Immunochromatographic Assay on Thread - Analytical Chemistry

Aug 13, 2012 - ... University and Génome Québec Innovation Centre, and §Department of Neurology & Neurosurgery, McGill University, Montreal, QC, Ca...
4 downloads 0 Views 1MB Size
Article pubs.acs.org/ac

Immunochromatographic Assay on Thread Gina Zhou,†,‡ Xun Mao,†,‡ and David Juncker*,†,‡,§ †

Biomedical Engineering Department, ‡McGill University and Génome Québec Innovation Centre, and §Department of Neurology & Neurosurgery, McGill University, Montreal, QC, Canada S Supporting Information *

ABSTRACT: Lateral-flow immunochromatographic assays are low-cost, simple-to-use, rapid tests for point-of-care screening of infectious diseases, drugs of abuse, and pregnancy. However, lateral flow assays are generally not quantitative, give a yes/no answer, and lack multiplexing. Threads have recently been proposed as a support for transporting and mixing liquids in lateralflow immunochromatographic assays, but their use for quantitative high-sensitivity immunoassays has yet to be demonstrated. Here, we introduce the immunochromatographic assay on thread (ICAT) in a cartridge format that is suitable for multiplexing. The ICAT is a sandwich assay performed on a cotton thread knotted to a nylon fiber bundle, both of which are precoated with recognition antibodies against one target analyte. Upon sample application, the assay results become visible to the eye within a few minutes and are quantified using a flatbed scanner. Assay conditions were optimized, the binding curves for C-reactive protein (CRP) in buffer and diluted serum were established and a limit of detection of 377 pM was obtained. The possibility of multiplexing was demonstrated using three knotted threads coated with antibodies against CRP, osteopontin, and leptin proteins. The performance of the ICAT was compared with that of the paper-based and conventional assays. The results suggest that thread is a suitable support for making low-cost, sensitive, simple-to-use, and multiplexed diagnostic tests. sample solution. Lateral flow assays only require the application of a sample (sometimes followed by the application of a buffer solution) and can yield a result within 5−15 min.2 Rapid tests for pregnancy,4 HIV,5 bacterial infection,6 drugs of abuse,7 and food contaminants,8 have been reported, and many are commercially available. Lateral flow assays are also being developed for global health applications, where devices that are inexpensive and easy-to-use are required. However, lateral flow assays are generally not quantitative and often only give a yes/ no answer. The test area is often several millimeter wide, which makes multiplexing difficult.2 Recently, paper (wood cellulose) has been introduced as an alternative material for the assembly of lateral flow tests and as a replacement for enzyme linked immunosorbent assay (ELISA) plates.9−14 Paper-based devices can be fabricated easily using wax printers to define hydrophilic and hydrophobic zones and using tapes to assemble sheets into multilayer structures,10 while also enabling control of fluid flow.12 Devices made of patterned paper sheets have been used in place of the

he lateral flow immunoassay, a type of sandwich assay, relies on a pair of antibodies to recognize two independent epitopes of a protein, and therefore it can achieve high specificity.1 A typical lateral flow assay strip is composed of (1) a sample-loading pad, (2) a glass fiber pad with detection antibody (dAb) conjugated to gold nanoparticles (AuNPs) or latex beads, (3) a nitrocellulose or polyvinylidene fluoride membrane with preimmobilized capture antibody (cAb) and control antibody for test validation, and (4) an absorbent pad used as a capillary pump to draw the sample solution.2 To perform a lateral flow assay, the sample containing the target analyte (antigen) is loaded on the sample pad and flows through the membrane by capillary effects. The liquid first dissolves the dAb−AuNP conjugates and the antigen binds to the dAb. As the antigen−dAb pair flows through the capture zone, the cAb will capture the labeled antigen. Further downstream, the unbound dAb−AuNP reacts with the control Ab, which binds specifically to dAb irrespective of the antigen. Both the capture and control lines may become visible due to the accumulation of the AuNPs that produce collective plasmonic effects and result in a red color.3 The color on the control line indicates the test is valid, and the color on the capture line suggests the presence of target analyte in the

