Composable Microfluidic Plates (cPlate): A Simple and Scalable Fluid Manipulation System for Multiplexed Enzyme-Linked Immunosorbent Assay (ELISA)

Abstract

Enzyme-linked immunosorbent assay (ELISA) is the gold standard method for protein biomarkers. However, scaling up ELISA for multiplexed biomarker analysis is not a trivial task due to the lengthy procedures for fluid manipulation and high reagent/sample consumption. Herein, we present a highly scalable multiplexed ELISA that achieves similar level of performance to commercial single target ELISA kits as well as shorter assay time, less consumption, and simpler procedures. This ELISA is enabled by a novel microscale fluid manipulation method, composable microfluidic plates (cPlate), which are comprised of miniaturized 96-well plates and their corresponding channel plates. By assembling and disassembling the plates, all the fluid manipulations for 96 independent ELISA reactions can be achieved simultaneously without any external fluid manipulation equipment. Simultaneous quantification of four protein biomarkers in serum samples is demonstrated with the cPlate system, achieving high sensitivity and specificity (~ pg/mL), short assay time (~1 h), low consumption (~5 μL/well), high scalability and ease of use. This platform is further applied to probe the levels of three protein biomarkers related to vascular dysfunction under pulmonary nanoparticle exposure in rat’s plasma. Because of the low cost, portability, and instrument-free nature of the cPlate system, it will have great potential for multiplexed point-of-care testing in resource-limited regions.

Quantitative detection of biomarkers is of great importance for both fundamental biological research and clinical diagnostics.^1-3 Enzyme-linked immunosorbent assay (ELISA) is the gold standard technique for the detection and quantification of protein biomarkers in complex matrices. Conventional ELISA is routinely performed in 96-well microtiter plates, enabling reliable, specific and versatile analysis. Compared to other immunoassay methods, such as lateral flow immunoassay and protein microarrays, ELISA offers higher sensitivity and specificity, due to its amplifying detection signals and multi-step sample/reagent manipulation for reducing non-specific bindings.⁴ However, conventional ELISA also suffers from several limitations. It requires high sample and reagent consumption (100 μL per well for each step), long incubation time (1-3 h for each step), and laborious loading and washing procedures.⁵ These issues make multiplexing especially difficult for ELISA as increasing the number of analytes will drastically increase system complexity and the number of operations as well as consume larger volumes of reagents and samples.

Since many diseases are complex and multifactorial, results from a single biomarker testing are not always adequate for the diagnosis or prognosis of a disease.^6-8 Therefore, multiplexed ELISA, which simultaneously detects different biomarkers from a single specimen, is highly desired to offer more reliable diagnostic results.⁹ To address the issues in conventional ELISA, microfluidics is a powerful platform due to its superior capability of manipulating small amount of fluids. Incorporation of microfluidics enables precise fluid control, short turnaround time and reduced reagent consumption. In addition, miniaturized working units and high integration capability facilitate the multiplexed testing.^10-12 To date, several microfluidic-based multiplexed ELISA systems have been reported, and multiplexing is mainly achieved through three strategies: (i) spatial separation of detection sites, (ii) regional separation of detection units by channel networks, and (iii) color-encoded beads.¹³ The spatial separation approach is always in a microarray format, which can be fabricated by wax-printing on μPADs^14-15, hydrophobic coating on plastic plates¹⁶, robotic spot printing in wells or channels^17-21, or DNA-directed assembly²². The microarray format is amenable to being scaled up and is straightforward to perform. However, its performance is often not at the same level of conventional ELISA due to the matrix effect, antibody cross-reactivity, and limited signal amplification.^23-25 Using an aqueous two-phase system with the microarray format could improve the performance by preventing antibody crosstalk. Efforts have been made to simplify the assay procedures for scalable ELISA.^26-28 For the regional separation strategy, multiple parallel units are integrated in a chip for different target analysis.¹³ Generally, this strategy exploits microchannels to facilitate the manipulation of small amount fluids for ELISA. For example, fluoropolymer microcapillary film²⁹, disposal sensor array^30-32, and 3D-printed pipette tip³³ systems all use parallel microchannel arrays to perform multiplexed testing. Lab-on-a-disc^34,35, “mChip”^36,37, SlipChip^38,39 and thermoelastic microvalve⁴⁰ platforms detect different analytes in discrete parallel chambers connected by microchannels. An exception is digital microfluidics, which confine different reactions in discrete droplets that are manipulated by insulated electrodes.^41,42 These methods streamline the ELISA procedures and often result in comparable performance to the conventional ELISA platform. However, these methods are often difficult to scale up further as they require navigating samples and reagents through microchannels to complete all the ELISA procedures, which is limited by the device footprint and/or external equipment (pumps, centrifuges or electrode controls). The number of ELISA reactions that can be run per device is often less than 12 in practical applications. The last strategy utilizes color-encoded beads for multiplexing, such as the digital ELISA system based on droplet microfluidics or microwell arrays.^43-45 Digital ELISA can achieve high sensitivity and has the capability for absolute quantification without calibration, but multiplexing is restricted by the number of distinguishable color labels and requires more complex and sophisticated optical detection systems.

