We have compared the effects of forskolin, N6,2'-O-dibutyryladenosine 3':5'-cyclic monophosphate (dibutyryl cyclic AMP, Bt2-cAMP), and butyrate on several aspects of neuroblastoma cell physiology. The morphology of Neuro 2A cells was similar after incubation with forskolin and Bt2-cAMP, which caused extensive neurite outgrowth, whereas in the presence of butyrate some rudimentary neurites were formed but they were not nearly as extensive. All compounds produced a dose-dependent inhibition of cell proliferation, but the effect of Bt2-cAMP was more marked than that caused by forskolin, thus showing that the effect of Bt2-cAMP is due partially to the butyrate released. Acetylcholinesterase activity was lower in the cells incubated with butyrate or Bt2-cAMP than in untreated cells or in forskolin-treated cells. This suggests that cyclic AMP does not play a role in the regulation of this enzyme. Bt2-cAMP produced histone acetylation, a well-known effect of butyrate in cultured cells, whereas forskolin did not affect this modification. Consequently, the levels of thyroid hormone receptor, a nuclear protein whose concentration is regulated by butyrate through changes in acetylation of chromatin proteins, were decreased in cells incubated with Bt2-cAMP or butyrate, but were unaffected by forskolin. Butyrate elevated the concentration of histone H1(0), a protein that increases in neuroblastoma cells as a result of different treatments that block cell division. The concentration of H1(0) in the cells treated with Bt2-cAMP was at a level intermediate between that found after treatment with butyrate and with forskolin. The present results clearly indicate that some of the effects of Bt2-cAMP on neuroblastoma cells can be attributed to the butyryl moiety of this compound rather than to the cyclic nucleotide itself.