T

© 2012 American Chemical Society

Received: April 29, 2012 Accepted: August 13, 2012 Published: August 13, 2012 7736

dx.doi.org/10.1021/ac301082d | Anal. Chem. 2012, 84, 7736−7743

Analytical Chemistry

Article

classical 96-well plate.9 Assay results of these tests can be read by eye or recorded using a scanner or a cell phone camera.13 We and others have proposed cotton yarn and thread as a support for making microfluidic circuits and rapid diagnostic tests.15−20 Using cotton yarn, we have developed microfluidic elements such as splitters, mergers, and mixers.15 We found that when using two strings, the knot topology can control the mixing ratios of the fluids. Since knots have little flow resistance, microfluidic circuits can be modeled using Kirchhoff’s law, analogous to electric circuits, and these concepts were applied to building a web functioning as a serial dilutor.19 Cotton threads were also used for the detection of small molecules and proteins in artificial urine.17,18 More recently, woven silk fabrics combined with wetting and nonwetting yarns were used for carrying out simple yes/no immunoassays,21 however their use for quantitative high-sensitivity immunoassays has yet to be demonstrated. These results suggest that thread is a promising support for immunoassays, but it will be important to establish whether it can yield reproducible and quantitative results, and whether it can be used in blood before it is applied to diagnostic tests. Here we present an immunochromatographic assay on thread (ICAT) that emulates the lateral flow assay format but uses cotton threads and nylon fiber bundles in place of pads and membranes. We introduce an ICAT cartridge made of a polyester frame with multiple cotton threads and a single nylon fiber bundle knotted across it. Using such a cartridge, multiple assays can be conducted on parallel threads separated by a few millimeters. Assay optimization experiments were performed, and the binding curves for C-reactive protein (CRP) in both buffer and serum were established and a limit of detection (LOD) in the pM range was calculated. We further demonstrate the possibility of multiplexing of ICAT by simultaneously measuring CRP, leptin (LEP), and osteopontin (OPN).

of 1.5 M NaCl.22,23 Aggregation in both tests was determined by visualizing the color change of AuNPs. Preparation of dAb−AuNP Conjugate. Polyclonal goat anti-rabbit IgG antibody (1.5 μg) was added to 1 mL of pH adjusted AuNP solution and incubated at room temperature for 1 h. Then 55 μL 10% BSA in PBS (to be diluted later) was added to the mixture, followed by 30 min incubation and 20 min centrifugation at 9000 rpm. After the supernatant was discarded, the precipitated dAb−AuNP pellet was resuspended in 100 μL of eluent buffer (PBS with 5% BSA, 2.5% Tween20, and 10% sucrose).24 For the preparation of dAb−AuNP conjugates against CRP, OPN, and LEP, a similar procedure was followed. A polyclonal goat IgG against human OPN, and a monoclonal mouse IgG against human CRP and LEP, respectively, were used as the dAb. Preparation of Threads. Cotton threads were rendered hydrophilic with an air plasma (Tegal Plasmaline 415) for 30 s with 250-mTorr pressure and 150-mW power. A capture zone and a control zone were defined on each thread. The capture zone was coated with 0.6 μL (applied as three aliquots of 0.2 μL, with 10 min drying at 37 °C after each step) of cAb against CRP, OPN, and LEP at a concentration of 1 mg/mL in PBS buffer containing 15% glycerol. The control zone was coated with anti-mouse IgG (0.6 μL at 1 mg/mL concentration), which directly binds the dAb irrespective of the antigen. The cotton thread was blocked with 1% BSA in PBS for 30 min, rinsed with PBS (0.1% Tween20) and deionized water, and quickly dried under a stream of N2. For the nylon fiber bundle, paraffin was applied on the bundle to create isolated zones to allow drying of multiple different dAb-AuNP on a single bundle without mixing. Three μL of dAb-AuNP against CRP, OPN and LEP, respectively, were applied to the nylon fiber bundle and the fiber bundle was vacuum-dried for 1 h at room temperature. Both Ab-coated cotton threads and nylon fiber bundles were stored at 4 °C and used within one week. Assembly of ICAT Cartridges and Assay Protocol. The frame of the ICAT cartridge was 25 × 75 mm2 and made from a 0.5 mm-thick polymer sheet. The center of the frame was cut and holes were drilled on the periphery. Ab-coated cotton threads were attached to a polyester frame, and one nylon fiber bundle was knotted across all the Ab-coated cotton threads. An absorbent pad (25 × 10 mm2) was placed at the end of the cartridge to serve as a capillary pump to draw the liquid. To perform an assay, 3 μL of sample was applied to the knot (where the dAb−AuNP conjugates had been dried), and the upstream end of the ICAT cotton thread was immersed in buffer (50 μL per thread) to allow flushing of the sample. The assay was conducted in a sealed Petri dish padded with a humidity sheet. After 20 min, the assay results were visually inspected. The absorbent pad was removed to prevent backflow and the cartridge was left to dry at room temperature in a ventilated area for 2 h prior to imaging. Imaging Analysis and Signal Quantification. The image of the assay results was recorded using a flatbed scanner, CanonScan LiDE 700E, and quantified using ImageJ software (NIH, Bethesda, MD, USA) based on the optical intensity of the capture zone. The binding curve was calculated using a four-parameter logistic curve in Systat SigmaPlot 12 (San Jose, CA, USA) and the limit of detection (LOD) for CRP was calculated as three standard deviation (SD) of the blank tests plus the average of the blank tests.



