Sensors & Actuators: B. Chemical 288 (2019) 266–273
Contents lists available at ScienceDirect
Sensors and Actuators B: Chemical journal homepage: www.elsevier.com/locate/snb
Double-layered microfluidic paper-based device with multiple colorimetric indicators for multiplexed detection of biomolecules
T
⁎
Fang Li , Xu Wang, Jiachang Liu, Yuting Hu, Jianbo He Anhui Province Key Laboratory of Advanced Catalytic Materials and Reaction Engineering, School of Chemistry and Chemical Engineering, Hefei University of Technology, Hefei, Anhui 230009, People’s Republic of China
A R T I C LE I N FO
A B S T R A C T
Keywords: Microfluidic paper-based device Glucose Uric acid Lactate Choline
In this work, a double-layered microfluidic paper-based device (μPAD) with multiple colorimetric indicators was presented for the sensitive simultaneous detection of four kinds of small biomolecules, including glucose, uric acid, lactate, and choline in mixture. The double-layered μPAD were composed of the top detection layer and the bottom auxiliary layer to construct 3D microfluidic channels. The detection zones on the top detection layer were modified with various colorimetric reagents, different kinds of oxidase (glucose oxidase, uricase, lactate oxidase or choline oxidase) and horseradish peroxidase (HRP). The immobilized chromogenic reagents could be oxidized by the generated H2O2 from the specific enzymatic reactions between the oxidases and their corresponding substrates to form different colored products with co-immobilized HRP as catalyst. As a result, four linear calibration curves were obtained for the determination of glucose (0.01–10.0 mmol/L), uric acid (0.01–5.0 mmol/L), lactate (0.04–10.0 mmol/L), and choline (0.04–24.0 mmol/L), respectively. The proposed double-layered μPAD with multiple colorimetric indicators exhibited good colorimetric performance, effectively improved sensitivity, and extended linear range, which was further applied to the simultaneously determination of glucose, uric acid, lactate, and choline in serum samples.
1. Introduction Recently, the development of simple, low-cost and practical diagnostic tools for point-of-care tests (POCT) have attracted lots of interests in various areas such as health care, food safety, and environment monitoring [1–3]. Paper is cheap, sustainable, abundant, biocompatible, easy to store, transport, use and modification, and has been extensively used as platform to fabricate detection tools for POCT and onsite diagnosis [4,5]. The most typical examples of paper-based detection tools are paper test strips for pregnancy and diabetes. Paper-based detection tools have attracted much attention due to its simplicity, lowcost, disposability and minimal sample requirement, but meanwhile suffer from some disadvantages. For example, most paper-based detection tools are “yes/no” or semi-qualitative detections which lack sensitivity, accuracy and cannot be used for multiplexed analysis [6]. Multiplexed analysis could not only save reagent, time, cost and labor, but also be benefit to give a comprehensive and reliable evaluation result for a single sample. Accordingly, microfluidic paper-based analytical devices (μPADs), combining the advantages of paper test strips and the utility of conventional lab-on-chip devices was developed by Whitesides’ group in 2007 [7]. The porous cellulose paper substrates of
⁎
μPADs could be patterned into desirable hydrophilic microchannels separated by hydrophobic walls to enable multiple distribution of samples for multiplexed analysis on one single microfluidic device. The hydrophilic microchannels and hydrophobic walls could be made by a variety of methods such as inkjet printing [8], screen printing [9], photolithography [10], laser treatment [11], and by cutting [12]. By combining sample separation, transportation, specific reaction, and detection together on one minimized device, the newly developed μPADs promoted the development of the next generation of paperbased analytical devices, which hold great application potential for POCT and on-site diagnosis [13]. Various μPADs integrated with different detection methods such as colorimetric [14], electrochemical [15], fluorescence [16] and chemiluminescence [17] have been established. Particularly, colorimetric detection methods have attracted lots of attentions due to its simplicity, versatility, straightforward detection results and applicability for instrument-free POCT analysis. Visual colorimetric analysis could be achieved by naked eyes [18], and more sensitive quantitative colorimetric analysis could be achieved based on the detection of color intensity by using professional colorimeters [19]. In addition, with the rapid development and widespread of smartphones, smartphones
Corresponding author. E-mail address:
[email protected] (F. Li).
https://doi.org/10.1016/j.snb.2019.02.116 Received 16 August 2018; Received in revised form 20 January 2019; Accepted 26 February 2019 Available online 27 February 2019 0925-4005/ © 2019 Elsevier B.V. All rights reserved.
Sensors & Actuators: B. Chemical 288 (2019) 266–273
F. Li, et al.
detection of each analyte. Finally, the practical application ability of the proposed colorimetric μPADs were tested by the detection of glucose, uric acid, lactate, and choline in serum samples.
