Sample-to-Answer Detection of SARS-CoV-2 Viremia Using Thermally Responsive Alkane Partitions

Abstract

The diagnosis of bloodborne viral infections (viremia) is currently relegated to central laboratories because of the complex procedures required to detect viruses in blood samples. The development of point-of-care diagnostics for viremia would enable patients to receive a diagnosis and begin treatment immediately instead of waiting days for results. Point-of-care systems for viremia have been limited by the challenges of integrating multiple precise steps into a fully automated (i.e., sample-to-answer), compact, low-cost system. We recently reported the development of thermally responsive alkane partitions (TRAPs), which enable the complete automation of diagnostic assays with complex samples. Here we report the use of TRAPs for the sample-to-answer detection of viruses in blood using a low-cost portable device and easily manufacturable cassettes. Specifically, we demonstrate the detection of SARS-CoV-2 in spiked blood samples, and we show that our system detects viremia in COVID-19 patient samples with good agreement to conventional RT-qPCR. We anticipate that our sample-to-answer system can be used to rapidly diagnose SARS-CoV-2 viremia at the point of care, leading to better health outcomes for patients with severe COVID-19 disease, and that our system can be applied to the diagnosis of other life-threatening bloodborne viral diseases, including Hepatitis C and HIV.

Graphical Abstract

Coronavirus Disease 2019 (COVID-19), caused by the SARS-CoV-2 virus, killed at least 7 million people between 2019 and 2023, with nearly one billion documented cases.¹ Although SARS-CoV-2 is a respiratory virus, it is known to enter the circulatory system if it evades the patient’s initial immune response. Numerous studies have shown that SARS-CoV-2 viremia (i.e., live virus particles in the bloodstream) is correlated with health outcomes.^2–8 A study by Cardeñoso Domingo et al. showed that COVID-19 patients with SARS-CoV-2 viremia were 11.2 times more likely to have poor health outcomes and 13.5 times more likely to die.² In fact, in their study, all patients in the cohort who died had viremia. Similarly, another study by Jacobs et al. showed that RNA levels in blood correlated strongly with whether a COVID-19 patient was hospitalized and whether the hospitalization included the intensive care unit (ICU).³ Li et al. reported that viremia predicted the severity of COVID-19 disease and death in a dose-dependent manner.⁴ These studies point to the significance of viremia as an indicator of health outcomes. In fact, Fajnzylber et al. showed that RNA load in blood is a better predictor of outcomes than nasopharyngeal swabs.⁵

While SARS-CoV-2 primarily infects the respiratory system, it is known to afflict the entire body and to persist in the form of post-acute sequelae of COVID-19 (PASC). In PASC, or long COVID, patients experience new, returning, or ongoing health problems four or more weeks following the initial SARS-CoV-2 infection.^9–11 Symptoms of PASC include post-exertional malaise, fatigue, brain fog, dizziness, gastrointestinal symptoms, palpitations, changes in sexual desire or capacity, loss of or change in smell or taste, thirst, chronic cough, chest pain, and abnormal movements.¹¹ Viremia appears to be the link between the initial respiratory infection and extrapulmonary multiorgan disease.⁴ In fact, Su et al. determined that circulating SARS-CoV-2 RNA is a predicting factor for PASC.¹² Moreover, one study found that PASC patients had circulating SARS-CoV-2 RNA levels equivalent to levels at hospitalization.¹³ Also, a study by Siddiqi et al. supported this as well, as 76% of patients with viremia suffered myocardial injury.¹⁴

It is evident that detecting SARS-CoV-2 virus (or RNA) in blood could be used as a predictor for short- and long-term health outcomes, as well as a diagnostic for PASC. Ideally, this could be implemented in the form of a point-of-care (PoC) diagnostic so that tests could be performed in walk-in clinics, emergency departments, and primary care provider offices. However, while PoC tests for the diagnosis of COVID-19 from nasal swabs are commercially available, PoC tests for viremia are unavailable. Viremia tests represent a significant challenge for PoC tests, as the viral load is much lower than nasopharyngeal swabs, and thus, RNA amplification is required; this implies RNA purification from a blood sample. While this can be done in a sophisticated lab setting with trained technicians and expensive equipment, PoC tests must be low in cost, portable, and sample-to-answer (S2A, i.e., no user interventions).