EXPERIMENTAL SECTION Reagents. 100% mercerized cotton threads were purchased from Coats & Clarks (Greenville, SC, USA), nylon fiber bundles (Ultra Floss) from Oral-B (Cincinnati, OH, USA), and humidity sheets from Heartfelt Industries (Dayton, NV, USA). Millipore (Billerica, MA, USA) generously provided cellulose absorbent sample pads (0.83 mm thick), and Diagnostic Consulting Network (Carlsbad, CA, USA) provided 40 nm AuNPs. Blue polystyrene beads (0.4 μm) were purchased from Spherotech (Lake Forest, IL, USA). Rabbit IgG and anti-rabbit IgG were purchased from Invitrogen (Burlington, ON, Canada) and Ab pairs against CRP, OPN, and LEP were from R&D Systems (Minneapolis, MN, USA). Single-donor normal human serum was from Golden West Biologicals (Temecula, CA, USA) and pooled serum was from Jackson ImmunoResearch Laboratory (West Grove, PA, USA). All other chemicals were from Sigma-Aldrich (St. Louis, MO, USA). AuNP Stability Test. To determine the optimal pH needed to stabilize the AuNPs, 100 μL of gold suspension with different pH (5.5 to 10 with 0.5 pH increments) was mixed with 10 μL of rabbit IgG at a concentration of 100 μg/mL.22,23 To determine the minimal amount of dAb needed to stabilize the AuNPs, 100 μL AuNP solution (pH = 7, optical density = 10.5) was mixed with 10 μL of anti-rabbit IgG at a series of concentration of 40, 20, 10, 5, 2.5, and 1.25 μg/mL. The solution was incubated for 15 min, and then mixed with 25 μL 7737

dx.doi.org/10.1021/ac301082d | Anal. Chem. 2012, 84, 7736−7743

Analytical Chemistry

Article

Figure 1. Cotton thread and nylon fiber bundle. Photograph and SEM image of (a) a cotton thread and (b) a nylon fiber bundle. (c) Cross-section of a cotton thread embedded in poly(dimethylsiloxane) (PDMS) and (d) a close-up of a microtome-cut cotton thread embedded in paraffin, showing the cellulose wall and lumen. (e) Cross-section of a nylon fiber bundle in PDMS, which is more porous and has a higher volume capacity than the cotton thread shown in (c).

