Handheld Purification-Free Nucleic Acid Testing Device for Point-of-Need Detection of Malaria from Whole Blood

Abstract

World Health Organization's aim to eliminate malaria from developing/resource-limited economies requires easy access to low-cost, highly sensitive, and specific screening. We present a handheld nucleic acid testing device with on-chip automated sample preparation to detect malaria (Plasmodium falciparum) infection from a whole blood sample as a feasibility study. We used a simple two-reagent-based purification-free protocol to prepare the whole blood sample on a piezo pump pressure-driven microfluidic cartridge. The cartridge includes a unique mixing chamber for sample preparation and metering structures to dispense a predetermined volume of the sample lysate mixture into four chambers containing a reaction mix. The parasite genomic DNA concentration can be estimated by monitoring the fluorescence generated from the loop-mediated isothermal amplification reaction in real time. We achieved a sensitivity of ∼0.42 parasite/μL of whole blood, sufficient for detecting asymptomatic malaria parasite carriers.

Keywords: loop-mediated isothermal amplification (LAMP); malaria; microfluidics; nucleic acid testing (NAT); point-of-need; purification-free.

Figure 1.

a) Extraction and purification-free Arcis sample preparation protocol performed in tubes. Step 1: Add Pf positive blood to Arcis 1 (1:3 v/v), step 2: Add blood lysate from step 1 to Arcis 2 (1:1 v/v), and step 3: use 1 μl of the mixture from step 2 in a 25 μl amplification reaction. b) Black curves represent the qPCR amplification curves (triplicates) with 1 μl 2-fold serially diluted Pf gDNA (stock concentration 50 ng/ μl) as standards in 25 μl total reaction volume. Ct values are used to generate the qPCR standard curve in c. Colored curves represent the qPCR amplification curves (triplicates) with 1 μl Arcis prepared Pf gDNA spiked whole blood. c) qPCR standard curve. Black circles represent mean Ct values for gDNA samples and are used to construct the standard curve. Colored crosses represent mean Ct values for prepared spiked whole blood samples and are used to determine the resulting gDNA concentration after ‘extraction’ or sample preparation. d) Measured gDNA concentration versus expected gDNA concentration in whole blood as a result of dilution. 0.039 ng/μl and 0.019 ng/μl gDNA in spiked whole blood are not amplified within 55 cycles due to dilution. e) LAMP curves (triplicates) for Arcis prepared Pf gDNA spiked whole blood samples, identical to those used in b. 0.039 ng/μl and 0.019 ng/μl gDNA in spiked whole blood are amplified by the LAMP assay. f) The correlation between LAMP times to positive and qPCR cycle thresholds for Arcis prepared Pf gDNA spiked whole blood samples.

Figure 2.

a) Details of each layer of the PMMA cartridge: The Top, Metering, and Bottom layers are 1 mm thick, the Microchannel layer is 3.17 mm thick, and the Reservoir layer is 5.65 mm thick. b) Assembled cartridge along with details of each chamber. Chambers for Lysate in Arcis 1 and Arcis 2 have a volume of 150 μl, reaction chambers have a volume of ~70 μl, and semi-circular metering chambers have a volume of ~4 μl. c) Use of elliptical structure to mix two sequentially loaded liquids. d) Top view of the sample preparation process on the microfluidic cartridge. e) Front view of the microfluidic cartridge showing various stages of the mixture dispensing step.

Figure 3.

a) Schematic of the instrument developed for streamlined, automated pressure-driven blood sample preparation for LAMP analysis on a microfluidic cartridge. b) Image of the assembled instrument along with the PMMA microfluidic cartridge. c) Detailed view of various modules of the instrument. d) Heating kinetics. 25 μl H₂O and 45 μl mineral oil were pipetted into the microfluidic cartridge’s LAMP reaction chambers. The heating block’s temperature is regulated to 71 °C by the onboard Arduino nano within five min. An external sensor was used to measure the temperature of the water by placing the thermistor inside the LAMP chamber. The LAMP chamber contents take another five minutes to reach ~62 °C and remain within the required LAMP temperature range at 64.61 +/− 0.83 °C. e) Color sensor characterization. 25 μl fluorescence (calcein) of varying concentrations was pipetted into all four chambers of the cartridge, and RFU (red channel counts) were recorded for five min. A linear relationship with increasing concentration and consistency among all four chambers is seen. The error bars represent the variation of one channel over five minutes. f) Pf gDNA amplification on the instrument. 10-fold serially diluted Pf gDNA (10² ng/ μl to 10⁻⁵ ng/μl) and LAMP mix were manually pipetted into all four cartridge chambers and amplified. g) Gel electrophoresis image of the amplicons confirms gDNA amplification on the instrument. The smear and banding pattern around the 200 bp rung confirm the LAMP amplification of Pf gDNA. As expected, the smear pattern is not seen for 10⁻⁵ ng/μl concentration and NC (water). The bands seen below 100 bp rung are due to the LAMP primers. h) Times to positive for each gDNA concentration. Error bars represent the standard deviation among the four chambers.

Figure 4.

a) Proposed test workflow: i) Collect 40 μl of the blood sample into a tube containing 120 μl of Arcis 1 reagent, shake well, and incubate for 1 min. ii) Transfer 150 μl of the blood lysate (denoted by red) to the microfluidic cartridge that is preloaded with Arcis reagent 2 (denoted by blue) and LAMP master mix (denoted by yellowish-green) topped with mineral oil. iii) Connect the microfluidic cartridge’s tubing to the piezo pump, place it in the recess and close the lid. Once the start button is pressed, autonomous sample preparation begins by mixing the lysate and Arcis reagent 2 and dispensing the mixture into the four LAMP reaction chambers, followed by LAMP for 60 min. iv) Real-time LAMP results can be plotted on a computer screen, or endpoint results can be displayed on the status LED bar. The numbers 1 through 4 represent the LAMP chambers. b) Amplification curves for contrived blood samples prepared by spiking 36 μl whole blood with 4 μl 10-fold serially diluted Pf gDNA (10² ng/μl to 10⁻⁴ ng/μl). Lysis with Arcis 1 was performed in a tube, followed by further autonomous processing of the resulting mixture in the cartridge on the analyzer. A quick evaluation warrants setting the threshold at 50 RFU. c) Times to positive vs. gDNA concentration in contrived blood samples. Parasite concentration was determined using the relation 1 parasite = 2.35×10⁻⁵ ng. A quick investigation suggests a sensitivity of 0.42 parasites/μl. d) The Pearson correlation between times to positive seen on the instrument using automated sample preparation and detection on the cartridge and benchtop thermal cycler using manual sample preparation steps (same as Figure 1e). e) Analytical specificity test using other Plasmodium gDNA spiked in whole blood. Contrived blood samples were processed with Arcis 1 and Arcis 2 in tubes, and LAMP was performed on the cartridge in the instrument simultaneously. Only the Pf gDNA spiked blood sample (10² ng/μl) was amplified, while the Pv and Po gDNA spiked blood samples, and the non-spiked blood remained negative.