Vortex‑ and Centrifugation‑Free Extraction of HIV‑1 RNA

Abstract

Background and Objective

HIV viral load measurements play a critical role in monitoring disease progression in those who are on antiretroviral treatment. In order to obtain an accurate measurement, rapid sample preparation techniques are required. There is an unmet need for HIV extraction instruments in resource-limited settings, where HIV prevalence is high. Therefore, the objective of our study was to develop a three-dimensional (3D) microfluidic system to extract HIV-1 RNA with minimal electricity and without complex laboratory instruments.

Methods

A 3D microfluidic system was designed in which magnetic beads bound with nucleic acids move through immiscible oil–water interfaces to separate HIV-1 RNA from the sample. Polymerase chain reaction (PCR) amplification was used to quantify the total amount of HIV-1 RNA extracted as we optimized the system through chip design, bead type, carry-over volume, carrier RNA concentration, and elution buffer temperature. Additionally, the extraction efficiency of the 3D microfluidic system was evaluated by comparing with a Qiagen EZ1 Advanced XL instrument using 20 HIV-1-positive plasma samples.

Results

Our method has near-perfect (100%) extraction efficiency in spiked serum samples with as little as 50 copies/mL starting sample. Furthermore, we report carry-over volumes of 0.31% ± 0.006% of total sample volume. Using the EZ1 Advanced XL as a gold standard, the average percentage HIV-1 RNA extracted using the microchip was observed to be 65.4% ± 24.6%.

Conclusions

From a clinical perspective, the success of our method opens up its possible use in diagnostic tests for HIV in the remote areas where access to vortexes and centrifuges is not available. Here we present a proof-of-concept device which, with further development, could be used for sample preparation at the point of care.

1. Introduction

Viral load testing is the cornerstone for monitoring HIV-infected patients on antiretroviral therapy. The extraction of HIV RNA from plasma is the first step in this process. Many commercial nucleic acid extraction kits have been developed that use pre-mixed reagents and magnetic beads or silica columns to bind RNA [1–3]. These techniques require repeated centrifugation and multiple washing steps. To meet the demand for high throughput of RNA extraction, automated systems have been designed to replace labor-intensive procedures [3]. Though these systems are accurate and reliable, they remain expensive, require a constant energy source, and are not scalable. Therefore, they are not applicable for low-resource settings in which a majority of HIV incidences occurs [4]. A simple, rapid, and accurate extraction technique is necessary for the sample-in, result-out platform in point-of-care settings.

Researchers have made some attempts to convert tube-based isolation kits to microfluidic platforms, which are more realistic for implementation in low-resource settings. Bordelon et al. [5] developed an RNA extraction cassette that encapsulates successive segments of precipitate, wash, and elute solutions in a tube separated by air. Forcing magnetic beads along the tube, they demonstrated sensitive and specific nucleic acid extraction within 3 min. Although the method of operation was claimed to be simple, it suffers from a time-consuming and complex initial preparation of long (61 cm) capillaries. Moreover, the method was not tested with clinical samples (swabs, blood, etc.). A similar idea has also been implemented on a microfluidic capillary valve by den Dulk et al. [6]. Sur et al. [7] employed a design involving a channel containing liquid wax which connects a lysis chamber to an elution chamber. Magnetic beads with captured nucleic acids pass from the lysis chamber, through an immiscible hydrophobic barrier and into an elution chamber. Additionally, recent reported techniques include immiscible filtration assisted by surface tension (IFAST), where an immiscible phase is used for enhanced separation of cells [8] and nucleic acids [9]. From this work, Berry et al. [8, 9] achieved promising results using an open-channel chip in which microfluidic channels are filled with Chill-out^™ Liquid Wax (Bio-Rad, Hercules, California). In our previous work [10–13] we have established isolation of nucleic acids using movable magnet beads in a closed microfluidic channel. The authors of these studies did not note a lack of centrifuge, vortex, or electricity, so we assume at least one of those recourses was used before or after the use of a microchip in their protocols; this makes them less than ideal for low-resource settings.

