Multivalent aptasensor array and silver aggregated amplification for multiplex detection in microfluidic devices

Talanta 188 (2018) 417–422

Contents lists available at ScienceDirect

Talanta journal homepage: www.elsevier.com/locate/talanta

Multivalent aptasensor array and silver aggregated amplification for multiplex detection in microfluidic devices ⁎

Xiaohui Liu, Hui Li , Yaju Zhao, Xiaodong Yu, Danke Xu

T

⁎

State Key Laboratory of Analytical Chemistry for Life Science, School of Chemistry and Chemical Engineering, Nanjing University, China

A R T I C LE I N FO

A B S T R A C T

Keywords: Microfluidic Microarray Aptamer Silver nanoparticles Multiplex detection

Herein, we developed a rapid and sensitive aptamers-based sandwich assay in microfluidic devices based on multivalent aptasensor array (MAA) chip and silver aggregated amplification (SAA) strategy for the detection of two biomarkers. Firstly, aptamers-modified silver nanoparticles were dotted in array to form MAA chip. Then PDMS was used to form a microfluidic device. After that, target proteins and two kinds of aptamer-modified silver nanoparticles (Tag-A and Tag-B) were rapidly injected into the microfluidic device. The aptamer on MAA chip recognized target, and the target also bound with Tag-A and Tag-B which could aggregate with each other to amplify fluorescence signal. Based on MAA chip and SAA strategy in microfluidic device, a linear response to PDGF-BB (r = 0.999) was obtained in the concentration range from 16 pg mL−1 to 250 ng mL−1, and the detection limit was 1.4 pg mL−1. In addition, a linear response to PDGF-BB (r = 0.992) was obtained in the concentration range from 16 pg mL−1 to 250 ng mL−1 in 10% blood serum with detection limit of 7.8 pg mL−1. Ultimately, this assay was used to simultaneously detect PDGF-BB and VEGF-165, and the results showed good specificity and sensitivity. This assay can also be expanded to sensitive and high-throughput detection of other protein biomarkers by coupling of various aptamers with nanoparticles.

1. Introduction Microfluidic system, also called micro-total-analysis-system (μTAS) or lab-on-chip (LOC), has received a growing number of attention in various fields. Compared to conventional techniques, microfluidic system has the advantages of rapid, high-throughput, miniaturization, automation, and parallel processing. Owning to these merits, the microfluidic systems have been applied widely in chemical and biological sensors [1], drug discovery [8], aptamer screening [23], medical diagnostic [39] and point of care testing [4]. Protein microarray chip with high-throughput capacity provides a promising way for multiplex detection in parallel [16,2]. In our previous work, we designed a structure switching aptamer-based silver microarray nanosensor for sensitive and selective detection of proteins [38]. We also reported a simple, ultrasensitive, and cost-effective electrochemical aptamer microarray sensor for specific determination of multiplied proteins using SPE array chip as sensing platforms [33]. Recently, Shivani Sathish developed a simple technology to create micro- and nanoarrays of biomolecules within microfluidic devices [31]. Antoine-Emmanuel Saliba developed a liposome-microarray-based assay (LiMA) that was capable of measuring protein recruitment into membranes in a quantitative and high-throughput manner [30]. María Díaz-González

⁎

Corresponding authors. E-mail addresses: [email protected] (H. Li), [email protected] (D. Xu).

https://doi.org/10.1016/j.talanta.2018.05.048 Received 17 January 2018; Received in revised form 4 May 2018; Accepted 11 May 2018 Available online 29 May 2018 0039-9140/ © 2018 Published by Elsevier B.V.

presented an automated electrical readout system incorporating microfluidic channels for low-cost glass-slide microarrays [7]. Although those methods provide simplicity, automation, high-throughput potentials for achieving multiplex detection in microarray based microfluidic devices, the solutions to enhance the sensitivity and reduce the detection limit are still limited. We aim to develop a totally new platform based on microarray in microfluidic devices to fulfill simple, sensitive, automated and multiplex detection. Platelet-derived growth factor-BB (PDGF-BB) [29] and vascular endothelial growth factor-165 (VEGF-165) [27] have been reported to be important biomarkers which play a vital role in diagnosis and prognosis of many diseases. Until now, a variety of analytical platforms for protein biomarkers detection have been reported, such as enzymelinked immunosorbent assay (ELISA) [31,46], electrochemical assay [32,43], optical methods [21,47], surface plasmon resonance (SPR) [28,3] and other method [9]. In addition, a new class of nucleic acid affinity elements called aptamers has been developed for high specificity and selectivity of protein assay [12,15]. In general, Aptamers are short single-stranded nucleic acids selected in vitro for binding certain molecules due to their high affinity and specificity, and they have been playing important roles in biosensing, medical diagnosis and diseases treatment [42,45]. Especially, when combined with nanomaterials,

