Hydrogel-Based Colorimetric Assay for Multiplexed MicroRNA Detection in a Microfluidic Device

Hydrogel-Based Colorimetric Assay for Multiplexed MicroRNA Detection in a Microfluidic Device This is an open access article published under a Creative Commons Attribution (CC-BY) License, which permits unrestricted use, distribution and reproduction in any medium, provided the author and source are cited. pubs.acs.org/ac Article Hydrogel-Based Colorimetric Assay for Multiplexed MicroRNA Detection in a Microfluidic Device Hyewon Lee, Jiseok Lee, Seung-Goo Lee,* and Patrick S. Doyle* Cite This: Anal. Chem. 2020, 92, 5750−5755 Read Online Metrics & More Article Recommendations sı Supporting Information ACCESS * ABSTRACT: Although microRNA (miRNA) expression levels provide important information regarding disease states owing to their unique dysregulation patterns in tissues, translation of miRNA diagnostics into point-of-care (POC) settings has been limited by practical challenges. Here, we developed a hydrogel-based microfluidic platform for colorimetric profiling of miRNAs, without the use of complex external equipment for fluidics and imaging. For sensitive and reliable measurement without the risk of sequence bias, we employed a gold deposition-based signal amplification scheme and dark-field imaging, and seamlessly integrated a previously developed miRNA assay scheme into this platform. The assay demonstrated a limit of detection of 260 fM, along with multiplexing of small panels of miRNAs in healthy and cancer samples. We anticipate this versatile platform to facilitate a broad range of POC profiling of miRNAs in cancer- associated dysregulation with high-confidence by exploiting the unique features of hydrogel substrate in an on-chip format and colorimetric analysis. mall noncoding RNAs, called microRNAs (miRNAs), have systems for nucleic acid hybridization, especially from S become increasingly important in disease diagnosis due to biologically complex samples. Better thermodynamics in their abnormal expression in many diseases such as cancer, the gel increases both sensitivity and specificity, and the 1−4 diabetes, neural, and heart diseases. Previous studies had solution-like environment of a hydrogel provides faster demonstrated that miRNAs have high stability and can be hybridization kinetics. Furthermore, it is possible to directly detected in various biological fluids such as blood serum and measure miRNAs in complex media such as cell lysate or 12,23 plasma. However, clinicians still lack proper tools for high- serum, without RNA extraction and target amplification. In confidence quantification of miRNA owing to their low a previous gel particle-based approach, a universal labeling 6,7 abundance and sequence homology. Most challenges arise scheme was developed, which allowed the use of a single label due to the grueling demands of an assay that could easily be for all captured miRNAs, and measured miRNA dysregulation 24,25 integrated into a point-of-care (POC) clinical setting. An ideal patterns. Here, we adapted the same hydrogel chemistry platform ought to provide high sensitivity, high specificity, and and miRNA capture approach into an on-chip assay for POC multiplexing while minimizing the use of external equipment, applications. and involving simple sample preparation and assay operation. POC diagnostics can facilitate fast and accurate identi- While quantitative real-time polymerase chain reaction (qRT- fication of diseases, which leads to better treatment of patients. PCR) is widely used as a gold standard for high sensitivity, it is To fit the POC criteria of cost-effectiveness, portability, and limited in practical application, due to its requirement of accessibility, miRNA assays are commonly performed with substantial sample preparation steps, such as RNA extraction, nonfluorescent labeling molecules for visualization. Since such expensive instruments, and intricately complex design of assay schemes offer only low to moderate detection sensitivity, primers to avoid sequence bias arising from target-based miRNA detection methods often rely on signal boosting step, 9,10 amplification. To address these limitations of the conven- 26−28 such as enzymatic and nucleic acid amplification. The tional method, several techniques have now been developed Mirkin group has developed an assay using gold nanoparticles, 11,12 for miRNA detection, fluorescent probes, isotachopho- with silver or gold deposition, which had been successfully 13,14 15,16 resis, and nanomaterial-based in vivo sensing. Despite advancements, the techniques do not satisfy all the require- Received: November 5, 2019 ments of POC regarding miRNAs, and the need for a more Accepted: March 24, 2020 clinically feasible approach still remains. Published: March 24, 2020 Hydrogel-based microfluidic approach can be advantageous in the development of reliable and affordable POC 18−20 diagnostics. The superiority of nonfouling hydrogels has been previously demonstrated in comparison to surface-based https://dx.doi.org/10.1021/acs.analchem.9b05043 © 2020 American Chemical Society Anal. Chem. 2020, 92, 5750−5755 5750 Analytical Chemistry pubs.acs.org/ac Article 29−31 24 applied to detect various biomarkers, including miRNAs. and let-7a: 50 μM). Probe-containing prepolymer solution Their method provides robust signals without bleaching issues was vortexed, centrifuged, and loaded into channels by because of high photostability of gold-labeled conjugates, and pipetting. An inverted microscope (Zeiss Axio Observer A1) very short amplification step with high sensitivity while and a CCD camera (Andor Clara) were used for UV-initiated minimizing equipment requirements. However, this micro- polymerization. With a desired photomask (Fineline Imaging) array-based method typically requires a long assay time placed in the fieldstop, the polymerization was performed at (overnight hybridization of target miRNAs) with complicated 100 ms exposure (Lumen 200, Prior Scientific) using a 20× fluidic steps. Thus, we aim to integrate on-chip hydrogel assay objective (Zeiss Plan-Neofluar) and a dichrioc filter for and gold deposition scheme to achieve sensitive POC excitation at 365 nm. After polymerization, channels were applications. rinsed with 1X TE buffer. For multiplexed assays, the We leverage the advantages of microfluidic channels, and the subsequent monomer solution with a different probe was chemical advantages of a polyethylene glycol (PEG) hydrogel loaded in the same channel and polymerization was repeated as scaffold for miRNA hybridization while optimizing an enzyme- described in earlier. The posts containing different probe free gold nanoparticle-based signal amplification scheme for sequences were immobilized at spatially distinct locations in a POC diagnostics. Particularly, we sought to use dark-field single channel. imaging for high sensitivity, which had previously been Fluidic Control. All assays were performed with gravity demonstrated for ultrasensitive colorimetric nucleic acid driven flow by appropriately inserting a 200 μL pipet tip into assay. Dark-field imaging-based methods boost the scattered PDMS inlet port. The flow rate was observed to be 1−5 μL/ intensity of gold conjugates with white light illumination, thus ́ min. Previously, we calculated the channel Peclet number (Pe) minimizing the requirement of complex and expensive to be approximately 7000. With the high values of Pe (Pe ≫ instrumentation for miRNA detection. With just a 15 min 1), target depletion is assumed to be negligible. The solution signal amplification step, we achieved a limit of detection of was refilled in the tip of inlet port every 10 min to maintain 260 fM and multiplexed miRNA quantification with minimal height differences for a steady flow. Since our system is highly sample input. We expect this assay platform to be beneficial in flexible and can be incorporated with various types of flows a wide range of clinical samples, including cellular lysate and such as pressure-driven flows and syringe pump-driven flows, serum for POC applications. the manual iterative operation can be avoided by simply using an external equipment for fluidics. EXPERIMENTAL SECTION MicroRNA Assay. Blocking solution of 3% (w/v) Pluronic F108 (Sigma-Aldrich) in nuclease-free water (Affymetrix) was Microfluidic Chip Preparation. Commercial chips with filled into channels containing hydrogel posts. After 30 min, straight channels (50 μm in height, 1 mm width, and 18 mm the process of target hybridization to specific probes was length) were purchased from Hilgenberg GmbH, Germany for performed with the mixture of synthetic RNAs (IDT) or total performing all assays. Connection ports were fabricated in RNA (BioChain, Newark, CA) in a TET buffer with a final