Figure 2. Illustration of ICAT and photograph of one ICAT cartridge. (a) ICAT device assembly and assay protocol. (a1) A nylon fiber bundle is coated with dAb-AuNP, and a section of the cotton thread is coated with cAb and control Ab. (a2) After vacuum drying, the cotton thread and the nylon fiber bundle are knotted and woven on a polymer frame. An absorbent pad is placed at the end of the thread. (a3) The assay starts when sample is applied to the knot and dissolves the dAb-AuNP conjugates. (a4) Immediately after, a buffer solution is applied upstream of the knot and flushes the sample down the length of the thread. (b) Photograph of one ICAT cartridge with six cotton threads in parallel, one nylon fiber bundle knotted across, and an absorbent pad placed at the end. The inset shows a close-up of the knot with a schematic representation of the reagents.



between individual fibers generate a capillary pressure that wicks the liquid into the fibers. The threads used in our experiments measure ∼600 μm in diameter and consist of 240 ± 12 fibers. The nylon fiber bundle used in our experiments is curly, nonwoven, and loosely entangled, resulting in a comparatively higher porosity than the cotton thread (Figure 1e). The fact that antibodies can be dried and then released from the nylon fiber bundle suggests that it has a nonfouling surface, which based on a patent application might be polyvinyl alcohol,27 a well-known nonfouling surface coating.28−30

RESULTS AND DISCUSSION Threads. The threads used in the ICAT are 100% mercerized cotton threads and nylon fiber bundles (Figure 1a and b). Cotton thread is made of twisted cellulose fibers (Figure 1c). Each fiber is composed of one outer waxy layer called the cuticle, three closely packed cellulose layers consisting of fine fibrils (small strands of cellulose), and a hollow center called the lumen (Figure 1d).25 The lumen collapses when desiccated, and presumably opens when wetted as for wood cellulose (paper).26 This lumen and the gaps 7738

dx.doi.org/10.1021/ac301082d | Anal. Chem. 2012, 84, 7736−7743

Analytical Chemistry

Article

Flow Properties within Threads. The cotton threads were rendered hydrophilic by plasma treatment to allow spontaneous wicking of aqueous solutions. To measure the flow rate in an ICAT device, we applied a trypan blue dye in PBS to one end of the thread in a saturated humidity and recorded the liquid flow using a video camera at 30 frames per second. We observed that the flow was fastest in the beginning and followed Washburn’s law until reaching the absorbent pad, after which it was constant because the pad generates a constant pressure and does not add significantly to the flow resistance owing to the large cross section, based on our previous experimental results.19 For a capture zone length of 20 mm and an average flow velocity of 0.3 mm/s over the cAb, the antigen−dAb complex interacts with the immobilized cAb for ∼1 min. ICAT Assay Procedure. The ICAT assay protocol mirrors the procedure used in lateral flow assays, as illustrated in Figure 2a. The cAb and control Ab were applied to the cotton threads at predefined capture and control zones, respectively. The Ab bound irreversibly to the surface of the cotton thread through a combination of electrostatic and hydrophobic interactions. The dAb−AuNP were dried with glycerol on the nylon fiber bundles to facilitate subsequent redissolution (Figure 2a-1). The coated cotton thread and the nylon fiber bundle were knotted perpendicularly and woven on the cartridge, and an absorbent pad was applied at the downstream end (Figure 2a−2). Sample was applied on the knot where dAb−AuNP had been dried. The sample solution dissolved dAb−AuNP and the flow was observable as it wicked along the thread (Figure 2a-3). While flowing, the dAb−AuNP conjugates first bound to the antigen, and then accumulated in the capture zone by binding the cAb. The intensity of the resulting color band should vary proportionally to the concentration of antigen. The excess dAb−AuNP conjugates continued to flow until they bound directly to the downstream control Ab and formed a second color band at the control zone, serving as a positive control, and any unbound conjugates were taken up by the absorbent pad (Figure 2a-4). The buffer upstream of the knot and the absorbent pad at the downstream ensured a continuous flow over time. We observed that 40−50 μL of buffer per thread was sufficient to flush the sample solutions and the dAb−AuNP conjugates. The absorbent pad has a volume capacity of 120 μL/cm2 and it was sized according to the number of tests per cartridge. Figure 2b shows an assembled ICAT cartridge with six threads that each can be used for one test. Image Analysis. The assay results were visible to the eye and were quantified using a flatbed scanner. A scanner is an economic alternative to a densitometer that is commonly used to read lateral flow tests, while providing a high resolution (9600 dpi) and a wide dynamic range (16-bit).31 The original RGB (red/green/blue) image (Figure 3a) was first converted to HSB (hue/saturation/brightness) using the ImageJ software. The saturation channel of the image stack was used for quantification, as it is the most sensitive channel (high contrast) compared with others. The intensity of capture and background signals, Icapture and Ibkgnd, respectively, were obtained by integrating the value in the saturation channel within the capture and background zone (Figure 3b), which is equivalent to averaging. The background was subtracted from the capture zone signal and normalized to a known antigen concentration of 100 ng/mL, yielding a normalized value of Inorm = (Icapture − Ibkgnd)/Iref, where Iref is the net signal given by 100 ng/mL. Other approaches were tested, including calculating the maximal signal intensity and calculating the gray scale value

