Development of a multiplex isothermal amplification molecular diagnosis method for on-site diagnosis of influenza

Influenza, which is an acute respiratory disease caused by the influenza virus, represents a worldwide public health and economic problem owing to the significant morbidity and mortality caused by its seasonal epidemics and pandemics. Sensitive and convenient methodologies for the detection of influenza viruses are important for clinical care and infection control as well as epidemiological investigations. Here, we developed a multiplex reverse transcription loop-mediated isothermal amplification (RT-LAMP) with quencher/fluorescence oligonucleotides connected by a 5′ backward loop (LF or LB) primer for the detection of two subtypes of influenza viruses: Influenza A (A/H1 and A/H3) and influenza B. The detection limits of the multiplex RT-LAMP assay were 103 copies and 102 copies of RNA for influenza A and influenza B, respectively. The sensitivities of the multiplex influenza A/B/IC RT-LAMP assay were 94.62% and 97.50% for influenza A and influenza B clinical samples, respectively. The specificities of the multiplex influenza A/B/IC RT-LAMP assay were 100% for influenza A, influenza B, and healthy clinical samples. In addition, the multiplex influenza A/B/IC RT-LAMP assay had no cross-reactivity with other respiratory viruses.


Introduction
Influenza, which is caused by the influenza virus, is a significant cause of morbidity and mortality that has major social and economic impacts throughout the world [1,2]. Iuliano et al. reported that 291,243-645,832 seasonal influenza-associated respiratory deaths occur annually (4�0-8�8 per 100,000 individuals) [3]. Influenza viruses belong to the Orthomyxoviridae family and have a single-stranded segmented RNA genome consisting of 7-8 segments encoding 10-11 proteins [4]. The influenza viruses are classified into types A, B, C, and D on the basis of their core proteins [5]. Among the four influenza types, influenza A viruses cause most of the global flu epidemics, influenza B viruses cause smaller localized outbreaks, influenza C viruses

Primer design
Influenza A and B LAMP primer sets were designed within conserved regions of segment 7 of influenza A and the nucleoprotein gene of influenza B. For internal control, actin beta LAMP primer set was newly designed within of conserved regions of human actin beta mRNA (NM_001101.5:c.287-c.498), which is commonly used as internal control [32]. All LAMP primers, including two outer primers (forward primer F3 and backward primer B3), two inner primers (forward inner primer FIP and backward inner primer BIP), and two loop primers (forward loop primer LF and backward loop primer LB), were designed using Primer Explorer software (Version 4; Eiken Chemical Co., Tokyo, Japan). For the multiplex LAMP assay, we designed the fluorophore probe oligomer (32 mer) at the 5 0 LF or LB primer and the quencher oligonucleotide (30 mer), which is the complementary sequence of the fluorophore probe oligomer, using Random DNA Sequence Generator (https://faculty.ucr.edu/~mmaduro/random. htm). All primers were assessed for specificity before use in the LAMP assays via a BLAST search of sequences in GenBank (National Center for Biotechnology Information [NCBI], Bethesda, MD). All LAMP primers and probes were synthesized by Macrogen, Inc. (Seoul, South Korea; Table 1).

Multiplex influenza A/B/IC RT-LAMP assay
The influenza A/B/IC multiplex RT-LAMP assay was performed with a Mmiso RNA amplification kit (Mmonitor, South Korea). The RT-LAMP reaction was prepared with 12.5 μL of 2x reaction buffer, 1.25 μL of influenza A LAMP primer mix, 0.625 μL of influenza B LAMP primer mix, 0.625 μL of internal control LAMP primer mix, 720 nM quencher solution, 2 μL of enzyme mix, and 2.5 μL of sample RNA (final reaction volume: 25 μL). The composition of the LAMP primer mix (influenza A and influenza B) included 4 μM of two outer primers (F3 and B3), 32 μM of two inner primers (FIP and BIP), 10 μM of loopB primer, 4 μM of loopF primer, and 6 μM of loopF probe primer. The composition of the internal control LAMP primer mix included 4 μM of two outer primers (F3 and B3), 32 μM of two inner primers (FIP and BIP), 10 μM of loopF primer, 4 μM of loopB primer, and 6 μM of loopB probe primer. The RT-LAMP assay was run on the CFX 96 Touch Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA, USA) at 60˚C for 60 min. In LAMP assay, negative control (human serum RNA and distilled water) were used for setting baseline. Positive signal was determined by checking whether signal is steep or gradual considering the baseline, because baseline of LAMP assay is not stable compared to qPCR or RT-PCR. The FAM, HEX and Texas red fluorescence channels was used for detecting Influenza A, influenza B and internal control (actin beta), respectively.