polydimethylsiloxane (PDMS) (Corning, Sylgard 184) with concentration of 350 mM NaCl for 90 min. The sequence holes, which were punched with 15-gauge needles. All information on all the probes and targets used in this study was connection ports were bonded onto inlets and outlets of summarized in Supporting Information Table S1. Total RNA glass chips by oxygen plasma treatment (25 s on RF = high, was stored at 100 ng/μLat −20 °C and 500 ng of total RNA Harrick Scientific, Pleasantville, NY). After a subsequent dissolved in 200 μL was used for assay. Before performing the incubation at 80 °C for 20 min, a 2% (v/v) solution of 3- assay, the solution with total RNA was incubated at 95 °C for 5 (trimethoxysilyl)propyl acrylate (Sigma) mixed in 24.5% (v/v) min in a thermoshaker for disrupting secondary structures, and 1X PBS (phosphate buffered saline, Corning), and 73.5% (v/v) cooled down at room temperature. The hybridization mixture ethanol was filled inside the channels for tight adhesion of the was loaded into the microchannel through the precut pipet tips gel pads on the glass walls. After 30 min, channels were rinsed in the injection port of the device. Target hybridization with ethanol and dried with argon gas. After bioassays, chips occurred at elevated temperature (55 °C) on a hot plate, and were cleaned by soaking in 1 M NaOH for 1 h and then all subsequent steps were performed at room temperature. In washed with DI water and ethanol. Next, they were dried with previous works, the temperature in channels was validated to argon gas; chips were stored at 80 °C until the time of usage. be constant based on a simple heat transfer calculation. For These reusable glass chips were used for several assays (more all assays, steady flows were maintained to deliver the than 10 times) by repeating the cleaning and activating molecules without depletion. In between assay steps, rinses procedures. were performed using 50 mM NaCl in TET for 30 s to impose On-Chip Hydrogel Synthesis. All chemicals were sufficient stringency for high specific miRNA measurement. purchased from Sigma-Aldrich (U.S.). Hydrogel posts were For ligation, the universal linker (IDT), T4 DNA ligase (800 synthesized using projection lithography from a polyethylene U/mL), ATP (250 nM), and 10× NEB2 buffer (New England glycol monomer mixture. The monomer mixture consisted of Biolabs, Ipswich, MA) were mixed in TET and incubated in 20% (v/v) PEGDA 700 (poly(ethylene glycol) diacrylate, MW the channels for 10 min as described in previous studies. = 700 g/mol), 40% (v/v) PEG 600 (poly(ethylene glycol), After washing, Nanogold-Streptavidin (80 μg/mL, Nanop- MW = 600 g/mol), 5% (v/v) Darocur 1173 photoinitiator, and 35% (v/v) 3x TE (Tris-EDTA, USB Corporation) buffer robes, Inc., Yaphank, NY) diluted in 1× PBS (0.1% v/v) was with food coloring dye, which was previously optimized for the loaded into channels for 30 min. For gold deposition, 35,36 diffusion and reaction in bioassays. The monomer solution GoldEnhance (Nanoprobes, Inc.) was used, which consisted was diluted 9:1 with the acrydite-modified probes (Integrated of Solution A (enhancer), Solution B (activator), Solution C DNA Technologies (IDT), Coralville, IA). The concentration (initiator), and Solution D (buffer). First, 200 μL of Solution A of probe molecules was adjusted from the coarse rate matching and B were mixed, and after 5 min, 200 μL of Solution C and as described in a previous study (miR-21:247 μM, miR-145, D were added into the mixture. After vortexing, the mixture 5751 https://dx.doi.org/10.1021/acs.analchem.9b05043 Anal. Chem. 2020, 92, 5750−5755 Analytical Chemistry pubs.acs.org/ac Article Figure 1. (A) Schematics of on-chip hydrogel post synthesis for multiplexing of small miRNA panels using projection lithography with the spatial encoding scheme. (B) miRNA assay scheme: target hybridization, universal linker ligation, gold nanoparticle labeling, and gold ion deposition- based signal amplification. (C) Dark-field images of posts after complete miRNA assay demonstrating the dose-dependent response of miRNAs. Scale bar represents 100 μm. was delivered to the microchannels with posts for an optimal RESULTS AND DISCUSSION amplification time of 15 min. Assay Development. For the development of a sensitive Data Acquisition and Analysis. The hydrogel posts were miRNA assay in point-of-care (POC) diagnostics, we sought to imaged using a Zeiss Axio Observer A1 microscope with a 10× apply a rapid signal amplification scheme by gold deposition objective and a dark-field condenser. Each image was analyzed onto a hydrogel-based on-chip miRNA sensor. Previously, we had demonstrated ultrasensitive measurement of miRNAs using ImageJ (National Institutes of Health). The color using oil-isolated hydrogel chambers in microfluidic chip, with images we obtained were split to the respective red, green, and a limit of detection (LOD) of 22.6 fM. We had achieved high blue image components. The region of interest (ROI) was Peclet ́ number in the previous work by using a gravity-driven identified as a post circle with a defined diameter. For each flow to eliminate mass transfer limits that could arise from ROI, the intensity of pixels in the red channel was only insufficient target molecule deliveries to the gel posts. averaged to analyze the wine-red color of miRNA-gold Importantly, this strategy eliminated the need for complicated conjugates. Background signal was calculated by averaging external flow controllers, making our system suitable for point of care applications. However, this assay was based on the red pixel intensities in a region outside the post, within the fluorescence measurements, which required expensive and microfluidic chip. Target signal was computed by subtracting complex imaging instruments. To overcome this issue, we used background signal from a raw signal with target incubation a colorimetric labeling approach based on gold deposition for (S = S − S ). Control raw signal was target target, raw background signal amplification and dark-field imaging. obtained from the posts after incubation with 0 fM of synthetic As done previously, we immobilized spatially encoded targets. After subtracting background signal (S = S control control, raw miRNA probe-bearing hydrogel posts in microfluidic channels − S ), the control signal serves as the assay background, using projection lithography to control their size and shape background and the noise was defined as the standard deviation of the (Figure 1a). Multiplexing can be achieved with a spatial encoding scheme by exchanging monomer solutions in the control signals (σ ). Control signal was subtracted from control device and polymerizing posts. The incorporated probes the desired target signal (S = S − S ) to obtain a target, net target control consist of two domains: a miRNA-specific domain and a net signal. This makes S = 0. Finally, the signal-to-noise control, net universal linker domain. The linker used for labeling was the ratio (SNR) was determined as the ratio of net signal to assay same for all measured miRNAs. A fully controlled steady noise (SNR = S /σ ). Limit of detection of the target, net control gravity-driven flow was used for all assay steps, thus eliminating system was defined as the miRNA concentration at which the the need for expensive and cumbersome flow controllers. SNR equals 3, as per previously published protocols. Target hybridization occurred at elevated temperatures (55 5752 https://dx.doi.org/10.1021/acs.analchem.9b05043 Anal. Chem. 2020, 92, 5750−5755 Analytical Chemistry pubs.acs.org/ac Article °C) on a hot plate, whereas all subsequent steps were performed at room temperature (Figure 1b). We ligated the biotinylated universal linker to the probe-target complex using the T4 DNA ligase for 10 min. Without target miRNA bound to the probe, the universal linker would be released during the stringent washing step with low salt (50 mM NaCl) buffer due to low binding affinity to probe cross-linked into the hydrogel. Next, the probe-target complex was labeled using a streptavidin-conjugated gold nanoparticle (SA-Au, 80 ng/mL, 30 min). Thereafter, the signal amplification scheme, based on gold ion deposition, was applied for 15 min. Figure 2. Optimization of amplification time. The signal-to-noise Both bright- and dark-field systems were considered for ratio (SNR) was continued to increase for up to 15 min of gold deposition-based signal amplification. After achieving a maximum imaging the gold-labeled target miRNAs. For system character- value at 15 min, SNR decreased and then saturated due to nonspecific ization, we used immobilized gel posts functionalized with binding of gold ions. Error bars represent standard deviation (n =8− biotinylated DNA probe (5 nM). Although bright-field