For decades, the field of microfluidics has pursued PoC diagnostics for nucleic acid amplification tests (NAATs) with an aim to miniaturize and automate diagnostic methodologies. However, commercial success has been elusive due to the challenges of meeting the requirements of low cost, portability, and S2A. Numerous groups have been able to implement most or all of the required steps on chip,^15–20 but many still require off-chip lysis, manual interventions, expensive microfluidic devices, and peripheral equipment. Two notable S2A NAAT demonstrations are Stumpf et al., who reported an S2A test for Influenza A using spinning disk microfluidics,²¹ and Tang et al., who reported an S2A test for SARS-CoV-2 in saliva.²² However, to our knowledge, there are still no demonstrations of S2A PoC systems that can detect viruses directly from whole blood samples using microfluidics.

In recent years, magnetofluidics has emerged as a potential solution for S2A diagnostics. Magnetic actuation of molecularly functionalized paramagnetic beads through various reagents may eliminate pipetting steps as well as the need for pumps. In the earliest realizations of magnetofluidics, ferrofluids on top of an open device surface were manipulated with magnets underneath the device.^23–27 The primary application for these devices was to bind DNA/RNA from a lysate and transport it through rinse solutions and into an elution solution for a NAAT. More recently, magnetofluidics has been implemented within enclosed cassettes or chips, which offer a more pragmatic implementation. Generally, these implementations use oil to separate aqueous reagent layers; paramagnetic beads can be pulled through the oil layers from one reagent to the next without disturbing the partitions. Multiple reports have demonstrated integration of some of the sample preparation steps with NAATs using microfluidic channels.^28–30 However, the manufacturing of microfluidic chips with integrated oil partitions may be challenging. In an alternative approach, Shin et al. have demonstrated the use of multiwell cassettes with an oil layer covering the wells to maintain partitioning; an external magnet transfers beads from each well to the next to perform sample preparation and qPCR.^31,32 Trick et al. incorporated a filter onto the cassette to enable compatibility with whole blood samples.³³

We have developed a magnetofluidic approach for S2A from complex samples that does not require pumps or microfluidics and that is stable to device agitation. This approach is enabled by thermally responsive alkane partitions (TRAPs), which we described and characterized in previous work.³⁴ TRAPs are solid at ambient temperature and liquid at a moderately elevated temperature, specific to the type of alkane. When melted in between two partitioned liquids in a millimeter-scale channel, if thin enough the TRAP will breach and the two liquids on either side will combine (i.e., a removable TRAP). This behavior enables automated reagent combination. At the same time, if the TRAP is sufficiently thick, the melted TRAP will continue to partition the two liquids (i.e., a stationary TRAP). In this state, the TRAP becomes permeable to magnetic microbeads but continues to partition aqueous reagents. We have proven that magnetically transferring beads (~1 μm diameter) across a TRAP causes minimal leakage that can be further negated with the addition of rinse layers.

We previously applied removable TRAPs to detect circulating histones in blood,³⁵ and we applied stationary TRAPs to create a portable PoC system to detect SARS-CoV-2 antibodies from whole blood.³⁶ The current work refines the TRAP system to be more robust, user-friendly, and more capable. Specifically, we demonstrate for the first time S2A detection of SARS-CoV-2 viremia in a low-cost portable format. To do this, we combine both TRAP behaviors described above (removable and stationary), as illustrated in Figure 1. Virus isolation from a whole blood sample is achieved by capturing virus onto aptamer-functionalized paramagnetic microbeads and magnetically pulling the beads through a melted stationary TRAP. We show effective virus particle lysis via incubation with Proteinase K. Automatic initiation of reverse transcription loop-mediated isothermal amplification (RT-LAMP) is accomplished by combining reagents upon melting a removable TRAP following lysis. Finally, we demonstrate the PoC system’s ability to distinguish SARS-CoV-2-positive and SARS-CoV-2-negative spiked whole blood samples as well as patient blood plasma samples collected during the height of the COVID-19 pandemic.

Figure 1.