embedded with camera function are regarded as convenient, portable and inexpensive detection devices for paper-based colorimetric assays. Hence, the development of colorimetric μPADs based assays combined with smartphone as detector have attracted increasing attentions [14,20]. Although noticeable improvements have been achieved, μPAD based colorimetric assays still suffer from certain restrictions. The major problem associated with μPADs based colorimetric assay is the unsatisfied color uniformity, which is mainly caused by the randomly diffusion of the colorimetric reagents outside of the detection zones, leading to the increase of experimental errors [21]. Recently, a rotational valves-based μPADs was reported for the simultaneous detection of Ni(II), Cu(II) and Cr(VI), which could reduce the diffusion of the colorimetric reagent outside of the detection zones by connection or disconnection of the detection zones with the fluidic channels [22]. However, the detection method was complex involving several rotating steps. Furthermore, the sensitivity of this method was not good and the detection limits were high. Another problem associated with μPADs based colorimetric assay is the complicated influence factors, including differences in ambient light and differences in color perceptions. In our previous work, a simple method by putting the test paper on the top of a flat panel lamp was used to provide constant illumination and prevent influence from ambient light. This method was proved to be able to enhance the distinguishability of the detection zones and effectively increase the assay accuracy and sensitivity [23]. In addition, it has been reported that greater colorimetric discrimination could be achieved when more than one color with different hues or intensities is presented as opposed to one single color [18]. Then, a μPADs based semi-quantitative colorimetric assay with multiple indicators was developed for the simultaneous detection of glucose, lactate, and uric acid [18]. Improvement in accuracy was obtained by the comprehensive evaluation of the results from multiple indicators for each analyte. However, the color uniformity in this report was poor. Furthermore, the detection limits of this semi-quantitative measurement were high that are 0.5, 1 and 0.1 mM for glucose, lactate, and uric acid detection, respectively. Thus, there is still an urgent need for the development of μPADs with good colorimetric performance, high sensitivity and good reproducibility for multiplexed analysis. Glucose, uric acid, lactate and choline are four kinds of important clinical small biomolecules, which all play critical roles in the metabolic processes. The abnormal of blood glucose is mainly associated with hypoglycemia and diabetes. The accumulation of uric acid is the major cause of gout, and the accumulation of lactate can cause muscle soreness, tiredness, headaches, and physical acidification. Meanwhile, inadequate of choline can affect cognitive decline, increase the risk of Alzheimer’s disease and affect infant's growth and development. The simultaneous determination of glucose, uric acid, lactate and choline can provide important information for health evaluation and early diagnosis of several diseases. In this work, a double-layered μPAD with multiple colorimetric indicators were developed for the sensitive simultaneous detection of glucose, uric acid, choline and lactate in mixture. The double-layered μPAD were composed of the top modified detection layer and the bottom unmodified auxiliary layer to construct three-dimensional (3D) microfluidic channels. Multiplexed analysis could be achieved by distributing sample mixtures from central sampling zone to multiple colorimetric detection zones through circuitous flowing route design by virtue of the 3D microfluidic channels. Hydrogen peroxide (H2O2) generated from the specific enzymatic reactions between the oxidases (including glucose oxidase, uricase, lactate oxidase and choline oxidase) and their corresponding substrates could oxidize the co-immobilized chromogenic reagents to form colored products with co-immobilized HRP as catalyst. The detection layer of the μPAD was put on top of a flat LED lamp. Then, the photographs of the detection layer were acquired by a smartphone. The average color intensities from two corresponding detection zones immobilized with different colorimetric indicators were measured for the quantitative
2. Experimental section 2.1. Materials Chitosan powder, glucose, uric acid, choline, and paraffin (melting points 48–50 °C) were bought from Shanghai Sinopharm Chemical Reagent CO. (Shanghai, China). Whatman chromatography paper (WCP#1, 20 cm × 20 cm), lactate, HRP, glucose oxidase, and uricase were bought from Beijing Solarbio Biotechnology Co. Ltd. (Beijing, China). Whatman grade 1 paper was size adjusted to 5 cm × 5 cm before use. Lactate oxidase and choline oxidase were purchased from Sigma-Aldrich (St. Louis, USA). 3,3′,5,5′-tetramethylbenzidine (TMB), 4-amino antipyrine (4-AAP), 3,5-dichloro-2-hydroxy acid sodium (DHBS), and N-Ethyl-N-(3-sulfopropyl)-3-methyl-aniline sodium salt (TOPS) were bought from Aladdin (Shanghai, China). All other reagents were of analytical grade. Ultrapure water was used throughout. 2.2. Design and fabrication of the double-layered μPAD Firstly, Auto CAD software (Autodesk Inc.) was used to design the pattern of the μPAD. The double-layered μPAD were composed of two layers of papers, including the top detection layer and the bottom auxiliary layer, which were in different patterns. As shown in Fig. S1A, the top detection layer consisted of one central sampling zone (10 mm in diameter) surrounded by eight microfluidic channels (3 mm in width, 8 mm in length) and eight detection zones (6 mm in diameter). Particularly, the eight detection zones did not connect with the eight microfluidic channels directly and they were separated by 2 mm. As shown in Fig. S1B, the bottom auxiliary layer consisted of one central sampling zone (10 mm in diameter) and eight microfluidic channels (3 mm in width, 10 mm in length) directly connected with eight smaller sampling zones (6 mm in diameter). Interestingly, the central sampling zone and the eight detection zones on the top detection layer were in the same shape with the central sampling zone and the eight small sampling zones on the bottom auxiliary layer. Thus, the bottom auxiliary layer could provide a solution connection between the microfluidic channels and the detection zones on the top detection layer by overlapping the two layers in alignment to construct 3D microfluidic channels, as shown in Fig. S1C. The hydrophilic microchannels and hydrophobic barrier on the detection layer and auxiliary layer were fabricated by using a traditional wax-screen-printing technology similar to our previous work [23]. The screen stencils were made by 200 mesh of nylon mesh printed with impenetrable photosensitive emulsion according to the designed μPAD patterns and were produced at a local printing shop. Then, the screen stencil was closely placed on top of the Whatman paper, and then brushed with a piece of paraffin. The paraffin penetrated the pores of the screen stencil, and form wax patterns on the surface of the paper. Then the patterned papers were put in the oven set at 100 °C and kept for 25 s, to allow the wax penetrating into the paper, forming hydrophobic barriers. The fabricated auxiliary layer was ready for use, while the detection layer need further modification. 2.3. Modification of the detection layer Firstly, 3.0 μL 1 mg/mL chitosan solution prepared in acetic acid (0.25%, v/v) were added to the eight detection zones. Then 3.0 μL chromogenic reagent were further added to the eight detection zones. The chromogenic reagents used here include TMB, mixture of 4-AAP and TOPS (4-AAP/TOPS), and mixture of 4-AAP and DHBS (4-AAP/ DHBS). The detection zones on the detection layer were numbered 1–8 as shown in Fig. S1A. Position numbered 1 and 6 were modified with 4267
Sensors & Actuators: B. Chemical 288 (2019) 266–273
F. Li, et al.
Fig. 1. Schematically illustration of the fabrication of the double-layered μPADs with multiple indicators for the simultaneous detection of four analytes.
