by the naked eye, using allele-specific RPA combined with SYBR green I 33 . For RT-RPA, the TwistDx Basic RT kit is now not available for sale. However, a Basic TwistDX kit for DNA amplification can be used by adding an initial step involving a reverse transcriptase reaction such as RT-PCR to convert RNA to cDNA. The modified multiplexed RT-RPA by adding Moley murine leukemia virus reverse transcriptase and RNase inhibitor to the RPA pellet (Basic TwistDx kit, UK) successfully amplified and differentiated avian influenza viruses (AIVS) subtypes. Moreover, the combination of RT-RPA with capillary electrophoresis were simultaneously detected four AIVs genes with high sensitivity and high-throughput 34 . Recently, a one step RT-RPA by only adding ProtoScript II Reverse Transcriptase to the RPA pellet (Basic TwistDx kit, UK) could rapidly and sensitively detected SAR-CoV-2 35 . A combination of FeLV provirus and RNA testing using RPA assays is valuable for the diagnosis of the infective stages of FeLV before vaccination or blood transfusion. FeLV testing should be performed prior to blood donation to reduce the unintentional horizontal FeLV transmission. Nesina et.al. 10 have demonstrated that naïve cats that received a blood transfusion from FeLV carrier cats developed FeLV-associated diseases, including nonregenerative anemia, T cell lymphoma, and secondary lymphoblastic leukemia.
We found that about 64% of asymptomatic cats visiting for vaccination become infected with FeLV. The existence of FeLV was revealed in 47.4% of healthy cats, using nested PCR 36 . In healthy cats not exposed to FeLV, 2% had regressive infections and 0.5% had progressive infections, while in a mixture of healthy and sick cats not exposed to FeLV 25% had regressive infections and 21% had progressive infections 37 . The administration of FeLV vaccine is ineffective in infected cats, and so makes cat owners become relatively incautious about preventing FeLV transmission. The rapid identification of the infective stage of FeLV would improve FeLV infection control by the timely isolation of an infected cats, maintaining separate food bowls and housing, and by promoting FeLV vaccination in high-risk cats, particularly in a multicat household.
Progressive FeLV infection has a poor prognosis, with a 80% fatality rate over 3 years 14 . Overall, FeLV infection was found in 87.5% of cats with lymphoma and leukemia, with 87.76% in progressive stages. The risk of lymphomas and leukemia was increased by up to 80% in cats infected with FeLV 18 . In a previous study, 15 out of 16 cats with myelodysplastic syndromes (MDS) were infected with FeLV. Nonregenerative anemia and thrombocytopenia were the most common hematologic abnormalities found in myelosuppression 13,38  www.nature.com/scientificreports/ In conclusion, the RPA, nRPA, and RT-RPA methods developed in this study were validated against the gold standard of PCR, nPCR, and RT-PCR, producing perfect agreement. Both RT-RPA and RPA had similar sensitivity and specificity to PCR-based methods but with the additional advantages of rapid amplification and no need for an expensive thermal cycler. The characterization of the infective stages of FeLV using RPA-based assays could be set up onsite in veterinary hospitals to enhance the efficiency of FeLV diagnosis, which is important for prognosis and disease control. Rapid FeLV detection is important for the ultimate goal of disease prevention.