Schematic of the mechanisms within a cassette during the diagnostic test. After a blood sample is loaded into the cassette, magnetic microbeads functionalized with aptamers capture virus particles present in the sample. The temperature is increased to melt the eicosane layer, allowing a magnet to pull beads through, isolating virus particles from the blood sample. Once through, the virus particles are lysed. After lysis, the hexacosane layer is melted, initiating the RT-LAMP reaction.

METHODS

Materials.

Cassettes were fabricated by cutting an acrylic sheet with 24 mm thickness (McMaster-Carr) into 6.35 mm diameter cylinders each with a channel 3.2 mm in diameter and 23 mm deep via CNC. The alkanes in this work include n-eicosane (melting point ~37 °C), 99%, and n-hexacosane (melting point ~57 °C), 99%, both from Thermo Fisher Scientific. Pierce Streptavidin Magnetic Beads (1.05 μm diameter) were purchased from Thermo Fisher Scientific. Antarctic thermomolecular uracil DNA Glycosylase (UDG) was purchased from NEB and is used throughout this work paired with DNA amplification reactions that result in nucleic acid products containing uracil bases. UDG helps reduce the risk of carryover contamination from previous DNA amplification reactions. WarmStart Multi-Purpose LAMP/RT-LAMP Master Mix with UDG was purchased from NEB. Thermolabile Proteinase K was also purchased from NEB. Mineral oil (light, lab grade) was purchased from VWR. γ-irradiated SARS-CoV-2 was supplied as a ~10⁹ copies/mL stock solution from BEI Resources. The biotinylated SARS-CoV-2 RBD aptamer³⁷ used in this work as well as LAMP primers for the N15 gene³⁸ were purchased from IDT and their sequences can be found in the Supporting Information.

RT-LAMP Reaction.

Based in part on the recommended protocol from the NEB website, our benchtop LAMP tests consisted of the following procedure. A solution containing 1X WarmStart Multi-Purpose LAMP/RT-LAMP Master Mix with UDG, 1X LAMP Fluorescent Dye, and LAMP primers (1.6 μM FIP/BIP, 0.2 μM F3/B3, and 0.4 μM LoopF/B) for the N15 gene is created using nuclease-free water to reach the appropriate concentrations. The volume made equals 20 μL times the number of samples being tested plus one to account for possible pipetting error. After vortexing the solution, 20 μL is loaded into PCR tubes. Then, 5 μL of each sample is added to each tube. Finally, each tube is inserted into a MiniOpticon (Bio-Rad) which heats the solutions to 65 °C for 60 min while recording the fluorescence of each solution (one reading every 30 s).

Portable Controller and Reader Device.

The device used in this work consists of an ArduCAM MT9M001 Camera with an ArduCAM USB2 Camera Shield, a 530 nm long-pass filter (Thorlabs), four 470 nm blue LEDs (one for each sample cassette), a DS18B20 digital temperature sensor, a pair of 30 mm × 40 mm polyimide heaters (DWEII) positioned about 8 mm from the inserted cassettes, and two 10 mm tall neodymium magnet cylinders 25 mm in diameter (K & J Magnetics) initially positioned below the inserted cassettes. Components are housed in a 3D-printed casing that fits in the palm of a hand. A photo of the device and an exploded view are shown in Figure 2.

Figure 2.

Photo and schematic of prototype device hardware. The photo includes a penny for scale. This device is capable of testing four samples at one time.

TRAP Demonstration.

To show the capabilities of the two-TRAP assay displayed in Figure 1, the following experiment was conducted. Aliquots of eicosane and hexacosane were melted on a hot plate set to 180 °C. 15 μL of a yellow-dyed solution containing 0.1% Tween 20 in water (a surfactant concentration similar to that of the LAMP master mix used in this work) was loaded into the bottom of a channel. 8 μL of hexacosane was loaded on top of the yellow solution to seal it in place. 10 μL of blue-dyed water was loaded on top of the hardened hexacosane. 35 μL of eicosane was placed on top of the blue solution to seal it in place. Finally, 25 μL of a red-dyed solution containing 100 μg of streptavidin magnetic beads was loaded onto the hardened eicosane. The assembled cassette was then placed onto a magnet with a polyimide heater attached. The heater increased the temperature of the magnet and cassette to about 65 °C while a video of the resulting behavior was recorded. To quantify mixing, the red color value of the bottom layer of the cassette was extracted.

