The function of the eosinophil in eosinophilic pulmonary syndromes and asthma is uncertain. To determine if eosinophils might play a harmful role in these conditions, we cocultured purified human eosinophils, eosinophil major basic protein (MBP), and chromatographically eluted eosinophil granule fractions with human A549 and rat type II pneumocytes. Damage to these target cells was measured as cell lysis and nonlethal cell detachment. We found that unstimulated intact eosinophils affected minimal lysis or detachment of either pneumocyte target, but eosinophils stimulated with phorbol myristate acetate and other activators produced time- and dose-dependent nonlytic detachment of both targets. In contrast, supernatants from activated eosinophils did not produce significant injury, suggesting that close apposition of the effector and target cells was required. Catalase and superoxide dismutase did not inhibit the detaching activity of eosinophils, suggesting that hydrogen peroxide and superoxide anion were not activity of eosinophils, suggesting that hydrogen peroxide and superoxide anion were not responsible for mediating this form of injury. In contrast to our findings with intact eosinophils, we observed that the addition of purified eosinophil MBP to pneumocytes caused marked cytolysis with little detachment. When sequential fractions of eosinophil granules separated by Sephadex G-50 chromatography were added to A549 and rat type II pneumocyte targets, it was found that different fractions produced distinct forms of injury. Higher molecular weight fractions containing lysosomal enzymes and eosinophil peroxidase produced predominantly detachment, whereas fractions enriched in MBP produced lysis. These results indicate that intact eosinophils can produce nonlytic detachment of alveolar pneumocytes that is probably not dependent on the generation of toxic oxygen radicals but rather appears to be mediated by granule-associated products, possibly lysosomal enzymes. Furthermore, although intact eosinophils are not capable of lysing alveolar epithelial cells under the conditions of our assay, MBP has the potential to do so when the protein is released in high enough concentrations. The in vivo relevance of these findings in eosinophilic lung diseases may be that eosinophils, by producing both desquamation and death of alveolar epithelium cells, may increase the permeability of the alveolus to fluid and cells. Moreover, these forms of damage might also enhance the ingress of inhaled antigens across the pulmonary epithelial barrier, thus increasing immunologic sensitization.