Poster Y5 - Multiplexed Islet Autoantibody Detection via Vertical Flow Assay for Early Type 1 Diabetes Diagnosis

Introduction: Over 57 million people globally suffer from diabetes mellitus, predominantly in resource-limited areas. Diabetic ketoacidosis (DKA), a life-threatening condition resulting from insufficient insulin remains a major concern, causing over 50% of child fatalities from Type 1 Diabetes (T1D) and costing the US $5.1 billion annually. Early diagnosis of T1D can significantly reduce risks by enabling timely interventions to prevent or mitigate DKA (Fig. 1A). Islet autoantibodies (IAb) have been identified as critical predictive biomarkers where the presence of two or more of these autoantibodies—insulin (Ins), islet antigen-2 (IA-2), zinc transporter 8 (ZnT8), and glutamic acid decarboxylase (GAD65)—signals a 70% likelihood of developing T1D within ten years (Fig. 1B). Currently, no point-of-care (POC) test exists for simultaneous detection of these islet autoantibodies in a cost-effective and rapid manner for population-level screening. To address this gap, we created a Vertical Flow Assay (VFA) for multiplexed detection of T1D-associated autoantibodies through a rapid, low-cost test, delivering results in less than 20 minutes. Coupled with a smartphone reader for data acquisition and automated analysis, the VFA T1D test provides T1D risk classification and is well-suited for the early diagnosis of T1D in at-risk populations. (Fig. 1C).

Materials and

Methods: To develop our T1D VFA test, we engineered the platform with specific autoantigens to detect the four IAbs, immobilizing each antigen in duplicate on a wax-patterned sensing membrane to enhance the test's reliability. Unlike lateral flow assays, the large sensing surface supports multiple immunoreactions, reducing variability by averaging technical replicates. Each VFA device includes additional paper layers, such as the Vertical Flow Diffuser and Inter Pad that were surface-modified with bovine serum albumin (BSA) and encased in 3D-printed structures to ensure uniform fluid distribution. Samples for testing were prepared by spiking rabbit autoantibodies into either phosphate-buffered saline (PBS) or healthy human serum, while T1D patient samples were sourced from a commercial repository with ground truth IAb values. We conjugated anti-IgG antibody to 40 nm gold nanoparticles (AuNPs) to enable detection of positive signals. Following the assay, a mobile phone-based reader is used to capture the read-out images of the activated sensing membrane for analysis, facilitating the quantification of signals by automated image analysis algorithms.

Results, Conclusions, and Discussions: We validated the T1D VFA test by successfully detecting three of the four IAbs, demonstrating the effect of capture antigen and spiked antibody concentration on signal intensity (Fig 2A). To simulate varying IAb titers in T1D(+) patients, we conducted a sweep of surrogate spiked rabbit autoantibodies, demonstrating the platform's ability to quantify signal and stratify disease progression stages. Increasing the capture antigen concentration on the sensing membrane resulted in higher signal intensities, enabling the ability to tune the assay to detect low IAb titers. We optimized the running buffer composition to ensure optimal flow of reagents and adequate washing of non-specific signals. Despite a drop in signal intensity, a wash buffer consisting of 1% Tween 20 and 1% bovine serum albumin (BSA) was found optimal to provide reliable signal (Fig 2B).
We demonstrate the assay’s robustness in detecting the IAbs by co-spiking rabbit autoantibodies in healthy human serum and concurrently detecting all three autoantibodies individually demonstrating the assay's capability for multiplexing and resilience against serum matric effects (Fig 3A). Finally, testing clinical human serum samples with the VFA showcased the assay’s capability to distinguish between diseased and healthy samples, particularly by adjusting antigen concentrations to enhance diagnostic specificity (Fig 3B).
In conclusion, initial results demonstrate the assay's capability to detect three critical islet autoantibodies, with high signal integrity and multiplexing capability. The simplicity of the assay, coupled with rapid results and automated analysis via a smartphone reader, highlights its potential for widespread adoption in population-level screening, potentially offering a reliable at-home test for early detection of T1D onset. Moving forward, the next steps include validating the performance of the assay in a larger cohort of human samples, ensuring that this platform can reliably identify all four key autoantibodies and provide comprehensive risk assessment for T1D.

Acknowledgements (Optional): We acknowledge the NIH and NSF PATHS-UP Engineering Research Center (ERC) and the NSF GRFP under grant no. DGE-2034835.