Methods
Ethical statement. The samples were obtained from The Small Animal Teaching Hospital, Faculty of Veterinary Science, Chulalongkorn University, Bangkok, Thailand, between January 2016 and December 2018. The study protocol was approved by Chulalongkorn University Animal Care and Use Committee (CU-ACUC 1831010). Written informed consent was obtained from each cat owner prior to the study. Blood was collected in EDTA tubes from 108 cats with clinicopathological abnormalities caused by FeLV infection, such as lymphoma, leukemia, hematologic disorders, FIV, and other infections, and from 14 healthy cats in a vaccination plan. All whole blood samples were tested for the FeLV p27 antigen using the Antigen Rapid FIV Ab/FeLV Ag test kit (BIONOTE, Korea) and kept at − 80 °C for further nucleic acid extraction.
All methods were carried out in accordance with relevant guidelines and regulations. The study was compliant with all relevant ethical regulations regarding animal research.
Nucleic acid extraction. Two hundred microliters of whole blood was used for RNA and DNA extraction using QIAamp cador Pathogen (QIAGEN, Hilden, Germany) according to the manufacturer's recommendation. Briefly, 20 µL of proteinase K and 100 µL of VXL buffer were combined and incubated at room temperature for 15 min. After adding 350 µL of ACB buffer, the solution was transferred to a QIAamp Mini column and centrifuged at 5000×g for 1 min. The column was washed twice with AW1 buffer and transferred to a new tube. The mixture of DNA and RNA was eluted by adding 70 µL AVE buffer, and centrifuged at 15,000×g for 1 min. The elution was aliquoted for DNA analysis and 30 µL was further purified for RNA using TURBO DNA-free kits (Ambion, Austin, TX, USA). All DNA and RNA samples were kept at − 80 °C for further analysis. Detection of FeLV DNA provirus by nRPA. The nRPA was performed using a Basic TwistDX kit (TwistDX, Maidenhead, Berkshire, UK). The first and second sets of nested RPA primers were the same as those used in the nPCR above, and yielded the same sized products of 145 bp and 101 bp, respectively. The amounts of primary RPA products, primer concentrations, incubation temperature, and time were first optimized in preliminary trials. The first RPA reaction consisted of 0.24 µM of the U3F1 and U3R1 primers, 29.5 µL rehydration buffer, and 20 ng/µL template DNA in a total volume of 50 µL. The reaction mixture was added in lyophilized pellets and 2.5 µL of 280 mM magnesium acetate was added, mixed by inversion 8-10 times, and then incubated at 37 °C for 20 min. The 2 µL of primary RPA products from negative samples were diluted 100-fold and subjected to the second RPA reaction using the same nPCR primer set. The RPA conditions for the second round were unchanged, except that 0.18 µM of each primer was used, and incubation was carried out for 10 min only. The RPA products were purified using NucleoSpin ® Gel and PCR Clean-up kits (MACHEREY-NAGEL, Düren, www.nature.com/scientificreports/ Germany) as per the manufacturer's recommendation, before analysis using 1.5-2% (w/v) agarose gel electrophoresis with visualization by a ChemiDoc XRS gel photo documentation system (Bio-Rad, Hercules, CA, USA).
Detection of FeLV RNA using RT-RPA. TwistDx Basic RT kits were used with 0.12 µM of U3F1 and U3R1 primers. The primer concentrations, temperature, and incubation time were first optimized in preliminary trials. The reaction mixture was prepared in a final volume of 50 µL composed of rehydration buffer, 10 ng/µL RNA template, and 100 unit of RNAse inhibitor. Lyophilized pellets were solubilized in the reaction mixture, and 2.5 µL of 280 mM magnesium acetate was added and mixed by inversion 8-10 times. The optimum conditions were 40 °C for 10 min. The RT-RPA products were purified using NucleoSpin ® Gel and PCR Clean-up (MACHEREY-NAGEL) and analyzed for the presence of the 145 bp product using 1.5% (w/v) agarose gel electrophoresis.
Clinical sample testing. All 122 clinical samples from symptomatic and vaccinated cats were tested using nRPA and RT-RPA, respectively. The TwistDx basic kit and TwistDx Basic RT kit were employed, using the conditions described above.

Data analysis.
The results of the nRPA were compared with those from the PCR and nPCR assays, while the results from the RT-RPA were compared to those from the RT-PCR and immunochromatography assays. Inconsistent results between RPA, nRPA, PCR, nPCR and RT-RPA, RT-PCR, and immunochromatography assay were repeated using all methods to confirm the results. The sensitivity and specificity were determined using the MEDCALC ® software (https:// www. medca lc. org/ calc/ diagn ostic_ test. php). To assess the concordance level, the kappa value was calculated with 95% CIs using QuickCals software (https:// www. graph pad. com/ quick calcs/ kappa1) 45 .

Data availability
All data generated or analysed during this study are included in this published article. The full-length gel in Fig. 1a was presented in supplemental file Fig. 1. www.nature.com/scientificreports/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.