Vortex- and centrifugation-free conditions for RNA extraction assays have not yet been studied. Here we use a three-dimensional (3D) polyurethane microchip system to provide insight into transport processes as well as report extraction efficiencies using HIV-positive patient plasma samples. Polyurethane is a commonly used material for 3D-printed devices due to its low cost, flexibility, low absorption of water, resistance to extreme temperatures and low degradation with many chemicals, including alcohol. The average water contact angle is 99° ± 1°. Each chip is comprised of a sample well, a connecting microchannel, and an elution well. In our system, the nucleic acid bound beads are moved from a sample well, through a castor oil phase, and into an elution well. Once in the final well, the nucleic acids are removed from the beads and the resulting sample can be used for amplification. Here, reverse transcription polymerase chain reaction (RT-PCR) was used to amplify and quantify our nucleic acid extraction efficiency. Using a printed 3D microfluidic chip and HIV-spiked serum or HIV-positive patient plasma samples, nucleic acid adsorption to and transport with magnetic beads is optimized. Our simplified method for sample preparation allows for extraction of viral HIV RNA. When amplified, this RNA can be used to detect an HIV-1 infection or to monitor an HIV-positive patient’s viral load. After the chip protocol was optimized, we compared our isolation microchip system with the commercially available Qiagen EZ1 Advanced XL Nucleic Acid Automated Purification System (Valencia, CA, USA) using plasma samples from individuals with an HIV-1 infection. We present here a proof-of-concept device that was optimized using spiked serum samples and compared to a gold-standard machine with plasma patient samples.

2. Materials and Methods

2.1. Microchip Fabrication and Design

The microchips were printed using Accura^® ClearVue^™ resin on a high-resolution SLA (stereolithography) process (Quickparts, Atlanta, GA, USA). As shown in Fig. 1, each microchip consists of a sample well (7 mm in diameter, 10 mm in height, volume 380 μL), a connecting microchannel (32 mm in length, 4 mm in width, 1 mm in height), and an elution well (7 mm in diameter, 10 mm in height, volume 380 μL). This architecture is different from two-dimensional microfluidics chips where oil is inserted from an in-plane side channel [14]. A removable chamber cap was designed to close the oil reservoir. We also designed four air openings to allow for the displacement of oil during bead movement within the channel. The air openings prevent any hydrodynamic flow of oil originating from bead drag force and minimizes incidents where the oil–water interface breaks due to magnetic bead transport. The microchannel width (= 2 mm), depth (= 1 mm), and length (= 32 mm) were optimized to maximize bead transport efficiency while minimizing carryover liquid. Both ends of the connecting microchannels were tapered to half depth to increase the interfacial stability. The forces exerted on the bead aggregate as it is moved from the sample well into the oil channel (the magnetic driving force $F_{\text{M}}$, the hydrodynamic drag force $F_{\text{D}}$, and the interfacial tension force $F_{\text{I}}$,) are shown in Fig. 2. The size of the magnetic bead aggregate through the interface perturbs the equilibrium between interfacial and hydrodynamic forces. The small holes in the top of the chip are used for oil displacement. For successful bead transfer, the beads are moved through the oil channel at approximately 1 mm/s. Previous studies in our microchip system determined this as the optimal speed for bead movement [14]. To ensure the bead aggregate is moved through the oil (as opposed to between the chip bottom and the oil) the magnet is moved by hand, alternating which side of the oil channel it makes contact with. This keeps the bead aggregate surrounded by oil and limits the interactions between the bead aggregate and the chip, while still moving the bead aggregate from one well to another. We can therefore neglect friction in our force model of the beads.

Fig. 1.

The microfluidic chip used in the experiments. a Side of chip showing dimensions of reservoirs and channels. The oil channel is 1 mm high and 2 mm wide, as shown in the cross-section. b The sample well is shown in blue, the oil channel in pink, and the elution well in green

Fig. 2.

