Catch-and-Display Immunoassay as an Accessible Platform for Digital Biomarker Detection

Abstract

Digital immunoassays provide exceptional analytical sensitivity for detecting low-abundance biomarkers, but their broad adoption is limited by practical barriers. Commercial platforms are prohibitively expensive for routine use by individual laboratories, and laboratory-scale concepts typically describe specialized biosensors and sophisticated workflows. Here, we introduce a nanomembrane-based Catch-and-Display Immunoassay (CAD-IA) as an accessible digital immunoassay for common laboratory settings. In CAD-IA, fluorescent nanoparticles are "captured" by the nanoscale pores of ultrathin silicon nitride membranes through a pipette powered filtration. The captured nanoparticles serve as optically isolated 'hotspots' for fluorescent immunocomplex formation when target antigen is present. Co-localization of the fluorescent particles and fluorescent immunocomplexes are then "displayed" and quantified by standard confocal microscopy to generate digital signals. CAD-IA is implemented using the μSiM-DX (microfluidic device featuring an ultrathin silicon membrane for diagnostics) platform, which is manually assembled from mass produced, cost-effective components. Using the traumatic brain injury (TBI) biomarker S100B as a model, we demonstrate that CAD-IA provides consistent digital outputs and linear quantification with a dynamic range of at least two orders of magnitude when digital and analog analysis are combined on the same image sets. We further demonstrate that the assay maintains linearity in serum matrices and achieves suitable sensitivity (LoD = 0.02 μg/mL) for clinically relevant diagnostic with the addition of tyramide signal amplification (TSA). While further optimization of CAD-IA is possible, these results constitute a proof-of-concept demonstration of a novel digital immunoassay that is accessible to most laboratory environments.

Figure 1.. Overview of the Nanomembrane-Based Immunoassay Platform.

(A) Ultrathin, permeable nanomembrane-integrated μSiM-DX device, consisting of a well and a microfluidic channel with inlet and outlet ports, separated by a nanoporous membrane (red dashed box). The nanomembrane (NPN) contains nanopores approximately 60 nm in diameter (blue solid box). (B) Conceptual illustration of surface-functionalized nanoparticle capture and molecular transport within the μSiM-DX device. (C) Digital signal readout strategy. Positive signals are detected by the colocalization of immunocomplexes (red channel, middle panel) with captured nanoparticles (left panel), as shown in the merged fluorescence image (right panel).

Figure 2.. Workflow of the Nanomembrane-Based Catch-and-Display Immunoassay (CAD-IA).

(I) Streptavidin-labeled nanoparticles captured on the nanomembrane surface. (II) Biotinylated capture antibodies immobilized onto the nanoparticle surface via streptavidin–biotin interaction. (III) TBI biomarkers bind specifically to the pre-conjugated capture antibodies. (IV) Alexa Fluor 647-labeled detection antibodies introduced to form fluorescent immunocomplexes for signal detection.

Figure 3.. CAD-IA for S100B Detection in 0.1% BSA.

(A) Representative fields of view for the blank sample (0 μg/mL) and S100B positive (0.2 μg/mL) samples. Pronounced colocalization between immunocomplex fluorescence and nanoparticle hotspots is observed only in the presence of S100B. (B) Digital analysis shows dose-dependent response with saturation at 0.75 μg/mL and linear detection from 0.02–0.75 μg/mL (R² = 0.9889). (C) Analog analysis based on total fluorescence enables linear quantification beyond 0.5 μg/mL, complementing digital readout and expanding the dynamic detection range. Data are presented as mean ± SD (N = 6 spots on a single device). Statistical significance was evaluated using one-way ANOVA with Dunnett’s test (****p < 0.0001, ***p = 0.0007).

Figure 4.. CAD-IA for S100B Detection in 10% FBS.

The assay showed consistent responses with a LoD of 1 μg/mL and a linear range from 1 to 4 μg/mL (R² = 0.9600). Data are presented as mean ± SD (N = 6 spots on a single device). Statistical significance was assessed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test. ****p < 0.0001.

Figure 5.. Workflow of TSA-Modified CAD-IA.

(I) Streptavidin-labeled nanoparticles captured on the nanomembrane surface. (II) Biotinylated capture antibodies immobilized onto the nanoparticle surface via streptavidin–biotin interaction. (III) TBI biomarkers specifically bind to the pre-conjugated capture antibodies. (IV) HRP-labeled detection antibodies introduced to form the sandwich complex. (V) Tyramide-labeled Alexa Fluor 647 molecules deposited onto tyrosine residues of the immunocomplex through HRP-catalyzed reaction.

Figure 6.. TSA-Modified CAD-IA for S100B Detection in 0.1% BSA.

(A) Representative fields of view of S100B-negative (no S100B input) and S100B-positive samples (0.1 μg/mL). In the presence of S100B, pronounced colocalization between immunocomplex fluorescence and nanoparticle hotspots is observed, indicating successful formation of target-specific immunocomplexes, whereas negligible colocalization is detected in the absence of S100B. (B) Digital Analysis of the TSA-Modified CAD-IA. Dose response analysis demonstrates sensitive and reproducible detection of S100B with a sensitivity down to 0.0002 μg/mL, at least 100-fold higher than CAD-IA not using TSA. Linear regression of the digital signal output shows a strong correlation between colocalization percentage and S100B concentration across the dynamic range of 0.0002–1 μg/mL (R² = 0.9663). Data are presented as mean ± SD (N = 6 spots on a single device). Statistical significance was evaluated using ordinary one-way ANOVA with Dunnett’s multiple comparisons test (****p < 0.0001, ***p = 0.0002).

Figure 7.. TSA-Modified CAD-IA for S100B Detection in 10% FBS.

The assay showed consistent responses with a limit of detection (LoD) of 0.002 μg/mL and a linear range from 0.002 to 1 μg/mL (R² = 0.9682). The projected LoD in 100% FBS is 0.02 μg/mL, exceeding the threshold for severe TBI diagnostic. Data are presented as mean ± SD (N = 6 spots on a single device). Statistical significance was assessed using ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test (****p < 0.0001).