Development of an early detection test for Type 1 Diabetes via NGS-based analysis of β cells-specific methylation patterns in cfDNA

PROJECT SUMMARY Type 1 diabetes mellitus (T1D) is a chronic autoimmune disease that affects ~1.5M Americans, characterized by persistent hyperglycemia requiring life-long administration of external insulin. T1D is caused by the immune system-derived death of the insulin-producing β cells in pancreatic islets. Because insulin production is initially sustained by a higher output from surviving β cells, overt symptoms arise when 80%+ of β cells have been lost, making diagnosis only possible at a stage of advanced disease. Islet autoantibodies, the only established T1D biomarkers, are valuable for assessing long-term disease risk. However, because they cannot inform on the rate of β cell death, autoantibodies cannot precisely predict when the disease will strike. Therefore, a technology capable of identifying β cell death during the asymptomatic stages of T1D would provide an invaluable diagnostic tool for the early detection of T1D, allowing earlier intervention with immunosuppressive therapies and delaying critical islet mass loss. To address this gap, Pupil Bio is developing an approach leveraging next-generation sequencing for methylation analysis (methyl-seq) of cell-free DNA (cfDNA). Each cell and tissue type carries unique methylation patterns, which can inform on the cellular origin of DNA molecules in a mixed sample such as cfDNA. cfDNA is released by dying cells as a result of cellular turnover or of an active injury occurring to a tissue. Thus, detecting methylation patterns specific to cell types that are not usually found in healthy cfDNA can reveal ongoing damage to a tissue during a pathogenic process, allowing for non-invasive, specific early detection. Available technologies for methyl seq-based early detection fail to achieve high sensitivity, and require high volumes of blood. We propose that high cumulative sensitivity can be obtained by (i) combining a high number of methylation markers with non-zero sensitivity, and (ii) implementing an amplicon-based NGS workflow that minimizes sample loss. In this Fast-track SBIR application, the Pupil Bio team proposes to first apply a custom hybrid capture panel to validate thousands of candidate methylation patterns that are ultra-specific to pancreatic islets. In parallel, we will build and characterize an initial amplicon-based panel implementing a novel algorithm for highly multiplexed primer design in methyl-seq. Next, we will hone the list of candidate markers by identifying patterns that are ultra-specific to β cells only via high-depth sequencing, and expand the amplicon-based panel to hundreds of amplicons. Finally, after analytical validation, we will characterize the clinical sensitivity of the panels by testing samples from recent-diagnosis T1D patients, and demonstrate we can achieve high sensitivity while requiring small blood amounts. Completion of these studies will allow Pupil Bio to establish a CLIA lab and to commercialize its test as a laboratory-developed test (LDT) for the early detection of T1D.