Forces acting on the bead aggregate at the sample oil interface. $F_D$ hydrodynamic drag force, $F_I$ interfacial rtension force, $F_M$ magnetic driving force

2.2. Microchip Operation

For nucleic acid extraction, 2 μL of carrier RNA (stock concentration 1 μg/μL) was added to 1 mL of lysis-binding solution concentration, both from the MagMax^™ Viral RNA Isolation Kit (catalog number AM1939, Ambion, Austin, TX, USA). This was followed by the addition of 1 mL of sample and mixed for 4 min at ambient temperature by aspirating and dispensing a 200 μL pipette. The solution was heated for 10 min at 70 °C in an incubator (Standard-Incubator Series BD, Binder, Bohemia, NY, USA), and then 1.25 mL of 100% ethanol (Pharmco-Aaper, Brookfield, CT, USA) was added. After incubation for 2 min at ambient temperature, 20 μL of bead mix (1:1 RNA binding beads: lysis/binding enhancer, included in MagMax^™ kit) is added and pipette mixed for 5 min, allowing for the precipitated RNA to bind to the magnetic beads. The beads were collected by placing a cylindrical magnet (Cyl0751, Super Magnet Man, Pelham, AL, USA) under the collection tube for 6 min. Supernatant was removed, leaving approximately 100 μL of sample lysate and all the collected beads in the collection tube. Next, 60 μL of castor oil (Sigma-Aldrich, Saint Louis, MO, USA) was added to the oil channel of the chip, which was capped after the addition of the oil. The choice of castor oil and more information on chip design is discussed in previous work [15]. This was followed by quickly adding 100 μL of elution buffer (included in the MagMAX^™ kit) into the elution well and 100 μL of sample lysate into the sample well. As each liquid filled its designated well, an interface was formed between the oil and the sample liquid, and between the oil and the elution liquid. This interface, along with hydrodynamic resistance, prevents the oil from entering the sample and elution wells. Using a magnet (D86-N52, K&J Magnetics, Inc., Pipersville, PA, USA), the paramagnetic beads from the sample lysate were collected at the interface and pulled through the oil channel and into the elution well. This is done by holding the sample-containing microchip in one hand and the magnet in the other hand. A half-circle motion from one side of the chip, under the channel, to the other side of the chip is traced to move the beads from the sample well to the elution well. The sample lysate/oil interface retains the protein and salt contaminants, which can inhibit downstream amplification of nucleic acids, in the sample well while only allowing the beads and bound nucleic acids to move into the elution well. This obviates the need for labor-intensive wash steps required by most nucleic acid isolation kits. Finally, the elution buffer and paramagnetic beads were removed from the elution well and heated for 10 min at 70 °C in an incubator in a 1.5 mL Eppendorf tube. The elution buffer was then pipette mixed for 2 min, which allowed for the RNA to desorb from the beads. The beads were collected with a magnet and the elution buffer, which contains the extracted nucleic acids, was placed in a new Eppendorf tube for storage or further processing. The chip is not reused after a run; hence, there was no issue of cross contamination in the cartridge. Table 1 shows a summary of the protocol for 100 μL and 1 mL samples. Unless otherwise noted, this protocol with the aforementioned reagents was used for all experiments in this study.

Table 1.

Sample preparation protocols

	Carrier RNA (μL)	LBS	100% ethanol	Bead mix (μL)	Castor oil (μL)	Sample lysate (μL)	Elution buffer	Wash 1 (μL)	Wash 2 (μL)
Sample preparation for on-chip
1 mL sample preparation	2	1 mL	1.25 mL	20
100 μL sample preparation	2	200 μL	250 μL	20
On chip					60	100	100
Conventional in tube sample preparation
100 μL sample preparation	2	200 μL	250 μL	20			100	300	450

