Rapid and sensitive detection of viral nucleic acids using silicon microchips

Laura Powell a, Rodrigo Sergio Wiederkehr b, Paige Damascus a, Maarten Fauvart b, Federico Buja b, Tim Stakenborg b, Stuart C. Ray a, Paolo Fiorini b and William O. Osburn *a
aLaboratory for Integrated Nanodiagnostics, Johns Hopkins University School of Medicine, Baltimore, MD, USA. E-mail: wosburn1@jhmi.edu
bDepartment of Life Sciences and Imaging, Imec, Leuven, Belgium

Received 23rd March 2018 , Accepted 30th April 2018

First published on 1st May 2018


Clinical laboratory-based nucleic acid amplification tests (NAT) play an important role in diagnosing viral infections. However, laboratory infrastructure requirements and their failure to diagnose at the point-of-need (PON) limit their clinical utility in both resource-rich and -limited clinical settings. The development of fast and sensitive PON viral NAT may overcome these limitations. The scalability of silicon microchip manufacturing combined with advances in silicon microfluidics present an opportunity for development of rapid and sensitive PON NAT on silicon microchips. In the present study, we present rapid and sensitive NAT for a number of RNA and DNA viruses on the same silicon microchip platform. We first developed sensitive (4 copies per reaction) one-step RT-qPCR and qPCR assays detecting HCV, HIV, Zika, HPV 16, and HPV 18 on a benchtop real-time PCR instrument. A silicon microchip was designed with an etched 1.3 μL meandering microreactor, integrated aluminum heaters, thermal insulation trenches and microfluidic channels; this chip was used in all on-chip experiments. Melting curve analysis confirmed precise and localized heating of the microreactor. Following minimal optimization of reaction conditions, the bench-scale assays were successfully transferred to 1.3 μL silicon microreactors with reaction times of 25 min with no reduction in sensitivity, reproducibility, or reaction efficiencies. Taken together, these results demonstrate that rapid and sensitive detection of multiple viruses on the same silicon microchip platform is feasible. Further development of this technology, coupled with silicon microchip-based nucleic acid extraction solutions, could potentially shift viral nucleic acid detection and diagnosis from centralized clinical laboratories to the PON.


1. Introduction

Detection of viral RNA or DNA by nucleic acid amplification tests (NAT) is the gold standard for diagnosis of an acute viral infection since antibodies are only detectable late in an acute infection or after viremia. Current commercially available molecular diagnostics are limited to use in clinical settings linked with a clinical laboratory, reliable electricity, and trained technicians precluding the use of these tests at the point-of-need (PON). Further, the long time-to-result of clinical laboratory-based NAT typically requires a follow-up visit with the patient to discuss the diagnosis and therapeutic options.

The need for PON viral NAT in low and middle income countries (LMIC) is well documented.1 The lack of critical laboratory infrastructure and cost precludes the use of most existing NAT platforms in these settings. Less recognized is the need for PON NAT in high-income countries (HIC). Despite an extensive clinical laboratory infrastructure in HIC, vulnerable populations still remain underserved. The limited efficiency of the hepatitis C virus (HCV) infection screening program in HIC illustrates this need. In spite of an extensive provider and patient education campaigns, only 40–50% of chronically-infected individuals in HIC are aware of their HCV infection status.2 While significantly higher than the awareness rate in LMIC (1–2%),2 the fact that a majority of individuals with chronic HCV infection in HIC are unaware of their infection status demonstrates that the current clinical lab-based screening approach is not sufficient. The availability of fast and sensitive PON viral NAT diagnostics could provide clinicians with greater diagnostic certainty at the PON, thereby increasing linkage of patients to appropriate care.

The last two decades have seen a rapid development of microsystems for performing complex clinical tests in small form-factor devices, known as Lab-on-a-Chip (LoC).3 Among LoC, microliter volume PCR microreactors occupy a prominent role. Their low thermal mass allows for fast temperature cycling, making them particularly suited for rapid PCR analysis.4 A negative aspect of miniaturization is the large surface-area-to-volume ratio of the microreactor, which sometimes adversely affects PCR performance, requiring the use of passivation layers5 or a re-optimization of benchtop protocols upon transfer to chip.5 Polymers,6–8 glass,9 and silicon10–12 have been used as microreactor materials. Although more expensive than polymers, silicon has obvious advantages when the goal is to go beyond the simple, isolated, proof-of-concept demonstration. Crucially, it enables seamless integration of other components (e.g. detectors, complex fluidic elements) using low-cost mass production processes. Reproducible amplification of multiple pathogen targets on the same silicon microchip design would increase the feasibility of using this technology for PON molecular diagnostic platforms.