As discussed above, existing multiplexed ELISA systems either suffer from sub-optimal performance (microarray format) or have limited scalability (microchannel format, color-encoded format). Here we report a novel multiplexed ELISA platform that brings “the best of both worlds” involving optimal performance, scalability, multiplexing, and ease of use into one system without using sophisticated equipment. This platform, composable microfluidic plates (cPlate), enables a unique flow control scheme that handles small volumes of fluids based on the assembly of 3D-printed microchannels and microwells with patterned superhydrophobic coatings (Figure 1). The cPlate enables complete physical separation of each individual ELISA reaction, eliminating antibody crosstalk, allowing one-step sample loading and single-phase fluid manipulation for reagent delivery and washing. Similar to the classic SlipChip concept,^39,46,47 the present method also shares the advantage of using the microdevice itself to drive fluid, making the system self-sufficient and scalable. However, SlipChip is based on horizontal movement, whereas the present method utilizes vertical movement. For applications like ELISA where repetitive fluid operation steps are needed, horizontal movement requires that all essential reagents are held in one plate, which largely restricts the total number of independent reactions performed simultaneously. In addition, the present method does not require the use of oil which further simplifies the operation and improves the robustness of the whole workflow. As shown in Figure 1, the main components of the cPlate are miniaturized 96-well plates and their corresponding channel plates, which are fabricated by stereolithographic 3D printing and coated with superhydrophobic layers on the outer surfaces. Spatially resolved fluid compartmentalization can be simply achieved by assembling the well plate with the channel plate and subsequently injecting fluids with a pipette. Combination of two preloaded well plates face-to-face allows fluids to merge in two wells at the corresponding positions, while fluid splits are achieved by plate disassembly. Thus, a magnetic bead-based multiplexed ELISA can be performed by a sequential assembling process of well plates preloaded with essential reagents for each ELISA step. This ELISA platform achieves highly multiplexed biomarker analysis with simple operations, short assay time (~1 h), small sample consumption (5 μL for each well) and high-throughput testing (96 reactions in a single run). Notably, the whole system does not require any external electric power or equipment. Only the microdevices, a pipette, a wash buffer container, and a magnet are needed to complete an ELISA experiment (Figure 1b). The cPlate based ELISA achieves comparable limit of detection (LOD) and dynamic range to commercial 96-well plate ELISA kits for simultaneous quantification of four protein biomarkers in serum samples. This platform has also been applied to examine plasma protein biomarkers in rat plasma after pulmonary nanoparticle exposures, to validate this novel technology in a biologic model.

Figure 1.

Working principle of the cPlate for multiplexed ELISA. (a) i. Spatially resolved reagent loading: miniaturized 96-well plate is assembled to their corresponding channel plate, and multiple fluidic paths are formed. Different reagents can be loaded through different inlets by a pipette, flow through the hydrophilic fluidic paths and be delivered to multiple wells; ii. Incubation step: two preloaded well plates are assembled face-to-face, and solutions in two corresponding wells contact each other, enabling the diffusion and reaction of the reagents; iii. Disassembly: two well plates can be disassembled after incubation, which disconnects two wells and allows further operation or detection. (b) The required apparatuses for a whole multiplexed ELISA process. (c) Working procedure of the magnetic bead-based multiplexed sandwich ELISA in the 3D-printed cPlate platform.