Talanta 188 (2018) 417–422

X. Liu et al.

2. Materials and methods

TAMRA and hybrid-B (Tag-PD-B) were used to form detection probe (Tag-PD-B). 1 mL of AgNPs-1 was mixed with Apt-PD-TAMRA (50 μL, 10 μΜ) and hybrid-A (50 μL ,10 μΜ)/ hybrid-B (10 μΜ). Then, 5 μL of Tween-20 and 64 μL of NaCl (2 M) were added and reacted for 4 h at 37 °C with gentle shaking. After standing overnight, excess reagents and unmodified Apt-PD-TAMRA, hybrid-A and hybrid-B were removed by centrifugation (10 min, 15,000 rpm for three times). Based on the same method, Tag-VE-A and Tag-VE-B were prepared for detection of VEGF165. The AgNPs-2 was functionalized with Apt-PD and 5′SH-oligo(d)A12NH2 to form capture probe (PD-AgNP). 1 mL of AgNPs-2 was mixed with Apt-PD (50 μL, 10 μΜ) and 5′SH-oligo(d)A12-NH2 (50 μL ,10 μΜ). Then, 5 μL Tween-20 and 64 μL of NaCl (2 M) were added and reacted for 4 h at 37°Cwith gentle shaking. After standing overnight, excess reagents and unmodified Apt-PD, 5′SH-oligo(d)A12-NH2 were removed by centrifugation (10 min, 15,000 rpm for three times). In addition, PD &VE-AgNP were prepared for multiplexed capture of PDGF-BB, VEGF165 detection.

2.1. Materials and reagents

2.4. Fabrication of MAA chip in microfluidic devices

SG-2506 borosilicate glass was purchased from Changsha Shaoguang Chrome Blank Co., Ltd. Normal human serum was purchased from Zhong Ke Chen Yu (Beijing) Trading Co., Ltd. Sylgard 184 elastomer base and curing agent for polydimethylsiloxane (PDMS) were purchased from Dow Corning (Midland, MI). The dechroming liquid was a mixture of 200 g L−1 ceric ammonium nitrate and 3.5% (v/v) glacial acetic acid. The etching solution contained 18.6 g L−1 NH4F, 4.64% (v/v) HNO3 and 5% HF. silver nitrate (AgNO3), sodium borohydride (NaBH4). Polyvinylpyrrolidone (PVP), sodium L-ascorbate (LSA), sodium citrate (SC) were ordered from Sigma-Aldrich, Co. LLC. Tween-20 was obtained from Nanjing Bookman Biotechnology Co. Ltd. Phosphate buffered saline (PBS) (Shanghai Sangon Biotechnology Co. Ltd.) was used for preparation of the following solutions: 1×PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 2 mM KH2PO4, PH 7.4), 1×PBSM (1×PBS + 1 mM MgCl2), BSAM (1×PBSM + 1 mg mL−1 BSA), Blocking solution (1×PBS + 10 mg mL−1 BSA). All the reagents were of analytical grade. All aqueous solutions were prepared using ultrapure water (≥ 18.20 MΩ) from a Millipore system. The oligonucleotides (Table 1 in supporting information) used in this work were synthesized and purified by Shanghai Sangon Biotechnology Co.

