Paper and flexible substrates as materials for biosensing platforms to detect multiple biotargets

Abstract

The need for sensitive, robust, portable, and inexpensive biosensing platforms is of significant interest in clinical applications for disease diagnosis and treatment monitoring at the point-of-care (POC) settings. Rapid, accurate POC diagnostic assays play a crucial role in developing countries, where there are limited laboratory infrastructure, trained personnel, and financial support. However, current diagnostic assays commonly require long assay time, sophisticated infrastructure and expensive reagents that are not compatible with resource-constrained settings. Although paper and flexible material-based platform technologies provide alternative approaches to develop POC diagnostic assays for broad applications in medicine, they have technical challenges integrating to different detection modalities. Here, we address the limited capability of current paper and flexible material-based platforms by integrating cellulose paper and flexible polyester films as diagnostic biosensing materials with various detection modalities through the development and validation of new widely applicable electrical and optical sensing mechanisms using antibodies and peptides. By incorporating these different detection modalities, we present selective and accurate capture and detection of multiple biotargets including viruses (Human Immunodeficiency Virus-1), bacteria (Escherichia coli and Staphylococcus aureus), and cells (CD4(+) T lymphocytes) from fingerprick volume equivalent of multiple biological specimens such as whole blood, plasma, and peritoneal dialysis effluent with clinically relevant detection and sensitivity.

Figure 1. Schematic of seamlessly integrated paper and flexible substrate-based platforms with broadly applicable electrical and optical sensing modalities.

(A) Flexible electrodes on polyester film-based electrical sensing platform for HIV detection. (i) Viruses are captured using streptavidin-coated magnetic beads conjugated with biotinylated polyclonal anti-gp120 antibodies in clinically relevant samples such as whole blood (ii) Sample is washed to remove electrically conductive solution. (iii) Captured viruses are lysed using 1% Triton X-100, releasing to release ions and biomolecules of the captured viruses. These biomolecules change the electrical conductivity of the solution. (iv) The impedance magnitude of the viral lysate samples decreases compared with that of control samples. Figure 1A is drawn by H.S. (B) Nanoparticle aggregation concept for bacteria detection on a cellulose paper and its incorporation with mobile phone camera. (i) Gold nanoparticles (AuNPs) are modified with 11-Mercaptoundeconoic acid (MUA), N-Ethyl-N'-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC), and N-hydroxysuccinimide (NHS) to form coupling groups on nanoparticle surface. AuNPs are further functionalized with specific recognition elements for bacteria pathogens. (ii) Bacteria-spiked samples are added to the modified nanoparticle solution. Bacteria samples cause nanoparticle aggregation and change the color of solution as detected with a mobile phone. Figure 1B is drawn by F.I. (C) Lensless imaging detection and counting of CD4⁺ T lymphocytes on polyester film-based platform with microchannels. (i) Surface chemistry on the platform are designed to immobilize biotinylated anti-CD4 antibodies using NeutrAvidin on chemically activated surface including N-g-Maleimidobutyryloxy succinimide ester (GMBS) and 3-Mercaptopropyl trimethoxysilane (3-MPS) modifications. (ii) CD4⁺ T lymphocytes are selectively captured from a whole blood sample on the surface of platform. (iii) Non-captured cells are washed away from the microchannels of platform. (iv) Captured CD4⁺ T cells are then detected and quantified using a lensless shadow imaging technology. (v) Shadows of captured CD4⁺ T cells on the substrate. Figure 1C is drawn and designed by H.S., W.A. and M.H.Z.

Figure 2. Impedance spectroscopy results for HIV-spiked DPBS samples.

Impedance magnitude (I.M.) (A) and phase spectra (B) of lysed HIV-1 subtypes A, B, C, D, E, G, and panel for frequencies between 100 Hz and 1 MHz measured on the flexible polyester film-based electrical sensing devices. (C) Impedance magnitude of HIV-1 lysate samples (HIV subtypes A, B, C, D, E, G, and panel) at 1,000 Hz and 1 V, where there was maximum impedance magnitude shift. (D) Normalized impedance magnitude change of HIV-1 lysate samples with respect to control samples. (E) Repeatability of the measured impedance magnitude for HIV subtypes A, B, C, D, E, G, and panel. The system demonstrated a repeatability between 88% and 99%. The control samples were virus-free DPBS samples. Statistical assessment on the results was performed using ANOVA with Tukey's posthoc test for multiple comparisons. Statistical significance threshold was set at 0.05 (p < 0.05). Error bars represent standard deviation of the mean (n = 3). Asterisk indicates statistically significant impedance magnitude change compared to all other groups. The viral loads of these HIV-1 subtypes were 1.74 × 10⁸, 1.2 × 10⁸, 1.17 × 10⁸, 2.9 × 10⁸, 8.39 × 10⁸, 6.53 × 10⁸, and 1.49 × 10⁹ copies/mL for HIV subtypes A, B, C, D, E, G, and panel, respectively.

Figure 3. Evaluation of specificity characteristics for the flexible material (i.e., polyester film)-based electrical sensing platform.