In the current study, we describe the first step in developing molecular diagnostic devices that integrate a modular workflow (e.g. centrifuge-based plasma separation, column-based nucleic acid extraction, and PCR reaction in a microplate well) into single device process where plasma separation, nucleic acid extraction, and PCR amplification occurs on the microchip by demonstrating that newly developed, highly sensitive and rapid RT-qPCR and qPCR assays for detection of viral RNA and DNA targets can be transferred to silicon microchips with 1.3 μL PCR microreactors. RT-qPCR and qPCR assays for detecting HCV, human immunodeficiency virus (HIV) and Zika virus (ZIKV) RNA and human papilloma virus (HPV) 16 and HPV 18 DNA were initially developed on a benchtop real-time PCR instrument. These assays were then successfully transferred to silicon microchips while maintaining assay sensitivity down to 4 copies (cp) RNA or DNA/reaction. These on-chip assays were performed in under 25 min and the efficiency of the on-chip assay was similar to the efficiency of the assay performed on a bench scale real-time PCR instrument. Taken together, these results demonstrate the utility of pairing robust NAT with silicon microchip technology for development of a flexible molecular diagnostic platform for detection of viral nucleic acids with potential applications for diagnosis of viral infections at the PON.

2. Experimental

2.1 Oligonucleotides and standards

All primers and hydrolysis probes were designed using Primer-BLAST13 (NCBI) and publically-available sequences obtained from GenBank14 (NCBI) and synthesized by IDT technologies (Table 1). In vitro transcribed (IVT) viral RNA (Amsbio, Massachusetts) was used as a standard for each RNA target (Table 2). For the DNA viruses, linearized pHPV-16 plasmid DNA (clone 45113D, ATCC, Virginia) and pHPV-18 plasmid DNA (clone 45152D, ATCC, Virginia) were used as standards. Purified human splenocyte total RNA and human genomic DNA (Promega, Wisconsin) were used as negative controls to test assay specificity.
Table 1 Primer and probe sequences used in development of both bench scale and on-chip RNA and DNA assays
Oligo name Sequence (5′–3′)
HCV forward CCCCTGTGAGGAACTACTGT
HCV reverse GACCACTATGGCTCTCCCG
HCV probe ATTO633-AGCCATGGCGTTAGTATGAGTGTCG-IAbRQ
HIV forward CCCTACCAGCATTCTGGACAT
HIV reverse CTGGGTTCGCATTTTGGACC
HIV probe ATTO633-AGAGCCGAGCAAGCTTCACAGGAGGT-IAbRQ
ZIKV forward TGTCAATATGCTAAAACGCGGA
ZIKV reverse AAGGCTAGAATCGCCAAGAC
ZIKV probe ATTO633N-AGCCGGACTTCTGCTGGGTCAT-IAbRQ
HPV 16 forward CAGACGACTATCCAGCGACC
HPV 16 reverse GCAGTGAGGATTGGAGCACT
HPV 16 probe ATTO633-AGCCAGACACCGGAAACCCCTGCCA-IAbRQ
HPV 18 forward GTGCCAGAAACCGTTGAATC
HPV 18 reverse GAATGGCACTGGCCTCTAT
HPV 18 probe ATTO633-CGACGATTTCACAACATAGCTGGGCA-IAbRQ


Table 2 Sequences of IVT RNA used as standards for development of both bench scale and on-chip RNA assays
Virus Sequence (5′–3′)
HCV ACCTGCCCCTAATAGGGGCGACACTCCGCCATGAATCACTCCCCTGTGAGGAACTACTGTCTTCACGCAGAAAGCGCCTAGCCATGGCGTTAGTATGAGTGTCGTACAGCCTCCAGGCCCCCCCCTCCCGGGAGAGCCATAGTGGTCTGCGGAACCGGTGAGTACACCGGAATTGCCGGGAAGACTGGGTCCTTTCTTGGATAAACCCACTCTATGCCCGGCCATTTGGGCGTGCCCCCGCAAGACTGCTAGCCGAGTAGCGTTGGGTTGCGAAAGGCCTTGTGGTACTGCCTGATAGGGTGCTTGCGAGTGCCCCGGGAGGTCTCGTAGACCGTGCACCATGAGCACAAATCCTAA
HIV CCAAGGGGAAGTGACATAGCAGGAACTACTAGTACCCTTCAGGAACAAATAGGATGGATGACAAATAATCCACCTATCCCAGTAGGAGAAATCTATAAAAGATGGATAATCCTGGGATTAAATAAAATAGTAAGAATGTATAGCCCTACCAGCATTCTGGACATAAGACAAGGACCAAAGGAACCCTTTAGAGACTATGTAGACCGGTTCTATAAAACTCTAAGAGCCGAGCAAGCTTCACAGGAGGTAAAAAATTGGATGACAGAAACCTTGTTGGTCCAAAATGCGAACCCAGATTGTAAGACTATTTTAAAAGCATTGGGACCAGCAGCTACACTAGAAGAAATGATGACAGCATGTCAGGGAGTGGGAGGACCCAGCCATAAAGCAAGAGTTTTGGCDGAAGCAATG
ZIKV ATGAAAAACCCAAAAAAGAAATCCGGAGGATTCCGGATTGTCAATATGCTAAAACGCGGAGTAGCCCGTGTGAGCCCCTTTGGGGGCTTGAAGAGGCTGCCAGCCGGACTTCTGCTGGGTCATGGGCCCATCAGGATGGTCTTGGCGATTCTAGCCTTTTTGAGATTCACGGCAATCAAGCCATCACTGGGTCTCATCAATAGATGGGGTTCAGTGGGGAAAA