The Apt-AgNP array was designed and manufactured by robotic printing. Glass slide was cleaned with ethanol and deionized water and dried in a drying oven. Each slide had 4 identical array which consisted a total of 36 (4 × 9) spots. For multiplex protein assay, each array consists of three kinds of capture probes. Glass substrate and PDMS layer were plasma treated under the optimal parameters (705 w, 1 min) to obtain a hydrophilic surface. After plasma treatment, the slides were immediately spotted. PDMS architecture was designed and fabricated by standard soft lithography techniques. Briefly, the laser printing film with customized patterns was transferred to the borosilicate glass, and UV exposure was immediately used for 35 s, followed by rinsing for 1 min with NaOH (0.5% (w/v)) 5 min with dechroming liquid. For etching, the glass with micropatterns was fist etched by solution at 37 °C for 30 min and then rhomboidal chamber was protected by tape during the etching process. The resulting glass molds were then washed with ultrapure water and dried at 80 °C. A 10:1 mixture of the PDMS pre-polymer and the curing agent was degassed in a vacuum chamber and then poured onto the mold and cured at 80 °C for at least 1.5 h. After curing, the PDMS replica was removed from the mold to produce 5 mm thick PDMS layer. The PDMS layer was cut, punched, and plasma treated. After the plasma treatment, bonding between the PDMS layer and the glass slide were immediately performed.

aptamers bring great improvement of performance in molecular recognition [25,5]. Nanomaterials are used to increase the binding ability of aptamers [22,35] as well as amplify signal [14], especially amplifying fluorescence signal [44]. Our group have developed a series of methods based on aptamers and silver nanoparticles for sensitive detection of proteins [19–21,37,6]. However, it was the first time that we integrated aptasensor array, silver amplified signal and microfluidic devices into a new platform. Herein, we developed a rapid and sensitive aptamers-based sandwich assay in microfluidic devices based on multivalent aptasensor array (MAA) chip and silver aggregated amplification (SAA) strategy for the detection of two biomarkers at the same time. Based on MAA chip and SAA strategy in microfluidic device, the detection limit for PDGF-BB was 1.4 pg mL−1 in buffers and 7.8 pg mL−1 in 10% serum. Additionally, this assay was used to simultaneously detect PDGF-BB and VEGF-165 with good specificity and sensitivity.

2.2. Apparatus A microfluidic pump (Model LSP04-1A, Longer pump Corp., Baoding, China) was applied for liquid manipulation. Plasma Cleaner (Model PDCMG, Chengdu Ming Heng Science & Technology Co., Ltd., China) was applied to bond glass slide with PDMS piece. The aptamer-silver nanoparticles array was made by SmartArray 48 arrayer (Beijing CapitalBio Co. Ltd., China) and scanned by Luxscan-10K/A microarray Scanner (488 nm laser source for TMRNA, Beijing CapitalBio Co. Ltd., China). The images were collected and analyzed by LuxScan 3.0 software (CapitalBio Ltd., China). UV–Vis spectrum was performed on a Synergy Hybrid Reader (BioTek, USA). MIKRO 220 R refrigerated centrifuge from Hettich (Andreas Hettich GmbH & Co. KG, 78532 Tuttlingen, GERMANY) was used for centrifugation. Scanning electron microscopy (SEM) (S-4800, Japan) and Transmission electron microscope (TEM) (JEM-200CX, Japan) were used for collecting SEM and TEM images.

2.5. Analytical procedure The Apt-AgNP array were immobilized at 37 °C for 1.5 h. Then, 10 mg mL−1 BSA was used to block the nonspecific binding sites of AptAgNP array for 1 h followed by 3 min of washing step using 1xPBST at a flow rate of 25 μL min−1. After that, The MAA chips were ready to use or stored at 4 °C. The sample of proteins, Tag-A and Tag-B were injected into three inlets of the microfluidic device with a flow rate of 10 μL min−1 for 20 min followed by 3 min of washing step using 1xPBST at a flow rate of 20 μL min−1. Finally, A Luxscan-10K/A Microarray Scanner (λex = 532 nm, λem = 570, PMT = 560) was used to scan the slide and collect data. An MAA chip containing capture probe of PD-AgNP, VE-AgNP and PD&VE-AgNP was used for multiplex proteins detection.

2.3. Preparation of the aptamer-silver nanoparticle 3. Results and discussion Two different sizes of AgNPs (AgNPs-1 and AgNPs-2) were synthesized according to our previously reported method with some modification [18,20]. The AgNPs-1 functionalized with oligonucleotides of Apt-PD-TAMRA and hybrid-A to form detection probe (Tag-PD-A) according to our previous protocol [17]. And oligonucleotides of Apt-PD-

3.1. The principle of MAA chip and SAA strategy in microfluidic devices In this work, a new platform based on multivalent aptasensor array and silver aggregated amplification strategy in microfluidic devices was 418