Virus Particle Lysis.

In our proposed assay, which captures and isolates virus particles from complex samples, it is necessary to lyse the virus particles to release the virus genome for amplification and detection. In this work, we use thermolabile Proteinase K to lyse virus particles. To determine how effective Proteinase K is at lysing virus particles, first a pure (free of extra-viral RNA) virus particle solution must be made. This was done by incubating a 50 μL solution including PBS and 10,000 copies/μL γ-irradiated SARS-CoV-2 with 1 μL (20 μg) Monarch RNase A (NEB) for 5 min at 56 °C. 1.6 μL (64 units) of RNase Inhibitor (Murine, NEB) was added to the solution after the incubation. Six of these extra-viral RNA-free SARS-CoV-2 solutions were made. In three of them, 2.5 μL (0.3 units) of thermolabile Proteinase K was added. These three solutions were incubated at 37 °C for 15 min to allow Proteinase K to digest proteins. Then, the solutions were incubated at 55 °C for 10 min to inactivate the Proteinase K. To prepare the RT-LAMP detection reaction, nine 20 μL solutions were created as described above in the RT-LAMP Reaction section. 5 μL of each solution that underwent Proteinase K incubations as well as each solution that did not was added to LAMP reagents to create six separate LAMP reactions. Three more reactions were prepared with 5 μL of PBS used as a negative control. Each solution was then incubated at 65 °C for 60 min, with fluorescence measurements taken every 30 s in the MiniOpticon.

Virus Viability across a Melted TRAP.

In order for our proposed assay to work properly, captured virus particles must remain intact and attached to magnetic beads while subjected to the increased temperature of a melted TRAP. To determine whether viruses remain viable after crossing a TRAP, the following experiment was conducted. First, 900 μg of streptavidin magnetic microbeads were rinsed with wash buffer then incubated with 225 pmol biotinylated capture aptamers in 18 μL of PBS for 60 min. Following the incubation, 12 μL of the mixture was added to a 178 μL solution containing PBS and 10000 copies/μL γ-irradiated SARS-CoV-2. Simultaneously, the remaining 6 μL of the aptamer-bead mixture was added to 84 μL of PBS (to act as the no-target control). Both mixtures were incubated for 15 min. Eicosane was melted on a hot plate set to 180 °C. Six cassettes were loaded initially with 30 μL of the aptamer-bead-virus mixture, while three cassettes were loaded with 30 μL of the no-target control mixture. 40 μL of melted eicosane was added to the three cassettes loaded with the no-target control and to three of the cassettes that were loaded with the aptamer-bead-virus mixture to form TRAPs. To the three cassettes that remained, mineral oil was used in place of eicosane. 50 μL of water was then added to each cassette, with careful deposition on top of the oil to ensure no mixing occurred between the layers. The six cassettes containing TRAPs were placed horizontally on a hot plate set to 65 °C for 5 min to melt the TRAPs. Once melted, a neodymium magnet was used to transfer the beads across each melted TRAP, as well as across the oil partition in the cassettes (not on the hot plate). After bead transfer, the cassettes with TRAPs were removed from the hot plate to allow eicosane to harden. Once the eicosane hardened, 40 μL of the mixture on top of each TRAP and oil partition (now containing the magnetic beads) was pipetted out and into separate tubes each containing 1 μL of thermolabile Proteinase K. The nine tubes were incubated at 37 °C for 15 min to allow Proteinase K to break down proteins in any virus particle that made it across the TRAP or oil partition. Then, a 10 min incubation at 55 °C inactivated the Proteinase K. An RT-LAMP detection reaction was prepared in the meantime: nine 20 μL solutions were created as described above in the RT-LAMP Reaction section. 5 μL of the supernatant of each mixture that underwent Proteinase K incubations (obtained by moving the beads out of the way of a micropipette via a magnet) was transferred to each 20 μL LAMP solution. Each solution was then incubated at 65 °C for 60 min while fluorescence data was recorded in the MiniOpticon.

Sample-to-Answer SARS-CoV-2 Blood Test.