μPAD was investigated to demonstrate the multiplexed sample distribution ability of the 3D μPAD. As shown in Fig. S2, after the addition of 120 μL blue dye solution to the central sampling zone, the blue solution flowed quickly outward to the eight microfluidic channels and detained in the microfluidic channels. Then the detained blue solution flowed slowly from the top layer downward to the microfluidic channels on the bottom auxiliary layer and continued flowing to the eight small sampling zones on the bottom auxiliary. Finally, the blue solution diffused upward from the eight small sampling zones on the bottom auxiliary to the eight detection zones on the top detection layer via capillary forces in 150 s and detained in situ. It demonstrated that the hydrophilic microchannels and hydrophobic barrier of the doublelayered μPAD were successfully fabricated. It was found that 120 μL solution was just enough to fully fill the detection zone with no more flowing solution. Thanks to the support of the clamps, there was space between the bottom auxiliary layer and the desktop to prevent contamination of the μPAD. Interestingly, the top detection layer could cling closely to the bottom auxiliary layer after the diffusion of the solution from the top layer to the bottom layer due to the surface tension effect. The self-adhesion of the two layers could create an unimpeded 3D microfluidic channels for sample migration. Thus, the solution added to the central sampling zone on the detection layer could migrated to the eight detection zones smoothly through the 3D microfluidic channels, which could prevent the random diffusion of reagents outside of the detection zones.
AAP/DHBS, 2, 4 and 8 were modified 4-AAP/TOPS, 3, 5 and 7 were modified with TMB. Then, 3.0 μL mixture of HRP and a specific oxidase dissolved in pH 7.0 phosphate buffer were further added to the detection zones. The positions numbered 1 and 2 were modified with glucose oxidase for glucose detection, 3 and 4 were modified with uricase for uric acid detection, 5 and 6 were modified with choline oxidase for choline detection, 7 and 8 were modified with lactate oxidase for lactate detection. The concentrations of chromogenic reagents, HRP and oxidase enzymes on each detection zone were optimized in the following experiments. After the modification procedures, the detection layer was ready for the further use. 2.4. Procedures for multiplexed colorimetric detection The modified detection layer and the auxiliary layer were overlapped in alignment and fixed in tight contact with four clamps to obtain the double-layered 3D μPAD. Then, the fabricated μPAD was ready for the multiplexed colorimetric determination of glucose, uric acid, choline, and lactate in mixture. Firstly, standard solutions containing series concentrations of glucose, uric acid, choline, and lactate dispersed in pH 7.0 phosphate buffer and a blank solution containing only the buffer were prepared. Then, 120 μL sample solution was added to the central sampling zone on the detection layer and the solution flowed outward to the eight microfluidic channels, and then diffused downward from the top layer to the bottom layer, finally diffused upward from the eight small sampling zones on the bottom layer to the eight detection zones on the top detection layer via capillary forces to initiate the colorimetric reactions. The μPAD was kept at room temperature for 15 min. Then, the top detection layer of the μPADs was placed on top of a flat LED lamp and the color photographs of the detection layer were acquired by an iPhone 6 smartphone. The color intensities represented by Gray-scale value, Red-channel value, Green-channel value, and Blue-channel value of interested regions located at the center of the detection zones while avoiding the edge of the detection zones were measured by Image J software. The color intensity represented by hue value were measured by Photoshop software (Adobe Inc.). ΔGray-scale value, ΔRed-channel value, ΔGreechannel value, and ΔBlue-channel value were calculated by subtraction the Gray-scale value, Red-channel value, Green-channel value and Bluechannel value acquired in the presence of blank buffer with that acquired in the presence of the analytes, respectively. The relationship between the concentrations of the four analytes and the color intensities represented by ΔGray-scale value, ΔRed-channel value, ΔGreen-channel value, ΔBlue-channel value and hue value, respectively, were investigated. It was found that the ΔGray-scale value of the detection zones exhibited the best correlation ship with the concentrations of the four analytes. Thus, based on the relationships between the ΔGray-scale values and the concentrations of the analytes, glucose, uric acid, choline, and lactate were quantitatively analyzed.