EXPERIMENT SECTION

Workflow of the Multiplexed ELISA.

The cPlate enables a convenient workflow for running multiplexed ELISA without the need of sophisticated fluid handling equipment such as pumps and/or robotics. In this work, we used bead-based ELISA as it offers more flexibility in designing the multiplex format and allows reuse of the plates by simply flushing with water to remove the residuals and drying with air flow. All the required components for completing a multiplexed ELISA assay with the cPlate are shown in Figure 1b. After aligned with their corresponding channel plates, microwell plates are loaded with desired reagents/samples by a single pipette through the designed fluid paths, generating patterned arrays containing capture Ab-coated magnetic beads, protein standards/samples, enzyme-labeled detection antibodies, or chromogenic substrates. Different biomarkers can be tested in different regions of the well plate. To perform a multiplexed sandwich ELISA, a sequential plate assembly process is conducted (Figure 1c). The magnetic beads plate is first assembled with the standards/samples plate, then the beads are transferred into the corresponding standard/sample wells by a magnet, enabling the interactions between the capture antibodies on the bead surface and antigens in solution. After incubation, the two plates are disassembled with a magnet placed below the bead plate, which allows for retention of the magnetic beads and avoids bead loss during the separation. The whole bead plate is then immersed into a washing buffer with a magnet underneath, to rinse the unbound substances. After thorough washing, the bead plate is subsequently assembled with the detection Ab plate, allowing for the recognition between detection antibodies and antigens, then rinsed again with washing buffer after incubation. Finally, by assembling with the substrate plate, enzymes on the beads can catalyze the substrate to generate optical signals for detection. Disassembly of the two plates on a magnet separates the beads from substrate, which stops the reaction and eliminates the influence of magnetic particles on optical detection.

Materials and reagents, detailed fabrication, cPlate validation process, single and multiplexed biomarker analysis procedures, and rat sample processing can be found in the Supporting Information.

RESULTS

Design Principles of the cPlate.

To address the scalability, we choose the microwell format for each individual ELISA reaction. To manipulate small volume of fluids in different wells simultaneously and conveniently, the cPlate is designed with the following features: 1) small-volume wells for low consumption and easy washing; 2) superhydrophobic coating facilitating the merging and splitting of solutions in two wells; 3) microstructures for preventing cross-talk among wells; 4) plates with 3D microchannels for convenient and spatially resolved delivery of specific reagents.

The detailed design and dimensions of the well and channel plates are demonstrated in the Supporting Information and Figure S1. The radius of each well is 1 mm. After printing, a commercial superhydrophobic coating reagent was applied to the outer surfaces of channel plates and well plates by an air powered sprayer. To avoid undesired coating inside the wells, well plates are assembled with corresponding 3D-printed shield structures (Figure S1c) before the spray coating.

For initial reagents/sample delivery, we utilize the flexibility of 3D printing and design the channel plates, which contain discrete microchannels. As shown in Figure 1a, when the channel plate is assembled to the well plate, microwells are connected by the microchannels, forming multiple fluidic paths. Reagents can be loaded through inlets using a pipette, and because of the superhydrophobic coating on both surfaces, these aqueous solutions will only flow through the hydrophilic fluidic paths and be delivered to multiple wells. After loading, the two plates are disassembled, generating a patterned array of reagents. This design simplifies fluidic handling and eliminates the requirement for pumps or valves in conventional microfluidic systems. Owing to the flexibility of channel design by 3D printing, different reagents can be delivered via different flow paths, and then spatially compartmentalized into different wells, achieving highly multiplexed capability. For channel plates, different fluid paths are designed depending on the assay needs, and the inlet and outlet of the discrete microchannels are all designed as zigzag paths (Figure S1e), which can avoid the contamination of channels by the superhydrophobic coating.

For subsequent reagent mixing and incubation, two preloaded well plates are assembled face-to-face, allowing solutions in two corresponding wells to contact with each other and mix. This process allows for manipulation of multiple independent reaction compartments simultaneously for reagent incubation with one simple operation. To enable easy and precise alignment of two corresponding plates, concave or convex aligned structures are designed on the four corners of the plates. After the incubation step is complete, the two plates can be disassembled, which disconnects two wells without affecting the solution volumes, thereby allowing for further operations.