To demonstrate that our S2A system can detect viruses in blood, the following experiment was conducted. 100 pmol biotinylated capture aptamers were attached to 400 μg streptavidin magnetic beads via a 15 min incubation in 8 μL of PBS. A LAMP solution and a lysis solution were prepared by combining and mixing the following reagents. (1) LAMP solution: 50 μL WarmStart Multi-Purpose LAMP/RT-LAMP 2× Master Mix with UDG, 2 μL 50× LAMP Fluorescent Dye, and 8 μL water. (2) Lysis solution: 4 μL (~0.5 units) Thermolabile Proteinase K, 10 μL LAMP primer mix containing 16 μM FIP/BIP, 2 μM F3/B3, and 4 μM LoopF/B for the N15 gene, 2 μL (2 units) Antarctic Thermolabile UDG, and 24 μL water. 15 μL of the LAMP solution was added to the bottom of four acrylic cassettes, followed by an 8 μL hexacosane seal, then 10 μL of the lysis solution was added after the hexacosane hardened, then 38 μL of the eicosane, then 2 μL of the bead mixture after the eicosane hardened. 20 μL of whole blood (obtained via finger stick) with 7 μg sodium polyanethole sulfonate (to prevent blood from coagulating around the magnetic beads), spiked with 5 μL of varying amounts of γ-irradiated SARS-CoV-2 in PBS was added to each cassette. The cassettes were capped and then inserted into the portable fluorescence reader to commence the assay reactions.

The cassettes were subjected to 45 °C for 6 min to melt the first TRAP (made of eicosane); after it melted, the beads were pulled through the eicosane by a magnet at the bottom of the cassettes. After 5 min, the magnet positioned below the cassettes was moved away to remove the beads from its influence. At this point, the magnetic bead complexes were in the lysis solution, isolated from the blood sample (Figure 1). The heating element within the fluorescence reader was shut off for 6 min to allow Proteinase K to degrade any virus envelope proteins, compromising the stability of virus shells and releasing viral genome from virus particles present in the mixture. The heating element was turned on again to deactivate Proteinase K, melt the second TRAP (made of hexacosane) to combine LAMP reagents, and initiate the LAMP reaction. The space within the device stabilizes to 65 °C after about 15 min. This temperature was held until 60 min passed since the insertion of the cassettes. Fluorescence measurements were recorded as normalized grayscale pixel values collected by the ArduCAM. These measurements began 23 min into the test as that is the point in which the hexacosane TRAP melts, initiating the LAMP reaction.

SARS-CoV-2 Patient Plasma Test.

Plasma samples (deidentified) from COVID-19 patients were obtained from the University of Maryland Medical Center. Samples were acquired from hospitalized COVID-19 patients between the dates of 2020–09-18 and 2022–02-15 under approved IRB protocols. Each sample was tested by the MiniOpticon benchtop RT-qPCR system. Using the QIAamp Viral RNA Kit (Qiagen), 140 μL of each plasma sample was filtered into 60 μL of elution buffer containing extracted virus genomes that may have been present in the plasma sample. 5 μL of the resulting solution was added to a one-step RT-qPCR master mix (Promega). Each solution underwent a 15 min reverse transcription step at 45 °C for 15 min, followed by 2 min in a 95 °C initial denaturation step, then 45 PCR cycles with each cycle being 95 °C for 3 s and 55 °C for 30 s. The threshold cycle in this work is defined as the first PCR cycle that resulted in a fluorescence measurement higher than 0.011 AU. To estimate the number of copies in each solution, dilutions of SARS-CoV-2 (BEI) were subjected to the same sample treatment and qPCR reaction, and a standard curve was created (Supporting Information). Cycle thresholds from tested patient samples were compared to those of the standard curve. Plasma samples were also tested with the S2A diagnostic. The experimental setup above in the Sample-to-Answer SARS-CoV-2 Blood Test section was repeated, but instead of adding 20 μL of whole blood spiked with 5 μL of SARS-CoV-2, 25 μL of plasma from COVID-19 patients was loaded into the cassettes.

RESULTS AND DISCUSSION

TRAP Operation.

