To investigate functional and chemical properties of anatomically characterized corticotropin-releasing factor-41 (CRF-41) producing neurons in vitro, hypothalamic slices of 6-day-old rats were maintained in culture for up to 6 weeks using a modified roller culture technique. This technique yields thick (100 microns) slices that contained an average of 300-400 CRF-41-immunostained neurons. The majority of CRF-41-positive cells were of small size (12-15 microns in diameter), and contained CRF-41-labeled dense core vesicles of 100 nm diameter as detected by electron microscopic postembedding immunocytochemistry. These cells represented the only CRF-41-positive cell population in the culture. Light microscope double immunolabelling of colchicine-treated cultures kept in a serum-containing media (SCM) indicated that about 60% of these CRF-41-positive neurons contains detectable levels of vasopressin-associated neurophysin (VP-NP). Culturing slices in serum-free, chemically defined media (SFM) resulted in an increased VP-NP immunostaining: parvicellular neurons labeled for both CRF-41 and VP-NP could be detected without colchicine treatment, and practically all CRF-41-positive neurons expressed VP-NP immunoreactivity. At the electron microscopic level there was a significant increase in VP-NP labeling density in the dense core vesicle compartment of CRF-41-positive varicosities. Adding dexamethasone (10 nM) to the SFM restored the staining pattern originally observed in SCM. Hence, the increased VP-NP and CRF-41 immunostaining after culturing CRF-41 neurons in SFM is most likely due to the absence of inhibitory glucocorticoids. The capacity of cultured paraventricular cells to release CRF-41 was assessed using an immunoassay. Unstimulated (basal) secretion of CRF-41 was not altered by five successive samplings at 2-hour intervals and stimulation of the same culture with 56 mmol K+ significantly increased (2-3 times) the CRF-41 content in the medium. The presence of dexamethasone (10 nM) in SFM induced a 6-fold reduction of K(+)-stimulated CRF-41 release and a 5 times reduction in tissue content in relation to cultures maintained in SFM without dexamethasone. In summary, we have demonstrated that cultured CRF-41 cells display morphological and biochemical features, as well as responsiveness to glucocorticoids, that is reminiscent to the situation in vivo. Thus, the model is well suited for studies of hypophysiotrophic CRF-41 cell functions.