EBV- and HIV-spiked DPBS samples were used in the experiments. Control samples were virus-free DPBS. Average impedance magnitude (A) for lysed HIV-1 subtype A, EBV, and mixture of HIV subtype A and EBV for frequencies between 100 Hz and 1 MHz. (B) Impedance magnitude of the lysed HIV-1 subtype A, EBV, and mixture of HIV subtype A and EBV samples at 1,000 Hz and 1 V. (C) Repeatability of the impedance magnitude shift for HIV-1 subtype A, EBV, and mixture of HIV and EBV samples. The system demonstrated repeatability between 79% and 99%. Average impedance magnitude of control and lysed HIV-1 subtype C spiked in whole blood (D) and plasma (E) at 1,000 Hz and 1 V. The control samples were virus-free blood and virus-free plasma for experimental results reported in (D) and (E), respectively. Error bars represent standard error of the mean (n = 3). (F) Normalized impedance magnitude change of HIV-1 lysate samples with respect to control samples at 1,000 Hz and 1 V in spiked whole blood and plasma samples. Statistical assessment on the results was performed using ANOVA with Tukey's posthoc test for multiple comparisons. Statistical significance threshold was set at 0.05 (p < 0.05). Error bars represent standard error of the mean (n = 3). Brackets connecting individual groups indicate statistically significant impedance magnitude difference (p < 0.05). The viral loads of the HIV-1 subtype A and C used here were 1.74 × 10⁸, 1.17 × 10⁸ copies/mL, respectively.

Figure 4. Cellulose paper-based nanoparticle aggregation for the detection of bacteria with a mobile phone system.

Bacteria-spiked samples were first added into the modified gold nanoparticle solution, and then, transferred to a cellulose paper for distribution of nanoparticles due to capillary effect. (A) The images of sample spots were taken using a mobile phone camera for image acquisition, and analyzed using an image processing MATLAB code in the computer. Red (R) pixel intensity value was used in the data analysis. (B) E. coli-spiked into DPBS samples ranging from 8 to 1.5 × 10⁶ CFUs/mL were evaluated. The limit of detection (LOD) was observed to be 8 CFUs/mL (n = 5). (C) E. coli-spiked into serum samples ranging from 8 to 1.5 × 10⁶ CFUs/mL were evaluated. The LOD was observed as 8 CFUs/mL (n = 5). (D) E. coli-spiked in PD fluid samples ranging from 8 to 1.5 × 10⁶ CFUs/mL were evaluated. The LOD was observed as 8 CFUs/mL (n = 5). E. coli-free solutions were used as control samples, and the corresponding signals were represented in line and dashes in the plots to show the mean and standard error of mean, respectively (n = 5). (E) Three sets of experiments were designed for multiplexing and specificity experiments. S. aureus (10⁵ CFUs/mL) were applied into anti-LTA antibody-modified gold nanoparticle solution for multiplexing. E. coli (10⁵ CFUs/mL) were sampled into anti-LTA antibody-modified gold nanoparticle solution for specificity. A mixed bacteria solution was examined with anti-LTA antibody-modified gold nanoparticle solution for nonspecific interactions. Statistical analysis demonstrated that S. aureus and mixture experiments were statistically different than E. coli samples (n = 5, p < 0.05), and there was no significant difference observed between S. aureus and mixture samples (n = 5, p > 0.05). (F) Evaluation of image code with other mobile phone cameras. S. aureus experiments were further assessed with iPhone 4, Samsung i9300 Galaxy S3 and HTC Vivid. There was no statistical difference between various mobile phone cameras (n = 5, p > 0.05). Statistical analysis was described in Supplementary Information. Brackets connecting individual groups indicate statistically significant difference. Error bars represent standard error of the mean.

Figure 5. A flexible material (i.e., polyester film)-based platform for the capture of CD4⁺ T lymphocytes for HIV-1 diagnosis.

In the experiments, glass was determined as a standard substrate, and compared with polyester film (3915 300ga Hostaphan®). (A) Intensity plots for glass and Hostaphan®. Images are taken using various fluorescent filters (DAPI, GFP, and CY5) and exposure time. (B) Fluorescent intensity produced by surface immobilized NeutrAvidin conjugated FITC at various concentrations. (C) Fluorescent images of captured cells on these flexible film-based microfluidic devices. (i) Bright-field image, (ii) DAPI image staining nucleated cells, and (iii) Alexa647 stained CD4⁺ T lymphocytes cells. (D) Comparison of CD4⁺ T lymphocytes capture efficiency from whole blood using the platform at various anti-CD4 capture antibody concentrations. There was no statistical difference observed between the groups (n = 3, p > 0.05). (E) Comparison of capture specificity from whole blood using the platform at various anti-CD4 capture antibody concentrations. There was no statistical difference observed between the groups (n = 3, p > 0.05). Statistical assessment was performed using ANOVA with Tukey's posthoc test for multiple comparisons. Statistical significance threshold was set at 0.05 (p < 0.05). Error bars represent standard deviation of the mean.