ECE2024 Poster Presentations Late-Breaking (77 abstracts)
1University College London, UCL Division of Medicine, United Kingdom; 2University College London, United Kingdom
Background: Type 1 diabetes mellitus (T1DM) involves the autoimmune destruction of pancreatic islet beta cells, with recent research implicating a role for viral infections in its onset. The rise in T1D cases during the COVID-19 pandemic has sparked investigations into potential causes. Molecular mimicry between SARS-CoV-2 antigens and pancreatic islet antigens has been suggested as a plausible mechanism, yet no attempts have been made to explore this phenomenon as comparing functional protein structure remains challenging. To address this, we used pcDelta, a near-coincidence statistic originally developed for evaluating T-cell receptor (TCR) clonotype skewing following antigen exposure. Through the application of pcDelta to two sets of TCR sequences, we quantified near-sequence similarity, with higher pcDelta values indicating greater sequence similarity. The presence of similar TCR sequences suggests similar target antigens and is thus a surrogate for the degree of molecular mimicry.
Methods: CDR3 sequences of the beta chain of TCR specific to various viral antigens were sourced from the VDJ database, a meticulously curated repository of TCR sequences with annotated antigen specificities. Concurrently, islet antigen-specific CDR3 were extracted from a comprehensive review consolidating TCR clonotypes isolated from individuals diagnosed with T1D, cross-validated across multiple studies. Using pcDelta, we quantified near-sequence similarity at the amino acid level between TCR specific to islet pancreatic antigens and viral antigens. To account for different number of epitope-specific TCRs, subsampling was performed and a pcDelta value calculated for each sample.
Results: The analysis of TCR specific to Islet antigens and SARS-CoV-2 revealed a pcDelta value of (4.1310^-5), significantly exceeding the pcDelta values observed between Islet TCRs and influenza A (2.4210^-5) and HIV-1 (2.19*10^-5). This indicates a higher degree of near-sequence similarity between pancreatic islet and SARS-CoV-2 TCRs compared to other viral antigens (P-value <0.001).
Conclusion: The significant elevation in pcDelta values between SARS-CoV-2 and islet antigens TCR compared to other viral antigens indicates increased near sequence similarity between the two TCR groups. This suggests the TCRs might be targeting antigens that might be structurally similar, implying molecular mimicry between SARS-CoV-2 and islet antigens. However, the reliance on existing data sources for TCR sequences may introduce biases or incomplete representation of the T1D landscape. In addition, the study simplifies TCR binding dynamics, although the assumption has been widely used in the literature. Future investigations could employ structural modelling or functional assays to elucidate the impact of molecular mimicry in T1D and consolidate the validity of the approach.