mode is 15). widely used in colorimetric assay, its high background signals often generate large variations. More importantly, there was no difference between the presence and absence of probe after conjugates over 24 h, after Nanogold-Streptavidin labeling and labeling streptavidin-conjugated gold nanoparticles and signal gold ion deposition steps. As described earlier, the net signal amplification with catalytic gold deposition (Supporting was determined as control (no biotinylated probe)-subtracted Information (SI) Figure S1). Conversely, with dark-field target signal. As expected, the net signal did not decrease, as illumination we observed measurable signals from 5 nM shown in SI Figure S2. Importantly, it was noted that hydrogels probe. Thus, for sensitive measurement of surface plasmon and gold nanoparticles are stable for several months without 21,39 resonance (SPR) scattering from gold nanoparticles, we used degradation, which is expected to allow long-term storage dark-field microscopy with the appearance of wine-red color. and reimaging. Therefore, this assay platform would provide As shown in Figure 1c, in an initial trial with miRNA assay with high stability in analysis, without photobleaching, and can be the analysis using dark-field microscopy, the red dots were stored for a long time or be transported, if necessary, rendering gradually increased as the concentration of target miRNA it ideal for POC diagnostics. increased from 0 to 10 pM. Considering simplicity of the assay, Detection Sensitivity. Using the gold labeling scheme, we further analysis was performed with red channel in RGB examined the signal changes in 10 pM spike-in of let-7a as a imaging. function of hybridization time. The target capture increased for Optimization and Assessment of Signal Amplifica- 90 min and then saturated, as observed in previous studies (SI 24,25 tion. Since previous studies had used the gold deposition Figure S3). To minimize the RNA input requirement while scheme in a microarray format, we first needed to optimize the retaining high sensitivity, we decided to use 90 min of signal amplification step in the on-chip hydrogel-based assay. hybridization. The 90 min hybridization was recommended While previous studies used three rounds of 5 min gold due to the stringent buffer conditions (high temperature and deposition to maximize the signal-to-noise ratio (SNR) by low salt concentration), which we optimized earlier for specific increasing target-binding signal, we sought to run a single- miRNA measurements to distinguish between even 1−2 24,25 step amplification to minimize the assay complexity for POC mismatches. If necessary, we could reduce the assay applications. We hypothesized that the solution-like and time, which would still give us a reasonable signal. nonfouling environments of hydrogel posts would be superior We next investigated the sensitivity of our hydrogel-based to a microarray platform (rigid, planar surfaces) for gold colorimetric detection scheme for detecting miRNAs. The deposition on targets without nonspecific background. For the synthetic miRNA spike-ins (from 100 fM to 10 pM let-7a) optimization of the signal amplification step, we measured the were incubated in each channel of a microfluidic chip. To SNR as a function of gold ion deposition time. Using account for the assay background, the net signal was immobilized gel posts functionalized with either biotinylated considered−the control signal (0 fmol spike-in) was subtracted DNA (final concentration of 5 nM) or no biotin (serving as from thetargetsignals.Weshowthe dose-dependent control), we loaded Nanogold-Streptavidin (80 ng/mL, PBS) responses from let-7a spike-ins in Figure 3. As shown in SI for 30 min and, after washing, GoldEnhance mixture for 5−60 Figure S4, we then calculated the limit of detection (LOD), min. Then, we calculated SNR, which is defined as the net which was previously defined in Experimental Section. Our control-subtracted signal divided by the standard deviation of assay scheme provided a LOD of 260 fM, which is relatively control measurements (assay-derived noise). As shown in sensitive compared to other colorimetric assays (pM to sub- 40−42 Figure 2, SNR increased up to a time of 15 min, and then pM). Although some recent studies on colorimetric 10,43,44 decreased due to high background signal. This indicated that a miRNA assays show high sensitivity (fM to sub-fM), hydrogel-based system can achieve high sensitivity, simply with their methods rely on target-based amplification with the a single-step amplification without the need of multiple rounds. possible high risk of sequence bias. In addition, we might Next, we examined the assay stability, which is one of the achieve better sensitivity if we run multiple rounds of signal important considerations for reliable measurement in POC amplification as a previous study optimized to perform the diagnostics. An unstable signaling label would induce test three rounds of gold deposition to maximize SNR. errors. Also, sometimes assay platforms might need to be Additionally, the size and geometry of hydrogel posts can be transferred from remote areas to a core facility for analysis by optimized to improve the sensitivity of miRNA measurements. experts for accurate diagnostics. With 5 nM of the biotinylated In a miRNA assay with the high concentration of targets such probe, we analyzed the net signal from the gold-labeled as 10 pM, we observed that more targets were attached to the 5753 https://dx.doi.org/10.1021/acs.analchem.9b05043 Anal. Chem. 2020, 92, 5750−5755 Analytical Chemistry pubs.acs.org/ac Article Figure 3. Assay sensitivity. The colorimetric miRNA measurements in microfluidics provided limit of detection (LOD) of 260 fM without expensive and complex instrument (see SI Figure S4 for SNR calculation). Error bars represent standard deviation (n =5−14). edge of the post, rather than evenly distributed (Figure 1c). Previously, we optimized the pore size of gel scaffold to allow the fast diffusion of molecules. However, the rapid deposition of gold ions seems to affect the transport rate of labeling Figure 4. Multiplexed measurements of three miRNA targets from molecules, especially with the high level of probe-target total RNA derived from tissue. (A) The signal-to-noise ratio (SNR) of conjugates. To consider flux into porous hydrogel posts, we three miRNAs in tumor and healthy samples were plotted, and each could use ring structures instead of disk shapes as in a recent of the tumor and healthy samples was measured 10 times. Error bars publication. By analyzing a ring-area around the edges, we represent the standard deviation of targets normalized by assay noise (n = 10). (B) Dysregulation ratios of three miRNA targets in lung expect that we could achieve higher mean signal and better tumor versus healthy tissue were as expected from the previous sensitivity, which would be beneficial in POC diagnostics. studies, validating our assay multiplexing scheme. Error bars represent Multiplex Detection. After characterization, we measured the standard deviation of miRNA expression measurements in tumor the cross-reactivity of three microRNAs. Three clinical miRNA normalized by background-subtracted average miRNA signal in targets relevant in lung tumor were considered in this study: normal and by the ratio of tumor to normal miRNA expression. let-7a, miR-145, and miR-21. We immobilized three types of hydrogel posts, bearing each miRNA probe using the spatial encoding scheme. As shown in SI Figure S5, there was no CONCLUSIONS significant interference among three different miRNA targets Here, we present a technique for miRNA quantification in (∼20% of cross-reactivity). This minimal cross-reactivity of our point-of-care (POC) diagnostics by using gold deposition- hydrogel-based miRNA assay scheme enabled multiplexing based signal amplification scheme with the on-chip hydrogel analysis of small panels of miRNAs. sensor platform. Unlike previous studies, the colorimetric assay For multiplexing, each miRNA probe concentration was developedheredemonstratedhighlystableand reliable adjusted by coarse-rate matching (Experimental Section). All miRNA measurement without the need of expensive instru- three miRNAs were expected to follow the same rate under the ments. This assay provided a limit of detection of 260 fM, same hybridization condition (salt and temperature). To verify which is relatively sensitive compared to other colorimetric this hypothesis for the hydrogel-based colorimetric assay, we assays. Moreover, our system enabled multiplex analysis of measured the detection limit for each microRNA target based small panels of miRNAs with relatively simple assay steps. We on the calibration curve (SI Figure S4). As expected, all three