Talanta 188 (2018) 417–422

X. Liu et al.

Fig. 1. (A) Apt-AgNP were dotted in array to form MAA chip. (B) PDMS layer with 3D rhomboidal chamber were bond to the MAA chip to form (C) microfluidic device. (D) Photograph of microfluidic device. (E) Schematic illustration of the SAA strategy with microfluidic devices for the detection of protein.

form capture probe (PD-AgNP) is about 91 nm (Fig. 2B). Fig. 2C represents the 30 times magnified of MAA chip whose average dimension is 420 ± 25 µm. Fig. 2D represents an array of MAA chip at 150 times magnified. We find that the diameter of the array is about 430 µm, and the silver nanoparticles show good distribution on glass substrate. Fig. 2E and Fig. 2F represent the SEM images of MAA chip before and after the detection of 10 ng mL−1 of PDGF-BB with microfluidic devices. The diameter of AgNPs-2 changed from 91 nm (inset image of Fig. 2E) to 208 nm (inset image of Fig. 2F) apparently, which reflects that Tag-A and Tag-B was hybridized with each other on Apt-AgNP. The aggregation of Tag-A and Tag-B increase the number of TAMRA and broaden of the AgNP aggregate LSPR peak. We hypothesized that the hybrid probes-proteins composites aggregated to the Apt-AgNP in a cooperative manner with a multivalent effect resulting in enhanced detection efficiency. As reported in many articles, multivalent binding could increase the binding ability of aptamers [13,24,34]. In this work, silver nanoparticles were used to load aptamers and amplify fluorescence signal. The target bound with Tag-A and Tag-B in a silver aggregated amplification (SAA) strategy could aggregate with each other on multivalent aptasensor array (MAA) chip in microfluidic device which could amplify fluorescence signal and increase the binding ability of aptamers to their targets. To generate sufficient ability of the platform, we had carried out a series of optimization studies involving flow time, flow rate, as well as concentration of Apt-AgNP, Tag-A and Tag-B. As shown in Fig. S4A, the fluorescence value increased from 5 min to 20 min but there is no significant increase from 20 min to 40 min. Taking both signal intensity and experimental time into considerations, the optimal fluorescence value was reached at a flow time of 20 min. The flow rate was also investigated as shown in Fig. S4B. The fluorescence signal firstly increased and then decreased. It demonstrated that rate that higher than 10 μL min−1 can reduced the capture efficiency of Apt-AgNP, resulting

developed. The produced PD-AgNP was dotted on glass slide to fabricate Apt-AgNP array on glass substrate (Fig. 1A) in microfluidic device for the capture of PDGF-BB. Then the plasma-treated PDMS layer was bonded with the Apt-AgNP array chip (Fig. 1B) to form a microfluidic device (Fig. 1 C). The microfluidic devices (Fig. 1D, Fig. S3) contain 4channels, and each channel had three inlets, a micro mixer [10], a 3D rhomboidal chamber, and an outlet. The rhomboidal channel was often acted as an important unit in microfluidic device. However, there is often poor fluid distribution, serious laminar flow especially in lower flow velocity with traditional rhomboidal channels. To solve the problems, we design a fluid separator with the height of 50 µm in the rhomboidal channels (Fig. 1B) to improve the fluid distribution and enhance mass transport. Thus, with the micro hybrid unit and 3D rhomboidal unit microfluidic device, the detection efficiency was significantly increased compared to conventional rhomboidal channel (Fig. S5). The standard deviation of fluorescence intensity was 8.6% of 10 ng mL−1 PDGF-BB detection in a microfluidic device with 3D rhomboidal channel (Fig. S5C, Fig. S5D), whereas 21.4% with conventional rhomboidal channel in microfluidic device (Fig. S5A, Fig. S5B). The principle for PDGF-BB detection was shown in Fig. 1 E. Firstly, Tag-A, Tag-B and PDGF-BB sample were simultaneously injected into the microfluidic device to from the sample inlets for on-line mixing. Then, MAA chip on 3D rhomboidal chamber recognized the target while Tag-A and Tag-B bound with targets and aggregated with each other to amplify fluorescence signal. The entire process of multiple sample mixing and recognition could be accomplished in 25 min. To better understand the reasons of the high sensitivity, we observe the SEM images of silver aggregated on Apt-AgNP array. The average diameter of AgNPs-1 which was functionalized with oligonucleotides of Apt-PD-TAMRA and hybrid-A or hybrid-B to form detection probe TagPD-A or Tag-PD-A is about 24.3 nm (Fig. 2A). The diameter of AgNPs-2 which was functionalized with Apt-PD and 5′SH-oligo(d)A12-NH2 to 419