Figure 3. Scanned image of a completed ICAT and its signal quantification. (a) ICAT cartridge with ten threads and the three assay zones: “background”, “capture zone”, and “control zone”. Red color from the AuNPs is visible at the signal and control zones. The absorbent pad on the right also appears red due to the absorption of the unbound AuNPs. (b) Process flow of ICAT image analysis. First, the acquired red-green-blue (RGB) composite image (left panel) is converted to hue-saturation-brightness (HSB). Then, the value in the saturation channel is integrated over the background and capture zone, respectively (middle panel). Finally, the intensity, I, is extracted for both zones (right panel) and the net assay signal is obtained by subtracting the Ibkgnd from Icapture.

directly, but they were found to be less sensitive than the current method. Optimization of the ICAT. Several parameters were identified to be important variables that can improve the ICAT performance, reduce the reagent consumption, and save material cost. These parameters include (i) stability of AuNPs, (ii) cAb spotting conditions and application procedure, and (iii) dAb−AuNP buffer volume. We used the same polyclonal anti-rabbit IgG as both cAb and dAb, and rabbit IgG as antigen, to optimize the assay conditions. First, the capture zone length was optimized. The capture zone should be short to minimize antigen depletion during the assay,32 and at the same time, sufficiently long to average out the local variations because thread is a nonuniform and nonhomogenous material. We thus applied three aliquots of 0.2 μL each to minimize the wicking of liquid along the thread (0.2 μL is the smallest volume that conventional low volume pipettes can deliver). Glycerol was added to the buffer to increase the viscosity to obtain a short capture zone, and to serve as a preservative for the cAb during the ICAT storage. 15% of glycerol was determined to be the optimal value as it yielded the best assay signal (Figure 4a). A higher concentration (20%) reduced the flow rate, and occasionally stopped the sample flow from reaching the absorbent pad, resulting in an incomplete assay. Second, different volumes of dAb−AuNP were tested and the minimal amount needed to coat the nylon fiber bundle while obtaining a good assay signal was found to be 3 μL. Smaller volumes yielded a weaker signal at the capture zone, while larger volumes resulted in an increase in background, consequently reducing the net signal (Figure 4b). Finally, using the optimized conditions (1.5 μg of dAb per 1 mL of AuNPs, adjusted to pH of 7.4, 15% glycerol in cAb buffer, and 3 μL dAb−AuNP conjugates), we measured higher color intensity as the rabbit IgG concentration increased from 10−250 ng/mL (Figure 4c). 7739

dx.doi.org/10.1021/ac301082d | Anal. Chem. 2012, 84, 7736−7743

Analytical Chemistry

Article

Figure 4. ICAT assay condition optimization. Assay results for optimization experiments of (a) glycerol concentration in cAb buffer, and (b) volume of dAb−AuNP solution dried on the thread. (c) Sandwich assay results of using rabbit IgG as antigen and the same polyclonal anti-rabbit IgG as both cAb and dAb, performed under the optimized conditions. A concentration range of 10−250 ng/mL was measured and a straight line was drawn as a guide to the eye. The error bar is the standard error from three experiments.

Figure 5. Assay results for CPR detection. (a) Scanned image of the thread and capture zone showing the color intensity for PBS spiked with CRP concentration ranging from 0 to 100 ng/mL. The scale bar represents 20 mm. (b) Binding curve obtained by fitting the data obtained in (a) using a four-parameter logistic function. An LOD of 9.82 ng/mL was calculated as three SD of the blank tests above the average of the blank. (c) Binding curve of CRP in diluted serum (1:4 v/v serum in PBS). The error bar is the standard error from three independent experiments.

