Joint analysis of the nPOD-Virus Group data: the association of enterovirus with type 1 diabetes is supported by multiple markers of infection in pancreas tissue

Abstract

<p>Aims/hypothesis: Previous pathology studies have associated enterovirus infections with type 1 diabetes by examining the enterovirus capsid protein 1 (VP1) in autopsy pancreases obtained near diabetes diagnosis. The Network for Pancreatic Organ Donors with Diabetes (nPOD) has since obtained pancreases from organ donors with type 1 diabetes (with broad age and disease duration) and donors with disease-associated autoantibodies (AAbs), the latter representing preclinical disease. Two accompanying manuscripts from the nPOD-Virus Group report primary data from a coordinated analysis of multiple enterovirus indices. We aimed to comprehensively assess the association of multiple enterovirus markers with type 1 diabetes. Methods: The nPOD-Virus Group examined pancreases from 197 donors, recovered between 2007 and 2019, classified into five groups: donors with type 1 diabetes, with residual insulin-containing islets (T1D-ICI group, n=41) or with only insulin-deficient islets (T1D-IDI, n=42); donors without diabetes who are AAb-negative (ND, n=83); and rare donors without diabetes expressing a single AAb (AAb<sup>+</sup>, n=22) or multiple AAbs (AAb<sup>++</sup>, n=9). We assessed the overall association of multiple indicators of enterovirus infection, case-by-case and between donor groups, as well as assay agreement and reproducibility, using various statistical methods. We examined data from 645 assays performed across 197 nPOD donors. Results: Detection of enterovirus indices by independent laboratories had high reproducibility, using both enterovirus-targeted and unbiased methods. T1D-ICI donors had significantly higher (p<0.001) proportions of positive assay outcomes (58.4%) vs T1D-IDI (10.3%), ND (17.8%) and AAb-positive donors (AAb<sup>+</sup> 24.6%; AAb<sup>++</sup> 35.0%). Head-to-head comparisons revealed increased proportions of donors positive in two independent assays among T1D-ICI vs ND donors (VP1/HLA class I [HLA-I], p<0.0001; VP1/enterovirus-specific RT-PCR (EV-PCR), p=0.076; EV-PCR/HLA-I, p=0.016; proteomics/HLA-I, p<0.0001; VP1/proteomics, p=0.06). Among 110 donors examined for three markers (VP1, EV-PCR and HLA-I), 83.3% of T1D-ICI donors were positive in two or more assays vs 0% of ND (p<0.001), 26.7% of AAb<sup>+</sup> (p=0.006), 28.6% of AAb<sup>++</sup> (p=0.023) and 0% of T1D-IDI (p<0.001) donors. Conclusions/interpretation: The nPOD-Virus Group conducted, to date, the largest and most comprehensive analysis of multiple indices of pancreatic enterovirus infections in type 1 diabetes; these were more prevalent in T1D-ICI and AAb<sup>++</sup> donors than in other groups. Their preferential detection of these indices in donors with residual beta cells and autoimmunity implicates enterovirus infections across disease progression stages and supports a contribution to beta cell loss, directly or indirectly, even after diagnosis. The relatively small number of infected cells and the low amount of viral RNA support the existence of non-acute, low level, possibly persistent enterovirus infections in the pancreas.</p>