Talanta 188 (2018) 417–422

X. Liu et al.

Fig. 2. SEM of AgNPs-2 (A) and AgNPs-1 (B) conjugated with aptamers, (Inset: particle diameter distribution pictures of AgNPs-2, AgNPs-1). (C) SEM of MAA chip with a magnification of 30 times (Scale bar: 1.00 mm). (D) SEM of MAA chip array with a magnification of 150 times (Scale bar: 300 µm). (E) SEM of MAA with a magnification of 40.0k times (Scale bar: 1.00 µm), (Inset: Apt-AgNP Apt-AgNP on MAA chip before assay). (F) SEM of MAA with a magnification of 40.0k times (Scale bar: 1.00 µm), (Inset: Apt-AgNP on MAA chip after SAA strategy for detection of 10 ng mL-1 PDGF-BB).

Fig. 3. (A) Fluorescence images of different PDGF-BB concentrations from 16 pg mL−1 to 250 ng mL−1. (B) The linear relationship between log F and log C. (C) Fluorescence images of different PDGF-BB concentrations from 16 pg mL−1 to 250 ng mL−1 with 10% blood serum. (D) The linear relationship between log F and log C.

in decreased fluorescence value. Therefore, 10 μL min−1 was chosen as optimal flow rate and was used in next experiments. The concentration of Apt-AgNP on the channels may have some significant effects on capture efficiency of target. Based on the formula C = CAg+ M/ ( 4 πr3ρNA), the concentration of initial AgNPs-1 and AgNPs-2 was cal3 culated as 4.83 nM and 0.17 nM. Fig. S1 shows the UV–vis spectrum obtained from AgNPs-1(b1), AgNPs-2 (a1), Tag-A(b2), Tag-B(b3) and Apt-AgNP (a2). It is illustrated that the peak absorbance wavelength of AgNPs-1 and AgNPs-2 had a slight spectral red shift from 425 nm to 429 nm and 494 nm to 499nm separately, suggesting that the aptamers, which may affect the silver surface plasmon resonance, has been

successfully modified onto the surface of AgNPs-1, AgNPs-2. And according to the Beer–Lambert law, the concentration of Tag-A and Tag-B was calculated as 4.09 nM and 3.81 nM by the ratio of maximum absorbance of Tag-A, Tag-B to AgNPs-1. The concentration of Apt-PDAgNP was calculated as 0.13 nM by the ratio of maximum absorbance of Apt-PD-AgNP to AgNPs-2. Low concentration of Apt-PD-AgNP is not enough for providing enough aptamers to capture target protein hence reducing the fluorescence signal. As shown in Fig. S4C, different concentrations of Apt-PD-AgNP were investigated, and the fluorescence signal increased when the concentration of Apt-PD-AgNP was increased from 3.9 nM to 23.4 nM, but then decreased a little. Thus, 23.4 nM was 420

Talanta 188 (2018) 417–422

X. Liu et al.

detect PDGF-BB in 10% blood serum and the fluorescence images were shown in Fig. 3C. A linear response to PDGF-BB (r = 0.992) was obtained in the concentration range from 16 pg mL−1 to 250 ng mL−1, and the detection limit was 7.8 pg mL−1 (Fig. 3D). Our assay showed a competitive sensitivity when it was compared with other optical approaches for detection of PFGD-BB (Table 2 in supporting information), such as FRET sensor [21], fluorescence [36,37,48], SERS [11,40], colorimetric [41], and luminescence [26]. In a word, the high sensitivity of our assay was due to three ways, including silver aggregated amplification, multivalent binding from aptamer-modified silver nanoparticles, and enhanced mass transport capability inherent from microfluidic devices. 3.3. Specificity for detection of PDGF-BB Fig. 4. Detection selectivity and specificity of the MAA microfluidic device for 10 ng mL−1 PDGF-BB in 10% human serum, 10 ng mL−1 PDGF-BB, 10 ng mL−1 PDGF-AA, 50 μg mL−1 IgG, 1 mg mL−1 BSA. The inset is the fluorescence images of PDGF-BB and other proteins.