Limit of Detection (LOD) tests of the multiplex influenza A/B/IC RT-LAMP assay
pTOP Blunt V2 plasmids, including the segment 7 partial sequences of Influenza A, were used to test the LOD of the influenza A LAMP assay. For the LOD of the influenza B LAMP assay, pTOP Blunt V2 plasmids, including the nucleoprotein gene partial sequences of influenza B, were used. pTOP Blunt V2 plasmids, including the beta actin partial sequences of human, were used to test the LOD of the Internal control LAMP assay. All plasmids were constructed by Macrogen, Inc. The plasmids were serially diluted 10-fold from 1.0 × 10 8 copies/μL to 1.0 × 10 1 copies/μL to determine the LOD of the multiplex influenza A/B/IC RT-LAMP assay.
In addition, the LOD of the multiplex influenza A/B/IC LAMP analysis was tested using clinical samples of influenza A/H1, A/H3, and B. Clinical specimens were diluted 10 times (as much as 10 −1 to 10 −6 times) based on the original samples. The LOD of the multiplex influenza A/B/IC RT-LAMP analysis for clinical influenza samples was compared to those of the WHO influenza A/B primer set and RealStar1 Influenza RT-PCR Kit 2.0.

Optimization of the multiplex influenza A/B/IC LAMP primer set
Before optimization of multiplex LAMP assay, each Influenza A, B and internal control (IC, actin beta) LAMP primer set were tested with clinical samples (Influenza A H1, H1N1, H3N2, Influenza B, human serum RNA and distilled water). All three LAMP primer sets (Influenza A, B and IC) showed the no cross-reactivity (S1 Fig).
For optimization of the multiplex influenza A/B/IC LAMP primer set, different concentration ratios of the influenza A, influenza B, and internal control primer sets (1:1:1, 1:1:0.5, and 1:0.5:0.5, respectively) were tested using human serum RNA samples spiked with influenza A and B plasmids (10 7 copy, ratio of 1:1) ( Fig 1A). As a result, three signals (influenza A, influenza B, and internal control) were detected in the ratios of 1:0.5:0.5, whereas two signals (influenza B and internal control) were detected in the ratio of 1:1:1 and 1:1:0.5. Among three ratio of LAMP primer set, the ratio of 1:0.5:0.5 showed the lowest Ct values of influenza A and influenza B were 12.58 and 10.41, respectively. Therefore, the ratios of influenza A, influenza B, and internal control LAMP primer set (1:0.5:0.5) was determined as optimum ratio for multiplex influenza A/B/IC LAMP assay. Next, temperature gradient tests (58-65˚C) were performed to determine the optimal temperature of the multiplex influenza A/B/IC LAMP assay ( Fig 1B).

