We have shown that exposure to bovine insulin (BI) in cow's milk (CM) formula induces an insulin-specific immune response in infants. Here we studied the role of human insulin (HI) in breast milk as a modulator of the immune response to insulin. In a group of 128 children participating in the TRIGR pilot study, maternal breast milk samples were collected 3-7 days and/or 3 months after delivery. After exclusive breast-feeding, the children received either CM formula or casein hydrolysate during the first 6-8 months of life. Insulin concentration in breast milk and immunoglobulin G (IgG) antibodies to BI in plasma samples were measured by EIA. The levels of insulin in breast milk samples were higher in mothers affected by type 1 diabetes than in non-diabetic mothers (p = 0.007 and p < 0.001). The concentration of insulin in breast milk correlated inversely with the plasma levels of IgG antibodies to BI at 6 months of age in children who received CM formula (r = -0.39, p = 0.013), and at 12 months of age in all children (r = -0.25, p = 0.029). The levels of breast milk insulin were higher in the mothers of nine children who developed beta-cell autoimmunity when compared with autoantibody-negative children (p = 0.030); this holds true also when only children of diabetic mothers were included (p = 0.045). BI in CM induces higher levels of IgG to insulin in infants than does HI in breast-fed children. Instead, HI in breast milk seems to be tolerogenic and may downregulate the IgG response to dietary BI. However, our results in infants who developed beta-cell autoimmunity suggest that in this subgroup of children breast milk insulin does not promote tolerance.