2.2 Bench scale assay development for quantification of RNA and DNA viruses

All bench scale one-step RT-qPCR and qPCR assays were developed using a LightCycler 480 instrument (Roche, Switzerland) and IVT RNA or plasmid DNA standard for each target in a 10 μL reaction using a 96-well PCR microplate. Amplification was performed using 10 μL reactions containing 50 mM Tris pH 8.3, 75 mM KCl, 200 μM dNTP Mix (Invitrogen, California), 200 nM forward primer, 400 nM reverse primer, 200 nM hydrolysis probe, 0.2 mg mL−1 BSA (Thermo Fisher, Massachusetts), 3 mM DTT (Thermo Fisher, Massachusetts), 1.25 units of AmpliTaq360 polymerase (Thermo Fisher, Massachusetts) and 50 units SuperScript III (Thermo Fisher, Massachusetts) in RT-qPCR reactions. The MgCl2 concentrations differed by target – HCV and HIV, 2.5 mM; ZIKV, 3 mM; HPV 16, 1.5 mM; and HPV 18, 4.5 mM MgCl2. For the RNA targets, the PCR cycling conditions were preceded by an RT step of 5 min (HIV, ZIKV) or 15 min (HCV). The PCR cycling conditions consisted of 3 min at 95 °C followed by 50 cycles of 5 s (HIV, ZIKV, HPV 16, HPV 18) or 10 s (HCV) at 95 °C and 10 s (HIV, ZIKV, HPV 16, HPV 18) or 30 s (HCV) at 60 °C (62 °C for HPV 18). The standard amounts used in the experiments ranged from 4 × 100–4 × 105 cp per reaction (standard concentration, 1 × 100–1 × 105 cp μL−1). No template controls (NTC, 10 mM Tris pH 7.5) and non-specific template controls (RNA targets, human splenocyte total RNA; DNA targets, human genomic DNA) were included in every experiment.

2.3 Bench scale data analysis

For bench scale assays, Ct values were determined using the Lightcycler 480 software. Although 50 cycles were performed for each assay, the reaction fluorescence was not monitored during the first ten cycles. Therefore, at the end of each run, 10 cycles were added to the Ct values calculated using the LightCycler 480 software. Average Ct values for each standard were plotted against the log[thin space (1/6-em)]10 standard concentrations and a linear regression line was fitted. The slope of the line was used to determine the reaction efficiency (efficiency = ((10(−1/slope)) − 1) × 100) and the back-calculated concentrations of each standard replicate were used to estimate sample reproducibility (standard deviation) of each standard dilution.