successfully analyzed the dysregulation of miRNAs in lung microRNA targets showed similar LOD (SI Table S2). tumor with respect to that in healthy tissues. In the future, it As a proof of concept, we compared the miRNA expression might be possible to integrate smartphone-based imaging in tumor and healthy tissue from total RNA samples using the system for immediate processing and analysis, with the help of colorimetric platform developed here. With the posts bearing smart algorithms. We envision that our system to have wide- three miRNA probes (let-7a, miR-145, and miR-21), we ranging applications in POC clinical settings for various targets performed the multiplexed assay with total RNA samples, and such as miRNAs, RNAs, DNA, and proteins. we observed the dysregulation patterns of the three miRNA targets in healthy and tumor tissue (Figure 4). The patterns are ASSOCIATED CONTENT consistent with prior studies the literature, since the expression sı * Supporting Information of both let-7a and mir-145 are known to be decreased in lung The Supporting Information is available free of charge at 24,25,45,46 tumor tissue, whereas that of miR-21 is elevated. Also, https://pubs.acs.org/doi/10.1021/acs.analchem.9b05043. by using the same total RNA sample, signal from this A list of reagents, additional supporting tables, and colorimetric assay were comparable to those in previous figures are available as noted in the text (PDF) studies using fluorescent labels (phycoerythrin-conjugated streptavidin reporter, SA-PE) (SI Figure S6), which were AUTHOR INFORMATION previously validated with qRT-PCR. This consistency with prior work demonstrates the high performance and reprodu- Corresponding Authors cibility of our new miRNA measurement without using Patrick S. Doyle − Department of Chemical Engineering, expensive fluorescence detection. Massachusetts Institute of Technology, Cambridge, 5754 https://dx.doi.org/10.1021/acs.analchem.9b05043 Anal. Chem. 2020, 92, 5750−5755 Analytical Chemistry pubs.acs.org/ac Article (16) Ryoo, S. R.; Lee, J.; Yeo, J.; Na, H. K.; Kim, Y. K.; Jang, H.; Lee, Massachusetts 02139, The United States; orcid.org/0000- J. H.; Han, S. W.; Lee, Y.; Kim, V. N.; Min, D. H. ACS Nano 2013, 7, 0003-2147-9172; Email: pdoyle@mit.edu 5882−5891. Seung-Goo Lee − Synthetic Biology and Bioengineering (17) Cheng, Y.; Dong, L.; Zhang, J.; Zhao, Y.; Li, Z. Analyst 2018, Research Center, Korea Research Institute of Bioscience and 143, 1758−1774. Biotechnology, Daejeon 34141, Republic of Korea; Department (18) Yetisen, A. K.; Butt, H.; Volpatti, L. R.; Pavlichenko, I.; Humar, of Biosystems and Bioengineering, KRIBB School of M.; Kwok, S. J.; Koo, H.; Kim, K. S.; Naydenova, I.; Khademhosseini, Biotechnology, University of Science and Technology, Daejeon A.; Hahn, S. K.; Yun, S. H. Biotechnol. Adv. 2016, 34, 250−271. 34113, Republic of Korea; orcid.org/0000-0003-4539- (19) Wei, X.; Tian, T.; Jia, S.; Zhu, Z.; Ma, Y.; Sun, J.; Lin, Z.; Yang, 9812; Email: sglee@kribb.re.kr C. J. Anal. Chem. 2016, 88, 2345−2352. (20) Shapiro, S. J.; Dendukuri, D.; Doyle, P. S. Anal. Chem. 2018, 90, Authors 13572−13579. Hyewon Lee − Synthetic Biology and Bioengineering Research (21) Le Goff, G. C.; Srinivas, R. L.; Hill, W. A.; Doyle, P. S. Eur. Center, Korea Research Institute of Bioscience and Polym. J. 2015, 72, 386−412. Biotechnology, Daejeon 34141, Republic of Korea (22) Pregibon, D. C.; Doyle, P. S. Anal. Chem. 2009, 81, 4873− Jiseok Lee − School of Energy and Chemical Engineering, Ulsan (23) Lee, H.; Shapiro, S. J.; Chapin, S. C.; Doyle, P. S. Anal. Chem. National Institute of Science and Technology, Ulsan 44919, 2016, 88, 3075−3081. Republic of Korea (24) Chapin, S. C.; Appleyard, D. C.; Pregibon, D. C.; Doyle, P. S. Complete contact information is available at: Angew. Chem., Int. Ed. 2011, 50, 2289−2293. https://pubs.acs.org/10.1021/acs.analchem.9b05043 (25) Lee, H.; Srinivas, R. L.; Gupta, A.; Doyle, P. S. Angew. Chem., Int. Ed. 2015, 54, 2477−2481. Notes (26) Cheng, C. M.; Martinez, A. W.; Gong, J.; Mace, C. R.; Phillips, S. T.; Carrilho, E.; Mirica, K. A.; Whitesides, G. M. Angew. Chem., Int. The authors declare no competing financial interest. Ed. 2010, 49, 4771−4774. (27) Sun, Y.; Shi, L.; Wang, Q.; Mi, L.; Li, T. Anal. Chem. 2019, 91, ACKNOWLEDGMENTS 3652−3658. We acknowledge the support from NIH-NIBIB Grant (28) Kim, D.; Wei, Q.; Kim, D. H.; Tseng, D.; Zhang, J.; Pan, E.; 5R21EB024101-02, the C1 Gas Refinery Program funded by Garner, O.; Ozcan, A.; Di Carlo, D. Anal. Chem. 2018, 90, 690−695. the Ministry of Science and ICT (2018M3D3A1A01055732), (29) Elghanian, R.; Storhoff, J. J.; Mucic, R. C.; Letsinger, R. L.; National Research Foundation of Korea (NRF) grant funded Mirkin, C. A. Science 1997, 277, 1078−1081. by the Korea government (MSIT) (2018R1C1B6004086), and (30) Kim, D.; Daniel, W. L.; Mirkin, C. A. Anal. Chem. 2009, 81, 9183−9187. Korea Research Institute of Bioscience and Biotechnology (31) Alhasan, A. H.; Kim, D. Y.; Daniel, W. L.; Watson, E.; Meeks, J. Research Initiative Program. J.; Thaxton, C. S.; Mirkin, C. A. Anal. Chem. 2012, 84, 4153−4160. (32) Tang, L.; Li, J. ACS Sens 2017, 2, 857−875. REFERENCES (33) Li, J.; Kong, C.; Liu, Q.; Chen, Z. Analyst 2018, 143, 4051− (1) Lu, J.; Getz, G.; Miska, E. A.; Alvarez-Saavedra, E.; Lamb, J.; Peck, D.; Sweet-Cordero, A.; Ebert, B. L.; Mak, R. H.; Ferrando, A. (34) Hwu, S.; Blickenstorfer, Y.; Tiefenauer, R. F.; Gonnelli, C.; A.; Downing, J. R.; Jacks, T.; Horvitz, H. R.; Golub, T. R. Nature Schmidheini, L.; Luchtefeld, I.; Hoogenberg, B. J.; Gisiger, A. B.; 2005, 435, 834−838. Voros, J. ACS Sens 2019, 4, 1950−1956. (2) Zhu, H.; Leung, S. W. Diabetologia 2015, 58, 900−911. (35) Choi, N. W.; Kim, J.; Chapin, S. C.; Duong, T.; Donohue, E.; (3) Kumar, P.; Dezso, Z.; MacKenzie, C.; Oestreicher, J.; Agoulnik, Pandey, P.; Broom, W.; Hill, W. A.; Doyle, P. S. Anal. Chem. 2012, 84, S.; Byrne, M.; Bernier, F.; Yanagimachi, M.; Aoshima, K.; Oda, Y. 9370−9378. PLoS One 2013, 8, No. e69807. (36) Srinivas, R. L.; Johnson, S. D.; Doyle, P. S. Anal. Chem. 2013, (4) Sucharov, C.; Bristow, M. R.; Port, J. D. J. Mol. Cell. Cardiol. 85, 12099−12107. 2008, 45, 185−192. (37) Squires, T. M.; Messinger, R. J.; Manalis, S. R. Nat. Biotechnol. (5) Pritchard, C. C.; Cheng, H. H.; Tewari, M. Nat. Rev. Genet. 2008, 26, 417−426. 2012, 13, 358−369. (38) Schneider, C. A.; Rasband, W. S.; Eliceiri, K. W. Nat. Methods (6) Baker, M. Nat. Methods 2010, 7, 687−692. 2012, 9, 671−675. (7) Leshkowitz, D.; Horn-Saban, S.; Parmet, Y.; Feldmesser, E. RNA (39) Daniel, M. C.; Astruc, D. Chem. Rev. 2004, 104, 293−346. 2013, 19, 527−538. (40) Park, J.; Yeo, J. S. Chem. Commun. (Cambridge, U. K.) 2014, 50, (8) Chugh, P.; Dittmer, D. P. Wiley Interdiscip Rev. RNA 2012, 3, 1366−1368. 601−616. (41) Wang, Q.; Li, R. D.; Yin, B. C.; Ye, B. C. Analyst 2015, 140, (9) Mestdagh, P.; Hartmann, N.; Baeriswyl, L.; Andreasen, D.; 6306−6312. Bernard, N.; Chen, C.; Cheo, D.; D’Andrade, P.; DeMayo, M.; (42) Miao, J.; Wang, J.; Guo, J.; Gao, H.; Han, K.; Jiang, C.; Miao, P. Dennis, L.; Derveaux, S.; Feng, Y.; Fulmer-Smentek, S.; Gerstmayer, Sci. Rep. 2016, 6, 32219. B.; Gouffon, J.; Grimley, C.; Lader, E.; Lee, K. Y.; Luo, S.; Mouritzen, (43) Dong, J.; Chen, G.; Wang, W.; Huang, X.; Peng, H.; Pu, Q.; P.; et al. Nat. Methods 2014, 11, 809−815. Du, F.; Cui, X.; Deng, Y.; Tang, Z. Anal. Chem. 2018, 90, 7107−7111. (10) Feng, C.; Mao, X.; Shi, H.; Bo, B.; Chen, X.; Chen, T.; Zhu, X.; (44) Persano, S.; Guevara, M. L.; Wolfram, J.; Blanco, E.; Shen, H.; Li, G. Anal. Chem. 2017, 89, 6631−6636. Ferrari, M.; Pompa, P. P. ACS Omega 2016, 1, 448−455. (11) Ishihara, R.; Hasegawa, K.; Hosokawa, K.; Maeda, M. Anal. Sci. (45) Takamizawa, J.; Konishi, H.; Yanagisawa, K.; Tomida, S.; 2015, 31, 573−576. Osada, H.; Endoh, H.; Harano, T.; Yatabe, Y.; Nagino, M.; Nimura, (12) Chapin, S. C.; Doyle, P. S. Anal. Chem. 2011, 83, 7179−7185. Y.; Mitsudomi, T.; Takahashi, T. Cancer Res. 2004, 64, 3753−3756. (13) Garcia-Schwarz, G.; Santiago, J. G. Angew. Chem., Int. Ed. 2013, (46) Yanaihara, N.; Caplen, N.; Bowman, E.; Seike, M.; Kumamoto, 52, 11534−11537. K.; Yi, M.; Stephens, R. M.; Okamoto, A.; Yokota, J.; Tanaka, T.; (14) Shintaku, H.; Palko, J. W.; Sanders, G. M.; Santiago, J. G. Calin, G. A.; Liu, C. G.; Croce, C. M.; Harris, C. C. Cancer Cell 2006, Angew. Chem., Int. Ed. 2014, 53, 13813−13816. 9, 189−198. (15) Dong, H.; Ding, L.; Yan, F.; Ji, H.; Ju, H. Biomaterials 2011, 32, 3875−3882. 5755 https://dx.doi.org/10.1021/acs.analchem.9b05043 Anal. Chem. 2020, 92, 5750−5755 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Analytical Chemistry Pubmed Central

Hydrogel-Based Colorimetric Assay for Multiplexed MicroRNA Detection in a Microfluidic Device

Analytical Chemistry, Volume 92 (8) – Mar 24, 2020

Loading next page...