To investigate the specificity of the target protein in this assay, potential interfering substances: 10 ng mL−1 PDGF-AA，50 μg mL−1 IgG, 1 mg mL−1 BSA, and 10 ng mL−1 PDGF-BB + 10% Human serum were used as controls (Fig. 4). As we knew, PDGF-AA, and PDGF-BB were different dimeric isoforms of PDGF; however, only weak fluorescence signal could be observed in the presence of PDGF-AA. Whereas the fluorescence signals of the 50 μg mL−1 IgG, and 1 mg mL−1 BSA, were similar to the blank even though their concentration was thousand times higher than that of target protein, indicating no interference from these compounds. As for 10 ng mL−1 PDGF-BB + 10% Human serum, the signal was only a little lower than the signal of 10 ng mL−1 PDGFBB. These results indicated that the MAA chip and SAA strategy in microfluidic device possessed an excellent selectivity and anti-interference ability even in complex media. Results suggested that the assay had an excellent selectivity and specificity towards PDGF-BB, and the assay would have great potential in analytical areas.

taken as optimal concentration. The concentration of Tag-A and Tag-B also have some significant effects on fluorescence signal amplification. As shown in Fig. S4D, different concentrations of Tag-A and Tag-B were investigated. Fluorescence signal increased with the increased concentration of Tag-A and Tag-B and then almost kept constant when the concentration was larger than 1.2 nM. Taking the experimental cost into considerations, 1.2 nM was chosen as the optimal concentration and was used in next experiments.

3.2. Analytical performance for detection of PDGF-BB Based on the MAA chip and SAA strategy in microfluidic device, a series of different concentrations of PDGF-BB were detected in microfluidic platform under the optimal conditions. The fluorescence images were shown in Fig. 3A. As a result, a linear response to PDGF-BB (r = 0.999) was obtained in the concentration range from 16 pg mL−1 to 250 ng mL−1, and the detection limit was 1.4 pg mL−1 (Fig. 3B). To prove the high sensitivity and good specificity of our assay, we also

3.4. Multiplexed analysis of MAA and SAA in microfluidic devices Finally, we proved the multiplex capabilities of our assay for simultaneous detection of two types of proteins: VEGF-165 and PDGF-BB. As shown in Fig. 5A, The MAA chip was dotted with the four-different Apt-AgNP, (a): VE-AgNPs; (b): PD-AgNP; (c): VE&PD-AgNP; (d): A12-

Fig. 5. (A) Schematic illustration of SAA chip with different Apt-AgNP ((a): VE-AgNPs; (b): PD-AgNP; (c): VE&PD-AgNP; (d): A12-AgNP(Control) dotted for multiplex biomarkers (VEGF-165, PDGF-BB) detection at the same time. Fluorescent images (B) and fluorescence intensity (C) of multiplex biomarkers (VEGF-165, PDGF-BB) detection. 421

Talanta 188 (2018) 417–422

X. Liu et al.

AgNP (Control). Each channel had 3 different aptasensor arrays. In the first experiment, we introduced VEGF-165 (10 ng mL−1) and PDGF-BB (10 ng mL−1) sample to the channel 1, channel 2. The array 1 and array 3 in channel 1 indicated the fluorescence signal of VEGF-165, and array 2 in channel 1 indicated the fluorescence signal of background (Fig. 5B channel 1). The results showed that only VEGF-165 had fluorescence in channel 1. The results were the same for PDGF-BB in channel 2 (Fig. 5B channel 2). We could observe a highly positive fluorescence signal of protein coming from their correspond Apt-AgNP array, while there is no cross-reactivity with the other array. Next, we introduced a mixture sample with both VEGF-165 and PDGF-BB at 10 ng mL−1 to the channel 3. As expected, we could observe the fluorescence signal coming from both of array 1 and array 2 instead of array 3 (Fig. 5B channel 3). The results demonstrate that we could selectively detect both VEGF-165 and PDGF-BB simultaneously in single channel with no cross-reactivity. Furthermore, our platform was not limited to simultaneous detection of two proteins. It could be expanded to rapid, sensitive and highthroughput detection of other protein biomarkers by coupling of various aptamers with nanoparticles in microfluidic devices.