2.4 Chip fabrication and characterization

An important aspect of the chip design is the presence of thermal insulation trenches (Fig. 2A and B). These ensure that fast ramp rates can be achieved using a relatively limited power budget and that only the PCR microreactor of the silicon chip reaches high temperatures. Other chip parts, which might contain temperature-sensitive reagents to be used in later processing, remain close to room temperature. A previous incarnation of our PCR chip was heated using a thermoelectric cooler,5 in the current version an integrated heater has been implemented in order to improve miniaturization. The PCR reactor was fabricated using silicon-glass technology. Details of the fabrication have been reported previously15 and are briefly summarized here. First, the fluidic structures and the thermal insulation trenches were sculpted on the front side of the silicon by standard lithography and deep reactive ion etching. Then, a Pyrex wafer was anodically bonded to the silicon to seal the channels. It is important to note that before bonding, the silicon wafer is exposed to high temperature (600 °C) to remove fluorine-based compounds formed during the etching step, and that an oxide layer is deposited. The presence of silicon oxide on the surface reduces PCR inhibition observed in devices where the silicon surface is not passivated and is hence covered by a thin layer of native oxide. The heater, consisting of a meandering aluminum resistor, was deposited on the backside of the silicon and electrically insulated from it by a thin silicon oxide layer. Finally, a backside etch was performed to open access holes and etch the thermal insulation trenches completely through the silicon. The inlet/outlet ports were designed with a diameter of 750 μm, allowing a tight fit of standard pipette tips which creates pressure when loading and contributes to regular filling of the cavity. A resistance temperature detector (RTD) was fabricated in the same aluminum layer as the heater in order to monitor microreactor temperatures. The fabricated microreactor was mounted on a simple, custom-made PCB and contacts with the chip were wire-bonded to the PCB contacts. After mounting the chip on the PCB, the RTDs were calibrated in an oven to ensure correct measurement of the temperature during PCR. The PCB was then inserted in a holder connected to a dedicated custom-built instrument for temperature control. The holder was placed on the stage of an inverted fluorescence microscope (Olympus IX-73) equipped with a CMOS camera (Orca Flash 4.0, Hamamatsu, Japan) and fluorescent light source (X-Cite exacte, Excelitas Technologies). A script written in LabVIEW (National Instruments) was used to control temperature and acquire fluorescent images after each cycle of PCR amplification. Temperature uniformity of the microreactor was verified by melting curve analysis using dsDNA fragments with predicted melting temperatures close to the annealing temperature of the PCR oligos (Tm = 60 °C and 70 °C) (sequences available upon request). The ramp time from 60 °C to 95 °C was 2 s at a heater current of 0.1 A. The cooling time from 95 °C to 60 °C was 4 s and is fixed by the thermal design. Fluorescence images were taken at regular time intervals while increasing the temperature with a constant ramp rate. For this analysis only, the image of the microreactor was divided into 93 segments and the mean fluorescence intensity of each segment at each time point was determined using ImageJ analysis software (NIH). Melting temperature was determined using qpcR package16 in the R Studio software (RStudio, Boston, MA) and defined as the point at which the second derivative of the fluorescence intensity reached a maximum.

2.5 Transfer of bench scale assays to silicon microchip reactor

The optimized bench scale assays for both RNA and DNA targets were transferred to the 1.3 μL silicon microreactor with some modifications. All reagent concentrations were the same as the bench scale assay except the AmpliTaq360 concentration was increased to 5 units per reaction for all targets on-chip. Similar standards used to develop each bench scale assays were tested on-chip but less volume of each standard was added to the on-chip reactions resulting in the same number of copies analyzed on-chip and in the bench scale assays (range, 4 × 100–4 × 105 cp per reaction; standard concentration, 1 × 101–1 × 106 cp μL−1). Lower volumes of a higher concentration standard was used in these experiments in order to model the expected difference in sample processing in a microchip-based assay. Since the volume of the reactor is significantly lower than a microplate well, nucleic acids will have to be concentrated on-chip prior to PCR or RT-PCR analysis. Reactions were loaded into the reaction chamber and amplified according to the following cycling conditions. The HCV assay included a 5 min RT step at 55 °C, 3 min initial denaturation step at 95 °C, and 40 cycles of 5 s denaturation at 95 °C and 10 s amplification at 60 °C for a total assay time of 24.8 min. The cycling conditions for both the HIV and ZIKV assays included a 2.5 min RT step at 55 °C, 1.5 min initial denaturation step at 95 °C, and 50 cycles of 5 s denaturation at 95 °C and 10 s amplification at 60 °C for a total assay time of 25 min. The HPV 16 and HPV 18 assays included a 1.5 min initial denaturation step at 95 °C and 50 cycles of 5 s denaturation at 95 °C and 10 s amplification at either 60 °C (HPV 16) or 62 °C (HPV 18) for a total assay time of 22.5 min. Between runs, chips were cleaned by incubating the microreactor in 10% bleach at 95 °C for 5 min followed by one wash with water at 95 °C for 5 min and then two additional room temperature water washes. Three replicates of each standard were analyzed for each target. NTC replicates (10 mM Tris pH 7.5) were included in every experiment.

2.6 On-chip data analysis

During on-chip RT-qPCR and qPCR, images of the reaction chamber were captured using the inverted fluorescence microscope. The first image was captured at the start of the run, the second image was captured at the end of the RT step, and the remaining images were captured at the end of each amplification cycle. All images were then analyzed using ImageJ. The image of the reaction chamber was divided into 9 sections with each section encompassing a meander. To ensure the analysis was identical between chips, the 9-box configuration was saved, imported, and aligned for each experiment. The mean fluorescence intensity (MFI) was then calculated across all images for each of the nine meanders using the multi-measure function in the software package. The chip background MFI (post-RT for RNA; before start of run for DNA) was subtracted from all subsequent MFI values, for each section. The resultant chip-normalized MFI values were then divided by the average MFI across the first 10 cycles (assay background), for each section. The normalized fluorescence values at each cycle were averaged across the 9 meanders in order to account for local differences in fluorescence. These values were then used to fit amplification curves and determine the Ct of each standard.