LOD tests of the multiplex influenza A/B/IC multiplex LAMP assay
The analytical sensitivity of the multiplex influenza A/B/IC RT-LAMP assay was compared with monoplex LAMP primer sets [In A, In B and Internal control (IC, actin beta)] by testing synthetic RNA plasmids ranging from 10 8 to 10 1 RNA copies/μL (Fig 2, Table 2). In monoplex In A, In B and IC LAMP test showed all the detection limits of 1 × 10 2 copies/μL. In multiplex influenza A/B/IC RT-LAMP assay, both influenza A and B plasmids were detected up to 1 × 10 2 copies/μL and actin beta plasmid was detected up to 1 × 10 3 copies/μL. As a result, multiplex influenza A/B/IC showed comparable detection limits with monoplex LAMP assay, although detection limit of IC in multiplex was higher than that of IC monoplex LAMP. Furthermore, the LOD of the RT-LAMP assay was compared with the WHO RT-PCR primer set and commercial RealStar1 Influenza RT-PCR Kit 2.0 using serially diluted clinical influenza A/H1, A/H3, and B samples (range of 10 −1 to 10 −6 ; Table 3). As a result, the LOD of both the WHO RT-PCR primer set and RealStar1 Influenza RT-PCR Kit 2.0 for influenza A/H1 was 10 −3 while that of the multiplex influenza A/B/IC RT-LAMP assay was 10 −2 . For influenza A/ H3, the RealStar1 Influenza RT-PCR Kit 2.0 assay showed the highest sensitivity (10 −5 ), and the LOD of the other two assays was 10 −4 . For influenza B, the RealStar1 Influenza RT-PCR Kit 2.0 assay showed the highest sensitivity (10 −4 ), and the LOD of the other two assays was 10 −3 . Overall, the detection limits tested with clinical sample dilutions were the lowest in the 95.12%. Overall, the WHO RT-PCR primer set was found to have the highest sensitivity (97.84%) for all influenza A clinical samples (n = 93), followed by the multiplex influenza A/B/ IC RT-LAMP assay (94.62%), and finally the RealStar1 Influenza RT-PCR Kit 2.0 (88.17%). The specificities of the former two assays for influenza A clinical samples were 100% while that Table 2 of the latter was 95.69%. In addition, the multiplex influenza A/B/IC RT-LAMP assay showed 100% of sensitivities and specificities for H5 (n = 10), H7 (n = 10) and H9 (n = 10) subtypes of avian influenza viruses (S1 Table). For influenza B clinical samples (n = 80), the multiplex influenza A/B/IC RT-LAMP assay and WHO RT-PCR primer set showed the highest sensitivities (97.50% and 97.50%, respectively) while that of the RealStar1 Influenza RT-PCR Kit 2.0 was 86.25% (Table 4). The internal control channel of the multiplex influenza A/B/IC RT-LAMP assay showed 97.50% sensitivity for influenza B clinical samples. The specificities of all three assays against influenza B clinical samples were 100%. The specificities of the multiplex influenza A/B/IC RT-LAMP assay and WHO RT-PCR primer set for healthy clinical samples (non-infection; n = 100) were 100%, whereas that of the RealStar1 Influenza RT-PCR Kit 2.0 was 99% ( Table 4). The sensitivity of the internal control channel of the multiplex influenza A/B/IC RT-LAMP assay for healthy clinical samples was 100%.

Cross-reactivity test
To  (Table 5) [33,34]. All three molecular diagnostic tests showed no cross-reactivity with other infectious viruses, suggesting that these tests can specifically detect influenza viruses.

Discussion
Since many respiratory viruses, including influenza viruses, can cause similar symptoms, it is difficult for clinicians to distinguish one virus from another [35]. Given the annual morbidity Table 4

