Autoantibodies to the neuroendocrine protein insulinoma-associated pro
tein 2 (IA-2), a member of the tyrosine phosphatase family, have been
observed in individuals with or at increased risk for IDDM. Because th
is disease is thought to result from a T-cell-mediated autoimmune dest
ruction of the insulin-producing pancreatic beta-cells, we analyzed hu
moral and cellular immune reactivity to this autoantigen to further de
fine its role in the pathogenesis of IDDM. Peripheral blood mononuclea
r cells (PBMC) from individuals with newly diagnosed IDDM or at varyin
g levels of risk for the disease were stimulated in vitro with the ent
ire 42-kDa internal domain of IA-2 (amino acids 603-979), a series of
control antigens (glutathionine-S-transferase, tetanus toroid, Candida
albicans, mumps, bovine serum albumin), and a mitogen (phytohemagglut
inin). The frequency and mean stimulation index of PBMC proliferation
against IA-2 was significantly higher in newly diagnosed IDDM subjects
(14 of 33 [42%]; 3.8 +/- 4.5 at 10 mu g/ml) and autoantibody-positive
relatives at increased risk for IDDM (6 of 9 [66%]; 3.9 +/- 3.2) comp
ared with autoantibody-negative relatives (1 of 15 [7%]; 1.8 +/- 1.0)
or healthy control subjects (1 of 12 [8%]; 1.5 +/- 1.0). The frequenci
es of cellular immune reactivities to all other antigens were remarkab
ly similar between each subject group. Sera from 58% of the newly diag
nosed IDDM patients tested mere IA-2 autoantibody positive. Despite in
vestigations suggesting an inverse association between humoral and cel
lular immune reactivities against islet-cell-associated autoantigens,
no such relationship was observed (r(s) = 0.18, P = 0.39) with respect
to IA-2. These studies support the autoantigenic nature of IA-2 in ID
DM and suggest the inclusion of cellular immune responses as an adjunc
t marker for the disease.