2.3. Extraction of HIV Spiked in Serum Samples

For optimization experiments, AccuType^™ HIV-1 subtype A (Ghana 1–2496, SeraCare Life Sciences, Milford, MA, USA) was spiked into serum (Golden West Biologicals, Temecula, CA, USA). This is a clinically relevant subtype of HIV which is common in many HIV infections. Different brand paramagnetic beads were tested to see which yielded the highest RNA extraction. All bead experiments had a sample starting concentration of 28,000 HIV-1 copies per 100 μL and n = 3 replicates were performed. Ambion beads 250–500 nM in diameter, Chemagen beads of unknown size (catalog number M-PVA SAV1, Perkin Elmer, Waltham, MA, USA), and Chemagen 1 and 2 beads 3–5 μm in diameter (catalog numbers M-PVA C21, M-PVA C22, Perkin Elmer) were tested. Temperature optimization was performed (n = 3; 28,000 copies HIV/spiked serum sample, 100 μL starting volume) at 21, 30, 40, 50, 60, and 70 °C for 10 min for the elution buffer heat step. The impact of carrier RNA was tested with the addition of 0, 1, and 2 μL of carrier RNA in our system at low HIV starting copy levels (n = 3; 50 copies HIV/100 μL sample). The 100 μL starting sample protocol has the same steps as the 1 mL starting sample protocol already described, but with varied reagent volumes. The volumes are summarized in Table 1.

Five different concentrations of HIV-1 spiked serum (5.4 × 10⁵ copies/mL to 54 copies/mL HIV) samples were serially diluted and run through our device with a 1 mL starting sample (n = 15, three replicates per concentration). The extracted nucleic acids were analyzed with RT-PCR. A positive control from the spin column extraction kit (Carlsbad, CA, USA) was run with a starting concentration of 5.4 × 10⁵ copies/mL. Additionally, this created a high positive standard with a known starting concentration for our reference. Nuclease-free water was run as a no template control (NTC).

2.4. Extraction of HIV‑Positive Patient Plasma Samples

For comparison of our optimized chip to the gold standard for sample extraction, 20 plasma samples from HIV-1-infected individuals with viral loads ranging from 40 to 100,000 copies/mL were tested in both the chip and the EZ1 instrument. This allowed us to directly compare two different extraction methods. Five samples were tested from each of the following ranges of concentration: 40–2000 copies/mL (low), 2000–10,000 copies/mL (medium–low), 10,000–100,000 copies/mL (medium–high), and > 100,000 copies/mL (high). These samples were originally collected at the Miriam Hospital in Providence, Rhode Island, USA for two research projects between September 2003 and- September 2008 (IRB approval CMTT#2019–03 0709 and CMTT#2080–05). The samples were aliquoted and stored at − 80 °C prior to testing. These samples were previously tested in the HIV-1 NucliSens^® assay to determine viral load (bioMerieux, Durham, NC, USA). Additionally, five HIV-negative human plasma samples were run as controls. To test a plasma sample, 100 μL was taken from an aliquot and run in the chip while an additional 100 μL was taken from the same aliquot and run in parallel on the EZ1 system. To compare the two extraction methods, quantitative RT-PCR (RT-qPCR) was run in duplicate for each sample. Cycle threshold (Ct) values were obtained for the chip extraction eluant and the EZ1 extraction eluant for a given sample on the same RT-qPCR run. A workflow schematic is included in the Electronic supplemental Material. Additionally, there was a positive control (10,000 copies/mL, SeraCare Life Sciences) and an NTC. Extraction of the samples with the Qiagen EZ1 Advanced XL using the EZ1 DSP Virus Kit (Qiagen) was performed following the manufacturer’s recommendations.