Washing is critical in ELISA as it removes non-specific binding and reduces background signals. Our 3D printed microwells allow a simple bulk washing procedure because of the shallow depth of each well. One-step washing for multiple wells was achieved by directly immersing the whole bead plate into the washing buffer and shaking the washing chamber gently, while the beads are retained in their respective wells by a magnet. Because of the superhydrophobic coating, no residual solution is left on the plate surface except for inside the wells after pouring out the washing buffer. Collectively, all the steps that are necessary for a magnetic beads-based ELISA can be performed in a simple and scalable manner.

Validation and Optimization of the cPlate for ELISA.

Before using the cPlate for detecting biomarkers, we validated and optimized a series of key operations in ELISA. First, we studied the reagent loading performance of using a pipette and the assembled plates (Figure 2a). As shown in Figure 2b and 2c, simple pipette loading from different inlets, different food dyes were delivered into multiple wells without leakage, forming arbitrary patterned reagent arrays after disassembly. To examine the volume consistency of different wells after loading, fluorescent solutions (Alexa 488-strepavidin (SA) as a model for big molecules or resorufin as a model of small molecules) were loaded through different fluid paths. Figure 2d showed the fluorescence pattern of the wells, with green signals generated from Alexa 488-SA and red signals generated from resorufin, which are consistent with loading paths (Figure S2a). Fluorescence profiles along each lane were also plotted. As shown in Figure 2e and 2f, fluorescent intensities are uniform among the wells in each lane, with a coefficient of variation (CV) of 3.7% for Alexa 488-SA (lane 1) and 2.6% for resorufin (lane 2), indicating the consistent volume delivery to different wells without using any active volume control methods. We also examined the cross-contamination of this loading method. Figure 2g and 2h showed the green and red fluorescence profiles of lane 3, respectively. For Alexa 488-SA-loaded wells, there are high green signals but bare of red fluorescence, whereas the resorufin-loaded wells are the opposite. For wells loaded with the mixture solution, both fluorescence signals were observed, whereas for wells loaded with buffer, neither fluorescence can be detected. These results indicated that there is no fluid leakage or inter-well cross-contamination during the loading process, which is essential for the accuracy of multiplexed ELISA.

Figure 2.

Validation of the reagent loading process. (a) Schematic of the spatially resolved reagent loading. (b) (c) Different channel designs generate different patterned reagent arrays after disassembly, visualized by food dyes. (d) Fluorescence image of well plate loaded with different fluorescent solutions. Fluorescence profiles along the (e) lane 1 and (f) lane 2, and fluorescence profiles along the lane 3 under (g) green channel and (h) red channel were plotted for quantitively evaluating the reagent delivery consistency.

Second, we studied the performance of manipulating fluid by assembling and disassembling two well plates. By combining two plates respectively loaded with vertical and horizontal color gradient, we can observe the merging and mixing of the food dyes in paired wells, yielding an array with two-direction gradients after disassembly (Figure S4a). To validate the independence of each reaction compartment, as shown in Figure 3a, one well plate was loaded with amplex red/H₂O₂, and the other plate was loaded with HRP-coated beads or uncoated beads every other lane. After assembly, the HRP catalyzed the substrates and generated fluorescent product. Fluorescence measurement across the plate demonstrated uniform and high fluorescence in the wells incubated with HRP-coated beads, and no fluorescence in the wells incubated with uncoated beads (Figure 3b). These results indicated that reaction compartments in our system are isolated and the superhydrophobic layers on the plates and small groove structures between wells were sufficient to prevent cross-contamination during the fluid manipulation (Figure S3).

Figure 3.

Validation of the incubation and washing procedures. (a) Assembly of two preloaded well plates. One plate contains amplex red/H₂O₂, and the other plate contains HRP-coated beads or unmodified beads every other lane. (b) Fluorescence intensity of the well plate shown in (a) after incubation and disassembly. The number in x-axis represents the number of different lanes. (c) Change of fluorescence intensity for wells loaded with resorufin or Alexa 488-SA under green and red channels after different washing cycles, respectively. Each cycle lasts for 3 min. (d) Fluorescence changes of wells loaded with resorufin or Alexa 488-SA after different washing durations. *** indicates P≤0.001.