Clinical samples RT-LAMP assay WHO Influenza RT-PCR RealStar1 Influenza RT-PCR (Altona) In A (FAM) IC (Hex) In B (Tex) In A (Hex) In B (FAM) In A (FAM) In B (Cy5)
Inf A/H1N1 (n = 11) P/N 10/1 9/2 0/11 11/0 0/11 9/2 0/11 The sensitivities and specificities were calculated by taking the Anyplex™ II RV16 Detection as is standard. P/N: positive/negative ratio. and mortality caused by influenza viruses, there is an urgent need for sensitive and convenient laboratory methods to identify influenza virus subtypes in clinical care and infection control [36,37]. There are immunodiagnostic kits that can be utilized quickly, but their sensitivity and specificity are too poor; thus, PCR methods are currently used to make accurate diagnoses [38]. However, PCR-based target gene detection requires bulky and expensive equipment as well as highly skilled technicians [39]. Several isothermal amplification methods, such as HDA, RPA, and LAMP, have been developed for on-site diagnosis of infectious pathogens [40][41][42]. Among them, LAMP is a promising method that has been utilized to detect a variety of pathogens [43,44]. Recently, a variety of multiplex RT-LAMP methods have been developed to detect influenza viruses by using annealing temperature, nanoparticle hybridization, onepot colorimetric visualization and immunochromatographic strip etc [45][46][47][48]. However, their multiplex system amplified each type of influenza individually or consist of two steps, which are lamp assay and rapid test. Thus, it may take a more time when diagnosis a large number of clinical samples. Thus, it can take a more time to diagnose a large number of clinical samples. In addition, it is known that LAMP assay is easy to contaminate [49,50]. In this study, we developed the one tube-multiplex influenza A/B/IC LAMP assay, including an internal control (actin), for the detection of influenza A/H1, A/H3, and B using newly designed assimilating probes, which has advantages for reducing test time and risk of contamination. Our results showed that the multiplex influenza A/B/IC RT-LAMP assay had 100% analytical specificity for the identification of influenza A/H1, A/H3, and B viruses, and there was no cross-reaction with other genetically or clinically related control viruses tested in this study. The multiplex influenza A/B/IC RT-LAMP assay for influenza A clinical samples (n = 93) had a sensitivity and specificity of 94.62% and 100%, respectively. However, the multiplex influenza A/B/IC RT-LAMP assay showed a 90% sensitivity for influenza A/H1 because out of the 11 specimens that were positive for influenza A/H1, only 10 were determined by the assay to be positive. Further testing with additional clinical specimens is needed to evaluate the clinical performance characteristics of the multiplex influenza A/B/IC RT-LAMP assay to address the issue of the small sample size used (n = 10). In addition, the internal control signal showed in a very low sensitivity for influenza A clinical samples but not influenza B samples and negative Table 5

Virus
No clinical samples. This result might be that LAMP amplification reagents were consumed for the amplification of influenza A or influenza A amplification products interrupt the amplification of the internal control. Interestingly, the multiplex influenza A/B/IC RT-LAMP assay had a lower detection sensitivity for diluted clinical samples than the RealStar1 Influenza RT-PCR Kit 2.0 but higher sensitivity for original clinical samples. These results suggest that the multiplex influenza A/B/IC RT-LAMP assay had a higher sensitivity for various influenza genetic sequences while the RealStar1 Influenza RT-PCR Kit 2.0 had a higher sensitivity for specific influenza genetic sequences. RT-LAMP analysis is one of the most promising diagnostic tools for use in the field since it does not require sophisticated and expensive equipment or skilled personnel [51]. The multiplex method developed in this study can diagnose influenza A and B within 60 min using multiple fluorescence. In order to utilize the multiplex influenza A/B/IC RT-LAMP assay in the field, it is necessary to use an isothermal amplification device that detects portable multiple fluorescence. Most of the field isothermal amplifiers currently available have been developed as single channel. However, two-channel isothermal amplifiers (Genie III; OptiGene, West Sussex, UK) have recently been developed and marketed by Chayon Laboratories, Inc. Since there are no field isothermal amplifiers with three channels yet, the influenza A/B and internal control LAMP kits cannot be used. However, by excluding the internal control, influenza A and B can now be diagnosed in the field using commercially available isothermal amplifiers. In addition, conventional RNA extraction methods, which extract RNA using centrifuges from NP swabs or aspirate samples collected from suspected influenza patients, are time-consuming and may be potentially contaminated. Therefore, it is expected that influenza field diagnosis can be performed more effectively by using an RNA extraction chip [52] or the magnetic bead-nucleic acid extraction method [53].

RT-LAMP qRT-PCR (WHO) qRT-PCR (RealStar1) In A (FAM) IC (Hex) In B (Tex) In A (Hex) In B (FAM) In A (FAM)
In this study, we developed a multiplex real-time RT-LAMP assay that can diagnose influenza A and influenza B with one step. The multiplex influenza A/B/IC RT-LAMP assay using the probe-quencher that compensates for the disadvantages of LAMP, such as false positive diagnoses, shows similar sensitivity and specificity to the WHO RT-PCR primer set. Since LAMP takes less time (within 60 min) than conventional RT-PCR, the multiplex influenza A/ B RT-LAMP assay can be used as an efficient method in on-site molecular diagnostic kits.