Discussion
In this study, we developed a rapid and sensitive diagnostic assay for SARS-CoV-2 infections, which is based on RT-LAMP. The assay could detect 23.7 copies of viral RNA within 15 min from clinical specimens (Fig 1A, Table 6, and S1 Table). Currently, RT-qPCR is the gold standard assay of COVID-19 diagnosis and takes at least 2–3 hours to complete all processes including RNA extraction followed by a quantitative PCR step using a real-time thermal cycler. To shorten the assay time, several research groups developed RT-LAMP assays against SARS-CoV-2 [16–21]; however, the detection sensitivity and assay time can be improved. Huang et al. reported a novel RT-LAMP assay that could detect 2 copies of viral RNA after an additional gel electrophoresis step, requiring more assay time than the normal RT-LAMP method [21]. In addition to RT-LAMP, we also developed a remarkable rapid assay, Direct RT-LAMP, that does not require the extraction of viral RNA from clinical specimens. This assay could detect 203 copies of viral RNA in the virus-spiked swab solution within 10.5 min (Table 5). The LAMP method has an excellent specificity accomplished by 6 to 8 different primer binding regions in the target DNA sequence. Nevertheless, several viruses require multiple sets of primers to detect whole lineages/genotypes due to genetic diversity of their genome sequences. Lassa virus (LASV) has 6 lineages and has highly diverse genome even within the same lineage. There is a single report of a RT-LAMP assay for LASV, and the assay requires 3 sets of primers to detect only lineage II [22]. Zika virus also shows diversity in genome sequences between Asian and African genotypes, requiring the mixture of 2 primer sets to detect all strains in one reaction [13]. Fortunately, SARS-CoV-2 shows lower genetic diversity worldwide, and therefore would be suitable to be diagnosed using the simple RT-LAMP assay with one primer set (S1 Fig).
Before the emergence of SARS-CoV-2, two large outbreaks of novel coronavirus diseases were reported: severe acute respiratory syndrome (SARS) in 2002, and Middle East respiratory syndrome (MERS) in 2012 [23,24]. For the molecular diagnosis for these diseases, RT-PCR and RT-qPCR were used as gold standard assays with an immediate announcement of recommended protocols by WHO. Several years since the emergence of these coronavirus diseases, RT-LAMP assays were developed for faster diagnosis [25,26]. For SARS, Thai et al. developed a RT-LAMP assay which is 100-fold more sensitive than RT-PCR [25] and showed clear linearity between viral titer and detection time, and that could detect SARS-CoV even in RT-PCR negative samples. Similar to SARS, several RT-LAMP assays were developed for the detection of MERS-CoV. Shirato et al. developed a RT-LAMP assay for MERS with a sensitivity comparable to standard RT-qPCR, and with sufficient specificity to distinguish MERS-CoV from 20 other respiratory viruses [26]. These past reports indicate advantages in using RT-LAMP assays to detect coronavirus from clinical specimens with sufficient sensitivity.
A recent clinical study that evaluated respiratory tract specimens of COVID-19 patients with severe and mild symptoms showed the initial viral load of 6.17 and 5.11 log10 copies per mL, respectively [27]. Posterior oropharyngeal saliva, which might represent a non-invasive reasonable specimen acceptable by patients, contained 4 to 8 log10 copies of viral genome per mL in the first 5 days from the symptom onset [27]. Similarly, experimental SARS-CoV-2 infection of rhesus monkeys showed that the initial viral load in nasal or throat swabs also contained 4 to 8 log10 copies of viral genome per mL within the first 5 days [28]. Again, our RT-LAMP assay could detect 23.7 copies of SARS-CoV-2 viral RNA, approximately equivalent to 1,778 copies per mL in clinical specimens, effectively detecting SARS-CoV-2 in samples from patients with both severe and mild symptoms during the acute phase. Whereas, our Direct RT-LAMP is approximately 60-fold less sensitive than RT-LAMP in the assay for clinical specimens, indicating the reliable detection limit of approximately 5 log10 copies per mL (Table 6). Thus, although this Direct RT-LAMP assay detects SARS-CoV-2 in high viral titer samples, it should not be used as a sole molecular diagnostic method but used with another assay such as qPCR to avoid false negatives. Nevertheless, Direct RT-LAMP would significantly contribute to the rapid first screening of COVID-19 during an outbreak (1) in the high-risk population (e.g. pregnant women, individuals with a pre-existing illness) who needs an immediate care, and (2) in the resource-limited settings where a centrifuge is unavailable. We also found that HBSS affected the detection time of Direct RT-LAMP (Fig 2), indicating that HBSS contains inhibitors of RT-LAMP amplification. Therefore, in the Direct RT-LAMP assay, it is preferable to suspend swab samples in PBS.
Since the RT-LAMP assay can be performed with a portable battery-driven device, it may be available at point-of-care, even in poorly-equipped settings. Taken together, our results suggest that this new RT-LAMP protocol can be useful as a rapid and sensitive diagnostic assay for COVID-19, and that the Direct RT-LAMP version could also be made available as a much simpler, faster, and low-cost assay. These new assays may contribute to public health control in countries undergoing COVID-19 epidemics, not only for early disease detection, but also to monitor future expansion of viral infections.