2.7 On-chip Ct determination

C t values were determined for each standard replicate using the qpcR package16 in the R Studio software by calculating the 2nd derivative maximum for a 5-parameter sigmoidal fit line. Average Ct values for each standard were plotted against log[thin space (1/6-em)]10 standard concentrations and a linear regression line was determined. The slope of the line was used to determine reaction efficiency (efficiency = (10(−1/slope)) − 1) and the back-calculated concentrations of each standard replicate were used to estimate sample reproducibility (standard deviation) of each standard dilution.

3. Results and discussion

3.1 Bench scale one-step RT-qPCR and qPCR assays

To test the performance of each bench scale assay, three independent experiments were performed in which IVT RNA or plasmid DNA standard dilution series (4 × 100–4 × 105 cp per reaction) were tested on a Lightcycler480 instrument. The reaction MgCl2 concentration was optimized for each target as optimization experiments demonstrated that a single MgCl2 concentration was not optimal for all targets. Additionally, the optimal bench scale RT time for HCV was determined to be 15 min while for the other RNA targets a 5 min RT time was optimal. The two DNA targets required different amplification temperatures (60 °C, HPV 16; 62 °C, HPV 18) for optimal performance. Amplification curves are illustrated in Fig. 1. For all targets, each standard concentration was detected in all 3 independent experiments suggesting an assay sensitivity of 4 cp per reaction for each target. All NTC and non-specific amplification control replicates were negative for all assays. Standard curves were generated using mean Ct values calculated across the three independent experiments and the efficiency of each assay, derived from the slope of the linear regression line, was between 90–110% for all assays and the R2 values were all >0.99 (Table 3). To assess assay reproducibility, the Ct values of each standard replicate was used to determine a back-calculated RNA concentration from the standard curve across the three replicates. The average variability of all viral targets ranged from 0.05–0.35 log[thin space (1/6-em)]10 (Table 3). These data suggest that the newly developed bench scale assays are specific and sensitive with detection down to 4 cp per reaction.
image file: c8an00552d-f1.tif
Fig. 1 Benchtop and on-chip amplification curves for (A) RNA viral targets (HCV, HIV and ZIKV) and (B) DNA viral targets (HPV 16 and 18). RNA and DNA standards (4 × 100–4 × 105 cp per reaction) and no template controls were amplified on the LightCycler 480 and on-chip. Non-specific targets (RNA viral targets, human splenocyte total RNA; DNA viral targets, human genomic DNA) were amplified on the Lightcycler 480 only. Legend denotes the input copy number for each reaction (Lightcyler 480, 10 μL reaction volume; on-chip, 1.3 μL reaction volume). Curves denote the average normalized fluorescence values across three independent experiments.
Table 3 Performance characteristics of each assay
Target Benchtop On-chip
Slope Effa R 2 Repb Slope Effa R 2 Repb
a Efficiency, ((10(−1/slope)) − 1) × 100. b Reproducibility, average SD across three independent experiments.
HCV −3.11 109.8 0.991 0.05 −3.31 100.6 0.994 0.35
HIV −3.11 109.8 0.987 0.35 −3.43 95.7 0.998 0.41
ZIKV −3.35 98.8 0.999 0.10 −3.23 103.9 0.993 0.32
HPV 16 −3.13 108.9 0.994 0.16 −3.38 97.7 0.995 0.18
HPV 18 −3.26 100.9 0.999 0.08 −3.37 98.2 0.999 0.17


3.2 Characterization of the PCR microchip

Prior to on-chip assay development, a series of tests were performed on the microreactor chip designs in order to qualify its use in amplification experiments. The PCR microreactor was designed with a long, meandering microchannel with a resulting volume of 1.3 μL. The meandering shape was designed to ensure a uniform temperature across the reactor and to avoid trapping air bubbles during filling (Fig. 2A). The PCB-mounted microchip used in experiments is illustrated in Fig. 2B. Fluorescence generated by hydrolysis of the probe was detected by an external detector (Fig. 2C). The utility of the thermal isolation trenches was confirmed through the use of dsDNA probes. At 62 °C, temperature uniformity across the microreactor was 98.9% (0.7 °C difference between max and min Tm, Fig. 2D). Additionally, the efficiency of the thermal insulation trenches for containing microreactor heat generation was confirmed by monitoring the temperature of the microreactor and the bulk temperature of the chip surrounding the PCR microreactor. The insulating trenches performed as expected as the temperature of the bulk of the chip remained steady while the microreactor temperature was cycling between 60 and 95 °C (data not shown). These data indicate that the microchip design elements resulted in precise and uniform heating of the microreactor.
image file: c8an00552d-f2.tif
Fig. 2 Silicon microchip design and performance. (A) Microreactor layout. For clarity only front and back side etch are shown and the metal layer is omitted; (B) representative microchip mounted on PCB from backside showing integrated heaters and wirebonding; (C) representative pictures of 1.3 μL microreactor before (left) and after (right) amplification, boxes denote the segments used during quantitation of reaction fluorescence; (D) chip temperature distribution across the chip at a set temperature of 62 °C.

