Micro-a-fluidics ELISA for rapid CD4 cell count at the point-of-care

Abstract

HIV has become one of the most devastating pathogens in human history. Despite fast progress in HIV-related basic research, antiretroviral therapy (ART) remains the most effective method to save AIDS patients' lives. Unfortunately, ART cannot be universally accessed, especially in developing countries, due to the lack of effective treatment monitoring diagnostics. Here, we present an inexpensive, rapid and portable micro-a-fluidic platform, which can streamline the process of an enzyme-linked immunosorbent assay (ELISA) in a fully automated manner for CD4 cell count. The micro-a-fluidic CD4 cell count is achieved by eliminating operational fluid flow via "moving the substrate", as opposed to "flowing liquid" in traditional ELISA or microfluidic methods. This is the first demonstration of capturing and detecting cells from unprocessed whole blood using the enzyme-linked immunosorbent assay (ELISA) in a microfluidic channel. Combined with cell phone imaging, the presented micro-a-fluidic ELISA platform holds great promise for offering rapid CD4 cell count to scale up much needed ART in resource-constrained settings. The developed system can be extended to multiple areas for ELISA-related assays.

Figure 1. Design of m-ELISA for CD4⁺ T lymphocyte count.

(a) Overview of the m-ELISA based chip platform. The testing process can be divided into three steps consisting of sample collection, rapid cell-based m-ELISA, and cell phone based colorimetric detection. A droplet from a finger prick can be collected and loaded onto a micro-a-fluidic chip along with antibody-functionalized magnetic beads. The micro-a-fluidic chip is placed on a permanent magnet, which is fixed on a motorized stage. With the aid of a software program, the stage is used to control and complete the entire process of ELISA in an automated manner. As a result, the color development on-chip is imaged using a cell phone and the result is analyzed using an integrated mobile application. (b) Design of a micro-a-fluidic chip and cell-based CD4 ELISA. The chip is designed to contain five aqueous circular chambers (i, ii, iii and two wash chambers containing phosphate buffered saline plus 0.5% tween 20 (PBST)) and five elliptical chambers containing mineral oil for physical separation of the hydrophilic circular chamber liquids. In chamber i, functionalized magnetic beads conjugated with anti-CD4 antibody are used to selectively capture cells (CD4⁺ T lymphocytes and monocytes) expressing CD4 molecules on the surface. After a washing step in the intermediate chamber, the magnetic beads with captured CD4⁺ cells are moved to chamber ii and tagged with anti-CD3 antibody, which is conjugated with horseradish peroxidase (HRP). After another wash, the captured CD4⁺ T lymphocytes enable color development due to digestion of 3,3′,5,5′-Tetramethylbenzidine (TMB) by HRP in chamber iii. The photograph in (a) was taken by ShuQi Wang.

Figure 2. Characterization of m-ELISA for CD4 cell count.

(a) Capture of CD4⁺ T lymphocytes (red arrow) on magnetic beads (blue arrow). Magnetic beads were incubated with whole blood and the captured T lymphocyte was stained with anti-CD4 Alexa Fluor 488 and anti-CD3 Alexa Fluor 647. The fluorescent images of Fluor 488, Alexa Fluor 647, bright-field, as well as the overlaid image (by ImageJ software) demonstrate capture of CD4⁺ T lymphocytes by magnetic beads. The scanning electron microscope (SEM) image validates the capture of CD4⁺ T lymphocyte on magnetic beads. The scale bar is 10 μm. (b) Efficiency of magnetic beads for capturing CD4⁺ T lymphocytes at varying incubation times. 25 μL of functionalized magnetic beads (10 mg/mL) were incubated with whole blood containing 1,000 cells/μL CD4⁺ T lymphocyte. The unbound T lymphocytes in the supernatant were quantified by flow cytometry, which was used to calculate the capture efficiency. One-way ANOVA was performed to analyze and no statistical significance (p > 0.05) was found. (c) Efficiency of magnetic beads for capturing CD4⁺ T lymphocytes at varying concentrations of CD4⁺ T lymphocytes. The same procedure as mentioned in (b) was used to calculate the capture efficiency. One-way ANOVA analysis indicated that there was no significant difference in capture efficiency (p > 0.05). (d) Optimization of HRP-antibody in m-ELISA. The variable, HRP-conjugated antibody was diluted at different concentrations and tested with plasma samples, which served as negative control. 1:80,000 was chosen for the following experiments since this concentration was the highest concentration of HRP-antibody that avoided false positive results. A mobile application was used to extract red (R) pixel values from the color development region on-chip and normalized them according to the background as shown in the Y-axis. One-way ANOVA was performed to analyze the color intensity caused by excessive HRP (n = 3, * indicates the statistical significance, p < 0.05). (e) Whole blood from a healthy blood sample was serially diluted (undiluted, 1:3, 1:9, 1:27 and 1:81) used to construct the standard curve with different 3, 3′, 5, 5′ Tetramethylbenzidine (TMB) incubation periods. Each data point was performed in triplicates and the standard deviation was shown.

Figure 3. Computational modeling of driving forces exerted on magnetic beads.

(a) Magnetic forces on a single magnetic bead in z-direction for a range of vertical distance, z = 1, 2, and 3 mm. (b) Magnetic forces on a single magnetic bead in r-direction. (c) Critical height (h), i.e., the location where magnetic forces exerted on beads is comparable to Brownian forces, is calculated as a function of magnetic bead size (Rp) using Eq. 1–4 given in the supplementary information. Brownian motion of magnetic particles is not desirable because such motion would limit the bead transport. Results showed that Brownian forces are not effective in our experiments with the current settings and dimensions of the setup. (d) Capillary pressure difference sustained across the interface between mineral oil and PBS (dashed blue line), and magnetic forces (z = 2 mm) per cross-section area of magnetic beads (solid red line). (e) Simulation results of magnetic flux density (contours) and unit magnetic force along the horizontal plane (arrows) when a magnet moves horizontally from left to right. Results are plotted for a range of time points (top-to-bottom, left-to-right). Full results are given in the supplementary information. Dashed lines are the circular edges of the cylindrical magnet (top view).

Figure 4. Clinical validation of micro-a-fluidic ELISA for CD4⁺ T lymphocyte count.

(a) and (b) Correlation of flow cytometry with micro-a-fluidic ELISA for CD4⁺ T lymphocyte count in 17 AIDS patient samples and 18 immunosuppressed patient samples, respectively. (c) and (d) Bland-Altman analysis for the difference between flow cytometry and micro-a-fluidic ELISA for CD4⁺ T lymphocyte count in 17 AIDS patient samples and 18 immunosuppressed patient samples, respectively. The 95% confidence interval on the mean is shown on the graphs. (e) Accuracy of CD4⁺ T lymphocyte micro-a-fluidic ELISA at the WHO clinical cutoff of 350 cells/μL for prioritizing ART. The samples were defined as positive once flow cytometry count was below 350 cells/μL, otherwise negative.