 
/lp/pubmed-central/hydrogel-based-colorimetric-assay-for-multiplexed-microrna-detection-3tCfrPosS5
Publisher
Pubmed Central
Copyright
Copyright © 2020 American Chemical Society
ISSN
0003-2700
eISSN
1520-6882
DOI
10.1021/acs.analchem.9b05043
Publisher site
See Article on Publisher Site

Abstract

This is an open access article published under a Creative Commons Attribution (CC-BY) License, which permits unrestricted use, distribution and reproduction in any medium, provided the author and source are cited. pubs.acs.org/ac Article Hydrogel-Based Colorimetric Assay for Multiplexed MicroRNA Detection in a Microfluidic Device Hyewon Lee, Jiseok Lee, Seung-Goo Lee,* and Patrick S. Doyle* Cite This: Anal. Chem. 2020, 92, 5750−5755 Read Online Metrics & More Article Recommendations sı Supporting Information ACCESS * ABSTRACT: Although microRNA (miRNA) expression levels provide important information regarding disease states owing to their unique dysregulation patterns in tissues, translation of miRNA diagnostics into point-of-care (POC) settings has been limited by practical challenges. Here, we developed a hydrogel-based microfluidic platform for colorimetric profiling of miRNAs, without the use of complex external equipment for fluidics and imaging. For sensitive and reliable measurement without the risk of sequence bias, we employed a gold deposition-based signal amplification scheme and dark-field imaging, and seamlessly integrated a previously developed miRNA assay scheme into this platform. The assay demonstrated a limit of detection of 260 fM, along with multiplexing of small panels of miRNAs in healthy and cancer samples. We anticipate this versatile platform to facilitate a broad range of POC profiling of miRNAs in cancer- associated dysregulation with high-confidence by exploiting the unique features of hydrogel substrate in an on-chip format and colorimetric analysis. mall noncoding RNAs, called microRNAs (miRNAs), have systems for nucleic acid hybridization, especially from S become increasingly important in disease diagnosis due to biologically complex samples. Better thermodynamics in their abnormal expression in many diseases such as cancer, the gel increases both sensitivity and specificity, and the 1−4 diabetes, neural, and heart diseases. Previous studies had solution-like environment of a hydrogel provides faster demonstrated that miRNAs have high stability and can be hybridization kinetics. Furthermore, it is possible to directly detected in various biological fluids such as blood serum and measure miRNAs in complex media such as cell lysate or 12,23 plasma. However, clinicians still lack proper tools for high- serum, without RNA extraction and target amplification. In confidence quantification of miRNA owing to their low a previous gel particle-based approach, a universal labeling 6,7 abundance and sequence homology. Most challenges arise scheme was developed, which allowed the use of a single label due to the grueling demands of an assay that could easily be for all captured miRNAs, and measured miRNA dysregulation 24,25 integrated into a point-of-care (POC) clinical setting. An ideal patterns. Here, we adapted the same hydrogel chemistry platform ought to provide high sensitivity, high specificity, and and miRNA capture approach into an on-chip assay for POC multiplexing while minimizing the use of external equipment, applications. and involving simple sample preparation and assay operation. POC diagnostics can facilitate fast and accurate identi- While quantitative real-time polymerase chain reaction (qRT- fication of diseases, which leads to better treatment of patients. PCR) is widely used as a gold standard for high sensitivity, it is To fit the POC criteria of cost-effectiveness, portability, and limited in practical application, due to its requirement of accessibility, miRNA assays are commonly performed with substantial sample preparation steps, such as RNA extraction, nonfluorescent labeling molecules for visualization. Since such expensive instruments, and intricately complex design of assay schemes offer only low to moderate detection sensitivity, primers to avoid sequence bias arising from target-based miRNA detection methods often rely on signal boosting step, 9,10 amplification. To address these limitations of the conven- 26−28 such as enzymatic and nucleic acid amplification. The tional method, several techniques have now been developed Mirkin group has developed an assay using gold nanoparticles, 11,12 for miRNA detection, fluorescent probes, isotachopho- with silver or gold deposition, which had been successfully 13,14 15,16 resis, and nanomaterial-based in vivo sensing. Despite advancements, the techniques do not satisfy all the require- Received: November 5, 2019 ments of POC regarding miRNAs, and the need for a more Accepted: March 24, 2020 clinically feasible approach still remains. Published: March 24, 2020 Hydrogel-based microfluidic approach can be advantageous in the development of reliable and affordable POC 18−20 diagnostics. The superiority of nonfouling hydrogels has been previously demonstrated in comparison to surface-based https://dx.doi.org/10.1021/acs.analchem.9b05043 © 2020 American Chemical Society Anal. Chem. 2020, 92, 5750−5755 5750 Analytical Chemistry pubs.acs.org/ac Article 29−31 24 applied to detect various biomarkers, including miRNAs. and let-7a: 50 μM). Probe-containing prepolymer solution Their method provides robust signals without bleaching issues was vortexed, centrifuged, and loaded into channels by because of high photostability of gold-labeled conjugates, and pipetting. An inverted microscope (Zeiss Axio Observer A1) very short amplification step with high sensitivity while and a CCD camera (Andor Clara) were used for UV-initiated minimizing equipment requirements. However, this micro- polymerization. With a desired photomask (Fineline Imaging) array-based method typically requires a long assay time placed in the fieldstop, the polymerization was performed at (overnight hybridization of target miRNAs) with complicated 100 ms exposure (Lumen 200, Prior Scientific) using a 20× fluidic steps. Thus, we aim to integrate on-chip hydrogel assay objective (Zeiss Plan-Neofluar) and a dichrioc filter for and gold deposition scheme to achieve sensitive POC excitation at 365 nm. After polymerization, channels were applications. rinsed with 1X TE buffer. For multiplexed assays, the We leverage the advantages of microfluidic channels, and the subsequent monomer solution with a different probe was chemical advantages of a polyethylene glycol (PEG) hydrogel loaded in the same channel and polymerization was repeated as scaffold for miRNA hybridization while optimizing an enzyme- described in earlier. The posts containing different probe free gold nanoparticle-based signal amplification scheme for sequences were immobilized at spatially distinct locations in a POC diagnostics. Particularly, we sought to use dark-field single channel. imaging for high sensitivity, which had previously been Fluidic Control. All assays were performed with gravity demonstrated for ultrasensitive colorimetric nucleic acid driven flow by appropriately inserting a 200 μL pipet tip into assay. Dark-field imaging-based methods boost the scattered PDMS inlet port. The flow rate was observed to be 1−5 μL/ intensity of gold conjugates with white light illumination, thus ́ min. Previously, we calculated the channel Peclet number (Pe) minimizing the requirement of complex and expensive to be approximately 7000. With the high values of Pe (Pe ≫ instrumentation for miRNA detection. With just a 15 min 1), target depletion is assumed to be negligible. The solution signal amplification step, we achieved a limit of detection of was refilled in the tip of inlet port every 10 min to maintain 260 fM and multiplexed miRNA quantification with minimal height differences for a steady flow. Since our system is highly sample input. We expect this assay platform to be beneficial in flexible and can be incorporated with various types of flows a wide range of clinical samples, including cellular lysate and such as pressure-driven flows and syringe pump-driven flows, serum for POC applications. the manual iterative operation can be avoided by simply using an external equipment for fluidics. EXPERIMENTAL SECTION MicroRNA Assay. Blocking solution of 3% (w/v) Pluronic F108 (Sigma-Aldrich) in nuclease-free water (Affymetrix) was Microfluidic Chip Preparation. Commercial chips with filled into channels containing hydrogel posts. After 30 min, straight channels (50 μm in height, 1 mm width, and 18 mm the process of target hybridization to specific probes was length) were purchased from Hilgenberg GmbH, Germany for performed with the mixture of synthetic RNAs (IDT) or total performing all assays. Connection