3.3 On-chip one-step RT-qPCR and qPCR assays

To assess the performance of the on-chip viral RNA and DNA assays, three independent experiments were performed during which each IVT RNA or plasmid DNA standard across the standard range (1 × 101–1 × 106 cp μL−1 resulting in 4 × 100–4 × 105 cp per reaction) was tested on a microchip. Less volume of the standard was used in each on-chip reaction relative to the bench scale reaction but the number of nucleic acid copies in each reaction was equivalent to the bench scale experiments. A few minor changes to the bench scale protocol were needed to ensure optimal performance of the bench scale assay on the silicon microchips. The amount of Taq polymerase was increased in order to overcome binding of the enzyme to the silicon surface. Additionally, the length of the HCV RNA RT step was shortened to 5 min. Amplification curves for all on-chip assays are illustrated in Fig. 1 and the observed shape of the on-chip amplification curves is similar to the observed shape of the bench scale curves. The lowest concentration standard (4 cp per reaction) was detected in all three independents experiments suggesting an on-chip assay sensitivity of at least 4 cp per reaction. All on-chip NTC control replicates were negative for all assays. The mean Ct values calculated across the three replicates were plotted in the same way as the bench scale assays. Reaction efficiencies ranged from 95.7–103.9% and all R2 values were >0.99 (Table 3). As a whole, the on-chip reaction efficiencies were closer to ideal (100% efficiency) than the bench scale assays indicating that the on-chip conditions resulted in more optimal amplification. Some other differences in performance between the bench scale and on-chip assays were observed. The average reproducibility, based on the variability of back-calculated concentrations, across all standard concentrations ranged from 0.17–0.41 log[thin space (1/6-em)]10 (Table 3). This range was higher than the bench scale assays suggesting that the on-chip assays were less reproducible than the bench scale assays. For example, the HCV on-chip assay was less reproducible than the bench scale assay. However, the reaction efficiency of the HCV on-chip assay was more optimal than the bench scale assay. Additionally, the standard curves of the bench scale and on-chip assays for the RNA targets were almost superimposable with minor Ct differences mainly at the low-end of the standard curve (Fig. 3). In contrast, the Ct values of all DNA standards were consistently lower in the on-chip DNA assays compared to bench scale assays. Also, the on-chip ZIKV amplification curve was noisier with a less pronounced curve plateau compared to the bench scale assay; however, the on-chip ZIKV assay reaction efficiency was as close to optimal efficiency (100%) as the bench scale assay although the on-chip amplification curves were noisier. Therefore, it is likely that these minor differences are due to random variation or increased chip-to-chip variability relative to well-to-well variability on PCR microplates. These findings suggest that the silicon microchip technology may be adapted to detect a broad range of viral pathogens without the need for significant microchip redesign or assay optimization.
image file: c8an00552d-f3.tif
Fig. 3 Standard curves of benchtop (blue) and on-chip (red) experiments for HCV, HIV and ZIKV RNA and HPV 16 and 18 DNA standards (4 × 100–4 × 105 cp per reaction). The average Ct for each standard is plotted against the number of cp of RNA or DNA in the reaction. Error bars indicate the standard deviation of Ct values between three independent experiments. Lines denote a linear regression line fit to each average dilution series.

The limitations of clinical lab-based NAT decrease the clinical utility of these assays in all resource settings. One of these limitations is the long sample-to-result time which leads to the inability to diagnose viral infections at the PON. This diagnostic uncertainty is a significant contributor to inappropriate antimicrobial use, a significant global public health concern.17 The rapid thermal cycling of the silicon microchips used in this study allowed for all of our five on-chip assays to be performed in 25 min or less. The speed of our on-chip assays make them suitable candidates for incorporation in PON NAT that would provide sufficiently low sample-to-result times which should enable clinicians to more accurately diagnose viral infections at the PON and start appropriate therapies.18