Our previous work has shown that by melting a removable TRAP, we can control when prepartitioned reagents mix, simply by adjusting temperature.³⁹ We have also shown stationary TRAPs that enable a sample-to-answer immunoassay.³⁶ In this work, which presents an assay capable of detecting virus genomes from complex samples, both forms of TRAPs are utilized. A stationary TRAP is required so that, when melted, it will prevent the sample (e.g., whole blood) from interfering with detection chemistry while still allowing virus particles captured onto magnetic beads to transition to the lysis reagent. Meanwhile, a removable TRAP enables the sequential steps of virus lysis and RT-LAMP. The RT-LAMP reagents must be sequestered during proteolytic lysis (to avoid degradation of the polymerase) and then added following the heat-killing of the protease. Thermolabile Proteinase K is inactivated at a temperature lower than the melt temperature of the removable TRAP, and thus, when increasing the temperature to release the TRAP and combine the reagents, Proteinase K is inactivated. Once the removable TRAP is melted, the nucleic acid amplification reaction (RT-LAMP) commences.

Figure 3A illustrates the process described above using dyes so that it is observable. The red solution (representing the blood sample in the assay) remains separated from the blue solution (representing the lysis reagent layer) while the beads traverse the stationary TRAP. As the temperature increases to melt the removable TRAP, the blue solution and the yellow solution (representing RT-LAMP reagents) quickly mix. The sharp slope in Figure 3B indicates quick mixing upon melting the removable TRAP. A full video of the setup displayed in Figure 3 can be found in the Supporting Information. This demonstrates that the combination of the two types of TRAPs in a single cassette can maintain the partitioning of the blood sample away from the assay while sequentially enabling the assay steps, including the precise addition of reagents and rapid mixing.

Figure 3.

(A) Photos depicting each stage of the assay within a cassette with dyes in each solution. Magnetic beads traverse the eicosane stationary TRAP without red dye leaking through. When the hexacosane removable TRAP melts, the layers on either side quickly mix. (B) Data representing the speed at which two liquids mix upon melting a removable TRAP.

Thermolabile Proteinase K Enables RNA Release from Virus and Is Compatible with Sample-to-Answer LAMP.

It has been shown that Proteinase K can act as a viral lysis reagent and streamline sample preparation before RT-qPCR.⁴⁰ It works by breaking down proteins that are abundant in the envelope of many viruses. Here we utilize a thermolabile Proteinase K that is heat-inactivated below the operation temperature of LAMP. Notably, it is active at a temperature above the melt temperature of the stationary TRAP and below the temperature of the removable TRAP. Figure 4 demonstrates the effectiveness of Proteinase K in lysing the SARS-CoV-2 virus. The data compares the LAMP-based amplification of the SARS-CoV-2 genomes for three aliquots of virus treated with Proteinase K to three samples that were not treated. Note that no subsequent purification steps are performed after lysis; the sample is heated to inactivate the Proteinase K and then directly added to the RT-LAMP master mix. As Figure 4 shows, the sample treated with Proteinase K amplified quickly, while the untreated sample is essentially not detected.

Figure 4.

LAMP results from a sample of virus particles after incubation with Proteinase K (orange curve) compared to a sample of virus particles without a Proteinase K incubation (green curve), along with a no-target control (blue curve).

Aptamer-Beads Pull Virus Through a Melted TRAP.

In previous work, we demonstrated that antibody–antigen complexes attached onto magnetic bead traversed melted stationary TRAPs without releasing.³⁶ Here, however, the cargo is virus particles that are captured via aptamers instead of antibodies. Therefore, it was necessary to investigate whether our bead-aptamer-virus complex remains viable after traversing a melted TRAP. We compared the amplification of the virus that traversed a melted eicosane TRAP on aptamer-functionalized magnetic beads with the amplification of the virus that was similarly pulled through a mineral oil partition at room temperature. LAMP data from both experiments is presented in Figure 5. The timings of the amplification signal for the two cases are not statistically significantly different. We conclude that the elevated temperature of the TRAP results in minimal or no virus loss while beads are traversing a melted TRAP.

Figure 5.

LAMP results from a sample of virus particles after capture onto magnetic beads and transfer across a melted eicosane TRAP (orange curve) compared to transfer across an oil partition at room temperature (green curve), along with a no-target control (blue curve).