Sandwich Assay of CRP. Next, we used ICAT to quantify CRP, a 26-kDa acute inflammatory protein produced in the liver. CRP is a biomarker that has been used for cardiovascular disease diagnosis and prognosis, and an elevated CRP level (>3 μg/mL) in patient’s blood is considered high risk for developing coronary heart disease.33 CRP was chosen as it has been widely used as a model analyte for evaluating tests. We loaded a dilution series of CRP in PBS onto threads. After 20 min, a red color was visible on the thread at both the capture and the control zone. The control zone signal confirmed that the test results and valid, while the capture zone signal showed a more intense color when a high concentration sample was used (Figure 5a). The data were fit to a four-parameter logistic function to generate a binding curve (Supporting Information Table S1). The LOD was calculated as three SD of the blank tests above the average of the blank, and it was found to be 9.82 ng/mL or 377 pM (Figure 5b); a detailed calculation can be found in the Supporting Information. In clinical diagnostic

tests, the most common sample solution is serum. To test our device’s compatibility with such a complex sample matrix, we diluted CRP with pooled serum from healthy donors (1:4 v/v serum in PBS) for final CRP concentrations shown in Figure 5c (neglecting the native CRP concentration in the blood). The binding signal increases with the concentration, indicating that the ICAT can be used to quantify CRP in diluted serum. The binding curve also indicates that the ICAT can detect CRP well below the clinical cutoff (3 μg/mL) and into the physiological range of healthy individuals. Multiplexed Assays. After establishing the singleplex assay, we explored the potential of the ICAT format for multiplexed assays. We choose CRP, OPN, and LEP as they have been implicated in contributing to an increased risk of cardiovascular disease in obese patients.34 Plasma level of OPN, a 60−65-kDa protein, has been suggested to play a role in plaque formation and vascular disease progression.35 Leptin, a 16-kDa protein, has been suggested to stimulate vascular inflammation and up7740

dx.doi.org/10.1021/ac301082d | Anal. Chem. 2012, 84, 7736−7743

Analytical Chemistry

Article

Figure 6. Multiplexed ICAT. (a) Three threads were knotted together, and each was coated with cAb and dAb against CRP, LEP, and OPN, respectively. A cocktail of target proteins was prepared by diluting the proteins in diluted serum (1:4 v/v serum in PBS) for a final concentration of 100 ng/mL each. This solution was applied to the common node of the three threads. Three singleplex assays were conducted for comparison. (b) Assay results for the multiplex and singleplex tests. For CRP and OPN, the signal intensity between the singleplex and multiplexed assay are not statistically different. For LEP, there is a discrepancy between the two assay results, attributed to the flow rate difference in the central thread. **, P < 0.01. The error bar is the standard error from three independent experiments.

Figure 7. Comparison of the performance of ICAT with other CRP detection platform, and with low-cost paper-based detection. The selected CRP detection platforms include conventional lateral flow strips (diamond), diagnostic chips made of silicon or glass (cross), and immunoturbidity tests (circle). In addition, two 2D paper network devices and one paper-based ELISA plate are also presented here for comparison (triangle). The LOD and the dynamic range of the ICAT are comparable to other low-cost platforms.

that the flow velocity defines the antigen and dAb incubation time, and that as a result of shorter incubation time, a weaker signal is expected as shown in the experiments. This result indicates that the flow rate in each thread will need to be tightly controlled in multiplexed ICATs. Comparison of ICAT Assay Performance with Other Tests. We compared the LOD and dynamic range of the ICAT with other lateral-flow assays for CRP detection and low-cost paper-based assays reported in the literature (Figure 7). CRP detection has been performed on conventional membranes (NycoCard CRP), on silicon or glass chips,39,40 or in solution (immunoturbidity tests).41 For the commercial immunoturbidity CRP tests, we use values based on the manufacturer specifications and literature reports for comparison. To our knowledge, CRP detection has not been reported using paperbased assays, and only a few articles on paper-based devices have reported an LOD for sandwich immunoassays. Fu et al. used a 2D paper network card to quantify human chorionic gonadotropin (hCG) and malaria antigen in fetal bovine serum and showed a LOD of 1.2 mIU11 and 2.9 mIU,14 respectively. Cheng et al. used a paper-based ELISA device and measured 54 fmol in a 3 μL volume, which is equivalent to 2.3 μg/mL.9

