Type I, insulin-dependent diabetes (IDD) in man and the NOD (non-obese diabetic) mouse is believed to result from an autoimmune destruction of pancreatic beta (beta) cells. In both species the pathologic correlate of this destruction is a peri- and intra-islet immune cell infiltrate, referred to as insulitis, easily recognized histologically. Although histologic studies of insulitis have established the autoimmune nature of IDD, controversy remains concerning the phenotypes and especially function of the islet-infiltrating immune cells. In the present study, we reveal a new protocol which permits enumeration of islet-infiltrating leukocytes using flow cytometry (FACS). Using this technique, we have analyzed systematically the changes in major leukocyte phenotypes during the development of the insulitis lesion. Results indicate that: 1) the first islet-infiltrating leukocytes are class II+, Ig- monocytes and CD8+ T lymphocytes, 2) after a slight decrease in the CD8+ cell population, an influx of CD4+ T lymphocytes occurs, accompanied by increasing numbers of CD8+ T cells, as well as IgM+ and IgG+ B lymphocytes, 3) whereas all the IgM+ B cells appear to be CD5+, between 70-95% of IgG+ B cells express CD5, and 4) throughout the response, the class II+, IgG-cell population remains relatively constant. These data, together with our previous work showing that autoantibody binds to pancreatic beta cells prior to leukocytic infiltration, permit construction of a working model for the natural history of insulitis in the NOD mouse. Briefly, the initial cellular response against the beta cell, which begins between 5-7 weeks of life, appears to be a local inflammation in which class II-positive macrophages and CD8-positive cells (possibly with antibody-dependent cell-mediated cytotoxic potential) infiltrate the islet. Finally, T helper cells are activated, traffic briefly through the islet, and elicit a B lymphocyte and T killer cell invasion of the islet that ultimately leads to beta cell necrosis.