It has been hypothesized that secretin may act directly on gastrinoma through the adenylate cyclase system to cause stimulation of gastrin release. We studied gastrinoma cells in vitro to determine whether secretin would stimulate gastrin release directly and whether the gastrinoma cell membrane had a functional secretin receptor adenylate cyclase system. Fresh tumor was prepared in cell suspensions containing 1.5 X 10(6) viable cells and incubated for 2 hours with either 2 mM CaCl2 alone (control) or 2 mM CaCL2 and 0.025 U/ml secretin. The gastrin content of the cells in each incubation chamber and the medium were determined by radioimmunoassay and results were expressed as mean gastrin pg/microgram protein +/- SD. Under basal conditions the cellular gastrin content was 39.9 +/- 6.4 (control) compared with 16.7 +/- 2.1 (secretin). After 2 hours of incubation, cellular gastrin content increased in both groups: 68.5 +/- 11.9 (control) to 68.3 +/- 5.5 (secretin). However, the percent of gastrin released into the medium during incubation decreased by one half in both groups (control 37.3% +/- 4.0% to 22.2% +/- 3.0%; secretin 42.8% +/- 7.0% to 18.9% +/- 1.8%). Adenylate cyclase activity was assessed by measuring cAMP generation in fresh-frozen gastrinoma and cultured gastrinoma cell membranes. Isoproterenol (10(-5) M), PGE1 (10(-4) M), and GppNHp (guanine nucleotide) (10(-5) M) caused fivefold to 25-fold increases in cAMP generation. Secretin did not stimulate adenylate cyclase activity above basal (21.73 +/- 4.07 and 2.29 +/- 1.2 pmol cAMP/mg protein/min) for frozen and cultured gastrinoma, respectively. Secretin failed to stimulate gastrin release and adenylate cyclase in vitro. This suggests that secretin-stimulated gastrin release in vivo may not be due to a direct effect of secretin on the gastrinoma.