3.2. Principle of the double-layered μPAD for multiplexed analysis Fig. 1 schematically illustrated the fabrication of the double-layered 3D μPAD for multiplexed detection of the four analytes. Firstly, each detection zone on the detection layer was modified with chitosan, one kind of chromogenic reagent (TMB, 4-AAP/DHBS, or 4-AAP/TOPS), and HRP mixed with one kind of oxidase (glucose oxidase for glucose detection, uricase for uric acid detection, choline oxidase for choline detection, or lactate oxidase for lactate detection). Chitosan is a biocompatible substance with high specific surface area and good filmforming ability [24]. The added chitosan could form a thin film on the porous structure of cellulose paper through electrostatic interactions and physical absorption. The chitosan film on cellulose fibers could provide a better support to adsorb enzymes and chromogenic substrates, and construct a biocompatible microenvironment to maintain the reactivity of enzymes [25]. It has been reported that different detection limits and detection ranges could be obtained by using different chromogenic reagents as indicators [23,25]. Thus, every two kinds of optimized chromogenic reagents was used in combination as comprehensive indicator for each analyte in order to obtain the widest detection range with lowest detection limit. The chromogenic reagent used on detection zone numbered 1 and 2 was 4-AAP/DHBS and 4-AAP/ TOPS, respectively, for glucose detection, 3 and 4 was TMB and 4-AAP/ TOPS, respectively, for uric acid detection, 5 and 6 was TMB and 4AAP/DHBS, respectively, for choline detection, 7 and 8 was TMB and 4AAP/TOPS, respectively, for lactate detection. Then, the modified detection layer and the auxiliary layer were overlapped in alignment to form the double-layered 3D μPAD. After the addition of sample solution containing the four target biomolecules to the central sample zone of the double-layered μPAD, the sample solution was automatically
3. Results and discussion 3.1. Behavior of solution flow on this double-layered 3D μPAD In this work, a double-layered 3D μPAD was designed for multiplexed analysis. The behavior of solution flow on this double-layered 268
Sensors & Actuators: B. Chemical 288 (2019) 266–273
F. Li, et al.
reactions between the chromogenic reagents and generated H2O2 from oxidase enzymatic reactions catalyzed by HRP. Thus, the important detection conditions including the concentrations of chromogenic reagents, oxidase and HRP on each detection zone were optimized. For glucose detection, on detection zone numbered 1, the mixing ratio of 4AAP/DHBS, the concentrations of 4-AAP/DHBS, HRP and glucose oxidase were optimized (Fig. S3A–D). On detection zone numbered 2, the mixing ratio of 4-AAP/TOPS, the concentrations of 4-AAP/TOPS, HRP and glucose oxidase were optimized (Fig. S4A–D). For uric acid detection, on detection zone numbered 3, the concentrations of TMB, HRP and uricase were optimized (Fig. S5A–C). On detection zone numbered 4, the mixing ratio of 4-AAP/TOPS, the concentrations of 4-AAP/TOPS, HRP and uricase were optimized (Fig. S6A–D). For choline detection, on detection zone numbered 5, the concentrations of TMB, HRP and lactate were optimized (Fig. S7A–C). On detection zone numbered 6, the mixing ratio of 4-AAP/DHBS, the concentrations of 4-AAP/DHBS, HRP and choline oxidase were optimized (Fig. S8A–D). For lactate detection, on detection zone numbered 7, the concentrations of TMB, HRP and lactate oxidase were optimized (Fig. S9A–C). On detection zone numbered 8, the mixing ratio of 4-AAP/TOPS, the concentrations of 4-AAP/ TOPS, HRP and lactate oxidase were optimized (Fig. S10A–D). The optimized detection conditions as mentioned above were listed in Table S1.
divided into eight branched, and automatically migrated to their respective detection zones through circuitous flowing route by virtue of the 3D microfluidic channels to initiate the colorimetric reactions. The generated H2O2 from the specific enzymatic reactions could oxidase the immobilized chromogenic reagents with co-immobilized HRP as catalyst. Afterward, the detection zones immobilized with TMB, 4-AAP/ DHBS, and 4-AAP/TOPS changed from colorless to green-blue, red, and purple, respectively. Then, each analyte was quantified by the comprehensive evaluation of the colorimetric responses from two independent detection zones immobilized with different chromogenic reagents. In this work, double-layered μPAD with 3D microfluidic channel design was fabricated in order to improve the colorimetric performance. To prove it, one-layered 2D μPAD with eight microfluidic channels directly connected with eight detection zones (the same structure as the bottom auxiliary layer) which were commonly used in previous reports was fabricated [18,25]. Firstly, the detection zones of the one-layered μPAD were modified with chitosan, different kinds of chromogenic reagents, HRP and different kinds of oxidases according to the modification procedures of the detection layer of the double-layered μPAD. Then, the colorimetric response of the one-layered μPAD in the presence of 0.2 mM glucose, 0.15 mM uric acid, 0.4 mM choline, and 0.8 mM lactate were measured. The sample solution added to the sampling zone of the one-layered μPAD could flow directly to the eight detection zones through the microfluidic channels to initiate their respective colorimetric reactions. As shown in Fig. 2A, the eight detection zones of the one-layered μPAD exhibited colorfully responses. However, the colorimetric reagents on the one layered μPADs could randomly diffused outside the detection zones into the microfluidic channels, which might bring great experimental error to the quantification results. In comparison, the color of the detection zones of the doublelayered μPAD were quite uniform without reagents diffusion out (Fig. 2B). It demonstrated that the proposed double-layered 3D μPAD could effectively prevent the uncontrolled flow of colorimetric reagents outside the detection zones by disconnecting the detection zones and the microfluidic channels and transporting the sample solution to the detection zones indirectly through 3D microfluidic channels. The RSD of the color intensities were calculated to be below 9.54% for the commonly used one-layered μPAD and only below 2.56% for the fabricated double-layered μPAD in three repeated measurements at the eight detection zones. Thus, the difference in color intensities on the double-layered 3D μPAD was much smaller than that on the onelayered 2D μPAD, indicating that the double-layered 3D μPAD could effectively improve the reproducibility.
