In premature infants with necrotizing enterocolitis (NEC) the gastrointestinal mucosal barrier is immature, but little is known about the immune response of immature bowel. The aim of this study was to evaluate the intestinal mucosal immune response in experimental NEC. In general anaesthesia, NEC was induced in six newborn piglets by injection of bovine casein into terminal ileum. Six controls received an equal amount of saline. Four hours later, samples were taken from the macroscopically most affected part of the treated loop and from the macroscopically healthy untreated intestine. Monoclonal antibodies to porcine CD1, CD2, CD4, CD8, CD45 and IgM were used for immunohistochemical staining. Casein-treated bowel showed typical macro- and microscopic findings of NEC. No changes were found in the saline-treated bowel. In both groups the bowel outside the treatment sector was normal. In casein-treated animals, treated samples showed significant decrease in density of CD4+ cells when compared with saline-treated controls. Similar trend was found in CD2+ and CD8+ cells but without statistical significance. Macroscopically healthy proximal untreated samples showed significant decrease in densities of CD2+, CD4+ and CD8+ lymphocytes in casein-treated group when compared with control samples. In casein-treated animals the density of CD45+ cells in the non-injected bowel was also decreased, but this did not reach statistical significance. Densities of CD1+ and IgM+ cells did not differ between casein-treated and saline-treated groups. A significant T-cell decrease was found in the present NEC model. Surprisingly, this was most prominent in the macroscopically healthy bowel outside the casein injection segment. The reason for T-cell decrease remains unclear, but bovine casein is known to contain peptide fractions that can modulate immune function. These findings may have implications in the design of neonatal milk formulas.