The flexibility and scalability of silicon microchips allows for integration of these rapid and sensitive RT-qPCR and qPCR assays with other solutions (on-device plasma separation19 and nucleic acid extraction) in the development of PON diagnostic devices requiring limited sample preparation (e.g. collection of fingerprick blood sample) and end-user input. The standard concentration range used in this study was based on concentrations of viral nucleic acids in body fluids during acute and chronic viral infection20–23 and anticipated sample volumes (50 μL fingerprick blood sample) for use in hypothetical diagnostic devices incorporating these ultralow volume silicon microchips. The lowest concentration standard was consistently detected in all on-chip assays and represented four copies of each RNA or DNA target in each on-chip ultralow volume reaction. Using the standard concentration range in this study and a hypothetical device containing our described microchips coupled with microfluidic plasma separation and nucleic acid extraction, one would be able to detect down to 640 cp mL−1 specimen (e.g. blood plasma, biopsy in liquid preservative) assuming a 50 μL fingerprick blood specimen (25 μL of plasma), 50% on-chip recovery of plasma, and 50% on-chip nucleic acid extraction yield. While this sensitivity would be lower than existing clinical lab-based assays that require higher volume (≥200 μL) plasma specimens,24,25 it would be sufficient to detect acute or chronic viral infections. For example, non-human primate studies have demonstrated peak ZIKV RNA levels in plasma and urine (>105 cp mL−1 plasma) during acute infection at levels significantly higher than our theoretical sensitivity.23 Additionally, this estimated sensitivity would allow for detection of >99.5% of chronic HCV infections.20 Therefore, incorporation of these ultralow volume reactions in silicon microchip-based molecular diagnostics could potentially result in diagnostic devices with sufficient sensitivity to diagnose acute and chronic viral infections at the PON.

4. Conclusions

In the current study, we investigated the feasibility of detecting viral nucleic acids using newly developed RT-qPCR and qPCR assays and transferring these assays to a faster miniaturized format using silicon microchips. Bench scale one-step RT-qPCR assays for HCV, HIV, and ZIKV RNA and qPCR assays for HPV 16 and HPV 18 DNA were successfully transferred to silicon microchips with 1.3 μL PCR microreactors and the on-chip assay sensitivity (4 cp per reaction) and reaction efficiency was similar to the bench scale assay for each target following minimal optimization of reaction conditions. Taken together, the results of our study demonstrate that rapid and sensitive amplification of viral nucleic acids on silicon microchips containing microliter volume PCR microreactors is feasible and highlights an opportunity for developing PON NAT using scalable silicon microchip technology. Coupling these assays with on-chip plasma separation and nucleic acid extraction solutions could provide a device platform for rapid and sensitive molecular diagnostic tests suitable for use at the PON in any clinical setting that could potentially have significant positive impacts on global public health.

Conflicts of interest

miDiagnostics is a private company created to develop the miLab diagnostic platform. Johns Hopkins University and imec hold a Board of Directors position in miDiagnostics, owns equity in the company and are also eligible to receive royalty payments for commercial sales by miDiagnostics. Laura Powell, Paige Damascus, Stuart Ray and William Osburn may be eligible to receive distributions of royalty payments to JHU and/or payments related to JHU equity ownership interests in miDiagnostics. This arrangement has been reviewed and approved by the Johns Hopkins University in accordance with its conflict of interest policies.

Acknowledgements

Funding for the study described in this publication was provided by miDiagnostics.

