Comment on Manduchi et al. No Evidence for Persistent Enteroviral B Infection of Pancreatic Islets in Patients With Type 1 Diabetes and Prediabetes From RNA Sequencing Data. Diabetes 2024;73:1697–1704

The recent article by Manduchi et al. (1) reported no evidence of enterovirus (EV) RNA in the pancreas of autoantibody-positive organ donors with or without type 1 diabetes from the Human Pancreas Analysis Program (HPAP) repository. Building on these results, the authors conclude that persistent EV infections may not drive type 1 diabetes pathogenesis. We respectfully argue that this conclusion dismisses the technical limitations of their work and the available body of evidence supporting the opposite scenario.

First, Manduchi et al. (1) used RNA sequencing (RNA-seq) to detect EV RNA, but RNA-seq lacks sensitivity to identify the low-grade presence of EV that is repeatedly detected in the pancreas of donors with type 1 diabetes. A recent study (2) demonstrated that RNA-seq yields negative results even when RT-qPCR run in parallel detects EV RNA. These observations were made using the same donor tissue samples, replicated in two different laboratories, and confirmed by sequencing. Sensitivity is key, and even RT-qPCR protocols need to be carefully optimized to this end. Thus, the conclusions in the present study might have differed if RT-qPCR had been included.

Second, in the face of this technical shortcoming, the authors attempted to validate RNA-seq sensitivity by infecting HeLa cells in vitro at a low multiplicity of infection. However, this approach is not directly comparable to in vivo conditions: HeLa cells are not β-cells, and EV replication in vitro spreads much more rapidly and uniformly than in the human pancreas, even at a low multiplicity of infection (3). Moreover, RNA degradation in donor pancreas samples is greater than that in cultured cell lines.

Third, some caveats also apply to the cases of the 11 autoantibody-positive individuals analyzed, who may not be representative of progressive autoimmunity. They were all adolescents or adults (median age 21 years), they were mostly (9/11) positive for a single anti-GAD autoantibody, and at least three of them (including an individual who was double autoantibody positive) carried a protective HLA haplotype (DQB1*06:02 or DQB1*03:01 in combination with a neutral or protective haplotype). This information is not provided in the article, and it does aid in contextualizing the results.

Last, several other lines of evidence using diverse technologies, e.g., immunostaining (4), fluorescence in situ hybridization (5), and virus growth in cell culture, point to EV presence in the pancreas of donors with type 1 diabetes (6). Moreover, the The Environmental Determinants of Diabetes in the Young (TEDDY) study analyzed longitudinal samples from well-characterized at-risk living individuals and provided strong evidence for the association between islet autoimmunity and persistent or recurrent EV infections (7).

In conclusion, while publishing negative findings remains important, such findings should be particularly well supported to avoid misleading conclusions about a so-called lack of evidence, especially in the context of multiple previous reports containing such evidence. A comprehensive strategy covering multiple detection methods of appropriate sensitivity as well as consideration of virus-host interactions is indispensable to fully elucidate the relationship between EV infection and islet autoimmunity.