Butyrate may have paradoxical effects on epithelial cells of similar origin. This study aimed to examine the hypothesis that one mechanism that dictates a cell's response to butyrate is its state of activation. First, the responses to 24 h exposure to butyrate (1-2 mM) of normal and neoplastic human colonic epithelial cells activated by their isolation and primary culture, and of colon cancer cell lines, LIM1215 and Caco-2, were examined. In primary cultures of normal and cancer cells, butyrate had no effect on alkaline phosphatase activities but significantly suppressed urokinase receptor expression by a mean +/- SEM of 30 +/- 12% and 36 +/- 9%, respectively. Interleukin-8 secretion was suppressed by 44 +/- 7% in normal cells (P < 0.05) but was unchanged in cancer cells. In contrast, the cell lines significantly increased alkaline phosphatase activities by >50%, urokinase receptor expression >2-fold and interleukin-8 secretion >3-fold in response to butyrate. Secondly, the effect of butyrate on Caco-2 cells was examined with or without prior exposure to a specific activating stimulus [tumour necrosis factor alpha (TNF alpha)]. Interleukin-8 secretion increased by 145 +/- 23% and 132 +/- 17% on 24 h exposure to 2 mM butyrate or 0.1 microM TNF alpha alone, respectively. However, in cells pre-treated with TNF alpha, butyrate significantly inhibited secretion by 34 +/- 7% below unstimulated levels. The response to butyrate of urokinase receptor, whose expression was not stimulated by TNF alpha, was unchanged. These effects were mimicked by trichostatin A, an inhibitor of histone deacetylase, suggesting that butyrate's paradoxical effects may have been operating by the same mechanism. In conclusion, some of the paradoxical effects of butyrate do not appear to represent inherent differences between normal and transformed cells. Rather, the response may be determined by the state of activation of the cells.