2.5. Quantification and Extraction Efficiency

The following PCR protocol was used for amplification and quantification of all samples used in this study. Super-Script^™ III Platinum^™ One-Step qRT-PCR system with Platinum^™ Taq DNA Polymerase (Thermo Fisher Scientific, Waltham, MA, USA) and EvaGreen^® dye 20 × in water (Biotium, Fremont, CA, USA) were used with a Piko-Real^™ 24 RT-PCR instrument (ThermoScientific). The final concentration of the forward (GTGCAGGGGAAAGAATAATAGACA) and reverse (CTGTCCCTGTAATAAACCCGAA) primers was 0.2 μM. Primers were designed inhouse and cycling conditions can be found in the Electronic Supplementary Material. For a positive control, a spin column extraction kit was used on the highest concentration of HIV-spiked serum and the resulting eluant was included in each PCR run. Nuclease-free water was run as an NTC as well as negative serum/plasma samples. The PCR Ct values were plotted against the log₁₀ starting sample concentration using OriginPro9 (OriginLab, Northamptom, MA, USA). Ct values were compared between the recovered nucleic acids and the positive control (spiked samples) or from the paired EZ1 extraction (patient samples). Extraction efficiency was then calculated.

2.6. Carryover Volume and Bead Transfer Protocol

In a 5 mL conical tube (Eppendorf), 20 μL of bead mix was combined with 1 mL of serum, 1 mL of lysis buffer solution (LBS), and 1.25 mL of 100% ethanol, followed by 67 μL of 50 μM fluorescein. This was thoroughly mixed by vortexing, and the final fluorescein concentration was 1 μM. The protocol previously outlined was followed to run a sample through the chip. Once the run was complete, the sample and the beads were all removed from the elution well into a 0.5 mL conical tube, where the beads were separated from the supernatant using a magnet. The supernatant was used as the sample for quantification of fluorescein concentration. This was repeated three times with fluorescein in the sample, three times without fluorescein in the sample (as a measure of background) and TE (Tris EDTA) buffer (Integrated DNA Technologies Inc., Coralville, IA, USA), in lieu of serum with fluorescein. All of the samples were analyzed simultaneously with a fluorescein calibration curve in a plate reading fluorometer (PHERAstar^®, BMG Labtech, Offenburg, Germany). Using the calibration curve, the concentration of fluorescein in the elution well was determined, from which the volume of liquid carried over from the sample well could be quantified.

3. Data Analysis

Ct values from the EZ1 extraction and the chip extraction were directly compared for each paired sample. The percentage extracted was calculated using the following equation:

$$2^{(\text{EZ1 Ct}-\text{Chip Ct})}\times 100=\text{Percentage HIV recovered in chip}.$$

GraphPad Prism 5 (GraphPad Software, La Jolla, CA, USA) was used for linear regression analysis. The same equation was used for spiked serum samples compared to the positive control, adjusting for serial dilution.

4. Results

4.1. Carryover Volume and Bead Transfer

We calculated the amount of sample that was moved with the bead aggregate from the sample well into the elution well. We visually observed 100% bead transfer from the sample well into the elution well when the bead aggregate was pulled through the oil well. The calculated carryover volume for the fluorescein-spiked LBS samples was 0.31% with a standard deviation of 0.06%. There was little background found in the fluorescein-negative sample (0.012% carryover with 0.006% standard deviation), which permits the assumption that the lysis buffer, ethanol, serum, and carrier RNA created little background fluorescence. For comparison, we tested the carryover volume with 1 μM of fluorescein in TE buffer, which resulted in 0.112% of carryover volume (0.08% standard deviation).

4.2. Quantifying the Effect of Bead Types

The RNA yield for the Ambion beads and Chemagen beads was very similar. However, over time the binding capacity of the Chemagen beads declined. In a 3-month period, the percentage HIV RNA extracted dropped from 51 to 9.2% while the amount of unbound HIV RNA in the supernatant rose from 20% to almost 80%. Chemagen 1 and 2 beads had negligible percentage HIV RNA recovered (1.3% and 0%). Due to this decline of the Chemagen beads over time, the Ambion beads were incorporated into our optimized protocol.