[5] F. Charbgoo, F. Soltani, S.M. Taghdisi, K. Abnous, M. Ramezani, TrAC-Trend Anal. Chem. 85 (2016) 85–97. [6] Z. Chen, H. Li, W. Jia, X. Liu, Z. Li, F. Wen, N. Zheng, J. Jiang, D. Xu, Anal. Chem. 89 (2017) 5900–5908. [7] M. Diaz-Gonzalez, J.P. Salvador, D. Bonilla, M.P. Marco, C. Fernandez-Sanchez, A. Baldi, Biosens. Bioelectron. 74 (2015) 698–704. [8] E.W. Esch, A. Bahinski, D. Huh, Nat. Rev. Drug Discov. 14 (2015) 248–260. [9] G. Grasso, A.M. Santoro, A. Magri, D. La Mendola, M.F. Tomasello, S. Zimbone, E. Rizzarelli, J. Inorg. Biochem. 159 (2016) 149–158. [10] B. He, B.J. Burke, X. Zhang, R.J. Zhang, F.E. Regnier, Anal. Chem. 73 (2001) 1942–1947. [11] J. He, G. Li, Y. Hu, Anal. Chem. 87 (2015) 11039–11047. [12] L. Hong, F. Zhou, D. Shi, X. Zhang, G. Wang, Biosens. Bioelectron. 95 (2017) 152–159. [13] R.C. Huang, W.J. Chiu, I. Po-Jung Lai, C.C. Huang, Sci. Rep. 5 (2015) 10292. [14] X. Jia, S. Dong, E. Wang, Biosens. Bioelectron. 76 (2016) 80–90. [15] M. Kimoto, M. Nakamura, I. Hirao, Nucleic Acids Res. 44 (2016) 7487–7494. [16] H.J. Lee, A.W. Wark, R.M. Corn, Analyst 133 (2008) 975–983. [17] H. Li, W. Qiang, M. Vuki, D. Xu, H.Y. Chen, Anal. Chem. 83 (2011) 8945–8952. [18] H. Li, Z.Y. Sun, W.Y. Zhong, N. Hao, D.K. Xu, H.Y. Chen, Anal. Chem. 82 (2010) 5477–5483. [19] H. Li, M. Wang, W. Qiang, H. Hu, W. Li, D. Xu, Analyst 139 (2014) 1653–1660. [20] H. Li, M. Wang, C. Wang, W. Li, W. Qiang, D. Xu, Anal. Chem. 85 (2013) 4492–4499. [21] H. Li, Y. Zhao, Z. Chen, D. Xu, Biosens. Bioelectron. 87 (2017) 428–432. [22] H. Liang, X.B. Zhang, Y. Lv, L. Gong, R. Wang, X. Zhu, R. Yang, W. Tan, Acc. Chem. Res. 47 (2014) 1891–1901. [23] X. Liu, H. Li, W. Jia, Z. Chen, D. Xu, Lab Chip 17 (2016) 178–185. [24] Z. Liu, H. He, Acc. Chem. Res. 50 (2017) 2185–2193. [25] D. Lu, L. He, G. Zhang, A. Lv, R. Wang, X. Zhang, W. Tan, Nanophotonics 6 (2017) 109–121. [26] L. Lu, Z. Mao, T.S. Kang, C.H. Leung, D.L. Ma, Biosens. Bioelectron. 85 (2016) 300–309. [27] E.W. Ng, D.T. Shima, P. Calias, E.T. Cunningham Jr., D.R. Guyer, A.P. Adamis, Nat. Rev. Drug Discov. 5 (2006) 123–132. [28] L. Pi, P.Y. Chung, S. Sriram, M.M. Rahman, W.Y. Song, E.W. Scott, B.E. Petersen, G.S. Schultz, World J. Biol. Chem. 6 (2015) 379–388. [29] G.F. Pierce, J.E. Tarpley, J. Tseng, J. Bready, D. Chang, W.C. Kenney, R. Rudolph, M.C. Robson, J. Vande Berg, P. Reid, et al., J. Clin. Investig. 96 (1995) 1336–1350. [30] A.E. Saliba, I. Vonkova, S. Deghou, S. Ceschia, C. Tischer, K.G. Kugler, P. Bork, J. Ellenberg, A.C. Gavin, Nat. Protoc. 11 (2016) 1021–1038. [31] S. Sathish, S.G. Ricoult, K. Toda-Peters, A.Q. Shen, Analyst 142 (2017) 1772–1781. [32] M. Shamsipur, L. Farzin, M. Amouzadeh Tabrizi, F. Molaabasi, Biosens. Bioelectron. 74 (2015) 369–375. [33] W. Song, H. Li, H. Liang, W. Qiang, D. Xu, Anal. Chem. 86 (2014) 2775–2783. [34] N. Sun, M. Liu, J. Wang, Z. Wang, X. Li, B. Jiang, R. Pei, Small 12 (2016) 5090–5097. [35] M. Vorobyeva, P. Vorobjev, A. Venyaminova, Molecules 21 (2016) 1631–1633. [36] G. Wang, Y. Zhu, L. Chen, X. Zhang, Biosens. Bioelectron. 63 (2015) 552–557. [37] Y. Wang, H. Li, D. Xu, Anal. Chim. Acta 905 (2016) 149–155. [38] Y. Wang, D. Xu, H.Y. Chen, Lab Chip 12 (2012) 3184–3189. [39] K. Yamada, H. Shibata, K. Suzuki, D. Citterio, Lab Chip 17 (2017) 1206–1249. [40] S. Ye, X. Zhai, Y. Wu, S. Kuang, Biosens. Bioelectron. 79 (2016) 130–135. [41] H. Zhang, F. Li, H. Chen, Y. Ma, S. Qi, X. Chen, L. Zhou, Sens. Actuators B 207 (2015) 748–755. [42] L. Zhang, S. Wan, Y. Jiang, Y. Wang, T. Fu, Q. Liu, Z. Cao, L. Qiu, W. Tan, J. Am. Chem. Soc. 139 (2017) 2532–2540. [43] Z. Zhang, C. Guo, S. Zhang, L. He, M. Wang, D. Peng, J. Tian, S. Fang, Biosens. Bioelectron. 89 (2017) 735–742. [44] T. Zhao, T. Li, Y. Liu, Nanoscale 9 (2017) 9841–9847. [45] J. Zhou, J. Rossi, Nat. Rev. Drug Discov. 16 (2017) 181–202. [46] W. Zhou, D.Q. He, J.Y. Liu, Y. Feng, X.Y. Zhang, C.G. Hua, X.F. Tang, J. Tissue Eng. Regen. Med. 9 (2015) 1404–1416. [47] D. Zhu, W. Li, H.M. Wen, S. Yu, Z.Y. Miao, A. Kang, A. Zhang, Biosens. Bioelectron. 74 (2015) 1053–1060. [48] D. Zhu, R.X. Yang, Y.P. Tang, W. Li, Z.Y. Miao, Y. Hu, J. Chen, S. Yu, J. Wang, C.Y. Xu, Biosens. Bioelectron. 85 (2016) 429–436.