Sample-to-Answer Detection of SARS-CoV-2 in Blood.

To demonstrate the capability of our S2A system to detect viruses in whole blood, we spiked whole blood (drawn from finger pricks) with specific concentrations of the SARS-CoV-2 virus and tested it in our dual-TRAP cassette. Specifically, 20 μL of whole blood was spiked with 5 μL of PBS containing 5000, 500, 50, 5, or 0 copies of SARS-CoV-2 (according to the stock concentration reported by the BEI). Figure 6A shows the average amplification curves for three samples with 5000 copies (orange curve) and three blank samples (blue curve). The curves show rapid amplification for the positive sample with minimal variation in the time to positive (TTP); additionally, the curves show a significant amount of time between the TTP values of the positive and negative samples.

Figure 6.

(A) Fluorescence data from our portable reader device plotted on an annotated graph that includes the temperature profile inside the device (dotted line). (B) Compiled results from 35 blood tests using our device.

Figure 6B reports the results of 35 different blood sample tests, seven for each concentration of spiked SARS-CoV-2. If a blood sample resulted in a TTP of 22 min or less after the initiation of the LAMP reaction, the sample was deemed positive. The samples that had a TTP of greater than 22 min or never amplified for the duration of the test were deemed negative. Samples containing at least 500 copies of SARS-CoV-2 resulted in 100% accuracy. One false negative was reported for the seven samples with 50 copies, and two false negatives were reported for the seven samples with 5 copies. Thus, without any sample preprocessing, highly sensitive detection of viruses directly from a whole blood sample is made possible by our diagnostic system. This capability to detect virus directly from whole blood has the potential to significantly decrease time to diagnosis.

Detection of SARS-CoV-2 in COVID-19 Patient Plasma Samples.

To further validate the capability of our system to detect viremia, we used our dual-TRAP cassette and hand-held device to test plasma samples from patients with severe COVID-19 disease. The samples were first tested by conventional RT-qPCR; RNA in the samples was quantified using RT-qPCR. Three samples had quantifiable RNA while six samples appeared to have no RNA (reported as “not detected”). These nine patient samples were then tested with our S2A system. The correlation between our S2A system and conventional RT-qPCR is presented in Figure 7. The three patient samples that had quantifiable RNA had the lowest time-to-positive with the S2A system; each LAMP reaction had a TTP within 16 min of reaction initiation (i.e., when the hexacosane TRAP melted). All of the LAMP signals in our device that resulted from testing SARS-CoV-2-negative (i.e., not detected) plasma samples occurred at later time points compared to the positive samples, with many not appearing until after 30 min into the LAMP reaction. Two trials of the negative samples, however, resulted in an LAMP signal close to that of the positive samples. These results suggest that our S2A system is able to distinguish between viremia-positive and viremia-negative patient samples.

Figure 7.

COVID-19 patient plasma samples tested via RT-qPCR (x-axis) and via our S2A device (y-axis). Samples confirmed positive by RT-qPCR are colored orange, while those not detected by RT-qPCR are colored blue.

CONCLUSIONS

In this work, we have demonstrated the sample-to-answer detection of the SARS-CoV-2 virus in blood samples. Our system positively identified the virus in 100% of blood samples spiked with 500 copies/sample and >70% of blood samples spiked with 5 copies/sample. In addition, our system showed good agreement with conventional RT-qPCR in detecting SARS-CoV-2 viremia in banked samples from patients with severe COVID-19 disease. Notably, our S2A system utilizes an easy-to-manufacture cassette and an automated portable device that does not require microfluidics, a pump, or precise sample handling steps.

Future work will reduce the footprint of the portable device and include a motorized arm to slide the magnet in and out of the position under the cassette. Additionally, we can broaden the application of our system simply by changing the aptamer sequence and primer sequences. We aim to detect other life-threatening diseases such as Hepatitis C and HIV. Additionally, future extensions of this work should consider the specificity of the virus capture. As SARS-CoV-2 evolves, the aptamer sequence may become less specific and, thus, less effective at capturing the virus. In addition, as the system is adapted to other viruses, the capture molecule must be changed accordingly, and genotype variation must be considered. The optimal solution may be a magnetic bead functionalization that is capable of nonspecific virus capture.