3.4. Multiplexed colorimetric detection of four small biomolecules Under the optimal assay conditions, the analytical performance of this double-layered μPAD for the simultaneous colorimetric detection of glucose, uric acid, choline and lactate was investigated. Standard solutions containing series concentrations of glucose, uric acid, choline and lactate were added to the sampling zone to trigger the colorimetric reactions. Fig. 3 shows the color images of the detection layer of the μPAD in the presence of the blank buffer and the different concentrations of glucose, uric acid, choline and lactate. As shown in Fig. 3, the eight detection zones exhibited colorfully responses. Detection zones numbered 1 and 2 used for glucose determination generated red and purple color, respectively. Detection zones numbered 3 and 4 used for uric acid determination generated blue-green and purple color, respectively. Detection zones numbered 5 and 6 used for choline determination generated blue-green and red color, respectively. Detection zones numbered 7 and 8 used for lactate determination generated bluegreen and purple color, respectively. In comparison, the detection zones of the μPAD in the presence of blank buffer were almost colorless (Fig. 3A). The more corresponding analytes existed in the sample, the more H2O2 would be generated, leading to the increase in color intensities. Accordingly, the color intensities of the detection zones increased in proportion to the concentrations of the corresponding analyte existed in the sample solutions. Thus, multiplexed determination of
3.3. Optimization of detection conditions The colorimetric assay in this work was based on the chromogenic
Fig. 2. (A) Color photograph of one-layered μPAD. (B) Color photograph of detection layer of the double-layered 3D μPAD. Then concentration of glucose, uric acid, choline and lactate tested was 0.2, 0.15, 0.4 and 0.8 mM, respectively. 269
Sensors & Actuators: B. Chemical 288 (2019) 266–273
F. Li, et al.
Fig. 3. Color photographs of the detection layer of the μPAD in the presence of blank buffer (A) and different concentrations of glucose, uric acid, choline and lactate. The concentrations of glucose, uric acid, choline and lactate were 0.01, 0.01, 0.04, 0.04 mM (B); 0.02, 0.015, 0.07, 0.06 mM (C); 0.03, 0.02, 0.09, 0.08 mM (D); 0.08, 0.06, 0.2, 0.2 mM (E); 0.2, 0.15, 0.4, 0.8 mM (F); 0.5, 0.4, 0.9, 2 mM (G); 1.5, 0.9, 2, 4 mM (H); 4, 2, 4.5, 9 mM (I); 10, 5, 10, 14 mM (J), 12, 6.5, 11, 20 mM (K), 15, 8, 13, 24 mM (L), respectively. The detection zones were numbered 1–8. Position numbered 1–2: glucose detection zones, 3–4: uric acid detection zones, 5–6: choline detection zones, 7–8: lactate detection zones. The positions of the detection zones (A–I) were the same.
270
Sensors & Actuators: B. Chemical 288 (2019) 266–273
F. Li, et al.
Table 1 The comparison of the proposed μPAD with the reported μPADs for the colorimetric determination of glucose, uric acid and lactate.
Fig. 4. Representative calibration curves for the detection of glucose, uric acid, choline and lactate by using the double-layered μPAD.
Analyte
Indicator
Linear range (mM)
LOD (mM)
Reference
Glucose
4-AAP/DHBS 4-AAP/DHBS TMB KI/trehalose 4-AAP/DHBS, ODa, Yb, Bc, KI 4-AAP/DHBS and 4-AAP/ TOPS
0.1–1 2.0–12 1.0–6 0.5–10 0.5–20
0.023 0.7 0.057 0.5 –
[25] [27] [25] [28] [18]
0.01–15
0.003
This work
Uric acid
OD 4-AAP/DHBS 4-AAP/DHBS 4-AAP/DHBS, Y, B, KI TMB and 4-AAP/TOPS
1.0–5 0.1–1 1.0–5 0.1–7 0.01–8
0.15 0.037 0.3 – 0.005
[25] [25] [27] [18] This work
Lactate
4-AAP/DHBS, OD, KI TMB and 4-AAP/TOPS
1.0–25 0.04–24
– 0.03
[18] This work
a
the four small biomolecules were realized by calculating the averaged color intensities obtained from the two corresponding detection zones. For example, the averaged ΔGray-scale value for glucose quantification was calculated by the ΔGray-scale value measured from detection zone numbered 1 plus the ΔGray-scale value measured from detection zone numbered 2 and then divide by two. The averaged ΔGray-scale values for uric acid, choline and lactate quantification were calculated in the same methods by using their respective detection zones. As shown in Fig. 4, four linear calibration curves were obtained for the determination of glucose, uric acid, choline and lactate, respectively. The four calibration curves were I = 75.6 + 29.5 log C, R2 = 0.991, with linear range from 0.01 to 15 mM for glucose detection, I = 90.2 + 35.2 log C, R2 = 0.993, with linear range from 0.01 to 8 mM for uric acid detection, I = 92.6 + 37.9 log C, R2 = 0.995, with linear range from 0.04 to 13 mM for choline detection, I = 70.3 + 42.1 log C, R2 = 0.997, with linear range from 0.04 to 24.0 mM for lactate detection, where I was the averaged color intensity represented by ΔGray-scale value and C was the concentration of the corresponding analyte (mM). The relationship between the concentrations of the four analytes and the averaged color intensities represented by ΔRed-channel value, ΔGreen-channel value, ΔBlue-channel value and hue value, respectively were also investigated as shown in Fig. S11A–D. It can be seen that the averaged color intensities represented by ΔRed-channel value and ΔGreen-channel value also exhibited linear relationships with the concentrations of the four analytes, while the linear relationships obtained with ΔBlue-channel value and hue value were poor. As compared with Fig. 4, the linear relationships obtained with ΔGray-scale values were the best. Thus, glucose, uric acid, choline and lactate were quantified by using color intensity represented by ΔGray-scale value. The developed μPAD exhibited very wide linear ranges over two to three orders of magnitude. The detection limits calculated at a signal-to-noise ratio of 3 (S/N = 3) were 0.003, 0.005, 0.01 and 0.03 mM for glucose, uric acid, choline and lactate detection, respectively. As shown in Table 1, the linear ranges in this work are more than one order of magnitude wider and the detection limits are also more one order of magnitude lower than the previously reported μPAD based colorimetric assays for glucose, uric acid and lactate detection. Even though colorimetric assays have been reported for choline detection [26], as far as we known, there is still no report about paper-based colorimetric assay for choline detection and no report for the simultaneous colorimetric detection of glucose, uric acid, choline, and lactate. The relative standard deviations (RSD) were less than 1.96% within a day and less than 3.11% in different days for the determination of 0.2 and 0.5 mM glucose, uric acid, choline, and lactate (Table S2), indicating good repeatability and reliability of this μPADs based colorimetric assay.