4. Conclusion The assay based on multivalent aptasensor array (MAA) chip and silver aggregated amplification (SAA) strategy in microfluidic devices is a rapid, sensitive and high-throughput platform for detection of proteins. The assay showed excellent sensitivity and specificity even in complex medium of serum. The assay offered a wide linear range for detection of PDGF-BB and the detection limit was 1.4 pg mL−1 in buffers and 7.8 pg mL−1 in 10% serum. Additionally, this assay was also used to simultaneously detect PDGF-BB and VEGF-165 with good selectivity. Further more, the assay can also be expanded to other protein biomarkers by coupling of various aptamers with nanoparticles. Acknowledgments We acknowledge the financial support of National Natural Science Foundation of China (Grant nos. 21775068, 21475060 and 21405077), Natural Science Foundation of Jiangsu Province (BK20140591), and the Program B for Outstanding Ph.D. candidate of Nanjing University (201701B014). Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.talanta.2018.05.048. References [1] S. Ahmed, M.P. Bui, A. Abbas, Biosens. Bioelectron. 77 (2016) 249–263. [2] A. Atak, S. Mukherjee, R. Jain, S. Gupta, V.A. Singh, N. Gahoi, P.M. K, S. Srivastava, Proteomics 16 (2016) 2557–2569. [3] N. Cennamo, M. Pesavento, L. Lunelli, L. Vanzetti, C. Pederzolli, L. Zeni, L. Pasquardini, Talanta 140 (2015) 88–95. [4] H.N. Chan, M.J.A. Tan, H. Wu, Lab Chip 17 (2017) 2713–2739.

422