Back

Beta cell reactivity defines disease-relevant pancreatic CD8 T cells in type 1 diabetes

Anderson, A. M.; Landry, L. G.; Barra, J. M.; Wells, K. L.; Shilleh, A. H.; Castro-Gutierrez, R.; Ladd, A. M.; DeNicola, M.; Babon, J. A. B.; Mallone, R.; Kent, S. C.; Michels, A. W.; Russ, H. A.; Nakayama, M.

2026-06-10 immunology
10.64898/2026.06.08.729940 bioRxiv
Show abstract

Type 1 diabetes (T1D) is characterized by immune-mediated destruction of pancreatic beta cells, yet the properties that distinguish disease-associated CD8 T cells from other pancreatic resident T cells remain incompletely defined. In this study, we analyzed CD8 T cell receptor (TCR) clonotypes isolated from the pancreas of organ donors with and without T1D and assessed their reactivity to beta cells using stem cell-derived beta-like cells. We found that highly beta cell-reactive CD8 T cells were selectively present in the pancreas of T1D donors but were largely absent from donors without T1D. In contrast, virus-specific CD8 T cells were detected in pancreata of donors with and without T1D and showed no evidence of cross-reactivity to beta-like cells, indicating that pancreatic residency alone does not confer beta cell specificity. Among beta cell-reactive CD8 T cells in T1D, reactivity to native peptides from major islet proteins other than preproinsulin was rare. Thus, despite beta cell specificity as a hallmark of T1D, T cells reactive to native islet proteins other than preproinsulin do not infiltrate the islets. These results identify beta cell reactivity as a key functional feature separating T1D-associated CD8 T cells from other pancreatic T cells. This functional definition of pathogenic T cells offers a framework for understanding selective beta cell loss and for developing approaches to monitor and therapeutically target disease-relevant CD8 T cells.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.