b c
OD: o-dianisidine dihydrochloride. Y: acid yellow 34. B: acid black 1.
The wide detection ranges and low detection limits obtained with the developed μPAD were mainly attributed to the following reasons. Firstly, as compared in Fig. 2, the color intensity of the detection zones of the double-layered μPAD were greater than that of the one-layered μPAD for the determination of same concentration of analytes. Thus, improved sensitivity could be obtained by restricting the chromogenic reagents in the closed circular detection zones with the double-layered μPAD, rather than in the open-ended detection zones with the onelayered μPAD. Secondly, every two kinds of colorimetric reagents were used as comprehensive indicator for the quantification of each analyte in this work. The relationships between the ΔGray-scale values obtained from each detection zone (immobilized with one kind of colorimetric reagent) and the concentration of their corresponding analyte were investigated as shown in the Fig. S12. It can be seen that different calibration curves with different detection ranges were obtained for each analyte by using different kind of colorimetric reagent as indicator, which were in accordance with the reported work [23,25]. The difference in calibration curves was mainly due to the difference in degree of chromogenic reaction when the colorimetric reagent or substrate is different. For example, the linear range for the determination of glucose by using 4-AAP/TOPS and 4-AAP/DHBS as single indicator was 0.01–4 mM (R2 = 0.989) and 0.03–15 mM (R2 = 0.987), respectively. In comparison, the linear range for the determination of glucose by using 4-AAP/TOPS and 4-AAP/DHBS as comprehensive indicator was 0.01–15 mM (R2 = 0.991). Similar results were obtained with uric acid, choline and lactate. Thus, extended linear range with lowest detection limit could be obtained by using two kinds of optimized colorimetric reagents in combination as comprehensive indicator. Thirdly, it was found that by simply putting the detection paper on top of a flat LED lamp, the light of the LED lamp up illuminated penetrating and lighting the detection paper. Accordingly, as shown in Fig. 3, the color images of the detection papers were very bright with a high contrast. The LED lamp was frequently used as downward illuminating device in a cassette to provide constant illumination and avoid influence from ambient light. Then, the color images of the detection layer of the doublelayered μPAD by using a LED lamp as downward illuminating device were also captured. As shown in Fig. S13, the color of the detection zones by using a LED lamp as downward illuminating device were dark, blurred, weak, and some positions even become invisible when the concentration of the analytes were small. The relationships between the averaged ΔGray-scale value of the detection zones and the concentration of the corresponding analyte by using LED lamp as downward 271
Sensors & Actuators: B. Chemical 288 (2019) 266–273
F. Li, et al.
illuminating device was also investigated. As shown in Fig. S14, the ΔGray-scale values of the detection zones were about the same when the concentrations of glucose, uric acid, choline, and lactate was below 0.2, 0.06, 0.2, and 0.2 mM, respectively. Even though the ΔGray-scale values exhibited linear relationship with the logarithm of the concentrations of glucose (0.08–4.0 mM, R2 = 0.989), uric acid (0.06–2.0 mM, R2 = 0.971), choline (0.09–4.5 mM, R2 = 0.986), and lactate (0.08–9.0 mM, R2 = 0.973), respectively, the detection ranges were much narrower and the detection limits were much higher as compared with the results obtained in Fig. 4. Thus, enhanced distinguishability, extended detection range and increased sensitivity could be further obtained by simply using a LED lamp as up illuminating device.
Table 2 The recovery of glucose, uric acid, choline and lactate in spiked human serum samples by using the proposed μPADs. Analyte
Initially detected (mM)
Added (mM)
Total found (mM)
Recovery (%)
RSD (n = 3, %)
Glucose Uric acid Choline Lactate
4.54 0.42 0.02 1.21
1.0 0.1 0.1 1.0
5.62 0.51 0.11 2.18
108 90 90 97
2.08 1.59 4.78 2.67
3.6. Application in real samples The simultaneous detection of glucose, uric acid, cholesterol, and choline in serum samples have great clinical importance. To investigate the practical application ability of this μPAD for the simultaneous detection of glucose, uric acid, choline, and lactate in complex biological samples, the concentrations of glucose, uric acid, choline, and lactate in human serum samples were measured. Particularly, the low detection limits at around 0.01 mM of this μPADs enabled the detection of glucose, uric acid, choline, and lactate in diluted serum samples. Thus, only microliter grade serum sample was needed. The serum sample was diluted properly prior to the measurements by pH 7.0 phosphate buffer. The concentrations of glucose, uric acid, choline, and lactate in serum sample were determined by standard additional method and the detection results are shown in Table 2. The normal concentrations of glucose, uric acid, choline, and lactate in human serum sample are 2.5–5.3 mM [29], 0.1–0.4 mM [29], around 10 μM [30] and 0.5–1.7 mM [29], respectively. The determined concentrations of glucose, uric acid, choline, and lactate in the serum sample were in the normal ranges. The accuracy of the proposed μPADs was further evaluated by recovery experiments. The serum sample was further spiked with 1.0 mmol/L glucose, 0.1 mmol/L uric acid, 0.2 mmol/L choline, and 1.0 mmol/L lactate. The recoveries of spiked glucose, uric acid, choline, and lactate in the serum sample ranged from 90% to 108%, with RSD less than 4.25%. The satisfactory recoveries and RSD indicated that the proposed μPADs has good accuracy and is applicable for multiplexed analysis of glucose, uric acid, choline, and lactate in biological samples.