The washing efficiency was evaluated by monitoring fluorescence in each well after washing different times or with different time durations. As shown in Figure 3c, fluorescence signals in all wells were reduced to the background levels after only one washing cycle, confirming that the rinsing process was efficient with no evident carryover. Figure 3d demonstrated the effect of time duration, as 1 min was sufficient to thoroughly remove unbound molecules. After washing is complete, the volume consistency of solutions remaining in the wells was important for subsequent assembly process and the accuracy of the assay. By assembling the washed plate to a shallow, empty well plate (Figure S4b), extra solution in each well would fill into the corresponding empty well, and a more uniform volume distribution across the whole well plate was achieved after separating the two plates, with a CV of 5.2% (Figure S4c).

Single Biomarker Analysis.

We examined the performance of this method for running a complete ELISA assay. Interleukin-6 (IL-6), as an important biomarker for inflammation, was first tested as a model for protein biomarker.⁴⁸ IL-6 capture Ab-coated magnetic beads, IL-6 standards/samples, detection Ab with SA-HRP, and HRP substrate amplex red were preloaded into different well plates in certain patterns, respectively. The assay was conducted by sequentially assembling the preloaded well plates. Because of the high multiplexing capability of our system, important parameters that might impact the ELISA results, include the capture Ab concentration for bead coating, magnetic beads concentration, detection Ab and SA-HRP concentration, can be evaluated simultaneously in a single experiment (Figure S2b). As shown in Figure 4a, 4b, and 4c, considering the assay sensitivity, variation and reagent consumption, the capture Ab concentration of 0.25 mg/mL, the magnetic beads concentration of 0.05 wt%, and the detection Ab and SA-HRP concentrations of 1 μg/mL were determined as the optimal assay conditions. The incubation time for each step was also assessed (Figure 4d, 4e), and the results indicated that 15 min was sufficient to obtain distinguishable signal differences for samples in different concentrations. Further extending the incubation time did not show significant improvements. Therefore, the whole ELISA assay could be accomplished within 1 hour. The relatively short incubation time was due to the small sample/reagent volume (5 μL in each well) and the capability of moving the magnetic beads inside the wells by a magnet during incubation, which disturbed static condition and minimized diffusion limitation. Under the optimized conditions, a full IL-6 response curve was obtained (Figure 4f). The limit of detection (LOD) was 1.87 pg/mL and the full dynamic range was 2-1000 pg/mL, which were comparable to the commercial 96-well ELISA kit with the LOD of 0.7 pg/mL and a dynamic range of 3-300 pg/mL.

Figure 4.

Single biomarker analysis with the cPlate. Immunoassay results of IL-6 samples in three different concentrations (0, 10, 100 pg/mL) using (a) different capture Ab concentration for bead coating, (b) different magnetic beads concentration and (c) different detection Ab and SA-HRP concentration (both are 0.2, 0.5, 1 or 2 μg/mL). Signal responses of IL-6 samples in three different concentrations (8, 40, 200 pg/mL) under (d) different sample incubation times and (e) different detection Ab/HRP incubation times. (f) An IL-6 calibration curve. Each data point is the average of 3 parallel measurements. * indicates P≤0.05, ** indicates P≤0.01, and *** indicates P≤0.001.

Multiplexed Biomarker Analysis.

To demonstrate the multiplexing capability of the cPlate, four important biomarkers IL-6, C-reactive protein (CRP), carcinoembryonic antigen (CEA), prostate-specific antigen (PSA) were analyzed simultaneously. Instead of using antibody cocktails, as in many other multiplex ELISAs, different detection antibodies were loaded into different wells corresponding to the pattern of capture Ab-coated magnetic beads, which eliminated the cross-reactions among detection antibodies.^{19,22,29,34,38,49,50} Cross-reactivity was tested by incubating the magnetic beads coated with different capture antibodies, respectively, with each target antigen spiked samples, antigen mixture spiked sample, or blank buffer. Results demonstrated specific responses of the capture antibodies to their matched targets, and no significant signals for nonmatched proteins (Figure 5a), confirming the high specificity of our system. Four protein biomarkers were then quantified simultaneously in a single assay (Figure 5b-e), and their standard deviations for each concentration were less than 10%. The LODs and dynamic ranges were listed in Table 1, which were on the same level as the performance listed in the manuals of commercial ELISA kit for each single target. Only PSA showed a higher LOD compared to the commercial kit, which might be due to the imperfect matching between antibodies and protein from different vendors. Recoveries from spiked serum were measured to validate the accuracy of the multiplexed ELISA system. Serum was diluted 5-fold and spiked with four biomarkers each at three different concentration levels. The obtained recoveries were in the acceptable range of 100 ± 20%. These results demonstrated the high specificity, sensitivity and multiplexing capability of our assay, and that multiple samples can be tested simultaneously in a single run.