ports were fabricated in RNA (BioChain, Newark, CA) in a TET buffer with a final polydimethylsiloxane (PDMS) (Corning, Sylgard 184) with concentration of 350 mM NaCl for 90 min. The sequence holes, which were punched with 15-gauge needles. All information on all the probes and targets used in this study was connection ports were bonded onto inlets and outlets of summarized in Supporting Information Table S1. Total RNA glass chips by oxygen plasma treatment (25 s on RF = high, was stored at 100 ng/μLat −20 °C and 500 ng of total RNA Harrick Scientific, Pleasantville, NY). After a subsequent dissolved in 200 μL was used for assay. Before performing the incubation at 80 °C for 20 min, a 2% (v/v) solution of 3- assay, the solution with total RNA was incubated at 95 °C for 5 (trimethoxysilyl)propyl acrylate (Sigma) mixed in 24.5% (v/v) min in a thermoshaker for disrupting secondary structures, and 1X PBS (phosphate buffered saline, Corning), and 73.5% (v/v) cooled down at room temperature. The hybridization mixture ethanol was filled inside the channels for tight adhesion of the was loaded into the microchannel through the precut pipet tips gel pads on the glass walls. After 30 min, channels were rinsed in the injection port of the device. Target hybridization with ethanol and dried with argon gas. After bioassays, chips occurred at elevated temperature (55 °C) on a hot plate, and were cleaned by soaking in 1 M NaOH for 1 h and then all subsequent steps were performed at room temperature. In washed with DI water and ethanol. Next, they were dried with previous works, the temperature in channels was validated to argon gas; chips were stored at 80 °C until the time of usage. be constant based on a simple heat transfer calculation. For These reusable glass chips were used for several assays (more all assays, steady flows were maintained to deliver the than 10 times) by repeating the cleaning and activating molecules without depletion. In between assay steps, rinses procedures. were performed using 50 mM NaCl in TET for 30 s to impose On-Chip Hydrogel Synthesis. All chemicals were sufficient stringency for high specific miRNA measurement. purchased from Sigma-Aldrich (U.S.). Hydrogel posts were For ligation, the universal linker (IDT), T4 DNA ligase (800 synthesized using projection lithography from a polyethylene U/mL), ATP (250 nM), and 10× NEB2 buffer (New England glycol monomer mixture. The monomer mixture consisted of Biolabs, Ipswich, MA) were mixed in TET and incubated in 20% (v/v) PEGDA 700 (poly(ethylene glycol) diacrylate, MW the channels for 10 min as described in previous studies. = 700 g/mol), 40% (v/v) PEG 600 (poly(ethylene glycol), After washing, Nanogold-Streptavidin (80 μg/mL, Nanop- MW = 600 g/mol), 5% (v/v) Darocur 1173 photoinitiator, and 35% (v/v) 3x TE (Tris-EDTA, USB Corporation) buffer robes, Inc., Yaphank, NY) diluted in 1× PBS (0.1% v/v) was with food coloring dye, which was previously optimized for the loaded into channels for 30 min. For gold deposition, 35,36 diffusion and reaction in bioassays. The monomer solution GoldEnhance (Nanoprobes, Inc.) was used, which consisted was diluted 9:1 with the acrydite-modified probes (Integrated of Solution A (enhancer), Solution B (activator), Solution C DNA Technologies (IDT), Coralville, IA). The concentration (initiator), and Solution D (buffer). First, 200 μL of Solution A of probe molecules was adjusted from the coarse rate matching and B were mixed, and after 5 min, 200 μL of Solution C and as described in a previous study (miR-21:247 μM, miR-145, D were added into the mixture. After vortexing, the mixture 5751 https://dx.doi.org/10.1021/acs.analchem.9b05043 Anal. Chem. 2020, 92, 5750−5755 Analytical Chemistry pubs.acs.org/ac Article Figure 1. (A) Schematics of on-chip hydrogel post synthesis for multiplexing of small miRNA panels using projection lithography with the spatial encoding scheme. (B) miRNA assay scheme: target hybridization, universal linker ligation, gold nanoparticle labeling, and gold ion deposition- based signal amplification. (C) Dark-field images of posts after complete miRNA assay demonstrating the dose-dependent response of miRNAs. Scale bar represents 100 μm. was delivered to the microchannels with posts for an optimal RESULTS AND DISCUSSION amplification time of 15 min. Assay Development. For the development of a sensitive Data Acquisition and Analysis. The hydrogel posts were miRNA assay in point-of-care (POC) diagnostics, we sought to imaged using a Zeiss Axio Observer A1 microscope with a 10× apply a rapid signal amplification scheme by gold deposition objective and a dark-field condenser. Each image was analyzed onto a hydrogel-based on-chip miRNA sensor. Previously, we had demonstrated ultrasensitive measurement of miRNAs using ImageJ (National Institutes of Health). The color using oil-isolated hydrogel chambers in microfluidic chip, with images we obtained were split to the respective red, green, and a limit of detection (LOD) of 22.6 fM. We had achieved high blue image components. The region of interest (ROI) was Peclet ́ number in the previous work by using a gravity-driven identified as a post circle with a defined diameter. For each flow to eliminate mass transfer limits that could arise from ROI, the intensity of pixels in the red channel was only insufficient target molecule deliveries to the gel posts. averaged to analyze the wine-red color of miRNA-gold Importantly, this strategy eliminated the need for complicated conjugates. Background signal was calculated by averaging external flow controllers, making our system suitable for point of care applications. However, this assay was based on the red pixel intensities in a region outside the post, within the fluorescence measurements, which required expensive and microfluidic chip. Target signal was computed by subtracting complex imaging instruments. To overcome this issue, we used background signal from a raw signal with target incubation a colorimetric labeling approach based on gold deposition for (S = S − S ). Control raw signal was target target, raw background signal amplification and dark-field imaging. obtained from the posts after incubation with 0 fM of synthetic As done previously, we immobilized spatially encoded targets. After subtracting background signal (S = S control control, raw miRNA probe-bearing hydrogel posts in microfluidic channels − S ), the control signal serves as the assay background, using projection lithography to control their size and shape background and the noise was defined as the standard deviation of the (Figure 1a). Multiplexing can be achieved with a spatial encoding scheme by exchanging monomer solutions in the control signals (σ ). Control signal was subtracted from control device and polymerizing posts. The incorporated probes the desired target signal (S = S − S ) to obtain a target, net target control consist of two domains: a miRNA-specific domain and a net signal. This makes S = 0. Finally, the signal-to-noise control, net universal linker domain. The linker used for labeling was the ratio (SNR) was determined as the ratio of net signal to assay same for all measured miRNAs. A fully controlled steady noise (SNR = S /σ ). Limit of detection of the target, net control gravity-driven flow was used for all assay steps, thus eliminating system was defined as the miRNA concentration at which the the need for expensive and cumbersome flow controllers. SNR equals 3, as per previously published protocols. Target hybridization occurred at elevated temperatures (55 5752 https://dx.doi.org/10.1021/acs.analchem.9b05043 Anal. Chem. 2020, 92, 5750−5755 Analytical Chemistry pubs.acs.org/ac Article °C) on a hot plate, whereas all subsequent steps were performed at room temperature (Figure 1b). We ligated the biotinylated universal linker to the probe-target complex using the T4 DNA ligase for 10 min. Without target miRNA bound to the probe, the universal linker would be released during the stringent washing step with low salt (50 mM NaCl) buffer due to low binding affinity to probe cross-linked into the hydrogel. Next, the probe-target complex was labeled using a streptavidin-conjugated gold nanoparticle (SA-Au, 80 ng/mL, 30 min). Thereafter, the signal amplification scheme, based on gold ion deposition, was applied for 15 min. Figure 2. Optimization of amplification time. The signal-to-noise Both bright- and dark-field systems were considered for ratio (SNR) was continued to increase for up to 15 min of gold deposition-based signal amplification. After achieving a maximum imaging the gold-labeled target miRNAs. For system character- value at 15 min, SNR decreased and then saturated due to nonspecific ization, we used immobilized gel posts functionalized with binding of gold ions. Error bars represent standard