The mechanisms of leukocyte-mediated beta cell destruction during human type 1 diabetes remain unclear. It is also uncertain if immune-beta cell contact is always essential during this destructive process.
We determined the frequency and spatial distribution of leukocytes in the intra- and peri-islet regions and in relation to residual insulin cells in a cohort of 9 nPOD cases with T1D (duration 0.25 to 12 years) and in non-diabetic subjects without autoantibodies (n=3), with only GAD autoantibodies (n=2) and those with 2 autoantibodies (n=3). Pancreatic sections from the head, body and tail were co-immunostained for insulin, glucagon and total leukocytes. Islets were analyzed for the number and location of leukocytes in relation to the presence and absence of insulin. Selected sections were also co-immunostained for nitrotyrosine, insulin and glucagon.
In non-diabetic cases with either a single or two autoantibodies, all islets showed a normal complement of beta cells. The number and distribution of islet-associated leukocytes were similar to normal subjects. In contrast, diabetic cases showed significant differences in the number of islets with residual insulin cells in the three pancreatic regions and in the pattern and degree of leukocytic infiltration. In most diabetic cases, insulin-positive islets harboured various degrees of peri-islet leukocytes, independent of beta cell number, being present in considerably reduced numbers in intra-islet areas. Only a small number was seen closely juxtaposed to the remaining insulin cells. Nitrotyrosine was observed in residual beta cells in 2 cases with longstanding diabetes.
We conclude that in some T1D cases residual beta cells remain localized to islets clustered in specific microanatomical sites and independent of disease duration. Leukocytic aggregates had a heterogeneous pattern, with a predominant peri-islet location. Peri-islet leukocytes may be important contributors to beta cell injury and death through release of proinflammatory cytokines and free radicals.