T he studies reported here were designed to determine if there is an apical-basolateral asymmetry to the release of prostaglandins by or to the biochemical effects of prostaglandins on the renal collecting tubule. Canine cortical collecting tubule (CCCT) cells were isolated by immunodissection and seeded at supraconfluent densities on Millipore filters. The resulting confluent monolayer of CCCT cells: (a) developed and maintained a transcellular potential difference of 1 mV (apical side negative); (b) exhibited a permeability to inulin that was the same as that obtained with similar monolayers of Madin-Darby canine kidney (MDCK) cells; and (c) released adenosine 3',5'-cyclicmonophosphate (cAMP) in response to arginine vasopressin (AVP) added to the basolateral but not the apical surface of the monolayer. These results indicate that confluent monolayers of CCCT cells on Millipore filters have characteristics of asymmetry that are seen with intact collecting tubules. Moreover, PGE2 added to either side of the CCCT cell monolayer crossed the monolayer at the same slow rate as inulin, which demonstrated the feasibility of examining the sidedness of the effects of and the release of PGE2. Although AVP caused cAMP release only when added to the basolateral side of CCCT cells, AVP caused the release of PGE2 when added to either the apical or basolateral surface. This result implies that there are at least two AVP receptor systems, one coupled to cAMP synthesis and one to PGE2 formation. In contrast to the results observed with AVP, bradykinin caused PGE2 release only when added to the apical surface of CCCT cells, which suggested that urinary but not blood borne kinins elicit PGE2 formation by the canine collecting tubule. PGE2 was released in comparable amounts on each side of the monolayer in response both to AVP and to bradykinin. High concentrations (greater than or equal to 10(-8) M) of PGE2 added to either side of the monolayer caused the release of cAMP. However, at concentrations (10(-10) - 10(-12) M) at which PGE2 had no independent effect on cAMP release, PGE2 inhibited the release of cAMP, which normally occurred in response to AVP. This inhibition occurred with PGE2 added to either the apical or basolateral surface of the CCCT cell monolayer. PGE2 (10(-11) M) also inhibited the AVP-induced accumulation of intracellular cAMP by CCCT cells seeded on culture dishes. This inhibition was only observed when the cells were preincubated with PGE2 for greater than or equal to 20 min. Our results are consistent with the concept that inhibiton by prostaglandins of the hydroosmotic effect of AVP is due to inhibition of AVP-induced cAMP production. This inhibition does not appear to involve a direct physical interaction of PGE2 with the AVP receptor which is coupled to adenylate cyclase, since CCCT cells must be preincubated with PGE2 for 20 min for the inhibition to be observed, and since PGE2 added to the apical surface of CCCT cells inhibits cAMP release in response to AVP acting from the basolateral surface.