Previous studies have speculated that increased prostaglandin (PG) levels seen in several models of nephrogenic diabetes insipidus (NDI) might mediate the decreased aquaporin 2 (AQP2) expression which is thought to underlie the impaired urinary concentrating ability. On the other hand, in vivo experiments with non-steroidal antiinflammatory drugs, which inhibit PG synthesis, decreased AQP2 expression, suggesting that PGs may actually increase AQP2 expression (reviewed in Nielsen et al. 2002). These differences may reflect the effects of different prostaglandins and/or receptors in different situations. The mouse collecting duct cell line mpkCCDc14 seems to be the best existing model for the study of the regulation of AQP2 expression in vitro, since it has previously been shown to express the water channel AQP2 in response to vasopressin (VP) stimulation (Hasler et al. 2002) and hypertonicity (Hasler et al. 2005). We have therefore used it to investigate the direct effects of prostaglandins E₂ and F₂ on baseline and vasopressin-induced AQP2 expression. Cells were grown on Falcon cell culture inserts, until transepithelial resistance (measured using an EVOME voltometer) was maximal. They were then incubated for 48 h in serum- and hormone-free medium supplemented with 1 nM VP, 100 nM PGE₂, or 100 nM PGF₂, as appropriate: six filters received (i) nothing, (ii) VP, (iii) PGE₂, (iv) PGE₂ + VP, (v) PGF₂, and (vi) PGF₂ + VP. At the end of the incubation period, cells were scraped off the filters and dissolved in Laemmli sample buffer. The experiment was repeated four times, with cells from passages ranging from 25 to 35. Samples were run on 12% polyacrylamide gels, and Western blotted for AQP2, using an affinity purified antibody against the c-terminus of AQP2. Results were analysed using ANOVA, Student’s t tests and the false-discovery rate procedure. In the absence of vasopressin, levels of AQP2 in the cells were undetectable, and treatment with PGE₂ or PGF₂ alone had no detectable effect on this. In contrast, cells treated with VP showed substantial levels of AQP2. Concurrent treatment with PGE₂ decreased AQP2 expression to 43 ± 10% of the levels seen with VP alone, and treatment with PGF₂ reduced it to 57 ± 14% (each n = 4, p< 0.05). These results demonstrate that PGE₂ and PGF₂, which are the principal prostaglandins found in the collecting duct, decrease AQP2 expression, at least at this concentration, which is in the high physiological range. Thus increased prostaglandin synthesis, in pathophysiological conditions, may explain the decrease in AQP2 expression seen in a number of forms of acquired NDI.