Microdialysis was combined with a highly sensitive sequential multiple antigen radioimmunoassay to simultaneously measure extracellular cholecystokinin and neurotensin fragments from discrete regions of the rat brain in vivo. The assay was conducted in 96-well plates and provided a limit of detection for both peptides of 0.1 fmol. Dialysis membranes composed of polyacrylonitrile, Cuprophan and polycarbonate were evaluated in vitro using both radiolabelled peptides and radioimmunoassay. Polycarbonate probes were implanted in the posterior medial nucleus accumbens-septum, medial caudate nucleus or medial prefrontal cortex of halothane-N2O-anaesthetized rats. Cholecystokinin immunoreactivity levels were generally above the assay detection limits (0.1-0.7 fmol) in 30-min samples from all three regions under basal conditions. Recovered basal amounts of neurotensin immunoreactivity were detectable in the nucleus accumbens-septum in approximately 50% of experiments (0.1-0.2 fmol) but were not measured in the caudate nucleus or prefrontal cortex. In the nucleus accumbens-septum, a 10-min pulse of 200 mM K(+)-containing artificial cerebrospinal fluid in the perfusion medium during a 30-min sampling period increased the recovered cholecystokinin and neurotensin immunoreactivity to 9.7 fmol +/- 1.9 S.E.M. and 5.8 +/- 1.6 S.E.M., respectively. A second stimulation following a 2.5-h interval produced similar elevations with S2:S1 ratios of 0.62 +/- 0.07 and 0.68 +/- 0.07 for cholecystokinin and neurotensin, respectively. In a separate series of experiments the second stimulation of both peptides was prevented by perfusion of a 10 mM EGTA-containing medium. Similar results were obtained in the caudate nucleus for cholecystokinin, but K(+)-induced elevations in neurotensin immunoreactivity were much smaller (0.5 fmol) in this brain region and calcium dependency was not established. Sequential K+ stimulations at 50, 100 and 200 mM produced progressively greater increases in recovered cholecystokinin and neurotensin immunoreactivity from the nucleus accumbens-septum and of cholecystokinin immunoreactivity from the prefrontal cortex. No neurotensin immunoreactivity was detected in the prefrontal cortex following K+ stimulation. Large post mortem increases in the recovered amounts of cholecystokinin and neurotensin immunoreactivity were observed. This effect was significantly attenuated by EGTA although there was a large calcium-independent component of the cholecystokinin immunoreactivity. On reverse-phase high-performance liquid chromatography the major cholecystokinin-immunoreactive peak co-eluted with sulphated cholecystokinin octapeptide. Neurotensin-immunoreactive material co-eluted with neurotensin (1-13), neurotensin (1-12), neurotensin (1-11), neurotensin (1-10) and neurotensin (1-8). These results further demonstrate the potential of microdialysis for studying neuropeptide release and metabolism in vivo when combined with sufficiently sensitive assay procedures.