regulate CRP expression.36,37 For multiplexed assays, one option would be to mix all antibodies together, but this carries the risk of cross-reactivity which is detrimental to assay performance.38 The preferred option, as proposed here, is to have each cAb and dAb pair coated on a separate thread to avoid mixing (Figure 6a). Three threads were used: one for each of the three antigens. The three threads were knotted together to form a single inlet for the sample solution. The nylon fiber bundle, preloaded with the specific dAb-AuNP conjugates, was then knotted across the threads. To perform a multiplexed assay, 9 μL of serum sample (1:4 v/v in PBS) was spiked with CRP, LEP and OPN for a final concentration of 100 ng/mL for each of the protein. The sample was applied at the inlet, followed by the application of buffer. The sample solution was split among the three threads, and allowed to flow for 20 min. A singleplex assay was used as the control experiment, where each protein was tested separately. The results showed that for CRP and OPN, the signal intensity between the singleplex and multiplexed assay were not statistically different (Figure 6b). For LEP, a stronger signal was seen in the multiplexed test, which may be due to a different flow rate in one of the threads. We have observed a faster flow velocity on the middle thread. We also established 7741

dx.doi.org/10.1021/ac301082d | Anal. Chem. 2012, 84, 7736−7743

Analytical Chemistry

Article

Cost Estimation. The material cost of a single ICAT test was estimated to be ∼$3 with the most expensive component being the cAb and the cost of threads being negligible. The equipment costs, including a flatbed scanner (typically ∼$100) and a plasma chamber, were not included in this calculation. Thread has a smaller surface area than lateral flow assay tests, so it should be feasible to use less reagents than for larger structures. Currently, the cAb were manually applied at 0.6 μL per thread; with the use of automated liquid dispensing, it should be feasible to further reduce the cost by a factor of 10 by reducing the usage of antibody reagents 10-fold or more. In addition, the bulk prices for antibodies are much lower than the list prices used here, and we thus estimate that it is possible to reduce material costs to less than $1 for an ICAT cartridge. A more accurate estimation of manufacturing costs should also consider the economy of scaling such as the ones achieved for lateral flow assays, which are made for “$0.10−$3.00” per test in production.2

detection range of the ICAT was comparable to other lateral flow tests for CRP and to low-cost paper-based tests. The three proteins were detected in serum diluted with buffer, which is a common approach and is notably used in ELISA.42 Considering that serum is one of the most challenging samples, we anticipate that the ICAT can be extended to the analysis of other physiological fluids, such as saliva,43 tears,44−46 or sweat.47,48 The material cost for ICAT was estimated to be less than $3 per cartridge, and could be further reduced if automated reagent coating and assembly processes are used. Future developments may involve exploring alternative types of threads to further improve the LOD, using textile machinery to assemble the cartridge,21 and combining other “smart” textiles for biosensing.49 Paper-based and lateral-flow assay devices differ from ICAT on a number of key points. They follow a “top-down” assembly approach, need to be precoated with the reagents and depend on advanced manufacturing techniques to achieve multiplexing. Conversely, multiplexed ICATs can be assembled à la carte from threads precoated with antibodies against various proteins while maintaining a small footprint. The modularity of the ICAT may be useful for applications in low resource settings and for making tests on an “as needed” basis by the end users, potentially at the point-of-care.

Table 1. Material Cost Estimation of ICAT Per Test material Au nanoparticles cAb (1 mg/mL) dAb (15 μg/mL) control Ab cotton threads, nylon chemicals total

cost/volume

volume/test (μL)

$150/5000 μL 3 $300/100 μL 0.6 $300/100 μL 2 $100/500 μL 0.5 fiber bundles, polymer frame, and

cost per test $0.1 $1.8 $0.9 $0.1