3.5. Selectivity and stability study The biological samples containing various interference species other than glucose, uric acid, choline, and lactate. To evaluate the effects of the interference species on the target biomolecules, the interference study was carried out. Small biomolecules including urea, cholesterol, cysteine, tyrosine and glutamic acid, metal ions including K+, Na+ and Ca2+, and protein including human immunoglobulin G and bovine serum albumin were chose as typical interferences. The sample solutions containing the ten interferences, containing the four target analytes including glucose, uric acid, choline, and lactate, and containing the four target analytes mixed with the ten interferences were measured. The concentrations of the tested ten interferences and the four target analytes were 0.5 mM. The obtained ΔGray-scale values were compared as shown in Fig. 5. As expected, the target analytes mixed with interferences exhibited similar high ΔGray-scale values with the corresponding analytes. In comparison, samples containing only the interferences showed very small ΔGray-scale values. Thus, the μPADs based colorimetric assay in this work exhibited high selectivity and could be applied for the multiplexed detection of glucose, uric acid, choline, and detection in mixture. The good selectivity was mainly due to the high specificity of the oxidase enzymatic reactions. The storage stability of the developed μPAD was further studied in consecutive 30 days. The fabricated μPADs were stored in dark, dry, sealed condition at 4 °C and tested every day. No remarkable change in the colorimetric responses was observed. The results confirmed that the device can be stably maintained at 4 °C for at least one month. Thus, the storage stability of the μPAD is acceptable, and the device is suitable for clinical diagnostics and POCT.
4. Conclusion In this work, a 3D μPAD combined with double-layered structure design and multiple colorimetric indicator strategy was proposed for the simultaneous detection of four kinds of small biomolecules in mixture. The special double-layered structure design could effectively improve the colorimetric performance, sensitivity and reproducibility by preventing uncontrolled diffusion of colorimetric reagents outside of the detection zones. Every two kinds of colorimetric indicators were used for the detection of each analyte to extend the detection range. A LED lamp was used as up illuminating device and a smartphone was used as detector, resulting in further increased sensitivity and extended detection range. Finally, the developed double-layered μPAD was applied to the multiplexed detection of glucose, uric acid, choline, and lactate in mixed samples and serum samples with satisfied results. The linear ranges were much wider and detection limits were much lower than the previously reported colorimetric μPADs for glucose, uric acid and lactate detection. The four analytes can be regarded as model analytes, since the proposed μPADs can be readily expanded to multiplexed determination of other molecules such as cortisol, bilirubin and cholesterol by using their specific H2O2-producing oxidases. The present work shows the successful fabrication of a simple 3D μPAD with improved colorimetric performance, enhanced sensitivity and extended detection range for multiplexed colorimetric analysis, which hold great application potentials for simple, low-cost and rapid POCT and on-site diagnosis.
Fig. 5. Interference study of the double-layered μPAD. The ΔGray-scale values were obtained from the solutions containing the ten interferences, containing the four target analytes including glucose, uric acid, choline and lactate, and containing the four target analytes mixed with the ten interferences (n = 3). 272
Sensors & Actuators: B. Chemical 288 (2019) 266–273
F. Li, et al.
Conflict of interests [13]
The authors declare that there is no conflict of interests regarding the publication of this paper.
[14]
Acknowledgements [15]
This work was supported by the National Natural Science Foundation of PR China (No. 21605032) and the open project of Anhui province key laboratory of advanced catalytic materials and reaction engineering (No. 45000-411104/007).
[16]
[17]
Appendix A. Supplementary data
[18]
Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.snb.2019.02.116.
