Rapid isothermal amplification and portable detection system for SARS-CoV-2

Significance An important limitation of current assays for the detection of SARS-CoV-2 stems from their reliance on time-consuming, labor-intensive, and laboratory-based protocols for viral isolation, lysis, and removal of inhibiting materials. While RT-PCR remains the gold standard for performing clinical diagnostics to amplify the RNA sequences, there is an urgent need for alternative testing platforms that are rapid, accurate, simple, and portable. Here, we demonstrate isothermal RT-LAMP nucleic acid-based detection of SARS-CoV-2 with an additively manufactured cartridge and a smartphone-based instrument for testing that can be performed at the point of sample collection.


Using crude human nasal fluid samples
To characterize the robustness of our RT-LAMP assay, we spiked inactive SARS-CoV-2 viruses directly into purchased healthy human nasal fluid samples. Before the reaction, a thermal lysis step (95°C, 1 min) was performed.
To evaluate the effect of volume percentage of spiked nasal fluid per reaction on the detection limits and amplification times, we varied the spiked nasal sample per reaction from 12.5% to 50% of the total reaction volume (supplementary information, Figure S4). Surprisingly, we observed that the RT-LAMP reactions showed robust amplifications even for 50% nasal fluid per reaction. Moreover, we observed that higher sampling volumes of virus-spiked nasal fluid improved the detection limit of the assay from 5E5 copies/L (for 12.5% nasal sample/reaction) to 5E3 copies/L (for 50% nasal sample/reaction) in the final reaction. This is likely because the very viscous nasal fluid solution prevented effective pipette mixing and caused heterogeneous distribution of the viral target. Thus, higher nasal volumes allowed for sampling more viral particles from the inhomogeneous sample yielding better results. For a 12.5% sample, we sampled only 2 L of spiked nasal fluid in a 16 L final reaction, compared to 8 L nasal fluid added in a 50% sample per reaction. As expected, as volume of crude nasal sample increased in the final reaction, there was a delay in amplification times as the increase in the inhibitory components causes a delay in amplification times. Hence, the amplification time increased by 3 min. for the 25% sample and by 8 min. for the 50% sample compared to the amplification times of the 12.5% sample for 5.5E5 copies/L of viral concentration in the nasal sample. Finally, to evaluate whether sampling more of the spiked nasal fluid in a reaction improves the detection limit, we performed 96 L reactions, where the spiked nasal fluid sample volume was 48 L (50% nasal fluid/reaction). As shown in the supplementary information ( Figure S5), increasing the reaction volume did not improved the LOD. Together, these results highlight that our reactions can tolerate up to 50% of crude nasal fluid samples per reaction.

Use of VTM
For diagnostic testing of SARS-CoV-2, the current workflow includes collecting nasopharyngeal/nasal specimens using swabs, which are immediately transferred into sterile transport tube containing 2-3 mL of VTM until diagnostic assays can be performed [34]. In standard RT-PCR assays, this VTM sample with viruses undergoes an RNA purification step next [33]. In our RT-LAMP assay, to evaluate the direct use of VTM samples without any RNA purification, we spiked serial dilutions of inactive SARS-CoV-2 viruses in VTM and performed our RT-LAMP reactions. The VTM sample was either 12.5% or 50% of the final reaction volume, and the detection limit for both reactions was 50 copies/L (supplementary information, Figure  S6). These results highlight no loss in the detection limit of our assay in VTM compared to SARS-CoV-2 viruses in buffer. The reactions with 50% VTM showed a delay of ~20 min. compared to the 12.5%, likely due to the increase in inhibitory components in the VTM.

Supply chain of resources required for standard bench-top RT-LAMP assay compared to standard RT-PCR test to perform testing at scale
We quantified the resources required to scale up the RT-LAMP assay and compare it to the conventional RT-PCR test. For each test, we considered three scenarios in which the number of patient samples are 80, 800, or 8000 ( Figure S11, Table S3, Dataset S1). Different scenarios require different quantities of laboratory resources. Many laboratory instruments are capable of 96 parallel tests, and thermocycler loaded with 96 tests performing the CDC RT-PCR assay would have 80 patient samples plus eight positive controls and eight negative controls [33]. Thus, we select 80 as the smallest increment of patient samples and multiples of 80 for scale up. The model accounts for the time and cost to process 16 control samples for every 80 patient samples. For the 80 patient test scenario, we assumed the use of one thermocycler, for 800 patient tests we assumed 5 thermocyclers, and for 8000 patient tests we assumed 10 thermal cyclers. Other laboratory infrastructure is required as well including refrigerated sample storage, biohazard waste management along with sufficient space to work. The cost modeling includes the costs of the reagents and disposable supplies such as swabs, pipettes, and vials, and the time averaged cost of using laboratory instruments calculated as the instrument cost divided over 10,000 hours of useful lifetime. The model does not account for the cost of space or personnel as this could vary widely. The modeling also assigns an expected time required for each step in the process. All of the costs are tracked either to the CDC test for RT-PCR [33] or the steps described in the present study for RT-LAMP.
The list of resources required shows that RT-LAMP is faster and less costly than RT-PCR, and these advantages can be linked to two key differences between the assays. First, the amplification time for RT-LAMP is about half of that for RT-PCR. RT-PCR requires time to progress through successive thermal cycles while RT-LAMP requires only one isothermal step. Second, RT-LAMP does not require a separate step for RNA extraction, saving time and reducing the cost of consumables. While the elimination of the RNA extraction kit has advantages, a disadvantage is that additional controls are required to account for the presence of the swabs. The total cost could be reduced for both techniques if the use of swabs are eliminated. For instance, if other type of specimens such as saliva could be used for the assay. Likewise, the cost of the RT-PCR technique could be potentially also reduced as recent reports show direct RT-PCR with patient nasopharyngeal [44] and saliva [5] samples without nucleic acid extraction. 5x10 5 5x10 4 5x10 3 5x10 2 5x10 1 0 Thermal lysis step SARS-CoV-2 (copies/mL)