deviation (n =8− biotinylated DNA probe (5 nM). Although bright-field mode is 15). widely used in colorimetric assay, its high background signals often generate large variations. More importantly, there was no difference between the presence and absence of probe after conjugates over 24 h, after Nanogold-Streptavidin labeling and labeling streptavidin-conjugated gold nanoparticles and signal gold ion deposition steps. As described earlier, the net signal amplification with catalytic gold deposition (Supporting was determined as control (no biotinylated probe)-subtracted Information (SI) Figure S1). Conversely, with dark-field target signal. As expected, the net signal did not decrease, as illumination we observed measurable signals from 5 nM shown in SI Figure S2. Importantly, it was noted that hydrogels probe. Thus, for sensitive measurement of surface plasmon and gold nanoparticles are stable for several months without 21,39 resonance (SPR) scattering from gold nanoparticles, we used degradation, which is expected to allow long-term storage dark-field microscopy with the appearance of wine-red color. and reimaging. Therefore, this assay platform would provide As shown in Figure 1c, in an initial trial with miRNA assay with high stability in analysis, without photobleaching, and can be the analysis using dark-field microscopy, the red dots were stored for a long time or be transported, if necessary, rendering gradually increased as the concentration of target miRNA it ideal for POC diagnostics. increased from 0 to 10 pM. Considering simplicity of the assay, Detection Sensitivity. Using the gold labeling scheme, we further analysis was performed with red channel in RGB examined the signal changes in 10 pM spike-in of let-7a as a imaging. function of hybridization time. The target capture increased for Optimization and Assessment of Signal Amplifica- 90 min and then saturated, as observed in previous studies (SI 24,25 tion. Since previous studies had used the gold deposition Figure S3). To minimize the RNA input requirement while scheme in a microarray format, we first needed to optimize the retaining high sensitivity, we decided to use 90 min of signal amplification step in the on-chip hydrogel-based assay. hybridization. The 90 min hybridization was recommended While previous studies used three rounds of 5 min gold due to the stringent buffer conditions (high temperature and deposition to maximize the signal-to-noise ratio (SNR) by low salt concentration), which we optimized earlier for specific increasing target-binding signal, we sought to run a single- miRNA measurements to distinguish between even 1−2 24,25 step amplification to minimize the assay complexity for POC mismatches. If necessary, we could reduce the assay applications. We hypothesized that the solution-like and time, which would still give us a reasonable signal. nonfouling environments of hydrogel posts would be superior We next investigated the sensitivity of our hydrogel-based to a microarray platform (rigid, planar surfaces) for gold colorimetric detection scheme for detecting miRNAs. The deposition on targets without nonspecific background. For the synthetic miRNA spike-ins (from 100 fM to 10 pM let-7a) optimization of the signal amplification step, we measured the were incubated in each channel of a microfluidic chip. To SNR as a function of gold ion deposition time. Using account for the assay background, the net signal was immobilized gel posts functionalized with either biotinylated considered−the control signal (0 fmol spike-in) was subtracted DNA (final concentration of 5 nM) or no biotin (serving as from thetargetsignals.Weshowthe dose-dependent control), we loaded Nanogold-Streptavidin (80 ng/mL, PBS) responses from let-7a spike-ins in Figure 3. As shown in SI for 30 min and, after washing, GoldEnhance mixture for 5−60 Figure S4, we then calculated the limit of detection (LOD), min. Then, we calculated SNR, which is defined as the net which was previously defined in Experimental Section. Our control-subtracted signal divided by the standard deviation of assay scheme provided a LOD of 260 fM, which is relatively control measurements (assay-derived noise). As shown in sensitive compared to other colorimetric assays (pM to sub- 40−42 Figure 2, SNR increased up to a time of 15 min, and then pM). Although some recent studies on colorimetric 10,43,44 decreased due to high background signal. This indicated that a miRNA assays show high sensitivity (fM to sub-fM), hydrogel-based system can achieve high sensitivity, simply with their methods rely on target-based amplification with the a single-step amplification without the need of multiple rounds. possible high risk of sequence bias. In addition, we might Next, we examined the assay stability, which is one of the achieve better sensitivity if we run multiple rounds of signal important considerations for reliable measurement in POC amplification as a previous study optimized to perform the diagnostics. An unstable signaling label would induce test three rounds of gold deposition to maximize SNR. errors. Also, sometimes assay platforms might need to be Additionally, the size and geometry of hydrogel posts can be transferred from remote areas to a core facility for analysis by optimized to improve the sensitivity of miRNA measurements. experts for accurate diagnostics. With 5 nM of the biotinylated In a miRNA assay with the high concentration of targets such probe, we analyzed the net signal from the gold-labeled as 10 pM, we observed that more targets were attached to the 5753 https://dx.doi.org/10.1021/acs.analchem.9b05043 Anal. Chem. 2020, 92, 5750−5755 Analytical Chemistry pubs.acs.org/ac Article Figure 3. Assay sensitivity. The colorimetric miRNA measurements in microfluidics provided limit of detection (LOD) of 260 fM without expensive and complex instrument (see SI Figure S4 for SNR calculation). Error bars represent standard deviation (n =5−14). edge of the post, rather than evenly distributed (Figure 1c). Previously, we optimized the pore size of gel scaffold to allow the fast diffusion of molecules. However, the rapid deposition of gold ions seems to affect the transport rate of labeling Figure 4. Multiplexed measurements of three miRNA targets from molecules, especially with the high level of probe-target total RNA derived from tissue. (A) The signal-to-noise ratio (SNR) of conjugates. To consider flux into porous hydrogel posts, we three miRNAs in tumor and healthy samples were plotted, and each could use ring structures instead of disk shapes as in a recent of the tumor and healthy samples was measured 10 times. Error bars publication. By analyzing a ring-area around the edges, we represent the standard deviation of targets normalized by assay noise (n = 10). (B) Dysregulation ratios of three miRNA targets in lung expect that we could achieve higher mean signal and better tumor versus healthy tissue were as expected from the previous sensitivity, which would be beneficial in POC diagnostics. studies, validating our assay multiplexing scheme. Error bars represent Multiplex Detection. After characterization, we measured the standard deviation of miRNA expression measurements in tumor the cross-reactivity of three microRNAs. Three clinical miRNA normalized by background-subtracted average miRNA signal in targets relevant in lung tumor were considered in this study: normal and by the ratio of tumor to normal miRNA expression. let-7a, miR-145, and miR-21. We immobilized three types of hydrogel posts, bearing each miRNA probe using the spatial encoding scheme. As shown in SI Figure S5, there was no CONCLUSIONS significant interference among three different miRNA targets Here, we present a technique for miRNA quantification in (∼20% of cross-reactivity). This minimal cross-reactivity of our point-of-care (POC) diagnostics by using gold deposition- hydrogel-based miRNA assay scheme enabled multiplexing based signal amplification scheme with the on-chip hydrogel analysis of small panels of miRNAs. sensor platform. Unlike previous studies, the colorimetric assay For multiplexing, each miRNA probe concentration was developedheredemonstratedhighlystableand reliable adjusted by coarse-rate matching (Experimental Section). All miRNA measurement without the need of expensive instru- three miRNAs were expected to follow the same rate under the ments. This assay provided a limit of detection of 260 fM, same hybridization condition (salt and temperature). To verify which is relatively sensitive compared to other colorimetric this hypothesis for the hydrogel-based colorimetric assay, we assays. Moreover, our system enabled multiplex analysis of measured the detection limit for each microRNA target based small panels of miRNAs with relatively simple assay steps. We on the calibration curve (SI Figure S4). As expected, all three successfully analyzed the dysregulation of miRNAs in lung microRNA targets showed similar LOD (SI Table S2). tumor with respect to that in healthy tissues. In the future, it As a proof of concept, we compared the miRNA expression might be possible to integrate smartphone-based imaging in tumor and healthy tissue from total RNA samples using the system for immediate processing and analysis, with the help of colorimetric platform developed here. With the posts bearing smart algorithms. We envision that our system to have wide- three miRNA probes (let-7a, miR-145, and miR-21), we ranging applications in POC clinical settings for various targets performed the multiplexed assay with total RNA samples, and such as miRNAs, RNAs, DNA, and proteins. we observed the dysregulation patterns of the three miRNA targets in healthy and tumor tissue (Figure 4). The patterns are ASSOCIATED CONTENT consistent with prior studies the literature, since the expression sı * Supporting Information of both let-7a and mir-145 are known to be decreased in lung The Supporting Information is available free of charge at 24,25,45,46 tumor tissue, whereas that of miR-21 is elevated. Also, https://pubs.acs.org/doi/10.1021/acs.analchem.9b05043. by using the same total RNA sample, signal from this A list of reagents, additional supporting tables, and colorimetric assay were comparable to those in previous figures are available as noted in the text (PDF) studies using fluorescent labels (phycoerythrin-conjugated streptavidin reporter, SA-PE) (SI Figure S6), which were AUTHOR INFORMATION previously validated with qRT-PCR. This consistency with prior work demonstrates the high performance and reprodu- Corresponding Authors cibility of our new miRNA measurement without using Patrick S. Doyle − Department of Chemical Engineering, expensive fluorescence detection. Massachusetts Institute of Technology, Cambridge, 5754 https://dx.doi.org/10.1021/acs.analchem.9b05043 Anal. Chem. 2020, 92, 5750−5755 Analytical Chemistry pubs.acs.org/ac Article (16) Ryoo, S. R.; Lee, J.; Yeo, J.; Na, H. K.; Kim, Y. K.; Jang, H.; Lee, Massachusetts 02139, The United States; orcid.org/0000- J. H.; Han, S. W.; Lee, Y.; Kim, V. N.; Min, D. H. ACS Nano 2013, 7, 0003-2147-9172; Email: pdoyle@mit.edu 5882−5891. Seung-Goo Lee − Synthetic Biology and Bioengineering (17) Cheng, Y.; Dong, L.; Zhang, J.; Zhao, Y.; Li, Z. Analyst 2018, Research Center, Korea Research Institute of Bioscience and 143, 1758−1774. Biotechnology, Daejeon 34141, Republic of Korea; Department (18) Yetisen, A. K.; Butt, H.; Volpatti, L. R.; Pavlichenko, I.; Humar, of Biosystems and Bioengineering, KRIBB School of M.; Kwok, S. J.; Koo, H.; Kim, K. S.; Naydenova, I.; Khademhosseini, Biotechnology, University of Science and Technology, Daejeon A.; Hahn, S. K.; Yun, S. H. Biotechnol. Adv. 2016, 34, 250−271. 34113, Republic of Korea; orcid.org/0000-0003-4539- (19) Wei, X.; Tian, T.; Jia, S.; Zhu, Z.; Ma, Y.; Sun, J.; Lin, Z.; Yang, 9812; Email: sglee@kribb.re.kr C. J. Anal. Chem. 2016, 88, 2345−2352. (20) Shapiro, S. J.; Dendukuri, D.; Doyle, P. S. Anal. Chem. 2018, 90, Authors 13572−13579. Hyewon Lee − Synthetic Biology and Bioengineering Research (21) Le Goff, G. C.; Srinivas, R. L.; Hill, W. A.; Doyle, P. S. Eur. Center, Korea Research Institute of Bioscience and Polym. J. 2015, 72, 386−412. Biotechnology, Daejeon 34141, Republic of Korea (22) Pregibon, D. C.; Doyle, P. S. Anal. Chem. 2009, 81, 4873− Jiseok Lee − School of Energy and Chemical Engineering, Ulsan (23) Lee, H.; Shapiro, S. J.; Chapin, S. C.; Doyle, P. S. Anal. Chem. National Institute of Science and Technology, Ulsan 44919, 2016, 88, 3075−3081. Republic of Korea (24) Chapin, S. C.; Appleyard, D. C.; Pregibon, D. C.; Doyle, P. S. Complete contact information is available at: Angew. Chem., Int. Ed. 2011, 50, 2289−2293. https://pubs.acs.org/10.1021/acs.analchem.9b05043 (25) Lee, H.; Srinivas, R. L.; Gupta, A.; Doyle, P. S. Angew. Chem., Int. Ed. 2015, 54, 2477−2481. Notes (26) Cheng, C. M.; Martinez, A. W.; Gong, J.; Mace, C. R.; Phillips, S. T.; Carrilho, E.; Mirica, K. A.; Whitesides, G. M. Angew. Chem., Int. The authors declare no competing financial interest. Ed. 2010, 49, 4771−4774. (27) Sun, Y.; Shi, L.; Wang, Q.; Mi, L.; Li, T. Anal. Chem. 2019, 91, ACKNOWLEDGMENTS 3652−3658. We acknowledge the support from NIH-NIBIB Grant (28) Kim, D.; Wei, Q.; Kim, D. H.; Tseng, D.; Zhang, J.; Pan, E.; 5R21EB024101-02, the C1 Gas Refinery Program funded by Garner, O.; Ozcan, A.; Di Carlo, D. Anal. Chem. 2018, 90, 690−695. the Ministry of Science and ICT (2018M3D3A1A01055732), (29) Elghanian, R.; Storhoff, J. J.; Mucic, R. C.; Letsinger, R. L.; National Research Foundation of Korea (NRF) grant funded Mirkin, C. A. Science 1997, 277, 1078−1081. by the Korea government (MSIT) (2018R1C1B6004086), and (30) Kim, D.; Daniel, W. L.; Mirkin, C. A. Anal. Chem. 2009, 81, 9183−9187. Korea Research Institute of Bioscience and Biotechnology (31) Alhasan, A. H.; Kim, D. Y.; Daniel, W. L.; Watson, E.; Meeks, J. Research Initiative Program. J.; Thaxton, C. S.; Mirkin, C. A. Anal. Chem. 2012, 84, 4153−4160. (32) Tang, L.; Li, J. ACS Sens 2017, 2, 857−875. REFERENCES (33) Li, J.; Kong, C.; Liu, Q.; Chen, Z. Analyst 2018, 143, 4051− (1) Lu, J.; Getz, G.; Miska, E. A.; Alvarez-Saavedra, E.; Lamb, J.; Peck, D.; Sweet-Cordero, A.; Ebert, B. L.; Mak, R. H.; Ferrando, A. (34) Hwu, S.; Blickenstorfer, Y.; Tiefenauer, R. F.; Gonnelli, C.; A.; Downing, J. R.; Jacks, T.; Horvitz, H. R.; Golub, T. R. Nature Schmidheini, L.; Luchtefeld, I.; Hoogenberg, B. J.; Gisiger, A. B.; 2005, 435, 834−838. Voros, J. ACS Sens 2019, 4, 1950−1956. (2) Zhu, H.; Leung, S. W. Diabetologia 2015, 58, 900−911. (35) Choi, N. W.; Kim, J.; Chapin, S. C.; Duong, T.; Donohue, E.; (3) Kumar, P.; Dezso, Z.; MacKenzie, C.; Oestreicher, J.; Agoulnik, Pandey, P.; Broom, W.; Hill, W. A.; Doyle, P. S. Anal. Chem. 2012, 84, S.; Byrne, M.; Bernier, F.; Yanagimachi, M.; Aoshima, K.; Oda, Y. 9370−9378. PLoS One 2013, 8, No. e69807. (36) Srinivas, R. L.; Johnson, S. D.; Doyle, P. S. Anal. Chem. 2013, (4) Sucharov, C.; Bristow, M. R.; Port, J. D. J. Mol. Cell. Cardiol. 85, 12099−12107. 2008, 45, 185−192. (37) Squires, T. M.; Messinger, R. J.; Manalis, S. R. Nat. Biotechnol. (5) Pritchard, C. C.; Cheng, H. H.; Tewari, M. Nat. Rev. Genet. 2008, 26, 417−426. 2012, 13, 358−369. (38) Schneider, C. A.; Rasband, W. S.; Eliceiri, K. W. Nat. Methods (6) Baker, M. Nat. Methods 2010, 7, 687−692. 2012, 9, 671−675. (7) Leshkowitz, D.; Horn-Saban, S.; Parmet, Y.; Feldmesser, E. RNA (39) Daniel, M. C.; Astruc, D. Chem. Rev. 2004, 104, 293−346. 2013, 19, 527−538. (40) Park, J.; Yeo, J. S. Chem. Commun. (Cambridge, U. K.) 2014, 50, (8) Chugh, P.; Dittmer, D. P. Wiley Interdiscip Rev. RNA 2012, 3, 1366−1368. 601−616. (41) Wang, Q.; Li, R. D.; Yin, B. C.; Ye, B. C. Analyst 2015, 140, (9) Mestdagh, P.; Hartmann, N.; Baeriswyl, L.; Andreasen, D.; 6306−6312. Bernard, N.; Chen, C.; Cheo, D.; D’Andrade, P.; DeMayo, M.; (42) Miao, J.; Wang, J.; Guo, J.; Gao, H.; Han, K.; Jiang, C.; Miao, P. Dennis, L.; Derveaux, S.; Feng, Y.; Fulmer-Smentek, S.; Gerstmayer, Sci. Rep. 2016, 6, 32219. B.; Gouffon, J.; Grimley, C.; Lader, E.; Lee, K. Y.; Luo, S.; Mouritzen, (43) Dong, J.; Chen, G.; Wang, W.; Huang, X.; Peng, H.; Pu, Q.; P.; et al. Nat. Methods 2014, 11, 809−815. Du, F.; Cui, X.; Deng, Y.; Tang, Z. Anal. Chem. 2018, 90, 7107−7111. (10) Feng, C.; Mao, X.; Shi, H.; Bo, B.; Chen, X.; Chen, T.; Zhu, X.; (44) Persano, S.; Guevara, M. L.; Wolfram, J.; Blanco, E.; Shen, H.; Li, G. Anal. Chem. 2017, 89, 6631−6636. Ferrari, M.; Pompa, P. P. ACS Omega 2016, 1, 448−455. (11) Ishihara, R.; Hasegawa, K.; Hosokawa, K.; Maeda, M. Anal. Sci. (45) Takamizawa, J.; Konishi, H.; Yanagisawa, K.; Tomida, S.; 2015, 31, 573−576. Osada, H.; Endoh, H.; Harano, T.; Yatabe, Y.; Nagino, M.; Nimura, (12) Chapin, S. C.; Doyle, P. S. Anal. Chem. 2011, 83, 7179−7185. Y.; Mitsudomi, T.; Takahashi, T. Cancer Res. 2004, 64, 3753−3756. (13) Garcia-Schwarz, G.; Santiago, J. G. Angew. Chem., Int. Ed. 2013, (46) Yanaihara, N.; Caplen, N.; Bowman, E.; Seike, M.; Kumamoto, 52, 11534−11537. K.; Yi, M.; Stephens, R. M.; Okamoto, A.; Yokota, J.; Tanaka, T.; (14) Shintaku, H.; Palko, J. W.; Sanders, G. M.; Santiago, J. G. Calin, G. A.; Liu, C. G.; Croce, C. M.; Harris, C. C. Cancer Cell 2006, Angew. Chem., Int. Ed. 2014, 53, 13813−13816. 9, 189−198. (15) Dong, H.; Ding, L.; Yan, F.; Ji, H.; Ju, H. Biomaterials 2011, 32, 3875−3882. 5755 https://dx.doi.org/10.1021/acs.analchem.9b05043 Anal. Chem. 2020, 92, 5750−5755

Journal

Analytical ChemistryPubmed Central

Published: Mar 24, 2020

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create folders to
organize your research

Export folders, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off