References

  1. N. P. Pai, C. Vadnais, C. Denkinger, N. Engel and M. Pai, PLoS Med., 2012, 9, e1001 Search PubMed.
  2. C. E. McGowan and M. W. Fried, Liver Int., 2012, 32(Suppl 1), 151–156 CrossRef CAS PubMed.
  3. C. H. Ahn, G. Beaucage, J. H. Nevin, J. B. Lee, A. Puntambekar and J. Y. Lee, Proc. IEEE, 2004, 92, 154–173 CrossRef CAS.
  4. C. D. Ahrberg, A. Manz and B. G. Chung, Lab Chip, 2016, 16, 3866–3884 RSC.
  5. I. Erill, S. Campoy, N. Erill, J. Barbe and J. Aguilo, Sens. Actuators, B, 2003, 96, 685–692 CrossRef CAS.
  6. A. Sposito, V. Hoang and D. L. DeVoe, Lab Chip, 2016, 16, 3524–3531 RSC.
  7. H. O. Song, J. H. Kim, H. S. Ryu, D. H. Lee, S. J. Kim, D. J. Kim, I. B. Suh, Y. Choi du, K. H. In, S. W. Kim and H. Park, PLoS One, 2012, 7, e53325 CAS.
  8. M. Q. Bu, I. R. Perch-Nielsen, K. S. Sorensen, J. Skov, Y. Sun, D. D. Bang, M. E. Pedersen, M. F. Hansen and A. Wolff, J. Micromech. Microeng., 2013, 23, 074002 CrossRef.
  9. L. A. Legendre, J. M. Bienvenue, M. G. Roper, J. P. Ferrance and J. P. Landers, Anal. Chem., 2006, 78, 1444–1451 CrossRef CAS PubMed.
  10. L. Luo, W. Hou and J. Yu, Anal. Lett., 2010, 43, 12–21 CrossRef CAS.
  11. Y. K. Cho, J. Kim, Y. Lee, Y. A. Kim, K. Namkoong, H. Lim, K. W. Oh, S. Kim, J. Han, C. Park, Y. E. Pak, C. S. Ki, J. R. Choi, H. K. Myeong and C. Ko, Biosens. Bioelectron., 2006, 21, 2161–2169 CrossRef CAS PubMed.
  12. C. Consolandi, M. Severgnini, A. Frosini, G. Caramenti, M. De Fazio, F. Ferrara, A. Zocco, A. Fischetti, M. Palmieri and G. De Bellis, Anal. Biochem., 2006, 353, 191–197 CrossRef CAS PubMed.
  13. J. Ye, G. Coulouris, I. Zaretskaya, I. Cutcutache, S. Rozen and T. L. Madden, BMC Bioinf., 2012, 13, 134 CrossRef CAS PubMed.
  14. C. Burks, J. W. Fickett, W. B. Goad, M. Kanehisa, F. I. Lewitter, W. P. Rindone, C. D. Swindell, C. S. Tung and H. S. Bilofsky, Comput. Appl. Biosci., 1985, 1, 225–233 CrossRef CAS.
  15. B. Majeed, B. Jones, D. S. Tezcan, N. Tutunjyan, L. Haspeslagh, S. Peeters, P. Fiorini, M.O. de Beeck, C. Van Hoof, M. Hiraoka and H. Tanaka, Jpn. J. Appl. Phys., 2012, 51, 04DL51 CrossRef.
  16. C. Ritz and A. N. Spiess, Bioinformatics, 2008, 24, 1549–1551 CrossRef CAS PubMed.
  17. World Health Organization, Antimicrobial resistance: global report on surveillance, 2014 Search PubMed.
  18. A. M. Caliendo, D. N. Gilbert, C. C. Ginocchio, K. E. Hanson, L. May, T. C. Quinn, F. C. Tenover, D. Alland, A. J. Blaschke, R. A. Bonomo, K. C. Carroll, M. J. Ferraro, L. R. Hirschhorn, W. P. Joseph, T. Karchmer, A. T. MacIntyre, L. B. Reller, A. F. Jackson and A. Infectious Diseases Society of, Clin. Infect. Dis, 2013, 57(Suppl 3), S139–S170 CrossRef PubMed.
  19. A. Nabatiyan, Z. A. Parpia, R. Elghanian and D. M. Kelso, J. Virol. Methods, 2011, 173, 37–42 CrossRef CAS PubMed.
  20. J. R. Ticehurst, F. M. Hamzeh and D. L. Thomas, J. Clin. Microbiol., 2007, 45, 2426–2433 CrossRef CAS PubMed.
  21. C. Fraser, T. D. Hollingsworth, R. Chapman, F. de Wolf and W. P. Hanage, Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 17441–17446 CrossRef CAS PubMed.
  22. L. Del Rio-Ospina, S. C. Soto-De Leon, M. Camargo, D. A. Moreno-Perez, R. Sanchez, A. Perez-Prados, M. E. Patarroyo and M. A. Patarroyo, BMC Cancer, 2015, 15, 100–110 CrossRef PubMed.
  23. C. E. Osuna, S. Y. Lim, C. Deleage, B. D. Griffin, D. Stein, L. T. Schroeder, R. Omange, K. Best, M. Luo, P. T. Hraber, H. Andersen-Elyard, E. F. Ojeda, S. Huang, D. L. Vanlandingham, S. Higgs, A. S. Perelson, J. D. Estes, D. Safronetz, M. G. Lewis and J. B. Whitney, Nat. Med., 2016, 22, 1448–1455 CrossRef CAS PubMed.
  24. R. R. Al Olaby and H. M. Azzazy, Expert Rev. Mol. Diagn., 2011, 11, 53–64 CrossRef CAS PubMed.
  25. B. R. Cobb, J. E. Vaks, T. Do and R. A. Vilchez, J. Clin. Virol., 2011, 52(Suppl 1), S77–S82 CrossRef PubMed.

Footnote

These authors contributed equally to this work.

This journal is © The Royal Society of Chemistry 2018