Figure 5.

Multiplexed biomarker analysis in the 3D-printed composable microfluidic system. (a) The cross-reactivity among the different antibodies and antigens were systematically evaluated. (b) Calibration curves of (b) CRP, (c) IL-6, (d) CEA, (e) PSA obtained simultaneously from the multiplexed ELISA. Each data point is the average of 4 parallel measurements.

Table 1.

Immunoassay performance and spike-recovery results for the multiplex ELISA.

Biomarker	LOD (pg/mL)	Dynamic range (pg/mL)	Spiked concentration (pg/mL)	Found concentration (pg/mL)	(%) Recovery (±SD)	Commercial kit a
						LOD (pg/mL)	Dynamic range (pg/mL)
CRP	26	156-10000	500	532 (±74)	106 (±15)	22	800-5000
			1000	868 (±77)	87 (±8)
			2000	1761 (±10)	88 (±0.5)
IL-6	1.25	3.13-400	5	4.8 (±0.3)	95 (±6)	0.7	3.1-300
			10	8.2 (±0.4)	82 (±4)
			20	16.2 (±0.3)	81 (±2)
CEA	7.5	31.3-8000	100	139 (±16)	139 (±16)	9.1	46.9-3000
			200	217 (±26)	108 (±13)
			400	399 (±14)	100 (±4)
PSA	62	313-20000	500	623 (±23)	125 (±5)	9.4	31.25-2000
			1000	1121 (±3)	112 (±0.3)
			2000	2378 (±308)	119 (±15)

^aThe values are adopted from the commercial ELISA kits from R&D Systems.

Rat Plasma Sample Analysis.

Finally, we applied the cPlate method to an exploratory study on examining the levels of three protein biomarkers in rat plasma after the TiO₂ nanoparticle exposure. TiO₂ nanoparticles are widely used in industrial and consumer products due to their good photo catalytic properties.⁵¹ Exposure to nanoparticles results in complex responses of biological systems.⁵² It has been reported that the exposure of these nanoparticles negatively impact the lung and is related with an increased risk of cardiovascular dysfunction.^53-55 To study the complex influence of nanoparticle exposure, it is necessary to examine multiple biomarkers. The cPlate method offers a simple means to check multiple protein markers. Here we examined if nanoparticle exposure induces any changes at the global level of three biomarkers (activin, vascular endothelial growth factor (VEGF), and CRP) in pregnant rats. We first determined the analytical performance of the cPlate for these rat biomarkers. Response curves of the three proteins were shown in Figure S5, and the LODs were 7 pg/mL for activin, 13.4 pg/mL for VEGF and 0.19 ng/mL for CRP. Testing results of standard added commercial rat plasma obtained from the cPlate system demonstrated good correlations with that from commercial ELISA kits (R²=0.985 for activin and R²=0.978 for VEGF), indicating the good accuracy of the cPlate system (Figure S6). Rats were either exposed to nano-TiO₂ or filtered-air as control, and plasma were then collected using EDTA as an anticoagulant. The multiplexed analysis result was shown in Figure 6. For all the three markers, there was no significant change between the exposed rats and control rats (p > 0.05). Previous works had shown that inhalation of nano-TiO₂ impaired arteriolar dilation in rats, however, the levels of inflammation markers (e.g. IL-2, IL-8, IL-13) in plasma were mainly unchanged.⁵⁶ Our assay showed similar results, indicating that rats exposed to inhaled nano-TiO₂ does not induce a significant change at the global level. Further investigation with more experimental rats is needed to provide a conclusive determination for this biomarker.