4.3. Quantifying Effect of Elution Temperature

An analysis of variance (ANOVA) analysis was performed on the elution temperature data, with p = 0.0004, indicating the means of each temperature was not the same. As shown in Fig. 3, there is a significant difference in the percentage of RNA extracted when the elution buffer is heated for 10 min to between 30 and 70 °C when compared with ambient temperature (20 °C) elution buffer (two-tailed t test p = 0.002 for 20 vs. 70 °C). There was not a significant increase in RNA extracted between 40 and 70 °C (two-tailed t test p = 0.3669).

Fig. 3.

Elution buffer temperature versus the percentage HIV RNA recovered

4.4. Quantifying Effect of Carrier RNA

There was a significant difference in the yield with the addition of carrier RNA compared with no carrier RNA (two-tailed t test p = 0.0008 for 2 μL, p = 0.0198 for 1 μL) (Fig. 4).

Fig. 4.

Percentage HIV RNA extracted at varying carrier RNA volumes

4.5. Recovery Efficiency of HIV‑Spiked Serum Samples

Using the 1 mL protocol outlined in Sect. 2.2, five different concentrations of HIV-spiked serum samples were run through our device, and the isolated nucleic acids were analyzed with RT-PCR. From these data, we were able to report measurements of copies of RNA spiked into the sample versus copies of RNA extracted from the sample. Figure 5a shows the relative fluorescence units versus the cycle number for the serially diluted samples as a result of RT-PCR. The PureLink^™ kit had a very similar RT-PCR curve to the 540 K copy run (Fig. 5a). The calculated slope, in Fig. 5b, was – 3.177, which gives an average efficiency of 106.43% [16], indicating that there were no PCR inhibitors carried over from the sample or lysis-binding solution. Figure 5c shows the spiked HIV copy number versus the calculated extracted HIV copy number.

Fig. 5.

a Relative fluorescence units versus cycle number for polymerase chain reaction (PCR) plots for serially diluted 1 mL serum samples from 540,000 to 54 copies/mL. Additionally, a no template control and a positive control extracted using spin column kit were included in the run. b Cycle threshold versus the log copy number of HIV spiked into serum samples. c Input versus extracted HIV copy number with an R² value of 0.9995. The extracted copy number was extrapolated from the best fit line in b. Adj. adjusted, Ct cycle threshold, NTC no template control

4.6. Recovery Efficiency of Plasma Patient Samples

There was very little difference between the mean and median extraction efficiencies for the various viral load ranges: low (40–2000 copies/mL) = 68.7% (77.8%) [mean (median)], medium–low (2000–10,000 copies/mL) = 65.2% (66.9%), medium–high (10,000–100,000 copies/mL) = 58.2% (47.0%), and high (> 100,000 copies/mL) = 62.9% (68.1%), indicating that our chip’s efficiency is independent of viral load. There was no amplification observed during the 40 PCR cycles in the negative samples and all 20 of the positive samples were amplified. The comparison of extraction efficiency when testing 20 plasma samples using the microfluidic chip and the EZ-1 is shown in Fig. 6. The average percentage HIV RNA extracted in our chip was 65.4% for all 20 samples while the median percentage HIV RNA extracted in our chip was 68.7%. The average Ct on the EZ1 machine was 27.06 with a standard deviation of 3.19 while the chip average Ct value was 27.73 with a standard deviation of 3.03.

Fig. 6.

Viral load versus percentage of HIV RNA recovered using the microfluidic chip. Matched samples with the Qiagen EZ1 Advanced XL instrument were assumed at 100% extraction efficiency

5. Discussion

5.1. Results

Using our optimized chip protocol, our spiked serum samples show a linear relationship between the input HIV copy number and the extracted HIV copy number (Fig. 5c) with an R value of 0.9995. This demonstrates that we have optimized the adsorption of nucleic acids to our magnetic beads, had near 100% transport of beads through two oil–liquid interfaces, and appropriately adjusted our elution process. Additionally, these data show we are able to extract clinically relevant viral loads.