[19]
References
[20]
[1] J. Song, M.G. Mauk, B.A. Hackett, S. Cherry, H.H. Bau, C. Liu, Instrument-free point-of-care molecular detection of zika virus, Anal. Chem. 88 (2016) 7289–7294. [2] B.H. Park, S.J. Oh, J.H. Jung, G. Choi, J.H. Seo, D.H. Kim, et al., An integrated rotary microfluidic system with DNA extraction, loop-mediated isothermal amplification, and lateral flow strip based detection for point-of-care pathogen diagnostics, Biosens. Bioelectron. 91 (2017) 334–340. [3] L. Syedmoradi, M. Daneshpour, M. Alvandipour, F.A. Gomez, H. Hajghassem, K. Omidfar, Point of care testing: the impact of nanotechnology, Biosens. Bioelectron. 87 (2017) 373–387. [4] N.A. Meredith, C. Quinn, D.M. Cate, T.H. Reilly III, J. Volckens, C.S. Henry, Paperbased analytical devices for environmental analysis, Analyst 141 (2016) 1874–1887. [5] Y. Xia, J. Si, Z. Li, Fabrication techniques for microfluidic paper-based analytical devices and their applications for biological testing: a review, Biosens. Bioelectron. 77 (2016) 774–789. [6] P. von Lode, Point-of-care immunotesting: approaching the analytical performance of central laboratory methods, Clin. Biochem. 38 (2005) 591–606. [7] Andres W. Martinez, Scott T. Phillips, Manish J. Butte, George M. Whitesides, Patterned paper as a platform for inexpensive, low-volume, portable bioassays, Angew. Chem. Int. Ed. 46 (2007) 1318–1320. [8] K. Yamada, T.G. Henares, K. Suzuki, D. Citterio, Paper-based inkjet-printed microfluidic analytical devices, Angew. Chem. Int. Ed. 54 (2015) 5294–5310. [9] S. Wang, L. Ge, X. Song, J. Yu, S. Ge, J. Huang, et al., Paper-based chemiluminescence ELISA: lab-on-paper based on chitosan modified paper device and waxscreen-printing, Biosens. Bioelectron. 31 (2012) 212–218. [10] W. Dungchai, O. Chailapakul, C.S. Henry, Electrochemical detection for paperbased microfluidics, Anal. Chem. 81 (2009) 5821–5826. [11] Y.Y. Xia, J. Si, Z.Y. Li, Fabrication techniques for microfluidic paper-based analytical devices and their applications for biological testing: a review, Biosens. Bioelectron. 77 (2016) 774–789. [12] E.M. Fenton, M.R. Mascarenas, G.P. López, S.S. Sibbett, Multiplex lateral-flow test
[21] [22]
[23]
[24] [25]
[26]
[27]
[28]
[29] [30]
273
strips fabricated by two-dimensional shaping, ACS Appl. Mater. Inter. 1 (2009) 124–129. K. Yamada, H. Shibata, K. Suzuki, D. Citterio, Toward practical application of paper-based microfluidics for medical diagnostics: state-of-the-art and challenges, Lab Chip 17 (2017) 1206–1249. N. Lopez-Ruiz, V.F. Curto, M.M. Erenas, F. Benito-Lopez, D. Diamond, A.J. Palma, et al., Smartphone-based simultaneous pH and nitrite colorimetric determination for paper microfluidic devices, Anal. Chem. 86 (2014) 9554–9562. P. Rattanarat, W. Dungchai, D. Cate, J. Volckens, O. Chailapakul, C.S. Henry, Multilayer paper-based device for colorimetric and electrochemical quantification of metals, Anal. Chem. 86 (2014) 3555–3562. Z. Qiu, J. Shu, D. Tang, Bioresponsive release system for visual fluorescence detection of carcinoembryonic antigen from mesoporous silica nanocontainers mediated optical color on quantum dot-enzyme-impregnated paper, Anal. Chem. 89 (2017) 5152–5160. J. Yu, L. Ge, J. Huang, S. Wang, S. Ge, Microfluidic paper-based chemiluminescence biosensor for simultaneous determination of glucose and uric acid, Lab Chip 11 (2011) 1286–1291. W. Dungchai, O. Chailapakul, C.S. Henry, Use of multiple colorimetric indicators for paper-based microfluidic devices, Anal. Chim. Acta 674 (2010) 227–233. R. Mukhopadhyay, Cheap, handheld colorimeter to read paper-based diagnostic devices, Anal. Chem. 81 (2009) 8659. A. Roda, E. Michelini, M. Zangheri, M. Di Fusco, D. Calabria, P. Simoni, Smartphone-based biosensors: a critical review and perspectives, Trac-Trend. Anal. Chem. 79 (2016) 317–325. A.K. Yetisen, M.S. Akram, C.R. Lowe, Paper-based microfluidic point-of-care diagnostic devices, Lab Chip 13 (2013) 2210–2251. X. Sun, B. Li, A. Qi, C. Tian, J. Han, Y. Shi, et al., Improved assessment of accuracy and performance using a rotational paper-based device for multiplexed detection of heavy metals, Talanta 178 (2018) 426–431. X. Wang, F. Li, Z. Cai, K. Liu, J. Li, B. Zhang, et al., Sensitive colorimetric assay for uric acid and glucose detection based on multilayer-modified paper with smartphone as signal readout, Anal. Bioanal. Chem. 410 (2018) 2647–2655. M. Rinaudo, Chitin and chitosan: properties and applications, Prog. Polym. Sci. 31 (2006) 603–632. E.F.M. Gabriel, P.T. Garcia, T.M.G. Cardoso, F.M. Lopes, F.T. Martins, W.K.T. Coltro, Highly sensitive colorimetric detection of glucose and uric acid in biological fluids using chitosan-modified paper microfluidic devices, Analyst 141 (2016) 4749–4756. N. Nikzad, Z. Karami, Label-free colorimetric sensor for sensitive detection of choline based on DNAzyme-choline oxidase coupling, Int. J. Biol. Macromol. 115 (2018) 1241–1248. Pde Tarso Garcia, T.M. Garcia Cardoso, C.D. Garcia, E. Carrilho, W.K. Tomazelli Coltro, A handheld stamping process to fabricate microfluidic paper-based analytical devices with chemically modified surface for clinical assays, RSC Adv. 4 (2014) 37637–37644. E. Evans, E.F. Moreira Gabriel, T.E. Benavidez, W.K. Tomazelli Coltro, C.D. Garcia, Modification of microfluidic paper-based devices with silica nanoparticles, Analyst 139 (2014) 5560–5567. N.W. Tietz, Clinical Guide to Laboratory Test, second ed., W.B. Saunders Company, PA, 1990. I.H. Ulus, G. Özyurt, E. Korfali, Decreased serum choline concentrations in humans after surgery, childbirth, and traumatic head injury, Neurochem. Res. 23 (1998) 727–732.