Figure 6.

Multiplexed biomarker measurements from rat plasma samples. Box plot measurements of (a) activin, (b) VEGF and (c) CRP from five nano-TiO₂ PM exposed rats or four untreated rats using the multiplexing ELISA system. Each data point is the average of 2 independent measurements with 2 parallels. NS, no significance.

DISCUSSION

In clinical applications, multiplexed assay, in which multiple biomarkers are analyzed simultaneously from a single sample in a single assay, can provide more accurate and systemic results. ELISA is the clinical gold standard for protein biomarker analysis, but it’s difficult to scale up the number of reactions and scale down the reaction volume of ELISA due to its lengthy procedures for fluid manipulation. The fluid handling method reported here achieves microscale multiplexed ELISA without sacrificing the assay performance and ease of use. This method offers the following merits.

Scalability:

The present cPlate method offers a scalable means of manipulating fluid in each single well by leveraging microscale wells, the design flexibility of 3D printing, and the superhydrophobic coating. Different reagents can be loaded to their desired wells easily assisted by a microchannel plate, whereas other fluid operations can be achieved by simply assembling and disassembling two well plates. Thus, fluid in a large number of wells can be manipulated simultaneously. Figure S7 compares the typical manual ELISA assay procedures with the cPlate. As the number of wells increases, the number of operations increases linearly with a slope >10 for manual ELISA, whereas the number of operations largely remains the same for the cPlate except for one additional operation of loading each additional antibody with the microchannel. The number of wells is fixed at 96 in the present work, which is determined by the printing area of the 3D printer. Using newer generations of benchtop 3D printers with larger printing areas or a professional grade printer can easily overcome this limitation.

Low resource requirement:

Another notable feature of the cPlate is self-sufficiency. A well-known problem for many microfluidic devices is the need of bulky external equipment (e.g., pumps, power supply, gas system) despite the microdevice itself is extremely small. For the cPlate, the key component for fluid manipulation is the device itself. The loading volume is controlled by the dimension of the wells and the surface coating. Thus, operators do not need a secondary mechanism to control reaction volume for ELISA. Also, no oil is needed for the cPlate based protocol, which will reduce the operation complexity associated with controlling two phase liquids. Collectively, these features make the cPlate ELISA more robust to perform and less reliant on the operators’ experimental skills, thereby improving the testing results. The current fabrication cost per plate is <$1 using digital light processing (DLP) 3D printing. It is possible to further reduce the device cost through mass production using injection molding. The only procedure that needs special development is the superhydrophobic coating step, which will require investigating the optimal combination of coating material and procedure for mass production. Because of the minimal equipment requirement, the low cost and the ease of use, we believe the cPlate has great potential in point-of-care testing (POCT), especially considering performing multiplexed ELISA in POCT is still challenging for existing methods. In this report, signals are read out using fluorescent reporters with a fluorescence microscope. For field portable applications, the open well design of the cPlate make it easy to be adapted with colorimetric or chemiluminescence reporters, which can be detected using a cell phone camera or a CCD camera based portable imaging system.

High performance:

The cPlate based ELISA is highly multiplexed and scalable, easy to perform, and requires minimal external equipment and samples/reagents. Notably, these features are achieved without compromising the assay’s analytical performance. We evaluated the performance of the cPlate-based ELISA for 7 protein biomarkers from both human and rat in different matrices (serum, plasma). The results demonstrate that the cPlate based method achieves comparable LODs and dynamic ranges to the commercial ELISA kits. Further improvement of assay sensitivity will require optimizing the antibody pairs, buffers and antibody coating strategies in our cPlate system.

Because the cPlate based fluid manipulation is simple and scalable, we are able to perform complete ELISA protocols with high performance while still maintaining the assay’s low consumption and ease of use for multiple biomarkers. Future efforts will focus on expanding the printing area, integrating a portable optical detection system, improve the sensitivity, and performing on-site multiplexed biomarker testing. In addition, the cPlate based fluid manipulation is a generic method, other ELISA configurations, such as using well surface as solid support, can also be achieved in the cPlate system. The cPlate system is also not limited to ELISA applications, it could be beneficial for many bioassays that require multi-step and parallel fluid manipulations including PCR and drug screening.