The extraction yield was observed to be near perfect (100%) in the serum spiked samples. These results validate the high efficiency of the microchip extraction method. When testing 20 plasma samples using the microfluidic chip and EZ1, our method had lower efficiencies (~ 65.4%).

All circulating nucleic acids bind to the beads, so the isolation is not targeted specifically for HIV; however, due to the large surface area provided and the bead volume used, we do not believe non-specific binding is the rate-limiting step to the extraction efficiency. The magnetic beads used in these experiments are silica coated. In the presence of salt, all nucleic acids in the sample will bind to the beads [17], including non-HIV nucleic acids that are circulating in plasma [18]. If non-specific binding was competitively interfering with HIV-specific nucleic acid binding to the beads, lower extraction efficiencies would be expected at higher viral loads.

Differences in carryover volume or the amount of sample lysate that was transferred to the elution well were observed between samples in TE buffer and serum. This can be attributed to the difference in interfacial tension at the sample/oil interface. In a previous study [19], we found interfacial tension had a direct impact on carryover volume. In our model, we found that most of the carryover liquid is within a layer enveloping the bead aggregate. The calculated carryover volume from our 1 mL sample was 0.16% ± 0.028%. This is lower than other reported carryover volumes, which range from 0.7% ± 0.3% to 41.1% ± 1.0% depending on the starting volume, oil type, and number of oil barriers [20]. Minimization of carryover volume is important. The sample lysate contains salts and other contaminants that will inhibit downstream applications of nucleic acid extraction such as PCR. Our near-perfect PCR yield implies that the observed carryover volumes are not detrimental to PCR efficiency.

We did not find a significant increase in RNA extracted in temperatures above 40 °C. This is an important caveat because sources of heat in resource-limited settings may lack the capacity for very precise and consistent temperature control. Additionally, there are reports of electricity-free heating devices used in resource-limited settings [21, 22]. Future work will incorporate the use of such devices into our protocol.

We found that heating elution buffer and nucleic acid-bound beads after they were removed from the elution well significantly increased our percentage yield of HIV RNA, which aligns with available literature [23]. In the Electronic Supplementary Material, we modeled the dissociation reaction of the nucleic acids from the magnetic beads in terms of temperature of elution buffer using the Arrhenius Equation.

We observed that the addition of carrier RNA significantly increased the yield when the sample was at a low viral load (50 copies/mL). This aligns with the literature and previously discussed hypothesis [17], supporting that at low levels of target RNA, carrier RNA can increase the yield.

5.2. Limitations

Although our optimized method has successfully simplified the workflow for a rapid extraction of HIV-1 RNA, it still requires some resources, such as low electricity for heating and additional reagents such as carrier RNA and ethanol. We believe these resources are more available than those eliminated in our method and have presented electricity-free heating sources to confirm this. The focus of this study is the optimization of RNA extraction efficiency with little to no electricity; however, amplification is still required to quantify the nucleic acid extraction.

5.3. Conclusions

There is a critical need for a device that can detect and quantify HIV viral load in resource-limited settings. In this work, we have scientifically explored a 3D-printed chip that cuts down steps such as vortexing and centrifugation while achieving high-quality results. This work also demonstrates potential for electricity-free extraction and we present a sample preparation technique that replaces the numerous end-user wash steps required by commercial nucleic acid isolation kits. We have shown proof-of-concept in the clinically relevant sample matrix of spiked human serum with concentration as low as 50 copies/mL. Additionally, we have demonstrated the performance of the microchip using plasma samples from HIV-infected patients in comparison to a current gold-standard instrument. We have also studied the importance of mild local heating of the elution sample along with quantification of carryover liquid. Our chip module can be attached to any nucleic acid amplification method and, hence, could facilitate expanded HIV viral load monitoring in a fast and efficient way.

Key Points.

A three-dimensional printed microchip was developed to extract HIV-1 RNA with only minimal electricity and no complex laboratory equipment.

We present data on HIV-1 spiked serum samples, chip protocol optimization variables, as well as HIV-1-positive patient plasma samples.